Chapter 1

Basic Cell Culture

Jeffrey W. Pollard
1. Introduction
This article will describe the basic techniques required for successful
cell culture. It will also act to introduce some of the other chapters in this
volume. It is not intended, as this volume is not, to describe the establish-
ment of a tissue culture laboratory, nor to provide a historical or theoretical
survey of cell culture. There are several books that adequately cover these
areas, including the now somewhat dated but still valuable volume by Paul
(I), the mult’ r-authored Methods in Enzymology volume edited by Jakoby
and Pastan (Z), and the new edition of Freshney (3). Instead, this chapter’s
focus will be on the techniques for establishing primary rodent cell cultures
from embryos and adult skin, maintaining and subculturing these fibro-
blasts and their transformed derivatives, and the isolation of genetically
pure strains. The cells described are all derived from Chinese hamsters
since, to date, these cells, have proved to be the most useful for somatic cell
genetics (4,5). The techniques, however, are generally applicable to most
fibroblastic cell types.
I will only discuss growing fibroblastic cells in semidefined media. A
very detailed consideration of serum-free culture and the maintenance of
epithelial cells can be found in Chapter 21. Methods for culturing many
other nonfibroblastic cell types are described in Chapters 2 through 31.

1
2 Pollard

2. Materials
1. Alpha minimum essential medium (a-MEM): for economy, we buy
prepared medium as powder. A 44 g quantity of sodium bicarbonate
is added, the powder is made up to 20 L in deionized distilled water,
the pH adjusted to 7.4, and the media sterile-filtered through a 0.22 w
filter using a pressure vessel coupled to a filtration apparatus and
driven by a pressurized 95:5% CO, gas mix. This gas mix maintains
pH upon storage. The 500 mL bottles are stored at 4OCin the dark until
use (seeNotes 1 and 2).
2. Growth medium: a-MEM plus 15 or 7.5% (v/v) fetal calf serum. This
is made up as required and stored at 4OC.
3. Ca2+Mg2+-free Phosphate Buffered Saline (Dulbecco, PBS): NaCl, 8 g/
L; KCl, 0.2 g/L; KH,PO, 0.2 g /L; Na,HPO, l 7HZ0, 2.31 g /L pH 7.2.
4. PBS citrate: PBS + sodium citrate at 5.88 g/L.
5. Trypsin: One vial of lyophilized Difco Bacto-trypsin in 400 mL of PBS
citrate (0.125% trypsin) or 10 x this concentration for the isolation of
embryonic fibroblasts (seeNote 3).
6. Counting fluid; PBS + 0.2% (v/v) FCS.
7. Formalin fixative: 10% (v/v) commercial formaldehyde (comes as a
40% v/v solution).
8. Methylene blue stain: 0.1% (w/v> methylene blue in distilled water
filtered through a Whatman No. 1 filter.
9. Trypan blue: 0.5% (w/v) in PBS.
10. Colcemid: 10 pg/mL, stored at 4OC.
11. Karyotype fix: Methanol: acetic acid (3:l) made up on the day of use
and kept on ice in a tightly stoppered bottle.
12. Giemsa stain: Use commercial Giemsa concentrate diluted 3:47 parts
in commercial Gurr’s buffer (one tablet to 1 L distilled water). Alter-
natively, 10 mM potassium phosphate, pH 6.8, can be used as the
buffer. The diluted stain is only stable for 2-3 mo.

3. Methods
3.1. Establishment of Primary Chinese
Fibroblast CuZtures
3.1.1. Embryo Culture
1. Kill a 10-12-d pregnant Chinese hamster with ether.
2. Wash the animal in tap water and then with 70% ethanol.
Basic Cell Culture 3

3. Make a surgical incision on the dorsal side to expose the uterus using
sterile instruments (these can be dipped in ethanol and flamed to
maintain sterility during the operation).
4. Remove the uterus in toto, and transfer it to a sterile Petri dish. Dissect
out the embryos, and place them in a new sterile Petri dish (seeNote
4).
5. Mince the embryos very finely, and while still in the Petri dish, wash
the pieces with 5 mL of 0.125% Bacto-trypsin at 37OC.
6. Tilt the Petri dish so that embryo pieces go to the side. Remove the
pieces into a 50 mL centrifuge tube using a wide-bore pipet.
7. Add 40 mL of fresh 1.25% Bacto-trypsin, and incubate at 37OCfor 5
min.
8. Regain the embryo pieces by centrifugation at 100s for 3.5 min, and
discard the supernatant.
9. Resuspend the pieces in 40 mL of fresh 0.125% Bacto-trypsin, and in-
cubate at 37OCfor 25 min (this can be performed in a roller apparatus).
10. Neutralize the trypsin with 4 mL of FCS.
11. Deposit the supernatant through a 100 pm sterile mesh into another
centrifuge tube.
12. Centrifuge the supernatant for 5 min at 3008 at room temperature.
13. Resuspend the pellet in 10 mL a-MEM plus 15% FCS, and count the
cells in a hemocytometer at about l/100 dilution.
14. Lay down 1.5 x 1O’cells in 40 mL of a-MEM plus 15% FCSinto a 75 cm2,
and place it in a 37OCtissue culture incubator.
15. The next day, replace the medium with an equal volume of a-MEM
plus 15% FCS.
16. Forty-eight to72 h later, the monolayer should be confluent, and at this
point, the cells are ready for subculture. This is performed byincubat-
ing the monolayer with 4.5 mL of 0.125% Bacto-trypsin at 37OCuntil
the cells detach. Cell detachment can be visualized either by observ-
ing the cell monolayer in oblique light or directly under the micro-
scope. When the cells have detached (-go%), add 0.5 mL FCS (lo%),
pipet up and down five times, and transfer contents to a 15 mL
centrifuge tube.
17. Centrifuge the cells at 3008 for 3.5 min at room temperature.
18. Remove the supernatant, resuspend the cell pellet in 5 mL a-MEM
plus 15% FCS, and determine the cell concentration.
19. Resuspend the cells at 4 x 106/vial in a-MEM plus 15% FCS plus 10%
(v/v) sterile DMSO, and freeze at -120 or -176OC. The cells will remain
viable for several years.
4 Pollard

20. The cells may also be subcultured at one-third to one-tenth dilutions.

They have doubling times of approximately 36 h. At this point, start
to calculate the number of mean population doublings by keeping
careful records of subculture number and split ratio (seeChapter 3 for
details).
3.1.2. Skin Fibroblasts
(seeChapter 2 of this volume for human explants)
1. Kill and wash an animal as described for the isolation of embryonic
fibroblasts (points 1 to 3). In fact, it is often convenient to prepare skin
fibroblasts from the same animal as the one from which the embryos
were obtained.
2. Cut small pieces (l-2 mm2> of dermis from the exposed skin flaps
using sterile instruments and avoiding any fur.
3. Place several (5-10) small pieces (seeNote 4) into a 25 cm2 flask, and
allow them to adhere for 5 min in a very thin film of medium (0.5 mL).
4. Once adhered, add 5 mL of growth medium to the opposite surface of
the flask. Place the flask in the incubator in the upside-down position
for 24 h (the surface tension holds a thin film of medium to the upper
surface).
5. Once the explants are firmly stuck, gently invert the flask and return
to the incubator.
6. Thenext day, it is often advisable to change the medium to remove any
debris and unattached explants.
7. After several days, first “epithelial” type cells and then fibroblast will
grow out of the explants. Let this process continue until most of the
surface is covered with fibroblasts or until obvious necrosis is ob-
servedin theexplant. Itmaybenecessary tochange themediumevery
week until substantial outgrowth is observed.
8. Remove the explanted material (see Note 5) with a Pasteur pipet
attached to a vacuum line leaving the adherent fibroblasts.
9. At this stage, depending on the density, the fibroblasts can either be
trypsinized (>50% confluent) or allowed to continue to grow to form
a monolayer before they are trypsinized, subcultured, and frozen as
described above.

3.2. Maintenance and Subculture

of Transformed Cell Lines
Many transformed cell lines will grow both as monolayers and in
suspension culture. The CHO-S cell line is just such a line having been
Basic Cell Culture 5

selected for suspension growth by Thompson from the original Kl isolated

by Puck (seeNote 6, Ref. 5). Because CHO cells are transformed, they do
not require as much serum as normal diploid fibroblasts, and we routinely
culture them in 7.5% (v/v> FCS. Despite the relative ease with which
transformed cells can be cultured, however, unlike normal diploid fibro-
blasts, they do not enter a stationary phase of long-term viability (6,7). In
this phase, they rapidly lose viability, and therefore must be subcultured
during the exponential phase of growth and cannot be maintained as
arrested cultures in reduced serum.
1. CHO cells are stored frozen at -4 x lo6 cells/mL at -120 or at -176°C
(liquid nitrogen) in growth medium containing 7.5% FCS and 10%
(v/v) DMSO. A single vial is removed from the frozen stock, rapidly
defrosted in a 37°C water bath, and the cells regained by centrifuga-
tion at 3008 at room temperature.
2. The supernatant is discarded, and the pellet resuspended in 1 mL of
prewarmed medium and placed into a 25 cm2 flask or a 60mm diam-
eter dish containing 4 mL of growth medium.
3. Approximately 2 d later, the cells should be almost confluent and
ready for subculture (seeNote 7). They are trypsinized as described for
the primary diploid fibroblasts. After cell detachment, FCS is added
to 10% and the cells resuspended as single cells by pipeting up and
down about five times with a 5 mL pipet. An aliquot of this cell
suspension (up to a total of 10% of the recipient volume of the
medium) can be added directly to a new tissue culture vessel contain-
ing growth medium and returned to the incubator until the next
subculture. Alternatively, the cells may be regained by centrifugation,
resuspended, and the concentration/ml determined (see below).
Known concentrations of cells may then be subcultured by appropri-
ate dilution. In a 25 cm2 flask with 5 mL of growth medium, CHO cells
should yield about 2.5 x 105/cm2 but yields are variable depending on
serum batch and media used.
4. At this stage, cells may be transferred to a magnetically stirred spinner
flask (commercially available) containing pregassed (95% sir/5%
CO,) growth medium; usually a 250 mL spinner flask is seeded to give
a density of -8 x 104cells/mL. These flasks are then placed in a warm
room or in a temperature-regulated water bath (Heto), and stirred at
100 rpm (full details of spinner culture and scale up are described in
Chapter 6 of this volume; also see Note 8). CHO cells grown in sus-
pension should give -lo6 cells/mL at saturation density, at which
point the medium will be very yellow (acid).
6 Pollard

3.3. Determination of CeZZNumber

This can be performed either using an electronic particle counter (e.g.,
Coulter Electronics Inc.) or a hemocytometer. The former is the more ac-
curate and can be used to count low concentrations of cells (-lo3 cells/mL);
the latter requires higher density and is more prone to sampling error, but
allows a visual estimation of the “health” of the cells and, combined with
Trypan blue exclusion, can be used to estimate cell viability.
1. Resuspend cells to give a uniform cell suspension by pipeting up and
down against the side of the plastic centrifuge tube.
2. If the cells have been trypsinized, as described above, 0.2 mL of the cell
suspension to 7.8 mL of counting fluid in a 15-mL Falcon snap-cap
tube will give a statistically reliable cell count (1000-14,000 particles/
0.5 mL counted). Count three aliquots with the Coulter counter set to
count 0.5 mL, sum the three counts, divide by 3, and multiply by40 (for
dilution) and 2 to calculate the cells/mL (seeNote. 9).
3. The cells can then be appropriately diluted for the experimental setup
or subculture.
4. Alternatively, the cells can be counted on a hemocytometer. The cells
need to be resuspended at 3-5 x 105cells/mL. A drop of a cell sus-
pension is added to either side of the hemocytometer, taking care not
to overfill it and making sure that the coverslip is firmly in place.
5. Each large square on the hemocytometer (improved Neubauer type)
gives an area of 1 mm2 and a depth of 0.1 mm, i.e., the volume is 1W
mL. Count the cells in the square (usually using the one bounded on
each side with triple lines) on either side of the counter, average the
counts, and divide by 2 and multiply by lo4 to give the number of
cells/ml. If there are too many cells (>lOOO), just count the 5 diagonal
squares and multiply by 5 to give the number to be multiplied by 104.
If there are too few cells, count more than one complete square on each
side of the chamber, and divide the total cell number accordingly.
6. This procedure can also be used to determine cell viability, since prior
to placing the cells in the hemocytometer, they can be diluted 1:l with
0.1% Trypan blue. The number of cells that can then exclude the stain
(i.e., have intact cell membrane) can be determined by counting the
cells as described above.

X4. Isolation of Genetically Pure Cell Lines

Theisolationof somaticcell mutants is outside the scope of this article,
and thexeader is referred to Thompson (8) for the considerations necessary
Basic Cell Culture 7

to isolate such mutants successfully All cell lines will genetically alter over
time, however, and periodically the parental type will need to be purified
from variants or revertants. The easiestway to do this is to isolate a single
clone. This gives some potential problems, however, since a clone may it-
self be a variant, and thus several clones will need to be isolated and tested
to ensure the phenotype selected is the required one. To overcome this
problem of clonal variability, it is usually better to contract the cell popu-
lation to about 100 cells and then expand this to the mass culture. This
contraction should statistically remove any variants from the population.
It is worth remembering, however, that any variant that has a growth ad-
vantage over the parental type will soonovergrow the wholeculture. Once
a mass culture is obtained, it should be frozen in a large number of vials
(20-50) to provide a base for future experiments. This enables the investi-
gator to grow a culture for approximately 3 mo before discarding it, and
then to return to the frozen stock for the next set of experiments. This
protocol reduces the genetic drift in the culture and avoids the necessity of
frequent genetic purification using the methods described below.

1. Trypsinize a culture, recover the cells, and determine the cell number
as described above.
2. Dilute to 2.5 cells/mL with 20 mL of growth medium.
3. Plate out 0.2 mL/well into a 96-well tissue culture plate.
4. Incubate plates at 37°C in an humidified incubator for IO-12 d.
5. Examine every well with a microscope, and ring those that have a
single clone.
6. Trypsinize 2-3 of these individual clones with 0.2 mL of trypsin and,
once detached, transfer the well’s contents into 4 ml. of growth me-
dium in a snapcap tube.
7. Pipet this up and down to ensure a single ceI.t suspension, and then
plateitagainatone-tenthserial dilutions (i.e.,O.4-3.6mE)and0.2mL~
well into a 96-well tissue culture dish (seeNote IO).
8. Return these new plates to the incubators. Add medium fio’rn a di&
ferent batch to the trypsinized wells of the old plates, and also return
this to theincubator. This provides abackupincasethenewpIatesare
contaminated.
9, After lo-12 d, select individual clones in the new plates, and expand
them up to mass culture (remember to always keep abackup culture).
10. Freeze alarge stock (20-50 vials) as describedabove,sinceaatthisstage,
you will have a genetically pure line (except for the mutations that
may have occurred during the clone’s expansion). Split the frozen
8 Pollard

stock between a liquid N2 store (long-term) and a -70 or-120°C store

(short-term experimental stock).
11. Alternatively, the mass culture that needs to be genetically cleansed
can be plated into 60-mm dishes containing 5 mL growth at 100 cells/
dish.
12. Leave these to grow for approximately 10 d. Trypsinize the clones
from each plate and expand each to a mass culture in the normal way.
13. Freeze 20-50 vials of these cultures as described in item 10.

3.5. Karyotyping
It is often desirable to karyotype your cells. Full details for banding
and identifying karyotypes are given in Chapter 32 of this volume. This
chapter, therefore, will deal with a simple method, derived from Deaven
and Petersen (9) for producing karyotypes of Chinese hamster cells.
1. A culture growing in the exponential phase of growth (i.e., having a
high mitotic index) in a 10 mL suspension culture (2 x 105cells/mL)
or as a monolayer (lo6 tells/60-mm plate) is treated with colcemid at
0.06 pg/mL for 2 h to accumulate cells in mitosis.
2. For the monolayer culture, tap the plate and remove the medium con-
taining detached mitotic cells. Trypsinize the remaining monolayer,
pool with the medium, and proceed.
3. Regain cells by centrifugation at 3008 for 3.5 min at room temperature.
4. Resuspend cells in 1 mL of growth medium, add 3 mL of distilled
water, and invert to mix (do not pipet because the cells are fragile).
5. Leave for 7 min to allow the cells to swell (this can be altered if satis-
factory spreads are not obtained).
6. Add 4 mL of freshly prepared ice-cold fixative (methanohacetic acid,
3:l) directly to the hypertonic solution to avoid clumping.
7. Regain the cells by centrifugation at 300g for 3.5 min.
8. Disperse the pellet gently by agitation (do not pipet) in 10 mL of
fixative.
9. Repeat this procedure three times. At this point, the fixed cells can be
stored for a week at 4OC or slides can be made immediately.
10. Using a Pasteur pipet, drop 2-3 drops of the resuspended cells onto a
chilled slide from about 20 cm. Blow gently onto the surface, and place
the slide onto a hot plate at 60-65OC (just too hot to keep the palm of
one’s hand on the plate).
11. Leave the slide to dry for 5 min, and then place in a staining chamber
(a Coplin jar) ensuring that the surfaces do not touch’(see Note 11).
12. Stain the karyotypes with Giemsa for 3 min.
Basic Cell Culture 9

13. Wash the slides by dipping the slides through three additional Coplin
jars each containing 50 mL of water.
14. Dry the slides and count the chromosome number under the micro-
scope, or process for banding (seeChapter 32, this vol.).

3.6. Serum and Media Testing

Before a new batch of serum or media is purchased, it is advisable to
obtain a sample from the manufacturer and test its growth-supporting
characteristics. This is particularly important for serum. I usually select
two of the most used cell types in the lab-currently these are a human
diploid fibroblast strain and CHO cells-to test their growth and plating
efficiencies (seeNote 12).
1. Make up individual aliquots of growth media, all containing the same
media batch, but with the different test sera and including the serum
batch currently being used (or vice versa if you are testing media
batches).
2. Plate the cells into 15 dishes for each test media at 5 x 105cells/60 mm
tissue culture dish and containing 5 mL of the media.
3. Every day for 5 d thereafter, trypsinize the cells from triplicate plates
and determine the cell number/plate.
4. Plot a growth curve (log cell number vs time), and calculate the dou-
bling time and saturation density.
5. At the same time as setting up the growth curves, seed in triplicate 60-
mm dishes containing5 mL of the appropriate media with 100 and 200
cells (6 plates/ test).
6. After lo-12 d fix the culture for 15 min by flooding with formalin.
7. Tip the media and formalin down the drain, and stain the clones with
methylene blue.
8. Leave the stain for 15 min, and then wash it away with water.
9. Leave the plates stacked up against each other to dry in a 37OC room.
10. Count the colonies.
11. The three parameters of doubling time, saturation density, and plat-
ing efficiency should allow the section of a serum (or media) that gives
optimal growth (seeNote 13).

4. Notes
1. The shelf-life of a powdered medium is several years. Once reconsti-
tuted, however, this is reduced to 2-3 mo, mainly because glutamine
10 Pollard

is unstable. If older medium is used, the glutamine should be replen-

ished (292 pg/mL). The pH of a medium, upon storage, should not be
allowed to rise, and to achieve this, good plastic caps with close-fitting
rubber inserts should be used. I also find it useful to seal the caps with
a strip of Parafilm’, since this prevents condensation around the cap
rim and, thus, minimizes the risk of fungal contamination. Medium
containing Hepes can also be used to avoid bicarbonate buffering. I
have never been entirely happy, however, with the cell’s long-term
growth characteristics in this medium.
2. a-MEM is a rich, multipurpose medium developed by Stanners et al.
(10) to grow hamster cells. I have not had the experience of any mam-
malian cell type that will not grow in this medium, including hybrid-
omas. It is expensive, however, and many cells will tolerate less rich
and, therefore, cheaper media.
3. Purified trypsin can also be used and is sometimes necessary, e.g., for
macrophage cell lines (II), but it is much more expensive and usually
not necessary. The citrate chelates Mg2+ and Ca2+and replaces EDTA
(Versene) in the buffer.
4. It is advisable to keep fibroblast cultures from individual animals
distinct, since it may be required to distinguish between individuals
genetically.
5. If the explant is not necrotic, it is possible to remove it with sterile for-
ceps and transfer it to anew cultureflaskfor further outgrowthof cells.
6. The detailed derivation of the various CHO stains is given in Gottes-
man (5). It should be noted that CHO is a proline auxotroph and
should always be maintained in proline-containing medium.
7. CHO cells can maintain viability, providing the medium pH does not
become alkali at 4OCfor extended periods of time (7-10 d). Cultures
in capped bottles can therefore be moved to the cold room to avoid
subculture under desperate circumstances.
8. Primary cell cultures may also be grown on microcarriers in suspen-
sion culture. Full details of this technology are given in Chapter 6 of
this volume.
9. The Coulter counter should have a 140 yM aperture and the thresholds
set as described in the machine’s Instruction Manual. Serum in the
PBS prevents cells from aggregating and giving unreliable counts.
The counter sometimes gets partially blocked; only experience of the
time taken for each count and for the cell’s narticular displav on the
Basic Cell Culture 11

spectroscope will indicate problems with counting. Gentle brushing

of the orifice with a camel-hair brush will unblock the counter. The
Coulter counter can also give a visual display of cell volumes. This,
when combined with a pulse height analyzer, can be used quantita-
tively to measure cell volume or to determine cell viability by estimat-
ing the amount of cell debris in a sample.
10. To maintain genetically pure cell lines, it is absolutely essential not to
cross-contaminate cultures. To ensure this, fresh pipets must always
be used at every step. Do not re-enter a media bottle with a pipet that
has been near a culture. Similarly, never pour from a media bottle into
a culture. Splash-backs can occur. If you have more than one culture
at a time in a tissue culture hood, only one of these should be opened
at any one time. Meticulous attention to these small details will pre-
vent the cross-contamination scandals (e.g., HeLa cells in all cultures!)
that one so often reads about.
11. It is usual to prepare one slide and check it with phase contrast micro-
scopy so that adjustments can be made on subsequent slides. If there
are many nuclei without cytoplasm and a few metaphases, reduce the
swelling time. If there are many scattered chromosomes, blow less
vigorously. If the metaphase spreads are overlapping, either swell for
a longer time (up to 40 min) or blow more vigorously. All these para-
meters need to be adjusted according to the local environment condi-
tions and cell type (seeChapter 32 of this volume for greater detail and
ref. 12).
12. If many cell lines are being used, it is of ten impractical to test the serum
out on all the cell types. Usually the most difficult to grow are chosen
for the test, but caution needs to be exercised since I once had a batch
of serum that supported the cloning and growth of primary diploid
fibroblasts but failed to allow cloning of CHO cells!
13, This procedure need only be performed about once every year.
Enough serum can then be ordered for the next year, since the serum
is stable at -2O*C for at least 2 yr. We used to check our serum using
[3H]-thymidine incorporation 1 d after seeding the cells, but given the
hazard of using radioactive thymidine, we abandoned this procedure.
It is less labor-intensive, however, than measuring growth curves and
gives perfectly adequate results. Details of measuring radioactive iso-
tope incorporation into acid insoluble material may be found in
Chapter 9, Volume 1 of this series.
12 Pollard

Acknowledgments
I would like to thank Ms. D. Wylie, my tissue culture unit manager, for
all the expert help she has given over the years. I would also like to ac-
knowledge the training given to me by Dr. Cliff Stanners, many of whose
methods are represented in this chapter. This paper was written while my
research was supported by the Medical Research Council UK and the
Cancer Research Campaign UK.

References
1. Paul, J. (1975) CeEZand Tissue Culture (Churchill-Livingstone, Edinburgh and Lon-
don).
2. Jakoby, W. B. and Pastan, I. H. feds.) (1979) Methods Enzymd. VoZ 58. CeII Culture
(Academic, New York).
3. Freshney, R. I. (1987) Culture of Animal CeZZs:A Manual of Basic Techniques (Alan R.
Liss, New York).
4. Siminovitch, L. (1976) On the nature of heritable variation in cultured somatic cells.
CeZZ7, l-l 1.
5. Gottesman, M. M. (1985) Molecular Cell Biology (Wiley, New York and London).
6. Pollard, J. W. and Stanners, C. P. (1979) Characterization of cell lines showing
growth control isolated from both the wild type and a leucyl-tRNA synthetase
mutant of Chinese hamster ovary cells. J. Cell Physiol. g&571-586.
7. Stanners, C. P., Adams, M. E., Harkins, J. L., and Pollard, J. W. (1979) Transformed
cells have lost control of ribosome number through the growth cycle. 1. CeZZPhysiol.
100,127-138.
8. Thompson, L. (1979). Mutant isolation. Methods Enzymol. 58,308-322.
9. Deaven, L. L. and Petersen, D. F. (1974) Measurements of mammalian cellular DNA
and its localization in chromosomes, inMethods in CeZZBiology, vol. 8 (Prescott, D. M.,
ed.), Academic, New York and London, pp. 179-204.
10. Stanners, C. P., Eliceiri, G. L., and Green, H. (1971) Two types of ribosomesin mouse-
hamster cells. Nature (Nm BioZ.) 230,52-54.
12. Morgan,C., Pollard, J. W., and Stanley, E. R. (1987) Isolation and characterization of
a cloned growth factor dependent macrophage cell line, BACI.2F5. 1. Cell Physiol.,
130,420-427.
22. Worton, R. G. and Duff, C. (1979) Karotyping. Mefh. Enzpol. 58,321344.
 Chapter 2

Establishment, Maintenance,
and Cloning of Human
Primary Cell Strains

Gareth E. Jones
1. Introduction
My laboratory is involved in characterizing the behavior of cultured
fibroblasts established from skin biopsies of normal boys and those af-
fected with Duchenne muscular dystrophy. We are also interested in the
properties of fibroblasts obtained from female carriers of this X-linked
disease who are generally clinically unaffected (1). Following the argu-
ment of the hypothesis for random X-chromosome inactivation proposed
by Mary Lyon as a gene dosage compensation mechanism in all placental
mammals, it can be predicted that the dermis of female carriers of Duch-
enne dystrophy will be populated by fibroblasts mosaic for the Duchenne
genotype. By chance, some cells will be expressing the gene products of the
unaffected X chromosome, whereas others will have inactivated the nor-
mal X chromosome and be expressing the Duchenne gene product. In
theory, each carrier of Duchenne dystrophy should contain a roughly 1:l
ratio of normal and Duchenne-expressing fibroblasts, and techniques of
cell cloning applied to cultures of carrier biopsy material should produce
clones of cells exhibiting properties of either normal OYDuchenne cell
cultures in a similar ratio (2). In practice, things are not so simple, but the

13
14 Jones

theory outlined above has led to the development of a routine cell-cloning

method in my laboratory, which will be described here. Alternative meth-
ods do exist, which I have also tried and found to be quite successful, but
more time-consuming. One of these alternatives will be briefly outlined,
but I shall largely concentrate on the methods we use to establish a culture
of human fibroblasts from a biopsy and then to clone those cells. The
success of our cloning method is largely the result of the use of adsorbed
fibronectin as a seeding substratum; fibroblastic cells bind avidly to fibro-
nectin via a cell surface glycoprotein-receptor complex known as integrin,
resulting in high cell seeding efficiencies.
There are a few problems unique to collecting human biopsy mater-
ial. First, the tissue sample is usually taken for sound clinical reasons, such
as aiding diagnosis. The needs of the pathologist must be met first, and it
is essential that your wishes should have been communicated to the
clinician in charge well in advance of the time of biopsy. Secondly, there
is the problem of repeat sampling; multiple biopsy of one human is very
rare, since it often cannot be clinically justified. Similarly, one cannot
always be confident that the site of biopsy is consistent between patients.
For these and other reasons, it is virtually obligatory to obtain the active
participation of clinical colleagues who are well versed in the require-
ments of a prolonged research program plus the backing of the hospital
ethical committee for your project. Lack of attention to these early con-
siderations will probably render your research efforts useless.

2. Materials
1. Basal salts solution (BSS) such as Hanks’ saline (seeAppendix).
2. Growth medium containing 10% fetal calf serum (fetal serum can be
replaced by newborn for subcultures).
3. Calcium- and magnesium-free BSS (CMF).
4. 0.25% Crude trypsin in CMF.
5. For cell cloning, choose a rich medium; we use Ham’s F10 (see
Appendix) or MCDB 105 supplemented with 15% fetal calf serum.
6. Stock solution of human plasma fibronectin prepared following in-
structions of supplier, diluted to 10 pg/mL in phosphate-buffered
saline.
7. Stock solutions of cell lysis buffer: 10 mM Tris-HCl buffer (pH 7.6>,
containing 1 mM EDTA, 0.5% (v/v) NP-40,lO JJM NADP, and 1 mM
E-amino-n-caproic acid.
Human Primary Cell Strains I5

3. Methods
3.1. Collection of Biopsy
1. Provide a labeled container of medium to your clinical collaborator. A
15-mL sterile centrifuge tube with leakproof cap is ideal, about half
full of complete culture medium as given in the Materials section.
2. If it will not be possible to transfer the biopsy directly to the tissue
culture laboratory, provide a Thermos flask filled with ice. If kept at
4”C, biopsy samples will survive for periods up to 3 or 4 d, though
tissue degeneration may be expected.
3. It should not be necessary to ensure that the biopsy will be taken under
controlled sterile conditions. The site of collection will be steril-
ized(usually with 70% ethanol), perhaps following shaving to remove
hair, and many patients will ask for a local anesthetic (2-3 mL of
lidocaine or similar) (seeNote 1).

3.2. Primary Explant Culture

1. Transfer the biopsy to fresh sterile BSSin a centrifuge tube, and rinse
by gentle hand agitation for a few minutes. Transfer to a Petri dish.
2. Dissect off unwanted fat and necrotic material. Using scalpels, chop
up the material into blocks of about 2-mm cubes.
3. Wet the inside of a few wide-bore Pasteur pipets to stop subsequent
sticking of biopsy fragments to the glass. Transfer fragments using
Pasteur pipets to a 15-mL sterile centrifuge tube, and allow the pieces
to settle.
4. Wash twice in fresh BSS,aspirate off the BSS, rinse using a Pasteur
pipet and replace with fresh buffer. Allow the pieces to settle under
gravity.
5. Rinse the culture surface of a 25-cm2 culture flask with I mL of
complete growth medium. There should be just enough to cover the
whole surface with a film of liquid after a vigorous shake or two. This
will aid the attachment of the biopsy fragments to the culture surface.
6. Wet a clean Pasteur pipet. Transfer pieces of tissue to the culture flask,
placing them evenly over the whole available surface. Ideally, you
should aim to have some 20 fragments of material per flask, but less
is practicable and often inevitable with human material.
7. Cap the flask and invert the culture vessel so that tissue fragments will
be in a “hanging drop” configuration. Surface tension properties will
 Jones

retain a pool of medium around and over the explants, while gravity
will drain excess medium to the “upper” surface of the flask.
8. Transfer the flask to a tissue culture incubator set at 37OCand 5% CO,/
air mixture. A 37°C hot room is sufficient if media are HEPES-
buffered. Leave the flasks in a “hanging drop” configuration for up to
3 h. The tissue fragments should stick to the culture surface within this
time.
9. Return the culture flasks to a sterile cabinet, still inverted. Uncap and
remove excess medium (containing any explant material that failed to
adhere to the culture surface) using a Pasteur pipet. Reorientate the
culture flask, add 1-2 mL of growth medium, cap the flask, and return
to the culture incubator. Leave for 18-24 h.
10. Examine the cultures. If theexplants have adhered, makeup themedi-
urn volume to 5 mL, and then change weekly until a substantial
outgrowth of cells is seen (seeNote 2).
11. Remove the explants. These may either be picked off with a pair of
sterile fine curved (watchmakers) forceps or dislodged with suction
pressure generated down the capillary column of a fine-bore Pasteur
pipet.
12. Replace the medium in the culture flask, and monitor the outgrowth
of cells over the next few days (Fig. 1). Once some 75% of the culture
surface has been covered by cells, the culture is ready for passage (see
Notes 3 and 4).

3.3. Subculture and Maintenance (See Note 5)

1. Remove the medium, placing the culture flask on end and allowing
full drainage.
2. Add a wash of 5 mL of CMF to rinse the cells and remove traces of
serum. Remove the wash as above.
3. Add 2 mL of trypsin/CMF, and swirl the flask to cover the monolayer
with solution. Incubate in trypsin until the cells begin to round up. Do
not force the cells to detach by trituration, since this will cause cell
clumping. Gentle rocking of the flask is acceptable. It usually takes
some lo-20 min to detach 90% of the monolayer. Do not leave the cells
in trypsin longer than is necessary (seeNote 6).
4. Add 2 mL of growth medium to inactivate the trypsin. Disperse the
cells with gentle and repeated pipetting to provide a single-cell
suspension.
Human Primary Cell Strains 17

Fig. 1. A primary culture showing askin explant (dark area), rounded epithelial cells,
and stellate fibroblasts. The epithelial cells are the first to appear from the explant, but
after 5d, fibroblasts have migrated out to the substratum and are proliferating rapidly.
Magnification x 240.

5. Count the cells by hemocytometer or electronic particle counter.

Dilute to the appropriate seeding concentration by simply adding the
cell suspension to the total volume of medium required for distri-
bution to the culture flasks. This ensures that each flask will contain
the same concentration of cells (seeNote 7).
6. Cap the flasks and return to the incubator. Examine after 1 h to check
for pH changes. If needed, gas with sterile 5% CO,. Leave in an incu-
bator for 2 h. Open the caps of the flasks slightly to allow the escape
18 Jones

of air from the flasks, which will have expanded with rise in temper-
ature. Recap the flasks and leave in the incubator.
7. To determine when a medium change is required, check the culture
every day for a drop in pH. This is usually evidence of depletion of
medium by the growing culture. Examine cells under an inverted
microscope for signs of cytoplasmic vacuolation, and estimate cell
density.
8. If only feeding is required, remove and discard medium. Add an
equal volume of fresh medium that has been prewarmed to 37OC. Re-
turn the culture to the incubator (seeNote 8). Continue to feed cul-
tures until the flask culture surface is completely covered with cells (a
confluent culture). Then repeat the subculture protocol (seeNote 9).

3.4. CeZZ CZoning (See Note 10)

1. Clean glass coverslips in an acid wash (O.lN HCl for 10 min), rinse re-
peatedly in distilled water, and sterilize at 120°C for 2 h in a dry oven,
or use a microwave oven (2 min) if available.
2. Snap the coverslips to produce small shards of glass. The size will be
very variable, and you should select shards of approximately 4 mm2.
3. Place the glass fragments into wells of a 22-mm multiwell plate. Aim
to cover about 75% of the bottom surface of each well with glass.
4. Add a solution of plasma fibronectin to jus t cover the glass fragments.
Leave for 2 h at room temperature or 30 min at 37OC.
5. Prepare a cell suspension by trypsinization, following the routine
methods given for cell subculture.
6. Aspirate off the fibronectin solution (this can be reused for up to 5 h
if kept cold). Wash the wells in medium. Add the cells in completeme-
dium, plating at population densities of about 25 tells/22-mm di-
ameter plate. Keep the depth of medium to a minimum.
7. Place the culture in a 37OCincubator, and leave for 2-3 h to allow ad-
hesion and spreading of cells to fibronectin-coated glass.
8. Examine the culture with an inverted microscope. Identify those glass
shards with one cell attached. Using forceps, remove selected pieces
to 96-well plates.
9. Place a single fragment of glass into each well before adding medium
to the well. The glass will stick firmly to the underlying plastic as a re-
sult of surface tension (seeNote 11).
10. Add complete medium to the wells, again keeping the total volume to
a minimum. Place the multiwell plate in a tissue culture incubator
Human Primary Cell Strains 19

after checking again that each well contains only a single cell, and
leave for 18-24 h.
11. Examine cultures for signs of colony formation. At this early stage, it
is possible to check that each well contains a single colony. Remove
glass shards having multiple colonies using forceps (seeNote 12).
12. Attempt colony separation if feasible by snapping the glass and
replacing fragments in separate wells.
13. Allow the cells to proliferate until the base of the well is covered with
fibroblasts. Use standard trypsin-detachment procedure to isolate
cells. Seed clones into standard 25-cm2 tissue culture flasks in com-
plete medium and allow to grow up. Use the standard subculture pro-
tocol to passage cell clones (seeNotes 13 and 14).
14. Test clones for genetic contamination, i.e., a multiple-cell origin. We
clone cells from carriers of Duchenne dystrophy who are also hetero-
zygous for a second X-chromosome gene, glucose-6-phosphate dehy-
drogenase Mediterranean variant (G6PD Med) (3). GGPD phenotype
is determined according to the method of Ferraris et al. (4).
15. Trypsinize the cells, and pellet them in a conical centrifuge tube. Lyse
and sonicate the cells in cell lysis buffer.
16. Precipitate protein with acetone. Dissolve the protein precipitate with
1M NaOH containing 0.4% SDS. Assay for GGPD activity according
to Ferraris et al. (4) (seeNote 12).

4. Notes
1. This chapter deals with the culture of cells from dermal biopsy, but in
practice, fibroblastic cells can be grown from virtually any biopsy
material following this protocol.
2. You are likely to be provided with very small amounts of biopsy with
which to begin your culture. Explant culture is particularly suited to
this restriction, since there is a great risk of losing cells during enzymic
disaggregation. There are two major disadvantages to this technique.
Some tissue has very poor adhesiveness of explant to culture surface.
This is not a problem for skin or muscle biopsy. Two variations can be
tried to overcome this problem if need be. First, try to make your
explant size smaller; this is harder than it sounds and may call for the
use of a good binocular dissecting microscope. Secondly, increase the
percentage of serum in the growth medium from 10 to 20% or more to
aid the attachment of the explants to the culture surface.
 Jones

The second disadvantage relates to cell selection. The technique

will only allow culture of those cell populations with good migratory
powers; it does not provide a culture that represents themixture of cell
types to be found in the biopsy. Skin biopsy explants will show an ini-
tial migration of epithelial cells from the explants over the first few
days, these cells are then overgrown by fibroblasts (Fig. 1).
3. Always use fetal bovine serum in the culture medium of primary
explant cultures. This gives better survival figures than serum from
nonfetal sources. Some material proliferate better on culture plas-
ticware that has been precoated with denatured collagen. This can be
made in your own laboratory using rats tails as a source to “purify”
gelatin. However, I find that more consistent substrata can be made
using a solution of calf skin gelatin (such as Vitrogen), which can be
purchased from various sources. Avoid this added complication if at
all possible.
4. Despite the rarity value of good biopsy material, I always discard
explants that have been removed from the established primary cul-
ture. Many sources suggest that these explants can be reused to seed
a second round of cultures. Although this is true in theory, I feel that
problems over toxins released as a consequence of tissue necrosis
within the explant do not justify the effort.
5. Once a primary culture is subcultured, it becomes known as a cell line,
partially reflecting the heterogeneous lineages of the explant. As the
cell lines proliferate and are subcultured, a selection process will
usually occur that narrows the range of cell phenotype within the line.
For consistency between experiments and/or between biopsied
samples, it is essential to give a code to each cell line established in
your laboratory and a record of the number of subcultures per-
formed. For finite cell lines that are derived from biopsies, it has been
a routine to reduce the cell concentration at each subculture by two- or
fourfold. Under these circumstances, each passage is accompanied by
either a 1 or 2x population doubling prior to the next subculture. A
record of the number of passages and population doublings must be
kept with each cell culture. For example, a cell line will be designated
by NHSF (normal human skin fibroblast) upon the first subculture,
together with a number to indicate the source of the cell line (NHSF5
will indicate the fifth biopsy taken). We will also add a note of the
population doublings, so this will become NHSF5/1. This last num-
ber will increase by one for a split ratio of 1:2 (NHSF5/2, NHSF5/3,
Human Primary Cell Strains 21

and so on) or by a two for a split ratio of 1:4 (NHSF5/2,NHSF5/4, and

so on).
6. Many proteolytic enzymes are available in varying degrees of purity.
Crude trypsin, e.g., Difco 1:250, contain other proteases that may be
toxic to some cells. Start with the crude preparation, and move on to
purer grades only if necessary. You will also need much lower con-
centrations for shorter exposure periods than those given here for
crude trypsin.
7. Once a routine of medium change and subculture is underway, you
may wish to replace the 25-cm2 culture flasks with 75-cm2 versions. Do
this if the cells proliferate rapidly, and do it early in the life of the cell
line.
8. It is good practice to pretest a batch of serum for the capacity to
support your cell cultures and, once satisfied, to use the same batch in
all your subculture work for a particular experimental program. In
this way, you will avoid problems associated with the considerable
batch variation betweeen seemingly identical serum.
9. The subculture method given here will eventually provide a series of
replicate samples for setting up your experiment. You should not
necessarily adhere to this method when preparing a cell population
for an actual experimental assay.
10. Cloning of most primary cultures is a difficult and tedious procedure
that is only to be undertaken for strong scientific reasons. Normal
tissue cells only have a limited number of generations, and by the time
that sufficient cells have been produced from a clone, they may be
already near to senescence (seeChapter 3, this volume). Low plating
densities are needed to ensure successful isolation of clones, but this
results in low survival levels in all but a few cell lines. This method
attempts to overcome these difficulties by choosing a rich medium, by
seeding cells at high densities (some 5-6 times the recommended
concentrations) onto small fragments of easily manipulated glass, and
by pretreatment of the seeding surface with fibronectin to improve the
adhesiveness of the substratum. It is possible to use feeder layers of
irradiated cell monolayers, but I find the extra work involved offers no
improvement in the success rate over fibronectin coating.
11. I normally do not take much care over checking the orientation of the
glass fragment when it is taken to the well of the 96-well dish; the cell
can be either face up to the medium or sandwiched against the surface
of the well. Sometimes it seems that cells caught in the latter orien-
22 Jones
tation survive better than those facing up. The confines of the
microenvironment between the glass and plastic surfaces may allow
the cell to create a locally enriched environment that mimics a higher
cell density state. If you have problems with plating efficiency, it
would be worth experimenting with this observation, perhaps taking
care to place the cloning medium into the well before adding the glass
fragment in this case.
12. The visual identification of an isolated cell mass should not be taken
as conclusive evidence for its origin from a single progenitor cell. With
females, it is possible to use individuals who are heterozygous for
some X-chromosome-linked gene such as G6PD, clone their cells, and
then test individual clones for GGPD phenotype. You should either
follow this suggestion to check against genetic contamination of a
clone or find another gene that you can assay for in some way. If al2
you need is a clonal population of fibroblasts, you should use females
with a well-known allelic polymorphism of an X-chromosome gene.
13. If no cultures appear in the isolates by3 wk, it is very unlikely that they
will do so. I suggest that, if you get this result more than once, you
should consider using conditioned media in order to establish the
clones, Collect the medium from cultures of fibroblast cell lines,
centrifuge at 10,OOOgfor 30 min at 4°C to clarify, and filter the super-
natant through a 0.2 pm sterilizing filter. Add this conditioning
medium as a 40% v/v fraction to your cloning medium.
14. We have used cloning rings in combination with Petri dishes to gen-
erate clones (3). Isolated cells are plated at densities of between 25 and
50 tells/60-mm diameter Petri dish in Ham’s FlO medium and 15%
fetal calf serum. Individual clones were isolated using glass cylinders
that are dipped in silicone grease prior to pressing onto the base of the
Petri dish to encompass a growing clone and isolate it. Trypsin solu-
tion is then added by pipet to the clone, and the dissociated cells are
removed to a culture flask. Production of clones was less efficient
using this technique, and this led to the development of the method I
have described here.

References
I. Jones, G. E. and Witkowski, J. A. (1983) Membrane abnormalities in Duchenne
muscular dystrophy. 1. Neuro2.Sci.S&159-174.
2. Jones, G. E. and Witkowski, J. A. (1983) A cell surface abnormality in Duchenne
muscular dystrophy: intercellular adhesiveness of skin fibroblasts from patients
and carriers. Hum. Genet. 63,232-237.
Human Primary Cell Strains 23

3. Hillier, J.,Jones,G. E.,Statham,H. E.,Witkowski, J. A., and Dubowitz,V. (1985) Cell

surface abnormality in clones of skin fibroblasts from a carrier of Duchenne
muscular dystrophy. J. Med. Genet. 22,100-103.
4. Ferraris, A. M., Giuntini, P., Galiano, S.,and Gaetani, G. F. (1981) 2-deoxyglucose-
6-phosphate utilization in the study of glucosed-phosphate dehydrogenase mosa-
icism. Am. J. Hum. Genet. 33,307-313.
 Chapter 3

Aging of Cultured Human

Skin Fibroblasts

Calvin B. Harley
1. Introduction
There is substantial evidence that aging is related to the finite ability
of somatic cells to divide and repair damaged tissue. Since the seminal
observation of Hayflick andMoorhead (1) that cultured human fibroblasts
have a finite replicative lifespan, a great deal of basic biological research on
aging has been based on the model of cellular senescence in vitro (2).
There are three ways in which cultured cells can be used in geronto-
logical research. First, cells can be isolated from donors of various ages or,
in longitudinal studies, from one donor at various times. Molecular and
cellular features of the cells can then be compared at corresponding early
times in culture. Differences are related to the age of the donor. Similarly,
a second method of investigation involves comparison of cell cultures from
individuals with accelerated aging syndromes such as progeria or Werner’s
syndrome to cultures at corresponding passages in vitro from aged-
matched normal donors. These types of comparisons do not make use of
the in vitro senescence of cells. Thus, in the third model, cultures initiated
from normal donors are followed as a function of cell generations in vitro,
and comparisons are made between early and late passages of the culture
lifespan. Differences seen as a function of in vitro aging may be related to
cellular aging within the organism. Each of these methods has advantages
and limitations; it is probably best to use more than one (3).
25
26 Harley

Studies on cellular senescence in vitro demand careful attention to the

establishment of the primary culture and ongoing estimation of popu-
lation doublings. These are the points that will be emphasized in this
chapter. For example, periodically throughout the lifespanof a culture, the
number of viable cells used as the inoculum (NJ, the fraction that attaches
(f,>, the fraction of attached cells that divide (f,), and the cell number at con-
fluence (NC) should be measured (4). These numbers determine the incre
mental cell doublings within one passage, provided there are not dramatic
changes in other parameters such as cell size and interdivision time:
Doublings = logJ@( - (~-fd)f,N&)/f~dNol (1)
In practice, however, the fraction of normal skin fibroblasts that attach and
divide is close to 1 for most of the lifespan of the culture, and these effects
are ignored or assumed in the term MPD (mean population doublings):
MPD = 10g,(N,lNJ = log,0 /split ratio) (2)
Thus, if cells are split at confluence at a 1:4 or 1:8 ratio, for example, the
number of mean population doublings to refill the dish is simply 2 or 3, re-
spectively. It should be realized that this estimate of cell generations can
be substantially in error, especially toward the end of the culture when the
fraction of cells that attaches and divides becomes significantly less than
one, and the number of cells at confluence declines.
Other chapters in this volume provide greater detail on maintenance
of cell strains and special techniques relevant to specific cell types. Cells
other than fibroblasts have in fact been used in studies of aging in culture
(2). Themeth o d s outlined below for careful determination of mean popu-
lation doublings can be applied to any cell growing on a solid matrix.

2. Materials
1. Regular growthmedium (RGM): a-minimal essential medium (GIBCO)
supplemented with 15% fetal calf serum. Store the components as in-
dicated by suppliers. Prepare bottles of RGM as needed, and keep at
4°C (up to 2 mo) except during use, when it should be prewarmed to
37OC. It is important not to change lots of serum. Therefore, order a
small amount of two or three lots, asking for an appropriate number
of bottles of each to be kept on hold. Test for optimal growth of cells
at high and low cell density, and then place a large order for the best
batch (Notes 1,2).
Cellular Aging In Vitro 27

2. Ca2+-and Me-free phosphate buffered saline (PBS): 0.14M NaCl, 2.7

mM KCl, 1.5 mM KHJ?O,, 8.1 mM Na,Hl?O,.
3. Trypsin: Crude Difco or Sigma trypsin at 0.125% (w/v> in PBS. Pre-
pare 5-10 mL aliquots, and store at -2OOC. Thaw as needed. Unused
portions of the aliquot may be refrozen and used once or twice more
(Note 3).

3. Methods
3.1. Primary Culture
1. A standard 2-4 mm punch biopsy of epidermis plus dermis from the
abdominal or inner forearm surface is taken sterily by a qualified per-
son (Note 4).
2. Place the tissue piece into a loo-mm Petri dish containing a small
amount of RGM. Cut it in two, and place one-half in a vial containing
10 mL of RGM. Place this vial at 4OCas a backup sample. It will remain
viable for many days.
3. Dice the remaining half into pieces about 1 mm3 or less in size, and
place three pieces dermis-side down (if possible) in each of several 35-
mm Petri dishes. Use a siliconized Pasteur pipet to transfer the tissue.
Place a sterile 25-mm coverslip over the skin pieces as shown in Fig. 1,
add 2 mL RGM, and incubate at 37OC.
4. Monitor the explants for cell growth from around the edges of the tis-
sue every day or two. Refeed weekly. The first cells to appear are the
irregular, closely packed keratinocytes, which terminally differenti-
ate and die inRGM. Spindle-shaped fibroblasts appear within several
days.
5. When a dish has a total of about 1 cm20f fibroblast outgrowth (2-4 wk),
loosen and invert the coverslip, aspirate the RGM, rinse well twice
with PBS, and add 0.3 mL trypsin. Tilt the dish several times until
trypsin covers all cells. Set at 37OCfor 10 min or until cells start to de-
tach from the dish and/or coverslip. It may be necessary to give the
dish one or two gentle shakes during the digestion period.
6. Add 1.5 mL RGM, pipet up and down to suspend cells, remove the
coverslip and tissue debris, and count an aliquot of the cell suspen-
sion. Pool cells from several dishes.
7. Estimate the mean population doublings (MPD) in the primary cul-
ture as shown in Table 1.
28 Harley

- 25mm
coverslip

silicone
grease

tissue

Fig, 1. Establishing a primary culture. Tissue fragments are transferred with a silicon-
ized pipet containing a drop of medium into a dry 35mm dish. A dab of sterile silicone
grease (e.g., high vacuum lubricant) holds a 25-mm coverslip over the tissue fragments.
Medium is then carefully added to the dish. The tissue fragments anchor themselves to
the dish and/or coverslip, and cells begin to grow out onto these surfaces usually within
l-2 wk.

8. Place the remaining cells into fresh 35- or 60-mm dishes. Record on the
lid the date, strain designation, and expected MPD at confluence. The
latter is obtained by adding to the MPD of the primary culture 3.3*log
(NC/No), where Ncis th e expected number of cells at confluence (Table
2) and IV0 is the initial number of cells (Note 5).

3.2. Secondary Culture and Subsequent Passages

1. When the culture just reaches confluence, aspirate the medium, rinse
once with PBS, add trypsin (Table 2), and incubate at 37°C for 10 min
Cellular Aging In Vitro

Table 1
Estimated MPD During Primary Culture*
Cell number/dish MPD
30,000 6
60,000 7
120,000 8
250,000 9
500,000 10
*This estimation assumesthat about 500 cells initiate outgrowths from the tissue frag-
ments in one 35-mm dish. Some heterogeneity in population doubling is introduced into
the culture because of density-dependent inhibiton of growth (4).

Table 2
Tissue Culture Dishes
35 mm 60 mm 100 mm
Surface area (cm21 8 21 55
Cells at confluence (NJ 5x105 1.3 x lo6 3x106
Trypsin (mL) 0.3 0.6 1.0
RGM (mL) 1.7 3.4 9.0

or until cells loosen from the plate. It is often necessary to shake or rap
the dish gently.
2. Add RGM (Table 2) and pipet up and down until all cells are freed
from the dish and a single-cell suspension is achieved. Count an
aliquot of this suspension, and inoculate a fraction of it into a fresh dish
(or dishes) containing an appropriate volume of RGM. The inoculum
should represent l/8 to l/2 the number of cells required to form a
monolayer of cells in the new dish. Shake the dishes gently in an ir-
regular fashion to ensure a homogeneous dispersion of cells. Record
date, strain designation, and expected MPD at confluence.
3. Refeed the culture at least once per week until confluence is again
reached.

3.3. Senescence (Phase II.

1. Visually monitor cells at each passage and estimate growth rate by
days to confluence from a standard inoculum size (e.g. l/8 of final
confluent cell number). The culture is approaching its terminal
passage when growth rate drops and cells become larger and irregu-
lar in shape (Fig. 2).
30 Harley

Fig. 2. Cultured human fibroblasts from an adult forearm skin biopsy. The clean, spin-
dle appearance of cells aligned in parallel or spiral arrays at confluence during early pas-
sage (A) is gradually replaced in late passageby nonaligned, irregular-shaped cells con-
taining opaque degradation products (B). Early-passage cultures contain numerous
mitotic cells (A, arrow) during log phase and early confluence, whereas terminal passage
cultures have essentially no dividing cells and, thus, fail to reach confluence even after
numerous weekly refeedings.
Cellular Aging In Vitro 31

2. Senescence, also known as MI’&, Phase III, or terminal passage, is

the point at which one MPD takes longer than one week. For example,
if a culture split at a 1:8 ratio is not confluent in 3 wk, it may be consid-
ered terminal.

4. Notes
1. Do not use dialyzed or heat-treated serum for routine culturing of nor-
mal human fibroblasts. However, fibroblasts are fairly tolerant of
media changes for short-term culture, and it is often necessary to use
defined or special media lacking certain components for biochemical
studies. To ensure that cells are not adversely affected, conduct a
growth curve in regular vs experimental medium, or, if less than one
population doubling is involved, assay the rate of protein or DNA
synthesis. For incorporation assays, it is important that the specific
activity of the precursor be the same in each medium tested.
2. Avoid the use of antibiotics and antimycotics, since they may have un-
expected effects. They are not needed if sterile technique is diligently
practiced.
3. Crude preparations of trypsin may contain other proteases useful in
dissociating cells. However, they also vary in activity. Therefore, ad-
just the concentration up (or down) by a factor of 2 if cells take more
than 15 min (or less than 5 min) to detach from the dish.
4. Fibroblast cultures can be initiated from biopsies taken almost any
where on the body. Considerations include sun exposure, environ-
mental variation, nature of the study, and compliance from donor to
donor. Abdominal and the non-sun-exposed forearm surfaces are
commonly used, as are foreskins and other surgical pieces of skin.
5. If cultures or ampules of cells are supplied, one must estimate the
previous mean population doubling of the culture based on data from
the supplier. Assume 9 MPD for the primary culture and 1-3 MPD for
each passage depending on whether cells were previously split at 1:2,
1:4, or 1:8 ratios. Upon arrival, subculture the cells at a split ratio of 1:2
into fresh dishes and add an appropriate number of MPD to this esti-
mate based on how many cells appear to have attached after 6-12 h.
For example, if about 1/2of the cells attached, count this initial 1:2 split
as 2 MPD.
6. Biopsy fragments can be disaggregated with proteases, and cell sus-
pensions rather than outgrowths are then used to establish the pri-
mary culture. However, this often requires extensive proteolytic
32 Harley

treatment to liberate cells, and it is difficult to estimate the number of

viable founder cells of the fibroblast culture. The advantage of this is
that it eliminates the heterogeneity in doublings within the primary
cell population, which arises from density-dependent arrest of cells in
the outgrowth technique (4). However, this effect is reduced in the
technique described here by harvesting the primary culture before
outgrowths become large and heterogeneous.
7. Values of NC (Table 2) reflect an average value for human forearm skin
fibroblasts. Each laboratory should determine N, and N, periodically
throughout the lifespan of each strain studied and use these values for
MPD calculations, which are not based on valid “split ratio” methods.
8. In reporting comparisons of early and late passage cells, it is useful to
report culture age as “percent lifespan completed,” i.e., MPD/MPDmax ,
together with the value of MPDmax for the culture.
9. Various criteria have been used to define senescence, or Phase III. It
is wise to monitor the fate of the terminal passage cells to determine
if additional MPD occur. However, we have not been able to further
passage cells that fail to reach confluence at a 1:8 split ratio within 3 wk.
10. Other methodology relevant to the culture of human skin fibroblasts
has recently been described (5).

Acknowledgments
This article was prepared while the author’s work was supported by
MRC (Canada) and the Natural Sciences and Engineering Research Coun-
cil (Canada). I would like to thank Sam Goldstein and Elena Moerman for
introducing me to tissue culture techniques.

References
1. Hayflick,L.andMoorhead,P.S.(1961)Thelimitedinvitrolifetimeofhumandiploid
cell strains. EXQ. Cell Res. 25,585-621.
2. Stanulis-Praeger, B. M. (1987) Cellular senescence revisited: A review. Mech. Aging
Devel. 38,1-48.
3. Harley, C. B., Pollard, J. W., Chamberlain, J. W., Stanners, C. I’., and Goldstein, S.
(1980)Proteinsyntheticerrorsdonot increaseduringagingof culturedhumanfibro-
blasts. Proc. Natl. Acad. Sci. 77,1885-1889.
4. Harley, C. B. and Goldstein, S. (1978) Cultured human fibroblasts: Distribution of
cell generations and a critical limit. 1. CeEI. Physiol. 97,509-515.
5. Moerman, E. J. and Goldstein, S. (1986) Culture of human skin fibroblasts, Methods
in Diabetes Research, Vol. II (Clarke, W. L., Lamer, J., and Pohl, S., eds.), Wiley and
Sons, New York, pp. 283-312.
 Chapter 4

Separation and Maintenance

of Primary T and B
Lymphocytes

Derek Kinchington
and Eleanor Berrie
I. Introduction
Two distinct populations of lymphocytes have been identified: T
lymphocytes, which are thymus-dependent, and B cells, first observed in
the Bursa Fabricus of birds. Mammals do not have an equivalent structure,
and there are varying opinions as to the similarity of these cells between
species. In humans, current theories are that B lymphocytes differentiate
in the fetal liver and in the bone marrow of adults. Human T and B cells are
most easily obtained either from peripheral blood or from biopsy of
lymphoid tissues (lymph nodes, spleen, Peyer’s patches from gut, tonsils,
and adenoids).
To establish T and B lymphocytes from clinical material requires three
steps: separation from blood or other tissues, enrichment for either B or T
cells, and finally the maintenance of the primary cultures. It may be neces-
sary to carry out several enrichment steps to get greater than 90% purity.
These methods will be described in turn.

33
34 Kinchington and Berrie

2. Materials and Equipment

1. 2-aminotrylisotriouronium bromide hydrobromide (AET) treatment
of Sheep Red Blood Cells (SRBC): Dissolve 402 mg AET in 10 mL dis-
tilled water and adjust the pH to 9 with 1M NaOH. Do not overshoot.
Mix one packed volume washed SRBC in 4 vol AET, incubate at 37°C
for 25 min, and mix regularly. Wash five times in saline, and then
resuspend in RPM1 and Hepes buffer. This will keep for 1 wk at 4°C.
2. B Cell Growth Factor (BCGF).
3. Clostridium perfringens neuraminidase.
4. EBV suspension, 5 x lo4 Transforming Units (TFU).
5. F(ab’), fragment of goat anti-human IgM (p-chain specific).
6. Ficoll-paque.
7. RPM1 growth medium contains 5,10, or 15% fetal calf serum, 2 mM L-
glutamine, 100 pg/mL penicillin, and 100 mg/mL streptomycin.
8. Growth medium containing 10% FCS, and 2-Mercaptoethanol (2 x
lo4 M).
9. Hanks’ balanced salt solution pH 4 (seeAppendix).
10. Helix pomatia Lectin-Sepharose 6MB.
11. Heparin.
12. IgG: antihuman.
13. IgG: Pan T or Pan B.
14. Interleukin 2,10-40 U/mL final cont. (seeNote 3).
15. Nylon wool: sterile.
16. Pharmacia Column Kg/15 fitted with an BO-pm mesh net.
17. Phosphate Buffered Saline (Dulbecco A) (PBSA).
18. PBSA containing 0.2% human serum albumin and 0.02% sodium
azide.
19. Phytohemagglutinin (PHA) (1 mg/mL stock).
20. Protein A-Sepharose 6MB( Pharmacia).
21. Trypan blue 0.1%.

3. Methods
3.1. Separation of hmphocytes
from Blood Using Ficoll
Commercial products such as Ficoll-paque mixtures are available to
separate human mononuclear cell populations contaminated with RBC
and granulocytes.
1. Collect blood into tubes containing 10 l.tg/mL preservative free hepa-
rin.
Primary T and B Lymphocytes 35

2. Dilute blood 1:2 in serum-free medium.

3. Layer the diluted blood carefully onto the Ficoll-paque using a Pasteur
pipet to produce a clean interface between the two layers. To obtain
the maximum yield, it is advisable to keep the proportion of blood to
Ficoll in a ratio of 1:3. If small volumes of blood are being separated,
an ll-mL centrifuge tube is used with 3 mL blood layered onto 7 mL
Ficoll-paque.
4. Centrifuge at 2000 rpm for 25 min at room temperature.
5. Collect the white opaque mononuclear fraction from the interface be-
tween the diluent and the Ficoll-paque, and add at least 5 vol of serum-
free medium. Centrifuge at 1500 rpm for 10 min. Repeat for two more
times to remove the Ficoll-paque, which can be toxic to cultured cells.
6. Count the lymphocytes using Trypan blue exclusion and culture in
RPMI containing 10% FCS (Fig. 1).

3.2. Separation of Lymphocyte Suspension

fkom Biopsied Material
1. Collect biopsy material in a sterile manner.
2. Wash off any contaminating blood using sterile phosphate-buffered
saline (PBS). If the material is grossly contaminated, antibiotics can be
added at this point.
3. Remove any surface fat, and then chop the tissue into small pieces
about 1 mm diameter. Two sterile scalpels are used for this operation,
and make sure that the tissue is cut and not torn.
4. To prepare a cell suspension, two methods can be used: (a) place the
tissue into sterile stomacher bags with a small volume of medium.
Agitate for 10 s and then collect the suspension through a sterile gauze
filter, (b) press the soft tissue with a pipet tip against a sterile sieve, and
collect the resultant cell suspension.
5. If the cell suspension is grossly contaminated with blood, this can be
separated from the lymphocytes by density centrifugation using
Ficoll-paque (seesection 2.1.).
6. Count the lymphocytes using Trypan blue exclusion and culture in
RPM1 containing 10% FCS.

3.3. T and B Cell Enrichment Methods

Using Nonimmunological Substrates
Having obtained a separate mononuclear cell fraction, it is necessary
to distinguish and separate the T and B lymphocyte populations. Tech-
niques based on different adherence properties are used.
36 Kinchington and Berrie

Fig. 1. Phase contrast photomicrograph of human lymphocytes prepared using a

Ficoll-paque gradient. The preparation also contains some cell debris and a small number
of monocytes. (Mag: 440x).

3.3.1. Rosetting
T lymphocytes spontaneously form rosettes when added to sheep
erythrocytes, which are then separated from B cells by centrifugation.
1. Count viable lymphocytes using Trypan blue exclusion, and adjust to
5 x 10” lymphocytes/ml.
2. Adjust freshly prepared ART-treated sheep erythrocytes (seeMateri-
als section) to a 2.5% v/v suspension in PBS.
3. Mix lymphocyte suspension, fetal calf serum, and sheep erythrocytes
in a 1:l:l ratio, and centrifuge at 1000 rpm for 5 min at room tempera-
ture.
4. Incubate for 1 h at 4OC.
5. Resuspend this cell mixture by gently tapping the tube; mix 50 j.tL of
Trypan blue with 50 PL of the cell suspension and pipet sample into
a 0.2~mm deep hemocytometer.
Primary T and B Lymphocytes 37

6. Count 200 lymphocytes and determine the percentage of rosettes, i.e.

cells with three or more attached erythrocytes.
7. Layer the rosetted suspension onto Ficoll-paque (3 mL:7 mL, respec-
tively, see section 2.1.) and centrifuge at 2000 rpm for 10 min.
8. Harvest the nonrosetted B cells at the interface. Rosetted T cells will
be spun down to the base of the tube. These may be obtained by
removing and washing the rosettes; the erythrocytes are lysed with
sterile distilled water or high salt.
3.3.2. Lymphocyte Fractionation Using Helix Pomatia
Lectin-Sepharose 6 MB
T cells possess a receptor for Helix pomatia, and this property is used to
bind T lymphocytes to a Sepharose column (1).
1. Treat human peripheral lymphocytes (12.5 x lo6 cells/mL) with neu-
raminidase; 5 bg/mL in Tris buffered Hanks’ solution pH 7.4 for 45
min at 37OC (2).
2. Wash the lymphocytes free of neuraminidase with PBS containing
0.2% human serum albumin and 0.02% sodium azide.
3. Pipet lo8 cells in 1.5 mL of the PBS buffer onto a 3-mL column of Helix
pomatia lectin-Sepharose 6MB equilibrated with the same PBS buffer,
and leave for 15 min at room temperature.
4. Add 80 mL of the PBS buffer to the column at a flow rate of 6-10 mL/
min; this elutes unbound cells (enriched for B cells).
5. Then add 80 mL of PBS buffer containing 0.1 mg/mL N-ace+-A-D-
galactosamine; this releases a population of weakly bound cells (mix-
ture of T and B cells).
6. Finally, add 80 mL of buffer containing 1.0 mg/mL N-acetyl-a-D-ga-
lactosamine, which elutes the enriched T-lymphocyte fraction. Total
cell recovery is about 70%.

3.4. Nylon-Wool Columns

At 37°C in the presence of serum, B cells will bind preferentially to a
nylon-wool column. Most of the T cells and “null cells” remain unbound
and pass through the column. This method is rapid and is suitable for the
separation of large numbers of cells (3).
1. Pack 600 mg of sterile nylon wool into a 20-mL plastic syringe barrel
(a lOO-mL syringe can be used for large-scale preparation), and wash
with growth medium containing 5% FCS.
2. Seal the column and incubate at 37OC for 1 h.
38 Kinchington and Berrie

3. Rinse the column with 5 mL of warm growth medium containing 5%

FCS.
4. Add 2 mL of the cell suspension dropwise to the top of the column.
Then, add 1 mL of the warm growth medium, and ensure that there are
no air bubbles in the column.
5. Seal the column and incubate at 37OC for 45 min.
6. Wash the column with 25 mL of warm growth medium. Collect the
eluent, which will consist predominantly of T lymphocytes and null
cells. The lymphocytes are concentrated by centrifugation (1000 rpm
for 10 min at 4”C), and cell viability may be determined by Trypan blue
exclusion.
Some of the B-cells adhering to the column may be recovered by
mechanical elution:
1. Wash the column with 100 mL of warm growth medium and discard
this volume.
2. Seal the syringe tip with a plastic cap.
3. Add 2 mL of warm growth medium, and squeeze the nylon wool with
blunt stainless-steel forceps.
4. Remove the plastic cap, wash with 10 mL of warm growth medium,
and retain. Replace the syringe piston and expel the remaining
growth medium.
5. Concentrate these cells by centrifugation (1000 rpm for 10 min at 4OC).
6. The number of viable lymphocytes is assessedby Trypan blue exclu-
sion. These cells will be mostly B lymphocytes, but contaminated by
some T lymphocytes and “null” cells.

3.5. T and B Cell Enrichment Using Specific

Immunological Binding Methods
Lymphoid cells are separated from each other by exploiting differ-
ences in the molecules expressed on their surfaces; for example, immuno-
globulin molecules, histocompatibility or blood group antigens, and cell
surface receptors. Immunoabsorbents consisting of a specific antibody
coupled to some form of matrix, either macrobeads (column) or plastic, are
used.
3.5.1. Enrichment Using Antihuman IgG Coated Petri Dishes
1. Pour4 mL of 1% antihuman IgG in to a sterile Petri dish, and leave for
1 h at 4°C.
2. Rinse the coated Petri dish twice with PBS.
Primary T and B Lymphocytes 39

3. Add IO8 cells (maximum) in 4 mL of growth medium containing 5%

FCS.
4. Incubate for 1 h at 4°C.
5. Gently swirl the dish several times during the incubation, but keep it
level.
6. Remove the medium containing thenonadhering Tcells using a sterile
Pasteur pipet.
7. Gently rinse the dish with growth medium containing 5% FCS and
discard.
8. Add 4 mL fresh medium, and vigorously aspirate to obtain adherent
B cells.
9. Count thedifferent lymphocytesuspensionsusingTrypanblueexclu-
sion, and prepare them for culture.
3.5.2. Separation Using a Protein A-Sepharose 6MB Column
Protein A binds to the Fc portion of IgG and, when coupled to
Sepharose 6MB, will be adsorbed by cells that have been coated with a
specific IgG antibody. The immobilized cells are then released by adding
excess soluble IgG, which displaces the antibody-coated cells from thePro-
tein A-Sepharose 6MB column (4).
1. Take lo8 cells/mL from a separated lymphocyte culture (seesections
2.1 and 2.2) and treat either with pan B or pan T antibody (20 pg/107
cells/mL) at 4OC for 30 min.
2. Wash the cells three times with RPM1 containing 10% fetal calf serum.
Spin down and resuspend in 0.25 mL of the same medium.
3. Pipet the treated cells in to a small plastic column (0.5 cm in diameter)
containing 1.5 mL Protein A-Sepharose 6 MB, close the column, and
incubate at room temperature for 20 min.
4. Add 20 mL medium at a flow rate of about 2.5 mL/min to remove the
unbound cells.
5. Then add 2 mL of dog IgG (20 mg/mL), close the column, and incubate
at 37°C for 60 min.
6. Elute the cells by adding 3 mL of dog IgG (20 mg/mL) followed by 15
mL of buffer, both at 37°C. Approximately 60% of the bound cells are
recovered.
3.6. Establishment of Primary Lymphocyte Cultures
Primary lymphocytes can be maintained by mitogens, e.g., plant
lectins phytohemaglutinin (PHA), Concanavalin A (Con A), and allogenic
transplantation antigens and feeder layers; or by the addition of specific
growth factors to culture media, e.g., IL2 (seeNote 3).
40 Kinchington and Berrie

3.6.1. Maintenance of T-Lymphocytes Using IL2

1. Dilute the enriched T-cell suspension (with growth medium contain-
ing 10% FCS) to give a density of 5 x 1O5-1 O6cells /mL. Depending on
the volume, place either 200 PL aliquots in a microtiter plate, 1 mL/
well in a 24-well plate, or 3-7 mL in a 25-cm* flask.
2. Add PHA to give a final concentration of 2 pg/mL.
3. Incubate at 37OC in a humidified 5% CO,/95% air incubator.
4. Observe cell cultures daily for cell growth and evidence of cell death.
5. After 3-4 d, count the cells, using Trypan blue exclusion, harvest cells,
and wash (centrifuge at 1000 rpm for 5 min).
6. Gently resuspend the pellet in fresh growth medium (RPM1 plus 15%
FCS) supplemented with IL2 (lo-40 U/mL; see Note 3) at a cell
concentration of 5 x l@-lo6 cells/mL, and culture but do not add PHA.
7. Continue to adjust the cell concentration every three days, adding
fresh medium (plus IL2) each time.
8. After 2 wk, the individual T cells may be cloned and separated for
functional tests.

3.7. Maintenance of B-Lymphocytes Using EBV

to lkansform CelZs (See also Chapter 5)
1. Take 2 x 1O6of enriched B-lymphocytes and pellet (1100 rpm for 10 min
at room temperature).
2. Resuspend the cells in 1 mL of EBV suspension (5 x lo4 TFU), and
incubate at 37OC for 1 h in a humidified 5% CO,/95% air incubator.
3. Wash three times with PBSA to remove unadsorbed virus, and resus-
pend in growth medium containing 10% FCS.
4. Culture in 24-well plates at a cell concentration of lo6 cells/mL of
growth medium at 37OC in a humidified 5%CO,/95% air incubator.
5. Check for proliferation after 7 d.
6. Remove medium and cells from wells, spin down, and gently resus-
pend transformed cells in fresh medium. Continue to culture in 24-
well plates.
7. These cells can be grown inlimiting dilution for B-cell characterization
with monoclonal or polyclonal antibodies or functional assays.

3.8. Maintenance of B-CeZZs

Using B-Cell Growth Factor
Recently, B cells from peripheral blood from normal donors grown in
the presence of B-cell growth factor (BCGF) and anti-IgM has led to the
establishment of a long-term cultured human B cell line (5).
Primary T and B Lymphocytes 41

I. Pipet out 1 mL enriched B cells (106cells/mL) into 24-well plates with

growth medium (plus 10% FCS) containing 20 &mL F(ab’), frag-
ment and 10% BCGF.
2. Culture cells at 37°C in a humidified 5% CO,/95% air incubator.
3. Feed cells every 3-4 d by replacement of 50% of culture-well contents
with an equal volume of fresh medium containing 10% BCGF and 20
pg/mL anti-IgM.
4. Check for cell growth and viability using Trypan blue exclusion until
cell culture is established.

4. Notes
1. Optimal ‘growth conditions should be established with T cell lines
before embarking on experiments using the valuable material just
prepared, i. e., check that the sterile lx medium will maintain T cell
lines.
2. Fetal calf serum may vary from batch to batch: set up established T or
B cell lines in culture, count cells on a daily basis by the Trypan blue
exclusion method, and check which serum gives maximal cell growth,
or rH]-thymidine may be used to measure cell division rates.
3. IL2 (T-cell growth factor) also shows variation and ideally should be
assayed on established cell lines: increased growth is monitored by
increased [3H] thymidine uptake.

References
I. Cell Affinity Chromatography: Principles and Methods, Pharmacia Booklet (1984) Whar-
macia Ltd., Midsummer Blvd., Milton Keynes, MK9 3HP, England).
2. Hellstrom, U., Dillner, M. L., Hammarstrom, S., and Perlmann, I’. (1976) The
interaction of nonmitogenic and mitogenic lectins with T-lymphocytes: association
of cellular receptor sites. Scf272d.J. Immunol. 5,45-55.
3. Hudson, L. and Hay, F. C. (1980) Practical Immunology (Blackwell Scientific Publi-
cations), pp. 212,213.
4. Ghetie, V., Mota, G., and Sjoquist, J, I. (1978) Separation of cells by affinity chroma-
tography on SPA-Sepharose 6MB. J. Immunol. Methods 21,133-141.
5. Hayao, A., Rossio, J. L., Ruscetti, F. W., Matsushima, K., and Oppenheim, J. J. (1986)
Establishment of a human B cell line that proliferates in response to B cell growth
factor. J. Immunol. Methods 90,111-123.
 Chapter 5

Establishment of
Lymphoblastoid Cell Lines

A. Doyle
1. Introduction
The ability to establish long-term B lymphocyte cultures from patients
carrying particular genetic characteristics or with the ability to secrete spe-
cific antibodies (I) is an extremely valuable technique. However, there are
several basic principles to follow in the approach to this technology. It is
the purpose of this chapter to provide all the information necessary to run
an efficient and safe Epstein-Barr Virus (EBV) transformation system.
More detailed information regarding the use of this technique for theprep-
aration of human monoclonal antibodies is given elsewhere (2).

2. Materials
1. B95-8 marmoset cell line.
2. RPM1 1640 culture medium containing lo%, 5%, and 2% FBS.
3. Culture medium (without FBS) RPM1 1640 containing lOU/mL hepa-
arin (preservative-free), plus penicillin, streptomycin, and Polymixin
B sulfate.
4. 1 mL B95.8 cell culture supernatant.

43
44 Lymphoblastoid Cell Lines

5. 10% FBS RPM1 1640 culture medium, plus 0.5% PHA Wellcome.
6. Ficoll-hypaque or lymphroprep.

3. Methods
3.1. Production of Epstein-Barr Virus Stock (3)
1. The B95.8 cell line (Note 1) is cultured routinely in 5% FBS RPM1 1640.
Routine passage consists of diluting the cells between 1 in 2 and 1 in
5 at weekly intervals. A culture in the logarithmic growth phase
would usually yield lo6 cells/ml at 37°C.
2. For preparation of virus stocks, dilute from 106/mL to 0.2 x 106/mL in
2% FBS RPM1 1640.
3. Incubate at 33°C for 2 w without any medium additions or changes.
4. Allow the cells to settle out at 4OC.
5. Centrifuge the supernatant at low speed (1508 for 5 min) in order to
clarify.
6. Pass the supernatant through a 0.2 ym filter.
7. Aliquot the filtered supernatant in 2 mL sterile plastic ampules, and
store either short-term (1 mo) at -20°C, medium-term (6mo) at -7O”C,
or preferably in a gaseous phase liquid nitrogen container at -13OOC.
8. A sample of each batch should be examined for microbiological purity
(bacteria, fungi,mycoplasma) and also tested for transforming ability.

3.2. EB-Virus Transformation:

Blood Sample Preparation
1. The heparinized blood is mixed 1:l with RPM1 1640/heparin (see
Note 2).
2. Layer the diluted blood sample onto Ficoll-hypaque at a ratio of 21,
i.e., 8 mL Ficoll-hypaque/4 mL blood, in a 15-mL conical centrifuge
tube. Care must be taken to prevent the two solutions from mixing.
3. Centrifuge at room temperature at 300g for 20 min.
4. Carefully remove the buffy coat interphase with a Pasteur pipet. This
contains the mononuclear cell fraction.
5. Add RPM1 1640/heparin to the cells to make thevolume up to 15 mL.
6. Centrifuge at 1508 for 4 min.
7. Pour off the supernatant, and resuspend the lymphocyte pellet in
RPM1 1640/heparin.
8. Repeat this procedure twice more. IMPORTANT: It is essential to
ensure that the lymphocyte pellet forms after each centrifugation step.
If not, dilute the sample further (step 5 or recentrifuge). Avoid
centrifuging too hard; otherwise platelets will sediment.
Doyle 45

9. Resuspend the mononuclear cells in 5 mL of RPM1 1640/heparin.

10. Mix 0.2 mL of cell suspension with 0.2 mL Trypan blue (0.4% in PBS),
and count the viable cells on an improved Neubauer counting cham-
ber.
11. At this point, cells can be set up immediately for transformation or
stored in liquid nitrogen (seeNote 3).

3.3. Cell-Freezing Procedure

1. Centrifuge the cell suspension at 1508 for 4 min.
2. Resuspend the pellet in the correct vol of 91% newborn calf serum
(NBCS) + 9% DMSO to achieve 5 x lo6 cells mL.
3. Freeze in a programmable freezer at -3’C/min. Alternatively, freeze
overnight in a polystyrene container in a -8OOCrefrigerator and then
remove ampules to -196OC.

3.4. Virus Transformation

Either:
1. Carefully thaw the ampule in a 37OCwater bath, and transfer contents
to 10 mL of 10% FBS RPM1 1640 to remove DMSO. Centrifuge at 1OOg
for 5 min, or
2. Centrifuge cells in RPM1 1640/heparin at 1508 for 4 min.
Then:
3. Resuspend cell pellet in 1 mL EBV (B95.8) supernatant per 5 x lo6 cells.
4. Incubate at 37OC for l-1.5h. Agitate the suspension at least once
during the incubation period.
5. Centrifuge at 1OOgfor 5 min.
6. Discard supernatant.
7. Resuspend lymphocytes in an appropriate vol of 10% FBS RPM1 1640
+ penicillin/streptomycin + 0.5% PHA (Wellcome) to give lo6 cells/
mL (seeNote 4).
8. Transfer 1 mL of cell suspension to each well of a 24-well tissue cul-
ture tray (seeNote 5). Incubate at 37OCin 95% sir/5% CO, for a mini-
mum of 5 d before changing the medium.
9. With smaller cell numbers, 0.2 mL of cell suspension can be trans-
ferred to each well in a 96-well plate (seeNote 6).
10. Cells can be transferred sequentially from a 96-well tray to a 24-well
tray to a 25-cm2 flask over the next 14-21 d. This depends upon the rate
of appearance of cell growth in the wells. The technique can be
amended if particular problems are anticipated with cells. SeeNotes
4-7.
 Lyrnphoblastoid Cell Lines

4. Notes

1. A major consideration is the source of EBV. Most laboratories cur-

rently use the persistently infected Marmoset cell line B95.8 (4) as a
source of virus, although in some cases cocultivation with the cell line
QIMR-WIL has been employed. Unfortunately, many samples of the
B95.8 cell lines distributed informally between research laboratories
are mycoplasma contaminated. It is therefore of the utmost impor-
tance that the basic starting material for the technique, the virus
source, is obtained from a reliable source, such as an established cul-
ture collection.
2. Blood samples (5 mL min.) should be transferred from the syringe im-
mediately into preservative-free heparin (final concentration, 10 U/
mL). If the samples are to be stored prior to preparation, this must be
done at room temperature. Storage at 4OC leads to rapid cell death.
The maximum period of storage is 4 d.
3. There are two approaches to EBV-transformation: either cells can be
transformed immediately or, if this is not convenient, the cells, once
separated, can be stored in liquid nitrogen until transformation. It is
recommended that all procedures should be carried out in a Class II
containment cabinet.
4. The mononuclear cell suspension prepared by Ficoll-hypaque separ-
ation contains both B + T cells. There are several approaches to the re-
moval of T cells necessary to prevent specific cytotoxic T cell killing of
B cells carrying EBV antigens (seeChapter 4, this volume for details).
The alternatives are:
a. Rosetting with sheep red blood cells to deplete the T cells from
the suspension. This has the disadvantage of reducing the B
cell yield and contaminating T cells may remain.
b, Cyclosporin-A can be added to kill T cells. This has the dis-
advantages of requiring assay before use, has carcinogenic
properties, and needs to be dissolved in ethanol for use. It may
be difficult for some laboratories to obtain supplies. It can have
certain advantages in microtechniques.
c. PHA can be added. This results in the proliferation of T cells
and has a mitogenic effect on B cells. There is no evidence of
specific T cell killing following PHA stimulation. In our labor-
atory, this has proved to be the most efficient technique avail-
able: hence, it is reproduced here. (One possible drawback is
that mitogenesis leads to differentiation of committed B cells to
 plasma cells and then the loss of the C3 receptor, and thus, the
cells are not transformed.)
5. Mouse peritoneal macrophages can be prepared as feeder-layers (at 2
x lo4 cells/well) in a 24-well plate 24 h prior to cell transformation.
6. FCS concentration can be increased to 20% with small cell vol.
7. It is important to test the cell line for mycoplasma following estab-
lishment in culture. It has been noted inour laboratory that, even with
precautions taken, i.e., pretested virus and media components, my-
coplasma can be introduced with the original blood sample. Finally,
it is advisable to test staff for antibody to EBV (VCA test) prior to
commencement of work with the virus. If immunity is not present or
is at a very low level, additional caution should be exercised in the
handling of this agent. However, over 98% of adults aged 23+ have
antibody to VCA.

References
1. Stein&, M., Klein, G., Koskimies, S., and Makel, 0. (1977) EB virus-induced B
lymphocyte cell lines producing specific antibody. Nature 269,420.
2. Crawford, D. H. (1986) Use of the virus to prepare human-derived monoclonal
antibodies, in The Epstein-BarrVirus: RecentAdz~ances (Epstein, M. A. and Achong,
B. G., eds.), William Heinemann Medical Books, London, p. 249.
3. Adams, A. (1975) EBV Production, Concentration and Purification (Ablushi, D. V.,
Aalesed, H. G., and De The, G., eds.), IARC, Lyon, France, p. 129.
4. Miller, G. and Lipman, M. (1973) Release of infectious Epstein-Barr virus by trans-
formed marmoset leucocytes. Proc.Nutl. Acud. Sci.USA 70,190.
 Chapter 6

Scale-Up of Suspension
and Anchorage-Dependent
Animal Cells

J, Bryan Griffiths
1. Introduction
In this chapter, scale-up is described in a laboratory context (lo-20 L),
but the principles and techniques employed have been successfully
adapted so that cells are now grown industrially in unit volumes of up to
8000 L for vaccine, interferon, and monoclonal antibody production. The
need to scale-up cell cultures has been expanded from the historical re-
quirement for vaccine manufacture to include not only interferon and
antibodies, but many important medical products such as tissue plasmino-
gen activator and a range of hormones and blood factors. The low produc-
tivity of animal cells, resulting from their slow growth rate and low
expression of product, plus the complexity of the growth conditions and
media, led to attempts to use recombinant bacteria to express mammalian
cell and virus proteins. However, this has proven unsuitable for many
products, mainly because of incomplete expression and contamination
with bacterial toxins, and more importance is now being put on expression
of recombinant proteins from mammalian cells. This has allowed the use
of faster growing and less fastidious cell lines, such as CHO, and amplifi-
cation of product expression by multiple copies of the gene.

49
50 Griffiths

2. Principles of Scale-Up
Animals cells are grown in two completely separate systems. For
those cells that grow individually in suspension, the range of fermentation
equipment developed for bacteria can be readily modified. This is a great
advantage, since these culture vessels are economic in terms of space, the
environment is homogeneous and can be critically controlled, and scale-up
is relatively straightforward. Many cell types, however, will only grow
when attached to a substrate or, in some cases, will only produce signifi-
cant levels of a product when grown in this mode. Scale-up of substrate
attached cells is far more difficult to achieve and has given rise to a wide
range of alternative culture systems.
Two approaches to scale-up can be taken. The first is volumetric-a
simple increase in volume while retaining the sam&cell density or process
intensity. The second method is to increase the cell density/unit vol
lo-loo-fold by means of medium perfusion techniques. Cell densities of
over 108/mL can be achieved in a variety of systems, but they are difficult
to volumetrically scale-up because of difficulties in supplying media in
great enough, and constant enough, volumes. Compromise is possible
with large-scale (100-500 L) cultures operating at just 10-20 times above
the conventional cell densities (l-3 x 106/mL) by means of special perfu-
sion devices, such as the spin filter.
The environmental factors that can most readily be controlled are pH
(and redox) and oxygen. The limiting factor in scale-up, particularly in cell
density, is usually oxygen. Surface aeration used in small cultures soon
becomes inadequate, since the volume (and therefore depth) of medium
increases. Bubbling of air/oxygen mixtures into cultures, with turbulent
stirring/agitation, is the most efficient means of effecting mass transfer.
Unfortunately, cells are fragile, compared to bacteria and only slow stir-
ring and bubbling rates can be used, which are often inadequate for
maintaining a sufficient oxygen supply. To overcome this problem, most
cultures rely on several oxygenation methods, and many ingeneous meth-
ods have been developed for this purpose (1).
Two further points should be taken into account during scale-up. The
first of these is the increased risk of contamination and the proportionally
higher costs of culture failure. The second is a question of logistics in the
preparation of medium and particularly cell inocula. It is a small matter to
harvest 10Scells and inoculate them in a good physiological state into a new
culture. An inoculum of lOlo cells takes a long time to prepare, cells can be
left for long periods in damaging conditions, and media can lose its tem-
Scale- Up of Animal Cells

perature and set pH while these handling procedures are carried out. The
objective is to keep both the process and the culture system as simple as
possible, having everything well prepared and ready, and not to be over-
ambitious with regard to scale. This will ensure that cultures are initiated
with cells in good physiological condition and reduce the risks of microbial
contamination.

3. Methods
3.1. Suspension Culture
3.1.1. Culture Vessels (Fig. 1)
The simplest means of growing cells in agitated suspension is the
spinner culture vessel. The culture pot has a magnetic bar, usually placed
a few millimeters from the bottom of the vessel, and is placed on a magnetic
stirrer. As long as the bar is able to rotate freely and the stirrer is of suf-
ficient quality to maintain constant stirring speeds, and not overheat, this
methodology works extremely well for growing most cells up to densities
of l-2 x 106/mL. These glass spinner vessels are available from a wide
rangeof suppliers (e.g., Bellco, Wheaton) in sizes from 0.2-20 L (2). Amodi-
fication of this principal for shear-sensitive cells is the spinner vessel using
surface agitation as exemplified by the BR-06 Bioreactor (Techne). Spinner
vessels are only satisfactory up to a certain size-between 2 and 10 L de-
pending upon the cell line and its required use. Above 10 L, glass vessels
become inconvenient to handle and the progression to in situ stainless steel
vessels should be considered. The other reason for change is that, with
scale-up, the need to control the culture environment and carry out special-
ized manipulations (e.g., perfusion, media changes, and so on) increases.
For this level of sophistication, a fermentation system needs to be used. The
main differences are that (a) stirring is by a direct-drive mechanism with
a motor, and (b) the vessel has a complex top that allows the inclusion of
a range of sensors, probes, feed supply lines, and sampling devices for con-
tamination-free monitoring and control. Fermenter kits (laboratory scale)
are available in the range from lto 40 L and cost in the region of $4500-
37,500 (E3,000-25,000) (with control of stirring speed, temperature, pH,
and oxygen).
Culture vessels for animal cells should have the following features:
1. Curved or domed bottom to increase mixing efficiency at the low stir-
ring speeds that are used (loo-350 rpm).
52 Griffiths

I
’ 0’5;
0

0
0

0
0
0

Fig. 1. Range of culture apparatus for suspension cells (A) magnetic bar spinner cul-
ture; (B) Techne MCS stirrer; (C) surface stirrer (Techne BR-05/06); (D) small scalefermen-
ter with marine impeller; (E) airlift fermenter; (F) vibro-fermenter (Champ).

2. Water jacket temperature control to avoid the use of immersion

heaters that give localized high temperatures. Electrically operated
silicone pads are also suitable at volumes up to 5 L-the only dis-
advantage is the reduction in visibility into the vessel.
3. Absence of baffles and other sharp protrusions that cause turbulence.
The interior is finished to a high grade of smoothness to minimize
mechanical damage and for cleanliness.
4. An aspect ratio (height to diameter) of 2:l maximum, and preferably
no more than X5:1.
Scale- Up of Animal Cells 53

5. Suitable impeller to achieve nondamaging bulk flow patterns (e.g.,

modified marine or pitched blade impellers) with top drive, so that
there are no combinations of moving parts that would grind up the
cells.
Some animal cells, including many hybridoma lines, are very sensi-
tive to the mechanical effects of stirring. For such cells, there are two al-
ternative means of mixing besides stirring.
1. Vibromixer-this is a nonrotary device using a plate that vibrates in
the vertical plane a distance of 0.1-3 mm. Conical perforations in the
plate affect the mixing (Vibro-fermenter, Chemap).
2. Airlift-medium is circulated in a low velocity bulkflow pattern by
being lifted up a central draft tube by rising air bubbles, and recircu-
lated downwards in the outer ring formed by the draft tube. This sys-
tem forms a near ideal mixing pattern and allows near-linear scale-up
to at least 1000 L. Unfortunately, the apparatus, with a 121 aspect
ratio, is very high and a 30-L fermenter needs 3-m ceiling height.
Airlift fermenters are available commercially either as complete sys-
tems (e.g., LH Fermentation ) or as disposable 570 mL U (Cellift-
Fisher Scientific).
3.1.2. Culture Procedure for Suspension Cells
1. Inoculum should be prepared from a growing suspension of cells (i.e.,
in mid to late logarithmic phase). Stationary phase cells are either
slow to start in a fresh culture or do not grow.
2. Prewarm to 37OC,and equilibrate the pH of the culture medium with
the COJair gas mix, before inoculating the cells.
3. Inoculate cells at over 1 x 105/mL. Recommended level is 2-3 x 105/
mL for many cell (hybridoma) lines.
4. Stir the culture within the range 100-300 rpm. This speed depends
upon the individual type of culture reactor. Stir at a speed sufficient
to keep the cells in homogeneous suspension. Do not use speeds that
allow cells to settle out at the bottom of the culture,
5. Monitor cell growth at least daily by taking a small sample, either
through a special sampling device or removing the vessel to a laminar
flow cabinet and using a pipet, and carry out a viable cell count (Try-
pan blue stain and a hemocytometer).
6. pH-if the culture is closed (i.e., all ports stoppered with no filters),
then the pH will fall. You should remove the culture to a cabinet and
gas the head space with air. If the culture is very acid, sterile sodium
54 Griffiths

bicarbonate (5.5%) can be added or, when the cells have settled out, re-
move 50% of the medium and replace with fresh (prewarmed) me-
dium. Return culture to stirrer. It is preferable to have inlet and outlet
filters, so that there is continuous head-space gassing, initially (i. e.,
first 24 h) with 5% CO, in air, followed by air only. Suitable filters are
nonwettable with a 0.22 pm rating.
7. After 3-4 d, the saturation density of 1-2 x 106/cells/mL should be at-
tained. Most suspension cells will then die at a rapid rate unless har-
vested or maintained with medium changes.
3.1.3. Special Procedures
1. Airlift-follow the above protocol except, instead of stirring, a gas
flow rate of approximately 5-20 cc air/L/min is used for mixing.
2. Increase cell density by perfusion. To perfuse a culture (i.e., the con-
tinuous or semicontinuous addition of fresh medium and withdrawal
of an equal volume of spent medium) means that methods of separat-
ing the cells from the medium are needed. There are basically two
techniques that can be used, spin filter and hollow fibers.
3.L3.1. Spin FiZter. The problem with most filtration techniques is
that the filter rapidly becomes blocked with cells. A spin filter, so called be-
cause it is attached to the stirrer shaft, reduces the problem of blockage,
because as it spins it produces a boundary effect on its surface that reduces
cell contact. Also, they normally have a large surface area and, thus, have
a low flow rate at any one point. A porosity of about 6-10 pm is needed,
and stainless steel mesh can be used. This is thin enough to be cut to form
a cone in which the join can be double-folded and machine-pressed (Fig. 2).
This allows a perfusion rate in the order of l-2 vol/d. This device will allow
cell densities of over 107cells/mL to be achieved-higher if the culture sys-
tem has full pH, and oxygen monitoring and control (3).
3.1.3.2. Hollow-Fiber Cartridges. Hollow-fiber units are available at
both ultrafiltration and filtration grade. For the purposes of withdrawing
medium and returning the cells to the culture in a loop outside the vessel
filtration grade (0.22 pm) is sufficient. The scheme is shown diagramma-
tically in Fig. 3. Many fiber cartridges are not steam sterilizable, but poly-
sulfone and teflon fibers are. The quantity of medium withdrawn must be
balanced by adding fresh medium from a reservoir. Using flow rates of up
to 1 vol/h, cell densities in the region of 2-5 x 107/mL can be achieved, but
oxygen is a limiting factor and an additional filtration cartridge should be
put in the external loop as an oxygenator.
Scale- Up of Animal Cells 55

machine
press

direction
A

Fig. 2. Spinfilters: (A) construction using folded interleaving edges; (B) diagrammatic
representation of a spin filter; (C) open perfusion systems; (D) closed (recirculating
perfusion system).
3.1.4. Notes-Suspension Culture
1. Perfusion-media usually becomes limiting because of acid buildup
and low oxygen, and this leaves a surplus of the other nutrients, such
as glucose and amino acids. In the open perfusion systems described
above, the medium probably only yields 3-5 x 105cells/mL. If, how-
ever, the medium is circulated in a closed loop after passing through
a reservoir in which violent agitation/aeration and addition of NaOH
occurs (possible because it is a cell-free environment), then yields of
1-2 x 106/mL are achieved in media such as Eagles’ MEM (4).
2. Media-serum is a very high cost addition to culture media, but alter-
natives such as specialized serum-free media that include a range of
growth factors usually work out as expensive, or more so. Although
cultures may have to be initiated in a complex media, as cell density
56 Griffith

Fig. 3. Hollow-fiber cartridges in perfusion loop for (A) removing spent medium (H)
and returning cells to the culture and (B) oxygenating the medium.

increases, so cells become less dependent upon serum and growth

factors. Thus, at densities above 5 x 106-lO’/mL, serum concentration
can be drastically reduced (to 2%) or even excluded. If serum-free
medium is used, cells are more susceptible to damage by stirring and
air bubbles, but this can be offset to a certain extent by adding pluronic
F68 (polyglycol) at 0.1%. (Seereview (4) for other means of cutting the
costs of media.)
3. Contamination/sterility-the larger the scale, the more expensive a
culture failure becomes. Carry out stringent quality-control proce-
dures by testing growth media several days before it is to be used for
bacterial contamination. Do not take shortcuts on the support equip-
ment, but use specialized tubing connectors and sampling devices
supplied by fermenter equipment companies. Also, do not overuse
the air filter (6-10 sterilization cycles maximum), and do not allow
them to get wet (either during autoclaving or with media or condensa-
tion).
4. Suspension culture-many cells either attach to the surfaces of the
vessel or form unwanted clumps. Media for suspension culture
should have a reduced calcium and magnesium-ion concentration
(special formulations are commercially available) because of the role
of these ions in cell attachment. Attachment to the vessel can be dis-
couraged with a pretreatment of a proprietary silicone solution (e.g.,
Repelcote, Hopkins & Williams).
Scale- Up of Animal Cells 57

3.2. Anchorage-Dependent Culture

3.2.1. Materials
Anchorage-dependent culture systems are far more difficult to scale-
up than suspension cultures because of the additional requirement of pro-
viding the extra surface area in an economical (in terms of space) way and
still maintain homogeneity throughout the system. For this reason, sus-
pension culture, in which a 1-L stirred vessel is conceptually similar to a
1000-L vessel, is always the preferred culture method. The first step in
scale-up (Fig. 4) usually involves the change from stationary flasks (avail-
able in sizes up to 200 cm2) to roller bottles (sizes up to 1750 cm’). The larger
size of roller bottle will yield in the range of 2-5 x lo* cells, and therefore,
for most purposes, a mu1 tiplici ty of rollers has to be used. The next step in
scale-up is to use roller bottles that have an increased surface area resulting
from the inclusion of glass tubing (Belco-Corbeil, Chemap Gyrogen) or
plastic spiral films (Sterlin). By this means, the surface area within a roller
bottle can be increased to 8500 cm2 (spiral film) and 15,000 cm2 (glass
tubing). An alternative to investing in specialized, and costly, roller
culture equipment is to use plastic multi-tray units (Nunc). Each tray has
a surface area of 600 cm2 and units of 6,10, and 40 (24,000 cm2) plates can
be obtained. Two systems that allow a huge unit scale-up of substrate
attached cells are immobilized beds (e.g., constructed of glass spheres) and
microcarrier culture (cells growing on 200 ym spheres that are stirred in
suspension culture apparatus). Microcarriers can provide 5000 to 50,000
cm2/L and are currently being used in commercial production systems at
the 1000-L scale (a total of 15 x lo6 cm2 surface area, which has the potential
of supporting 15 x 10” cells).
3.2.1.1. Roller Culture. Reutilizable glass or disposable plastic roller
cultures are used. The most commonly used sizes are 750-850 cm2 and
1500- 1700 cm2. Complete modular systems holding up to 48 large bottles
can be purchased either free-standing, for use in hot rooms, or within an in-
cubator cabinet.
3.2.1 .l .1. PROCEDURE. The following procedure is based on a 1500
cm2 (24 x 12 cm) roller bottle.
1. Add 200-300 mL of medium.
2. Add 1.5 x 10’ cells (observe previously listed advice on preparation of
inocula and medium).
3. Revolve the culture at 15 rph.
58 Griffiths

(i&i
v

750 - 1750 cm*

Lf 3 II=

I w
6-250X103 Cm* 104cm21L
i
9 25X103
Y
Cld L

Fig. 4. Scale-up of anchorage-dependent cultures: (a) flask; (b) roller bottle; (c) plastic
spiral (Sterlin); (d) glass tubing (Belco-Corbeil); (e) stack plates; (0 multi-tray units
(NUNC); (g) immobilized bed (glass spheres); (h) microcarrier culture. The figures define
the surface area for each culture vessel.

4. Cell growth can be observed under an inverted microscope with a

long-distance objective.
5. After 3-5 d, the cell sheet will be confluent yielding from 1.5 x 10s
(human diploid) to 5 x lo5 (heteroploid cells, e.g., HeLa) cells/cm*.
6. Pour off the medium, wash the cell sheet with prewarmed phosphate
buffered saline, and add 50 mL trypsin (0.25%). Place culture back on
roller, and allow to revolve for lo-20 min. The cells will detach and can
be harvested, diluted in fresh medium and serum, and passaged on.
This outline protocol can be considerably modified. An advantage of
this method is that the medium volume:surface area ratio can be altered
easily. Thus, after a growth phase, and when a product is to be harvested,
the medium volume can be reduced to 100 mL in order to obtain higher
product concentration.
3.2.1.2. Glass Bead Immobilized Beds (3,5). Apparatus-this type
of culture system is easily fabricated in the laboratory (Fig. 5). A suitable
Scale- Up of Animal Cells 59

n A h
aiir
I

D
Fig. 5. Glass-sphere immobilize&bed culture: (A) immobilized bed; (B) reservoir with
stirrer, pH and oxygen probes, air spar&g; (Cl pump; (D) inlet for inoculum/trypsin; (E)
water at 37OC.
cylindrical or funnel-shaped glass container is packed with borosilicate
glass spheres (minimum diameter 3 mm, optimum diameter 4 or 5 mm).
Medium is perfused by means of a peristaltic pump from a reservoir
(which ideally is a spinner flask or small fermenter), which can be moni-
tored and controlled for pH, oxygen, and so on. The productive capabili-
ties of the system are give in Table 1.
3.2.1.2.1. PROCEDURE
1. Prepare growth medium, and add to reservoir (2 mL/cm2 culture sur-
face area).
2. Equilibriate the system for temperature and pH, and circulate the
media through the packed bed (allow sufficient time for the solid glass
spheres to reach 37°C).
3. Inoculate cells (l-2 x 104/cm2) into a volume of medium equal to the
void volume of the bed (250 mL/kg 5 mm spheres).
4. Allow the cells to attach (3-8 h depending upon cell type), but the cul-
ture can safely be left overnight (16 h) at this stage.
 Griffiths

Table 1
Phvsical Characteristics of Glass Sphere Beds
Bead diameter 3mm 5mln
Surface area (cm?
Total 7400 4600
Available (70%) 5200 3200
Void medium vol (cc) 295 250
Total vol (cc) 675 625
Cell count (x 105/cm2) 0.78 2.50
(x W/kg> 4.0 8.0

5. Start medium perfusion, initially at a rate of 1 linear cm of column/l0

min, but as the culture progresses, this rate is increased to a maximum
of 5 cm/min.
6. Cell growth can only be monitored by indirect measurement, and the
glucose utilization rate is the most convenient. An alternative is
oxygen utilization rate. Growth yields should be determined for the
particular cell line and nutrient, so that an approximation of cell num-
bers can be made (e.g., glucose utilization is usually within the range
of 2-5 x lo8 cells produced/g glucose).
7. When the culture is estimated to be confluent (after 5-7 d), drain the
medium, give the bed a phosphate-buffered saline wash, and add
trypsin/versene to harvest the cells. The efficiency of cell detachment
can be increased by intermittently draining and pumping back the
trypsin solution. The bed acts as a depth filter, and to recover a high
percentage of the cells, the bed should be washed through several
times with medium after the trypsin has been drained off. (NB., 5 mm
beads allow a better drainage and cell recovery than 3 mm.)
8. Wash out the culture bed immediately with a detergent, so that cell
debris does not become fixed onto the glass beads.
This culture method is basically very simple, and the apparatus is
cheap and reutilizable. It has a large potential scale-up and has been
proven at the 100 L vol scale (30 L bed of 3 mm beads) (5). It is perhaps most
suitable for harvesting a secreted cell product over a long period of time,
rather than acting as a source of cells resulting from the difficulties of re-
moving cells from the bed.
3.2.1.3. Microcarrier Czdture. The advantage of this methodology
is that the cells, when growing on small carriers, can be treated as a suspen-
Scale- Up of Animal Cells 61

sion culture with all the advantages of large unit scale-up, homogeneity,
and easily controlled environmental conditions. The range of microcar-
riers commercially available is extensive (Table 2) (2,6), and at least one
type is suitable for all cell types, however demanding. The decision of
which one to use is influenced by whether a dried powder or already-
prepared sterile solution is preferred, the cost/cm2, whether a special
derivitized surface is needed for a particular cell, or whether one wishes to
harvest cells by dissolving the carrier (gelatin, collagen) and thus produc-
ing a higher quality cell suspension. Experience has shown that it is worth-
while to evaluate several types in small-scale cultures for each particular
cell line, since significant differences in cell yield and longevity of culture
(before cell detachment) are seen.
3.2.1.3.1. CULTURE APPARATUS. Modifications of suspension culture
vessels are used. Spinner flasks with a magnetic bar are unsuitable, but ver-
sions are available with large paddle-type impellers (Bellco) or specially
modified stirring actions (e.g., Techne MCS). Stirring rates are much
slower (20-70 rpm) than for suspension cells, and thus more efficient mix-
ing at low speeds is required. Scale-up in laboratory fermenters can also
use the large-bladed paddles, but there are several modifications of the
marine impeller available that are very efficient (e.g., SGI ascenseur).
3.2.1.3.2. PROCEDURE. Microcarrier culture is not a difficult techni-
que, but it does require more critical attention to experimental detail than
most methods and the use of the correct culture vessels. The following
procedure is based on using Cytodex 3 (Pharmacia) or Dormacell 2.3
(Pfeifer & Langen) at 3 g/L. Prepare the microcarriers according to manu-
facturer’s recommendations.
1. It is essential that the medium with microcarriers be prewarmed and
stabilized before inoculating the cells. Cell attachment to moving
spheres requires conditions to be just right. It is even more important
with this method than with previously described ones to initiate the
culture with growing (logarithmic cells) and not stationary-phase
cells, and cells that have been rapidly prepared and are in good phys-
iological condition (i.e., have not been standing in trypsin, and so on,
for extended periods).
2. Inoculate at 2 x 104/cm2 into 30-50% of the final volume. Stir at the
minimum speed to maintain homogeneity (20-30 rpm) for 4-8 h.
3. When the cells have attached (expect 70-90% plating efficiency), the
volume can be increased to the full working volume. (Stirring speed
may also have to be increased to give complete mixing.)
 Griffiths

Table 2
Commercially Available Microcarriers
Name Manufacturer Type
Acrobead Galil Derivitized 5000
Bioglas Solo Hill Eng. Glass/latex 350
Bioplas Solo Hill Eng. Polystyrene 350
Biosilon Nunc Polystyrene 255
Cytodex 1,2 Pharmacia Dextran 6000
Cytodex 3 Pharmacia Collagen 4600
Cytosphere Lux Polystyrene 250
Dormacell Pfeifer & Langen Dextran 7000
Gelibead Hazelton Lab. Gelatin 3800
Mica Muller-Lierheim Polyacrylamide b

Micarcel G Reactifs 1BF Polyacrylamide 5000

Microdex Dextran Prod. Dextran 250
Superbeads Flow Labs. Dextran 6000
Ventregel Ventrex Lab. Gelatin 4000
Ventreglas Ventrex Lab. Glass/polystyrene 300
“A guideline only as different types vary Z-5-fold in cell yield/cm2.
bPorous matrix beads.

4. A great advantage of the microcarrier system is that samples can be

readily removed and microscopically examined. Unstained prepara-
tions will show whether or not cells have attached, spread out, and
then begun to grow. Cell counts can be made by the standard nuclei-
counting procedure, which releases the stained nuclei from the at-
tached cells.
5. As the culture progresses, the stirring rate can be increased to prevent
cell-to-cell attachment, bridging microcarriers and causing clumps to
form. A maximum of 75 rpm should be arrived at.
6. In nonenvironmental controlled cultures, the media will turn acid
after 3-4 d and a partial (50-70%) media change should be carried out.
Stop stirring, allow beads to settle (10 min), siphon off spent medium,
add fresh (prewarmed) medium, and start stirring, gradually increas-
ing the rate.
7. Cells can be harvested when confluent by allowing the carriers to set-
tle out, giving a serum-free wash, allowing the carriers to settle out
again, decanting off as much of the free fluid as possible, adding tryp-
sin, and restarting stirring, but at slightly faster speed (75-125 rpm).
After 20 min, allow the beads to settle out for 2 min. Cells can either
Scale- Up of Animal Cells

be removed by decanting, or the mixture can be filtered through a

coarse sintered glass filter that allows passage of the cells, not the
microcarriers. If gelatin or collagen carriers are being used, then cells
can be released by treatment with trypsin/EDTA (which solubilizes
gelatin) or collagenase.

3.2.2. Scaling- Up
This can be achieved by increasing the culture volume, and increasing
the microcarrier concentration from the suggested 3 to 5-15 g/L. If higher
concentrations are used, then it is imperative to have a perfused system
with full environmental control. The easiest means of perfusing is the spin
filter (as described for suspension cells), but a much larger pore size can be
used (60-100 pm). This allows much faster perfusion rates to be attained
(1-2 vol/ h). Perfusion from a reservoir that is adequately gassed is an effi-
cient means of oxygenating the culture. Spin filter systems are commer-
cially available (LH Fermentation, New Brunswick).

4. Conclusion
Microcarrier is undoubtably the most efficient scale-up method for
anchorage-dependent cells currently available. It allows volumetric (to
1000 L) and density scale-up with a spin filter (to 5 x lo7 cells/ml). The
equipment mus t be specially designed for the procedure, but it can also be
used for suspension cells, or as reservoir vessels for other high process
intensity systems (e.g., hollow fiber). A disadvantage is the high substrate
cost $1800-225O/kg (E120041500/kg) and the fastidious nature of the
technique.
References
I. Spier, R. E. and Griffiths, J. B. (1983) An examination of the data and concepts ger-
mane to the oxygenation of cultured animal cells. Develop0Bid. Sfandayd.55,81-92.
2. Griffiths, J.B. (1986) Scaling-up of animal cells, in Animal Cell Culture: AP~ucficuZAp-
proach (Freshney, I., ed.), IRL Press,Oxford, pp. 33-70.
3. Griffiths, J. B., Cameron, D. R., and Looby, D. (1987) A comparison of unit process
systems for anchorage dependent cells. Develop. Bid. Sfunduyd,66,331-338.
4. Griffiths, J. B. (1986) Can cell culture medium costsbe reduced? Strategies and pos-
sibilities. Trendsin Biofechnol.4,268-272.
5. Whiteside, J.P. and Spier, R. E. (1981) The scale-up fromO.l to lOOliter of aunit pro-
cesssystembased on 3 mm diameter glass spheres for the production of four strains
of FMDV from BHK monolayer cells. Biofechnol.Bioeng.23,551-565.
 Chapter 7

Mycoplasma Detection

J. M. Mowles
1. Introduction
Mycoplasma is the generic term used by cell biologists to denote or-
ganisms belonging to the Order Mycoplasmatales, which can infect cell
cultures. Of particular interest are those organisms that belong to the
Families Mycoplasmataceae (Mycoplasma) and Acholeplasmataceae
(Acholeplasma).
Mycoplasmas differ from other prokaryotes by their lack of a cell wall.
Unlike L-forms of bacteria, which under the right environmental condi-
tions are capable of producing a cell wall, mycoplasmas are unable to
produce even precursors of bacterial cell wall polymers. Another distin-
guishing feature is their size; mycoplasmas are the smallest self-replicating
prokaryotes with coccoid forms of only 0.3 pm diameter. Their genome
size is approximately one-sixth that of Eschevichia coli.
Mycoplasma infection of cell cultures was first observed by Robinson
et al. in 1956 (I). Since then, the incidence of mycoplasma-infected cell cul-
tures has been found to vary from laboratory to laboratory. Currently, 12%
of cell lines received by the ECACC are mycoplasma infected, which may
be an uncharacteristically low figure since many lines prior to deposition
are mycoplasma screened.

65
 Mowles

The importance of detecting mycoplasmas in cell cultures should not

be underestimated. In an infected cell culture, the concentration of myco-
plasmas in the supernatant can be typically in the region of 106-lo8 myco-
plasmas/ml. In addition, mycoplasmas will cytoadsorb to the host cells.
Mycoplasmas do not necessarily manifest themselves in the manner of
most bacterial or fungal contaminants, e.g., culture turbidity or pH change.
It is therefore important that an active routine detection procedure be
adopted. However, among the effects mycoplasmas have been shown to
elicit are interference in the growth rate of cells (2), induction of morpho-
logical alterations (including cytopathology) (3), induction of chromosome
aberrations (4), influencing amino acid (5) and nucleic acid (6) metabolism,
causing membrane alterations (7), and even cell transformation (8). Just as
importantly, the various regulatory bodies (9,lO) demand that cell cultures
used for the production of reagents for diagnostic kits or cell cultures used
to produce therapeutic agents are free from mycoplasma infection.
The majority of cell culture mycoplasma infections are caused by five
species. The organisms and their natural hosts areM. arginini (bovine), M.
fmmentuns (human), M. hyorhinis (porcine), M. orale (human), and A. laid-
lawii (bovine). The genus Acholeplasma differ from other members of the
order Mycoplasmatales by having no requirement for sterol for growth.
A wide range of assay techniques are available to detect mycoplasma
infection. Where possible, at least two methods should be used to remove
the risk of false-positives and false-negatives. At the ECACC, the two
methods of choice are DNA staining and culture.

2. Materials
Use analytical grade (Analar) reagents and distilled water through-
out.

1. Carnoy’s fixative: For each preparation, mix methanol (3 mL) and

acetic acid (glacial) (1 mL). Prepare fresh as required.
2. Hoechst stainstocksolution (100mL): (Bisbenzimide,Hoechst33258).
Add 10 mg Hoechs t 33258 to 100 mL water and filter sterilize through
a 0.2 ym filter. The container should be wrapped in aluminum foil and
stored in the dark at 4°C.
3. Hoechst stain working solution (50 mL): Add 50 PL stock solution to
50 mL water. Prepare fresh as required.
4. Mountant: Citric acid (O.lM) 22.2 mL, disodium phosphate (0.2M)
27.8 mL. Autoclave the above and mix with 50 mL of glycerol. Adjust
the pH to 5.5, filter sterilize, and store at 4°C.
Mycoplasma Detection 67

5. Fluorescencemicroscopeequipped for epi-fluorescence; excitation filter

340-380nm, suppression filter 430nm.
6. Sterile tissue culture dishes (33-mm) to which sterile coverslips (22 x
22 mm No 1.5) have been added.
7. Agar Media for mycoplasma culture: Prepare the following stock sol-
utions:
a. Mycoplasma agar base* (80 mL): Add 2.8 g of agar base to 80 mL
of water, mix, and autoclave at 15 lb/in2 for 15 min. Prepare
fresh as required.
b. Pig serum (10 mL): Dispense in IO mL aliquots in universal
containers, and then heat to 56OC for 45 min. Store at -30°C.
c. Yeast extract* (10 mL): Add 7 g to 100 mL water, mix, and auto-
clave at 15 lb/in2 for 15 min. Dispense in 10 mL aliquots in uni-
versal containers. Store at 4OC.
To prepare the agar media, autoclave the agar base, cool to 50°C and
then mix with the other two constituents (warmed to 5OOC). Dispense
8 mL/5-cm diameter Petri dish, and store at 4OCin sealed plastic bags.
Complete medium should be used within 10 d.
8. Brothmedia for mycoplasma culture: Prepare the following stock sol-
utions:
a. Mycoplasma broth base* (70 mL): Add 2 g broth base to 70 mL
water, mix, and autoclave at 15 lb/in2 for 15 min.
b. Horse serum (20 mL): Dispense in 20 mL aliquots in universal
containers. Store at -30°C. Do not heat inactivate.
c. Yeast extract* (10 mL): As for agar media.
To prepare the broth media, mix the three constituents, dispense in 1.8
mL aliquots in glass vials, and store at 4OC. Complete medium may be
stored for several weeks without deterioration.
9. Dienes stain stock solution:
Methylene blue 2.05 g
Azure II 1.25 g
Maltose 10.00 g
Sodium carbonate (Na,CO,) 0.25 g
Distilled water 100 mL
To prepare the Dienes stain working solution, filter sterilize (0.2 pm>
a 3% dilution of the stock solution in distilled water.
‘Oxford Ltd., Wade Rd, Basingsfoke, Hauts, RG24 OPW UK.

3. Method
The following general principles are common to both methods em-
ployed for the detection of mycoplasma:
68 Mowles

1. Cells for testing should have completed at least two passages in anti-
biotic-free medium prior to testing, because the presence of antibiot-
ics may mask infection.
2. Since cryoprotectants may also mask infection, previously frozen cell
cultures should not be tested until at least two passages are complete.
3.1. DNA Stain
1. Harvest adherent cells using a routine method of subculture, e.g., tryp-
sinization, and resuspend in the original culture medium to a concen-
tration of approximately 5 x 105 cells/mL. Test cell lines growing in
suspensions directly from culture at approximately 5 x 105cells/ mL.
Experience in working with any given cell line should eventually re-
move the necessity for an accurate cell count. Sufficient cells should
be added to dishes so that, at the time of observation (at 1 and 3 d post-
incubation), a semiconfluent spread of cells on the coverslip is ob-
tained. Similarly, if it is considered that after a further 3 d incubation
the media would be exhausted, cells should be resuspended in a mixture
of the media in which they were growing and fresh media.
2. Add 2-3 mL of test cells to each of two tissue culture dishes contain-
ing coverslips.
3. Incubate at 36 + 1°C in a humidified 5% CO,, 95% air atmosphere for
12-24 h.
4. Remove one dish, leaving the remaining dish for a further 48 h.
5. Prior to fixing the cells, examine on an inverted microscope for bacte-
rial and fungal contamination.
6. Cells are fixed by adding approximately 2 mL Carnoy’s fixative drop-
wise to the edge of the dish, which is then left for 3 min at room tem-
perature. The addition of fixative in this way is particularly important
for suspension cultures to avoid cells being swept to one side of the
dish.
7. Decant the fixative to a waste bottle (avoiding fumes), add a further 2
mL aliquot of fixative to the dish, and leave for 3 min at room tempera-
ture.
8. Decant the fixative to a waste bottle.
9. Allow the coverslip to air dry. Invert the lid of the dish and, using for-
ceps, rest the coverslip against the lid for approximately 30 min.
10. Add 2 mL of Hoechs t stain (gloves must be worn) to the coverslip, and
leave for 5 min at room temperature. At this point, shield thecoverslip
from direct light.
11. Decant the stain to a waste bottle.
12. Add 1 drop of mountant to a labeled slide, and place the coverslip cell-
side down onto slide.
Mycoplasma Detection 69

13. Examine the slide under UV epifluorescence at 100 x magnification

with oil immersion.

3.2. Microbiological Culture

In addition to passagein antibiotic-free and cryoprotectant-freemedia,
it is important that cultures for mycoplasma detection by microbial culture
should not receive a media renewal within 3 d of testing.
1. Harvest adherent cells using a sterile swab, and resuspend in the culture
medium to a concentration of approximately 5 x 105cells/mL. Test
suspension cells direct at approximately 5 x lo5 cells/mL.
2. Inoculate an agar plate with 0.1 mL of the test sample.
3. Inoculate broth with 0.2 mL of the test sample.
4. Incubate the test plate under anaerobic conditions (Gaspack)’ at 36 +
l°C for 21 d and the test broth anaerobically at 36 & l°C.
5. Subculture the test broth (0.1 mL) onto agar plates at approximately 7
and 14 d postinoculation, and incubate anaerobically at 36 + l°C.
6. Examine all agar plates after 7,14, and 21 d incubation under 40 x or
100 x magnification using an inverted microscope for detection of
mycoplasma colonies.
3.2.1. Dienes Stain
1. Add sufficient working solution to the agar plate to cover the suspect
colonies, and observe using an inverted microscope at 40 and 100 x
magnification.
2. Alternatively, if the agar plate is to be reincubated, remove a portion
of agar containing the suspect colonies using a sterile bacteriological
loop and place on a microscope slide, with colonies facing up. Add
Dienes stain and observe. Mycoplasma colonies stain blue (seeSection
4.2., Fig. 3).

4. Notes
4.1. DNA Stain
1. The fluorochrome dye, Hoechs t 33258, binds specifically to DNA. Unin-
fected cell cultures observed under fluorescent microscopy are seen as
fluorescing nuclei against a negative background (Fig. l), whereas cul-
tures infected with mycoplasma contain fluorescing nuclei plus extra-
nuclear mycoplasmal DNA (Fig. 2).

‘GasPack (BBL Microbiology Systems, Cockeysville, MD).

 Mowles

Fig. 1. Myoplasma-negative cell line.

2. Mycoplasma may appear as filamentous forms, some of which may be

branching,indicatingacultureinlogarithmicgrowth,orascocci,which
is typical of an older mycoplasma culture. Some slide preparations
may contain extranuclear fluorescence caused by disintegrating nu-
clei, and this should not be confused with mycoplasma. Such fluor-
escence is usually not of a uniform size and is often too large for myco-
plasma. Contaminating bacteria or fungi will also stain using this
technique, but will, however, appear much larger and brighter than
mycoplasmas.
3. The main advantages of the staining method are the speed at which re-
sultsareobtainedand that theas-yet-noncultivableM. hy~~!zinisstrains
observed in cell cultures can be detected.
4. An alternative to the method described is to use an indicator cell line
onto which is inoculated the test cells. The advantage of this system
is its standardization, which enables the mycoplasma screening of se-
rum andother cell culturereagents that may beinoculatedonto thein-
dicator line. Also, positive and negative controls may be included in
each assay. However, the growth of mycoplasma cultures to use as
Mvcoda-smu Detection 71

Fig. 2. Myoplasma-infected cell line.

positive controls may prove impracticable in many laboratories. The
disadvantages of using an indicator line are the extra time and effort
required in its preparation.
ArecommendedindicatorcelllineistheVeroGreenMonkeyKidney
line, which has a high cytoplasm/nucleus ratio. The indicator cells are
added to sterile coverslips in tissue culture dishes 10-24 h prior to in-
oculation of test sample at a concentration at which a semiconfluent
monolayer is formed at the time of observation (at 1 and 3 d postinoc-
ulation of test sample). Slides are prepared in the same manner as
those using the direct method.
5. Positive and negative slides may be kept in the dark for demonstration
purposes without deterioration.
4.2. Mycoplasma Culture
1. Quality control of each new batch of media ingredients should be per-
formed prior to agar or broth preparation. It is especially important
that new batches of pig and horse serum are shown to be capable of
supporting the growth of a representative sample of species found in-
72 Mowles

fecting cell culture, e.g., any two of A. Eaidlawii,M. auginini,M.feumentans,

M. hominis, M. hyouhinis, or M. ouaEe.Type strains may be obtained
from the National Collection of Type Cultures (11), or the American
Type Culture Collection (12), or wild-type strains may be used. Stock
positive control cultures may be kept frozen at -7OOC in mycoplasma
broth.
2, Each batch of complete broth media should be quality controlled prior
to use by the addition of at least two different strains of positive control
mycoplasmas. An uninoculated broth should also be incubated as a
negative control.
3. Mycoplasma agar medium should also be tested for its ability to support
mycoplasma growth. This can be most conveniently performed at the
time of test sample inoculation. An uninoculated plate should also be
included as a negative control.
4. It is necessary to be able to distinguish “pseudocolonies” and cell ag-
gregates frommycoplasma colonies on agar. Pseudocolonies are caused
by crystal formation and may even increase in size. They can be dis-
tinguished from genuine mycoplasma colonies by using Dienes stain,
which stains mycoplasma colonies blue, but does not stain pseudo-
colonies. Most bacterial and fungal colonies also appear colorless.
Cell aggregates, however, do not increase in size and so are more
easily distinguished. Also, by using a sterile bacteriological loop, cells
may be disrupted leaving the agar surface free of aggregates, but myco-
plasma colonies, because of the nature of their growth, will leave a
central core that is embedded in the agar. Another good indicator of
genuine mycoplasma colonies is their typical “fried-egg” appearance
on agar (Fig. 3); however, this is not always seen in primary isolates.

4.3. Other Detection Methods

1. Two commercially available detection methods recently introduced
are the Myco Tect system (13) and Gen-Probe (14), both of which are
indirect methods of detection. The Myco Tect system requires coculti-
vation of tes t cells with 6-me thylpurine deoxyriboside (6 MPDR), which
is nontoxic to mammalian cells, but which mycoplasmal phosphorylase
converts to6-methylpurine and6-methylpurineriboside,bothof which
are toxic to animal cells and, therefore, destroys them. Results are
available within 34 d (for a comparison of this method with DNA
staining and culture, seeref. 15).
2. Gen-Probe is a DNA hybridization assay that requires nocultureproce-
dure and claims to aive results within an hour. The protocol reauires
Mycoplasma Detection 73

Fig. 3. Myoplasma colonies on agar.

incubation of 3H-labeled DNA probe with either cell culture superna-
tant or the cells themselves. Hydroxyapatite is utilized to separate
bound from unbound probe prior to scintillation counting. The DNA
probe used is homologous to mycoplasmal r-RNA anD, therefore, hy-
bridizes with different species of mycoplasma, but not with mammal-
ian cellular or mitochondrial r-RNAs.

References
2. Robinson, L. B., Wichelhausen, R. B., and Roizman, B. (1956) Contamination of hu-
man cell cultures by pleuropneumonia-like organisms. Science124,1147,1148.
2. McGarrity, G. J., Phillips, D., and Vaidya, A. (1980) Mycoplasmal infection of lym-
phocyte cultures: Infection with M. sdivurium. In Vitro 16,346-356.
3. Butler,M.andLeach,R.H.(1964)Amycoplasma thatinducesacidityandcytopathic
effect in tissue culture. J. Gen. Microbial. 34,285-294.
4. Aula, P. and Nichols, W. W. (1967) The cytogenctic effects of mycoplasma in human
leucocyte cutures. 1. CeII Physiol. 70,281-290.
5. Stanbridge, E. J., Hayflick, L., and Perkins, F. T. (1971) Modification of amino acid
concentrationsinduced bymycoplasmasincellculturemedium. Nafure232,242-244.
 Mozules

6. Levine, E. M., Thomas, L., McGregor, D., Hayflick, L. V., and Eagle, H. (1968) Al-
tered nucleic acid metabolism in human cell cultures infected with mycoplasma.
Proc. Natl. Acad. Sci. USA 60,583-589.
7. Wise, K. S., Cassell, G. H., and Action, R. T. (1978) Selective association of murine
T lymphoblastoid cell surface alloantigens with M. hyorhinis. Proc. Natl. Acad. Sci.
USA 7544794483.
8. MacPherson, I. and Russel, W. (19661 Transformations in hampster cells mediated
by mycoplasmas. Nature 210,1343-1345.
9. EuropeanPharmacopoeia (1980)BiologicalTests. ZndEd.,Part l,V.2.1.3Mycoplas-
mas. France: Maisonneuve, S. A.
10. Code of Federal Regulations (1986) Animals and animal products. 9. Ch. 1. Part
113.28 p. 379. Detection of mycoplasma contamination. Office of the Federal Regi-
ster National Archives and Records Administration, Washington.
11. National Collection of Type Cul tures, Central Public Heal th Laboratory, 61, Colindale
Avenue, Colindale, London NW9 5 HT.
12. American Type Culture Collection, 12301 Parklawn Drive, Rockville, Maryland
20852, USA.
13. Myco-Tect, Bethesda Research Laboratories, Inc., Gaithersburg, Maryland 20877,
USA.
Gen-Probe, Inc., 9620 Chesapeake Drive, San Diego, California 92123, USA.
:;: McGarrity, G. J., Kotani, H.,and Carson, D. (1986) Comparativestudies to determine
the efficiency of 6 methylpurine deoxyriboside to detect cell culture mycoplasmas.
In Vitro Cellular and Developmental Biology 22,301-304.
 Chapter 8

Short-Term Chorionic Villi

and Amniotic Fluid Cultures

M. L. Murer-Orlando
and V. M. McGuire
1. Introduction
Prenatal diagnosis of genetic disorders associated with specific bio-
chemical, chromosomal, or molecular characteristics can be achieved from
amniotic fluid (AF) or placenta (chorionic villus: CV) samples. Chorion
material is usually obtained by sampling the placenta at the implantation
site, during the first trimester (i. e., 9-12 wk), using either the transcervical
or transabdominal route. The first method entails access to the placenta via
the cervical canal, with aspiration through a metal cannula or a flexible
catheter. Alternatively, the chorionic villi may be aspirated using a needle,
which is passed through the abdominal wall. In both these methods, the
positioning of the aspirating device must be made with the help of ultra-
sound scanning. Later pregnancies can only be sampled by the transab-
dominal route. Amniocentesis is usually performed at 16-18 wk of gesta-
tion by the transabdominal method.
In this chapter, we will describe the techniques we currently use to
obtain chromosome preparations from CV and AF samples. Chromosome
analysis from CV samples can be performed either on spontaneously di-
viding cells of the cytotrophoblast layer (direct preparation) or on cells

75
76 Murer-Orlando and McGuire

from cultured villi (culture preparation). The origin of cells cultured and
analysed in amniotic fluid samples is difficult to establish. Avariety of cells
can be isolated from uncultured amniotic fluid, and they are considered to
derive from fetal epidermis or periderm, amnion, trophoblast, mucosa of
digestive and urinary tract, and urogenital epithelia. Most of these cells are
enucleate or dead, and even those that are viable may not adhere to the
culture surface and proliferate.

2. Materials
2.1. Chorionic ViUus Samples
2.1.1. Direct Preparation
1. A biopsy of at least 10 mg is required for chromosome analysis.
2. Hepes buffered RPM1 1640 medium is used for tissue collection and
direct preparation, with appropriate supplements as follows:
a. Wash medium: 150 pg/mL streptomycin sulfate and 150 U/
mL penicillin (standard antibiotics) and 5 U/mL heparin. Can
be stored at 4OCfor 1 wk. This is used during sampling, in the
collection container, and for washing blood-stained tissue.
b. Transport medium: Standard antibiotics, 10% fetal bovine se-
rum (FBS) and 20 mM L-glutamine. Can be stored at 4°C for 1
wk. This medium is used for transporting washed villi to the
laboratory for culture.
c. Colcemid medium: Standard antibiotics, 10% FBS, 20 mM L-
glutamine and 0.5 pg/mL Colcemid. This is used for direct
chromosome preparations and is prepared as required.
3. Colcemid solutions: Stocksolution: 1 mg/mL in sterile distilled water,
stored at 4OC. Working sohtion: 50 pg/mL in sterile distilled water
stored at 4°C.
4. Hypotonic solution: 1% sodium citrate in distilled water. Can be
stored at room temperature for one week.
5. Fixative: 1:3 glacial acetic acid ARabsolute methanol AR. Made up
as required, but not stored.
6. Acetic acid in distilled water (60%): Made as required, but not stored.
2.1.2. Culture Preparation
1. A biopsy of at least 5 mg is required to set up chorionic villi cultures.
Accurate selection of villi is essential to avoid maternal contamination
in cultures.
CVS and Amniotic Fluid

2. Wash and transport media: As in protocols of direct preparation (see

above).
3. Incubation medium: Transport medium with an addition of 2.5 pg/
mL Amphoterycin B (Fungizone). This is used for the overnight incu-
bation of tissue, prior to culture, to prevent yeast infection.
4. Medium for primary cultures: Chang’s medium (Hana Biologic Inc.
Both the basal liquid medium and the lyophilized supplement are
marked with expiration dates). Standard antibiotics, 20 mM L-Gluta-
mine, and 20 mM Hepes buffer are added to the basal medium before
the addition of reconstituted supplement. This is a costly medium, the
reconstituted supplement may be stored in appropriate aliquots at
-2OOCup to the expiration date, and thawed for dilution with advised
volumes of basal medium before use. The complete medium may be
stored at 4OC for up to 10 d.
5. Medium for secondary cultures: Secondary cultures can be main-
tained in RPM1 1640 with standard antibiotics, 20 mM L-glutamine,
and 20% FBS.
6. Colcemid solution: Stock and working solutions as in protocols of di-
rect preparations (seeabove).
7. Trypsin solutions: Stock solution: 2.5% trypsin to be aliquoted and
stored at -2OOC. Working solution: 0.16% trypsin (diluted from stock
solution, with 0.02% EDTA, in PBS without calcium and magnesium).
Made up as required, but not stored.
8. Hypotonic solution. 1:6 calf serum in sterile distilled water. Made as
required.
9. Fixative. 1:3 glacial acetic acid AR: absolute methanol AR. Made as
required, but not stored.

2.2. Amniotic Fluid Samples

1. Amniotic fluid samples can appear: clear, cloudy (mostly from the
more advanced pregnancies), brown (where there has been some
bleeding or fetal distress prior to sampling), or blood stained (because
of bleeding during sampling). Approximately lo-20 mL of fluid will
be required for culture.
2. Culture medium. Hams FlO Hepes buffered with standard antibiot-
ics and t-Glutamine, supplemented with 5% FBS and 1% Ultroser G.
serum substitute (Life Science Labs. Cat. No. 50902).
3. Colcemid, trypsin, hypotonic solutions, and fixative, prepared as in
the chorion culture protocols.
78 Murer-Orlando and McGuire

3. Methods
3.1. Chorionic Villus Samples
3.1.1. Direct Preparations (See Notes 1-5)
1. Collect sample in wash medium. Wash at least once, or more if sample
is blood stained.
2. Examine villi under a stereomicroscope, and remove maternal de-
cidua and blood clots.
3. Leave selected material at room temperature in 5 mL of colcemid me-
dium for at least 2 h (max. of 4 h).
4. Take off the colcemid medium, and add 4 mL of hypotonic solution;
leave for 20 min at room temperature.
5. Take off the hypotonic solution, and add 4 mL of fixative. Two further
changes of fixative are recommended to cover a total of 20-min fixa-
tion time.
6. Store samples overnight at -20°C in fixative.
7. Individual villi are selected for chromosome preparations. Rehydrate
eachvillus by passage through decreasing concentrations of methanol
to water. Transfer to a clean slide, and add 2-3 drops of 60% acetic
acid. Tease the tissue gently with fine forceps to hasten disaggrega-
tion, and using a Pasteur pipet, transfer the resulting cell suspension
onto two or more slides. Allow to dry on a hot plate (4OOC).
3.1.2. Culture Preparations (See Notes 6-9)
3.1.2.1. Setting Up Cultures
1. Collect, wash, and select villi as described in protocols for direct prep-
arations. Leave overnight at 37OC in incubation medium.
2. Divide the sample into 30-mm Petri dishes (=5 mg/dish), add one
drop of Chang’s medium to moisten the tissue, and mince the villi
with a scalpel.
3. Transfer the fragments to a flat-sided Nunc tube. Add approximately
0.1 mL of medium to the edge of the dish, and to the base of the tube
to maintain humidity. Leave at room temperature for 2-3 h to allow
the tissue to adhere to the plastic.
4. Slowly add approximately 2 mL of Chang’s medium to each culture.
Transfer to a 2% CO, incubator, and leave undisturbed for 3-5 d.
5. Change medium, subsequently, every 2-3 d. Harvesting is usually
possible in 8-10 d.
CVS and Amniotic Fluid 79

3.1.2.2. Harvesting
1. Add colcemid to a final concentration of 0.5 pg/mL (0.01 mL of colce-
mid working solution for mL of medium), and incubate for 2 h at 37OC.
2. Discard the medium, and gently wash the cell monolayer with 1 mL
of 0.02% EDTA at 37°C.
3. Replace with 1 mL of trypsin working solution (0.16%) at 37°C. When
the cells start to detach (30-60 s), shake the container to aid even cell
suspension.
4. Add 2 mL of serum hypotonic at 37OC. Incubate for 15 min at 37OC.
5. Transfer the cell suspension to a centrifuge tube containing 0.5 mL of
fixative, and centrifuge for 5 min at 25Og. (This prefixation step en-
sures an even cell suspension.)
6. Discard supernatant, and resuspend the cells in 4 mL of fixative. Leave
for 5 min at room temperature.
7. Centrifuge, discard supernatant, and resuspend in 2 mL of fixative.
Leave for at least 30 min at room temperature.
8. Centrifuge, discard supernatant, and resuspend pellet in approximately
0.5 mL of fixative. Make slides from the cell suspension.

3.1.2.3. Subculturing
1. Discard the medium, and gently rinse the cultures with 1 mL of 0.02%
EDTA.
2. Replace with 0.5 mL of trypsin working solution (0.16%). When the
cells are detached, add 4 mL of culture medium, and pipet vigorously
to ensure even cell suspension.
3. Transfer half the cell suspension to a new tube. Return the cultures to
the CO, incubator.
4. If necessary, change the medium the following day. The highest mito-
tic activity is usually observed in 48 h.

3.2. Amniotic Fluid Samples (See Notes 10-14)

3.2.1. Setting Up Cultures
1. Mix the sample by inverting the container several times, and pipet
equal vols (5-7 mL) into two or more flat-sided Nunc tubes.
2. Centrifuge the tubes for 15 min at 25Og, and pipet off the supernatant
fluid, whichis retained for a-fetoprotein or other enzymatic, and meta-
bolic assays should they be requested.
80 Murer-Orlando and McGuire

3. Tap the tube sharply to resuspend the cell pellet, and add 1.5 mL of cul-
ture medium to each tube.
4. Transfer to a CO, incubator, and leave undisturbed for 7 d.
5. Change medium. Subsequent medium changes should be made every
2-3 d, with the last change on the day before harvesting. Harvesting
is usually possible in 10 d.
6. Harvesting and subculturing are the same as the chorion culture pro-
tocols.

4. Notes
1. Direct preparations have some technical disadvantages. The mitotic
rates vary widely between samples, there may be a high proportion of
incompletemetaphases, and chromosomes are difficult to band. How-
ever, following the outlined procedure, preparations will show, on
average, 10 complete metaphases per slide of sufficient quality to allow
analysis of 200-300 band stage karyotype (1) in at least two cells.
2. Semidirect protocols may also be used to obtain chromosome prepara-
tions. Samples can be incubated at 37°C in transport RPMI for 1 or 2
d; however, mitoses have been recovered from intact villi incubated
for as long as 4 d.
3. To improve the chromosome morphology of direct preparations, block-
ing agents, such as FudR, can be introduced during incubation in an
attempt to synchronize the dividing cells (2).
4. Semidirect protocols have been used to study the cell cycle of the cyto-
trophoblasts by BrdU incorporation. These studies have shown that
the cell cycle is long k36 h) with an extended late S phase of more than
10 h (3).
5. The processing of individual villi allows a more accurate assessment
in cases where the chromosome mosaicism is confined to the placenta.
6. Chromosome preparations obtained following the described culture
protocolareusually of good quality, withgoodmitoticrate andchromo-
some morphology, allowing analysis of 600-800 band-stage karyo-
types (1). Other methods have been reported, mainly in situ techni-
ques using cells growing on coverslips (4) or monolayer cultures
obtained from enzymatically dissociated villi (5).
7. If CVS material is not carefully selected, growth of cells from maternal
decidua can present a serious problem in chorion cultures, especially
when working with monolayer preparations from cell suspensions. It
CVS and Amniotic Fluid

is, therefore, advisable to set up separate cultures from discrete por-

tions of the sample whenever possible.
8. In order to ensure that chromosome analysis may be repeated if neces-
sary, it is advisable to harvest the cells from the tube only if there is
growth in the corresponding dish. If not, the tube must be subcul-
tured.
9. Use of Chang’s medium is recommended to establish chorion cul-
tures. Even so, the initial cell growth will still be slow (3-4 d) in com-
parison to cultures from fetal tissues. Secondary cultures may be
maintained in RPM1 with 20% FBS, but they deteriorate after a few
passages.
10. Following thedescribedprotocol,harvestingof AFculturesforchromo-
some analysis can be performed in 8-15 d. Preparations have an average
of 10 cells/slide, analyzable at the 600-800 stage karyotype (I).
11. Culture methods used for AF cells have been reviewed in detail by
Martin (6). Cells are harvested on the growth surface, or the mono-
layer is trypsinized before being processed as a cell suspension. Results
from in situ procedures are usually faster, but the quality of the chromo-
some preparation is less controllable. In situ procedures have also
been described where poly-L-lysine has been used to coat the growth
surface (7).
12. Amniotic fluid contains very few cells that are viable in culture. Ex-
ceptionally slow cell growth may be related to extraneous factors such
as toxic syringes or culture plastic ware, long delay in transport, or to
grossly blood-stained samples. The use of Ultroser G supplement is
recommended in cultures from blood stained fluids. Cultures from
cloudy samples do not vary from those of clear AE.
13. In AF cultures, different colonies are observed, and there is a tendency
to classify the cells in three major groups: “epithelioid,” “amniotic
fluid,” and “fibroblas tic” cells. Maternal cell growth can occur in AF
cultures.
14. It is advisable to use a simple protein test (Albym test-Boehring Cata-
log number 12 62 50) in all AF samples, to exclude the possibility that
the fluid is urine. In blood stained samples, the proportion of maternal
to fetal blood contamination can be determined using an histochemi-
cal test to differentiate cells carrying adult or fetal hemoglobin (Boe-
ringer Mannheim, Tes t combination-Catalog Number 124-249). This
test is recommended when a-fetoprotein analysis needs to be carried
out on the same AF sample, since fetal serum will give a falsely raised
aFP level.
82 Murer-Orlando and McGuire

References
1. Harnden,D. G.andKlinger,H.P.,eds.(1985)AnInternationalSysfemfovHumanCyto-
geneticNomenclature.Karger: Basel,Munchen, Paris, London, New York. Published
in collaboration with Cyfogenet. Cell Genet.
2. Gibas, L. M., Grujic, S.,Barr, M. A., and Jackson,L. G. (1987) A single technique for
obtaining higher quality chromosome preparations from chorionic villus samples
using FdU synchronization. Prenat. Diag. 7,323-327.
3. Zahed, L., Murer-Orlando, M. L., and Bobrow. M. (1988) Cell cycle studies in chor-
ionic villi. Hum. Genet.80,127-134.
4. Heaton, D. E.,Czepulkowski, B.H., Horwell, D. H., and Coleman, D. V. (1984) Chro-
mosome analysis of first trimester chorionic villus biopsies prepared by maceration
techniques. Prenat. Diag. 4,279-287.
5. Blakemore, K. J,,Watson,M. S.,Samuelson,J.,Breg, W. R., andMahoney,M. J. (1984)
A method for processing first trimester chorionic villus biopsies for cytogenetic an-
alysis. Am. 1, Hum. Genef. 36,1386-1393.
6. Martin, A. D. (1980) Characteristics of amniotic fluid cells in vitro and attempts to
improve culture techniques. Clinic. Obsfet. Gynaecol. 7,143-173.
7. Rajendra, B. R., Sciorra, L. J., and Lee, M. (1981) A rapid culture harvest protocol for
amniotic cell cultures. Hum. Genet. 59,416-418.
 Chapter 9

Keratinocyte Culture

Yvonne Barlow and Richard J. Pye

1. Introduction
Thein vitro growth of keratinocytes has proved to be an important tool
in the study of the normal biology and disease processes involving the skin,
e.g., the influence of extrinsic regulators of growth and differentiation,
effects of pharmacological agents, dermo-epidermal interactions, tissue
antigenicity, and models of carcinogenesis.
Fibroblasts adapt well to culture conditions. However, keratinocyte
culture has been hampered by the inadequacy of media, which had hith-
erto been optimized for fibroblasts, and the overgrowth of cultured kera-
tinocytes by more vigorous stromal elements. These difficulties were
initially overcome by the use of lethally treated 3T3 feeder layers, which
supported keratinocyte growth and inhibited overgrowth by fibroblasts
(2). This tech nique, with or without minor variations, is still widely used
today. It led to the supposition however, that undefined elements from the
dermal components were essential to keratinocyte growth. Later, Eisinger
et al. (2) demonstrated that, by altering seeding density, pH, and incubat-
ing conditions, keratinocytes could be grown in the absence of specialized
substrates, conditioned medium, or media supplements. These proce-
dures still required large seeding densities, and passage either on plastic or
3T3 feeder cells was limited. More recently, the development of the MCDB
series of media (3,4) has permitted serum-free, clonal growth of keratino-
cytes with as few as 400 cells/cm2 of substrate. The optimization of nutri-
ents, particularly the reduction in calcium concentration and the omission
of serum, favors proliferation rather than differentiation, extending the life

83
 Barlow and Pye

of the culture and permitting serial propagation. This model is particularly

useful because of the absence of undefined supplements in the medium.
The design or selection of the tissue culture system to be used does,
however, depend on the question to be explored. Indeed, central to our
understanding of epidermal-stromal interactions has been the develop-
ment of the dermal equivalent model (5,6) as a useful working alternative
to organ culture of skin. In this way, collagen lattices, seeded with living
fibroblasts, are placed on rafts. Keratinocytes are then either incorporated
into the lattice as an explant that grows over the surface, or seeded as a
single-cell suspension on top of the lattice. These simplified skin models
are cultured at an air liquid interface that promotes differentiation of kera-
tinocytes.
In this chapter, the techniques of keratinocyte culture on tissue culture
plastics, extracellular matrices, 3T3 feeder layers, and in the dermal equiv-
alent system are described.

2. Materials
Media and supplements may be purchased from any company sup-
plying tissue culture products and must be of tissue culture reagent grade.
1. Keratinocyte medium: Keratinocyte medium is a mixture of Dulbecco-
Vogt’s modification of Eagle’s Minimum Essential Medium and
Ham’s F12 (3/l, V/V)-3DMEM:F12.
(a) Basal medium-3DMEM:F12
DMEM powdered medium 10.03 g
Hepes 4.77 g
Sodium bicarbonate 0.80 g
Ham’s F12 powdered medium 2.65 g
Constituents are made up to 1 L with deionized, double dis-
tilled water, brought to pH 7.2-7.4 with 5M sodium hydroxide
and sterile filtered. Sterility checks should be made (seeChap-
ter 1, this volume). Alternatively, three parts single strength
liquid DMEM containing 2OmM Hepes buffer are mixed with
one part Ham’s F12 liquid medium and sterile sodium bicar-
bonate (0.8 g/L) and L-glutamine (2 mM> added before use.
Glutamine is unstable at 4°C and should be kept frozen. Its
half-life in medium stored at 4°C is approximately 3 wk, and
medium stored for long periods should be supplemented with
fresh glutamine.
(b) Supplements to 3DMEM:F12 keratinocyte medium. Stock so-
lutions.
Keratinocyte Culture 85

(1) Epidermal growth factor (EGF). Dissolve 0.1 mg EGF from

mouse submaxillary glands in 10 mL of distilled water (10 pg/
mL). Sterile-filter and store frozen in I-mL aliquots at -2OOC.
(2) Insulin. Dissolve 100 mg of bovine insulin in 20 mL of 0.005N
HCl (5 mg/mL). Sterile-filter and store in 1-mL aliquots at
-20°C.
(3) Hydrocortisone. Dissolve 5 mg in I mL of 95% ethanol (5 mg/
mL) and store this concentrated stock at 4°C. Add 0.4 mL of
concentrated stock to 9.6 mL of serum-free medium, sterile-fil-
ter, and store in 1-mL aliquots at -2OOC.
(4) Cholera toxin. Dissolve 1.0 mg of cholera toxin in 1.18 mL of
distilled water (lOsM), Store this concentrate at 4OC. Add 0.1
mL of concentrate to 9.9 mL of medium containing 10% serum.
Sterile-filter into I-mL aliquots and store at -2OOC.
(5) Transferrin/Triiodo-L-thyronine (T3). Dissolve 100 mg of hu-
man transferrin in 12 mL of phosphate buffered saline. Dis-
solve 13.6 mg of T3 in the minimum volume of 0.2N sodium
hydroxide, and make up to 100 mL in sterile distilled water
(2 x WM). Add 0.2 mL of the T3 concentrate to 12 mL of trans-
ferrin, and make the total volume up to 20 mL with distilled
water. Sterile-filter into I-mL aliquots, and store at -2OOC.
(6) Adenine. Dissolve 243 mg in the minimum volume of lNHC1,
and make up to 100 mL with sterile distilled water (1.8 x 10-2M).
Sterile-filter and store in lo-mL aliquots at -2OOC.
(7) Serum. Heat-inactivated fetal calf serum (FCS; 10%) is used to
culture epithelial cells. All batches of serum must be tested for
their ability to support keratinocyte growth at clonal densities
(seeChapter 1, this volume) (Table 1).
2. Feeder cell medium:
a. Basal medium-Dulbecco-Vogt’s modification of Eagle’s Mini-
mum Essential Medium (DMEM).
DMEM powdered medium
N-2 Hydroxye thy1 piperazine N-2 ethane 13.37 g
sulfonic acid (HEPES) 4.77 g
Sodium bicarbonate 0.80 g
Constituents are made up to 1 L with deionized, double-distilled
water, brought to pH 7.2-7.4 with 5M sodium hydroxide and sterile-
filtered. Sterile single-strength, liquid DMEM containing 20 mM
Hepes can be purchased ready for use. Sterile sodium bicarbonate (0.8
g/L) and L-glutamine (2 mM) are added separately.
b. Serum. Heat-inactivated newborn calf serum (lO%)is added to
DMEM before use.
86 Barlow and Pye

Table 1
Keratinocyte Seeding Density on Plastic,
Extracellular Matrix, and 3T3 Feeder Cells in Vitro
Substrate Keratinocyte seeding density/80 cm2 flasks
Tissue culture plastic 5-10x106
Extracellular matrix 3-5x106
3T3 Feeder cells 1-2x106

3. Antibiotics:
Antibiotic x 10 solution x 1 solution
Penicillin 1000 U/mL 100 U/mL
Streptomycin 1000 &mL 100 pg/mL
Fungizone 2.5 pg/mL 2.5 pg/mL
4. Dispase (Protease type ix from Bacillus polymyxu): Dissolve 250 mg
dispase in 100 mL serum-free DMEM. Filter sterilize and store in lo-
mL aliquots at -2OOC.
5. Trypsin:
a. Routine subculture of cells: Trypsin (1:250) 0.25% solution
containing 0.02% ethylene-diamine-tetra-acetic acid (EDTA)
is made up in Hepes buffered calcium and magnesium-free
Hanks’ balanced salt solution.
b. Keratinocyte suspensions. Trypsin (1:300) 0.25% containing
penicillin 200 U/mL, streptomycin 100 yg/mL, and 0.5 g/L
sodium bicarbonate in calcium and magnesium-free Hanks’
balanced salt solution are stored in 5-mL aliquots at 20°C. Be-
fore use, each 5-mL aliquot is made up to 20 mL with calcium
and magnesium-free Hanks’ balanced salt solution.
6. Trypan blue solution for cell viability counts. Dissolve 400 mg Trypan
blue, 810 mg sodium chloride, and 60 mg potassium dihydrogen
orthophosphate. Adjust the pH to 7.2-7.3 with 1M sodium hydroxide,
and make up to 100 mL with distilled water.
7. Dermal equivalents:
a. Medium for preparation of collagen rafts.
Eagle’s minimum essential medium powder 9.78 g
Hepes 5.0 g
Sodium bicarbonate 2.0 g
Distilled water 535 mL
b. Collagen. Collagen can be purchased from a number of com-
mercial sources. Pure Type I calf skin collagen (1 mg/mL) is
dialyzed against several changes of distilled water until the
pH is no lower than pH 4.5. Dialyzed collagen is stored in 5-
mL aliquots at -2OOC until required.
Keratinocyte Culture 87

c. Reconstitution of dermal equivaltents.

Eagle’s MEM 2.3 mL
Collagen 1.5 mL
Sodium hydroxide (0.W) 0.25 mL
Fetal calf serum 0.45 mL
Fibroblast cell suspension (8 x lo5 cells/mL) 0.5 mL

3. Methods
All procedures using human tissue are carried out in a Class II laminar
flow hood using aseptic technique.

3.1. Preparation of Keratinocytes

1. Human skin may be obtained from surgical specimens removed at
circumcision, mastectomy, apronetomy, or during cosmetic surgery.
Not all skin grows equally well in culture. In our experience, skin from
ears, hands, and feet are less successful than skin from breast, abdo-
men, or preputial skin. Growth rate is also affected by donor’s health
and age. Keratinocytes from children and young healthy adults un-
dergo more population doublings. The skin sample should be placed
in a dry sterile container and stored at 4OC until collection can be ar-
ranged. For maximum viability, skin should be processed as soon as
possible after excision. However, where necessary, skin can be stored
in this way for several days with a small loss of viability. If prolonged
storage is contemplated, a small quantity of phosphate buffered saline
containing penicillin and streptomycin should be added, to prevent
drying of the sample.
2. Fatty tissue should be removed from abdomen and breast samples;
otherwise tissues should be directly immersed in 10 x strength anti-
biotic solution for 30 min. After several washes in serum-free me-
dium, tissues are left to soak in single-strength antibiotic solution for
a further 30 min (seeNote 1).
3. The skin is transferred to a sterile 90-mm Petri dish with the sample
placed dermal side down. Using sterile scissors and forceps, the skin
is raised in the forceps, and small pieces of skin (0.5 x 0.5 cm) are cut
from the surface leaving behind as much of the connective tissue as
possible.
4. The pieces of tissues are placed in Dispase solution at 4OCovernight
with approximately 4-5 cm2 of tissue/ml of Dispase. For conveni-
ence, tissues may be left in Dispase solution for 48 h without signifi-
cant loss of viability.
88 Barlow and Pye

5. After 24 h, the tissue incubated in Dispase solution is transferred to a

sterile go-mm Petri dish. Using sterile forceps, the epidermis is peeled
very easily from the dermis. If the epidermis does not peel easily, the
tissue can be incubated for several hours at 37OC, or placed in fresh
Dispase and incubated for a further 24 h at 4°C. The pieces of peeled
epidermis should be kept moist in calcium and magnesium-free
Hanks’ balanced salt solution during this procedure.
6. The pieces of epidermis are transferred to a sterile universal contain-
ing 10 mL of prewarmed trypsin solution and incubated at 37°C for
30-45 min.
7. After this time, the contents of the universal container are shaken once
or twice toloosen the trypsinized cells, which are released into suspen-
sion. The solution is allowed to settle briefly, after which the pieces of
epidermis float to the surface while cellular debris falls to the bottom
of the universal.
8. The keratinocyte suspension is carefully pipeted into a centrifuge
tube. To each 5 mL of keratinocyte cell suspension, 5-mL of DMEM
containing 20% serum is added and the suspension centrifuged at
2008 for 5 min.
9. The supernatants are aspirated and 1 mL of fresh medium containing
10% serum added to each cell pellet. Using a plugged sterile Pasteur
pipet, the cell pellets are very gently resuspended to form a single-cell
suspension.
IO. More fresh, prewarmed trypsin solution is added to the pieces of epi-
dermis, and the process is repeated. Three trypsinizations may be per-
formed to maximize cell yield. The cells harvested in this way are
pooled and counted.
11. An equal volume of cell suspension and Trypan blue solution are
mixed and left at room temperature for 5 min. Using the improved
Nebauer hemocytometer, both chambers are filled by touching the
edge of the coverslip with the pipet and allowing the suspension to be
drawn up by capillary action. Avoid flooding the chamber.
12. Each hemocytometer chamber is divided into nine large squares deli-
neated by triple white lines. The center square is further subdivided
into 25 squares, which are again subdivided into 16 squares. Viable
cells, i. e., those not stained blue, are counted in the two central squares
of each chamber and the mean value for one large square calculated.
The number of cells in one large square is multiplied by the dilution
factor for Trypan blue, i. e., x 2 = n cells. Multiply the number of cells
nby104 = cells/mL. The number of cells/mL multiplied by the num-
ber of mLs of cell suspension = total number of cells harvested.
Keratinocyte Culture 89

Fig. 1. Human foreskin keratinocyte culture 3 d after plating (200x1.

3.1.1. Seeding of Keratinoeytes

Keratinocytes may be seeded into tissue culture flasks, onto 3T3
feeder layers, or onto an extracelluar matrix.
3.1.1.1. Tissue Culture Flasks
1. The suspension of cells is diluted in an appropriate volume and
seeded into plastic tissue culture flasks at a density of 6.25 x plastic 104
-1.25 x 10s cells/cm*.
2. After the cells have attached (1-2 d), cultures should be refed with
complete keratinocyte medium every 2-3 d. Cultures should be fed
with keratinocyte medium without EGF when first seeded into flasks.
When cells approach confluence, they may require to be fed every day.
Cultures are incubated at 37°C.
3. Cells may appear flattened a few days after plating (Fig. l), but as the
culture becomes denser, cells assume the typical pavement-like ap-
pearance of epithelium (Figs. 2,3). Cultures should approach conflu-
ence within 10-15 d and should be subcultured (seebelow) before they
reach confluence. Cultures held at confluence differentiate, and
blistering may occur.
90 Barlow and Pye

Fig. 2. Human foreskin keratinocyte culture 7 d after plating (200x).

Fig. 3. Human foreskin keratinocyte culture 12d after plating. Keratinocytes assume
typical pavement-like structure as cultures approach confluence (200x mag).
Keratinocyte Culture 91

Fig. 4. Swiss 3T3 cells 2 d after seeding at 1.5 x 106 cells/80 cm2 flask (100x).

3.1.1.2. 3T3 Feeder Layers

1. 3T3 Swiss albino transformed mouse embryo fibroblasts (ref. no. CCL-
92) may be obtained from the European Collection of Animal Cell
Cultures, PHLS Centre for Applied Microbiology, Porton Down,
Salisbury. 3T3 cells are grown in DMEM plus 10% newborn calf
serum. Cultures are routinely passaged two or three time/wk.
2. Prior to seeding 3T3 cells with keratinocytes, cells are passaged and
grown to SO-70% confluence (Fig. 4). The 3T3 cells are then lethally ir-
radiated with 6000 rads using any appropriate source or treated with
mitomycin C (seeNote 2). Cultures SO-70% confluent are incubated
with medium containing mitomycin C at 5-10 pg/mL (see Note 2) for
1 h at 37OC.
3. The cells are then thoroughly washed with serum-free medium before
keratinocytes are seeded onto the 3T3 cells.
4. Keratinocytes are seeded at a concentration of 2.5 x lo’ cells/cm2, fed
with keratinocyte medium without EGF, and incubated at 37OC.
5. After cells have attached, cultures should be fed with complete kera-
tinocyte medium. Clones of epithelial cells may not be apparent
92 Barlow and Pye

among the 3T3 feeder cells for several days, but with increasing incu-
bation, small islands of 20-50 cells should be easily visible (Fig. 5). The
cultures should approach confluence in lo-14 d and must be subcul-
tured before cells reach confluence. Keratinocytes may be subcul-
tured at seeding densities as low as 1.25-2.5 x 104cells/cm2 onto new
lethally treated 3T3 cells.
3.1.1.3. Extracellular Matrix
1. ECM made in this laboratory from 3T3 cells or human foreskin fibro-
blasts is less successful in promoting growth of keratinocytes than
commercially prepared bovine cornea1 endothelial cell matrix (7).
Once confluent, these cultured cells were lysed with 0.5% Triton-X-
100 in 0.0125M ammonium hydroxide for 5-10 min and washed in
balanced salt solution before use (Fig. 6).
2. Keratinocytes seeded onto bovine cornea ECM have up to two to
threefold greater plating efficiency than on tissue culture plastic, and
confluence may be achieved in up to 2 wk with seeding densities of
3.5-6.0 x lo4 cels/cm2 in 80 cm2 flasks.
3. The cultures are fed with keratinocyte medium without EGF for the
first day, after which cultures are refed as required with complete ker-
atinocyte medium. Cells should be subcultured before confluence is
reached.
3.2. Subculture
1. Nearly confluent flasks of keratinocytes are washed twice in calcium
and magnesium-free Hanks’ balanced salt solution.
2. To each 80 cm2 flask, 5 mL of trypsin-EDTA solution is added, and the
flasks incubated for 5-10 min at 37OC.
3. The cells are dispersed into solution by tapping the flasks.
4. Trypsinization is stopped by the addition of 5 mL of 3DMYEM:F12 +
20% FCS to each flask.
5. The cell suspensions are centrifuged at 200gfor 5 min and the superna-
tants aspirated.
6. The cell pellet is gently resuspended in 1 mL of 3DMEM:F12 + 20%
FCS and the cells counted in a hemocytometer. The cells are diluted
to an appropriate volume for subculturing onto various substrates
(Table 2).
7. Cells are fed with keratinocyte medium 3DMEM:F12 + 10% FCS plus
additives, but without EGF (Table 2). EGF is added to the cultures at
the first feeding (7). Proliferation rate is dependent on culture condi-
tions, age, health of donor, and the site of donor skin. Table 2 is a guide
to seeding densities used in this laboratory on different substrates.
Keratinocyte Culture 93

Fig. 5. Colonies of keratinocytes (K) growing on lethally treated 3T3 feeder cells, 7 d
after seeding at 2 x 106 keratinocytes/80 an2 flask (200x).

Fig. 6. Keratinocytes from 60-yr-old female after 3rd passage on extracellular matrix
from bovine cornea1 cells. The ghosts of the cornea1 cells can clearly be seen.
94 Barlow and Pye

Table 2
Complete Keratinocyte Medium
Supplement Vol Final concentration
3DMEM:F12 90 mL
Serum (HIFCS) 10 mL 10%
Adenine 1.0 mL 1.8 x 10-4M
EGF 0.1 mL 10 ng/mL
Insulin 0.1 mL 5 pg/mL
Hydrocortisone 0.2 mL 0.4 yg/mL
Cholera toxin 0.1 mL 10-‘OM
TransferrWT3 0.1 mL 5 pg/mL/2 x 10-gM

3.3. Dermal Equivalents

Keratinocytes attach to the surface of the dermal equivalent, divide,
and grow to cover the collagen lattice. Differentiation into a multilayered
epidermis occurs, and keratinization is much more complete in these cul-
tures incubated at an air-liquid interface (Fig. 7). Dermal equivalents are
made in this laboratory in 60-mm diameter Petri dishes, but can be scaled
UD or down to the required size.
;. Sufficient collagen, medium, serum, and human skin fibroblasts (as
described in the Materials section) are mixed and poured into 60-mm
diameter Petri dishes.
2. The dishes are incubated in humidified boxes at 37OC, and within a
few h, the collagen gels solidify.
3. At this stage, a skin plug approximately 2-3 mm in diameter can be in-
serted into the gel, which upon contraction will hold the skin plug
tightly in place.
4. Keratinocytes migrate from the explanted tissue across the surface of
the collagen/fibroblast matrix. The gels will normally contract within
2-3 d (Fig. B), but the degree and speed of contraction of the gel depend
partially on the seeding density of the fibroblasts.
5. The contracted rafts are raised onto transwell plates (Fig. 9). Rafts
without epithelial plugs may be seeded with keratinocytes at this
stage (3 x lo5 cells/ml). This procedure can be standardized using
stainless-s tee10 rings.
6. The cultures are then fed with enough keratinocyte medium to reach
just below the level of the raft. The medium is drawn up by capillary
action.
7. Cultures are refed every 2-3 d.
Keratinocyte Culture 95

Fig. 7. Keratinocytes cultured on dermal equivalent (x100).

Fig. 8. Contracted collagen-fibroblast matrix 3 d after seeding. Original diameter of

the gel was 60 mm (100x).
96 Barlow and Pye

Fig. 9. Collagen-fibroblast matrix raised to air-liquid interface on stainless-steelgrid

cmoxh

4. Notes
1. Several other antibiotics have been tested in this laboratory for use
with keratinocyte cultures. Vancomycin was found to be very effec-
tive when tested against ten methicillin-resistant strains of Staphy-
loccocus and has very low toxicity to cultured keratinocytes. Washing
in medium containing 1 mg/mL vancomycin for 30 min is recom-
mended for infected cultures. After several washes in serum-free me-
dium, cultures can be incubated with 100-200 pg/mL of vancomycin.
Vancomycin is only recommended in special circumstances and
should not be used routinely.
2. Mitomycin C, which disrupts microtubule formation in cells and pre-
vents completion of cell division, is often used as an alternative to
Keratinocyte Culture 97

lethal irradiation. Potency and toxicity of different batches of mitomy-

tin C varies, and each batch should be tested. In this laboratory, mito-
mycin C has been found to be effective at concentration of 5-10 pg/
mL. If none of the available flasks of 3T3 cells are at a suitable stage,
a fully confluent culture of 3T3 cells can be lethally treated, and subcul-
tured into a flask freshly seeded wi th keratinocytes the same day or the
following day with beneficial results.

Acknowledgments
Our thanks to Ms. Debbie Coakley for excellent technical assistance.
Our grateful thanks to Dr. J.K. Wright for assistance with photography and
to Dr. T. E. Cawston, Rheumatology Research Unit, Addenbrooke’s Hos-
pital, Cambridge for the collagen used in the dermal equivalent.

References
1. Rheinwalg, J. and Green, H. (1975) Serial cultivation of strains of human epidermal
keratinocytes: the formation of keratinizing colonies from single cells. Cell 6,331-
339.
2. Eisinger, M., Lee, J. S., Hefton, J. M., Darzynkiewicz, Z., Chaio, J. W., and Deharven,
E. (1979) Human epidermal cell cultures: Growth and differentiation in the absence
of dermal components or medium supplements. Proc. N&l. Acud., USA 76,5340-
5344.
3. Tsao, M. C., Walthall, B. J., and Ham, R. G. (1982) Clonal growth of normal human
keratinocytes in a defined medium. J. Cell Physiol. 110,219-229.
4. Boyce, S. and Ham, R. G. (1983) Calcium-regulated differentiation of normal human
epidermal keratinocytes in chemically defined clonal culture and serum free serial
culture. J. Invest. Dermatd. 81,533-540.
5. Bell, E., Ehrlich, H. P., Sher, S., Merrill, C., Sarber, R., Hull, B., Nakatsuji, T., Church,
D., and Buttle, D. J. (1981) Development and use of a living skin equivalent. J. Plastic
and Reconstructive Surgery 67,3&S-392.
6. Bell, E., Sher, S., Hull, B., Merrill, C., Rosen, S., Chamson, A., Asselineau, D.,
Dubertret, L., Coulomb, B., Lapiere, C., Nusgens, B., and Neveux, Y. (1983) The re-
construction of a living skin. J. lnvesf. Dermafol. 81,52-510.
7. Gospadarowicz, D., Mescher, A. L., and Birdwell, C. R. (1977) Stimulation of cornea1
endothelial cell proliferation by fibroblast and epidermal growth factors. Experi-
mental Eye Res. 25,75-89.
 Chapter 10

Establishment of Mouse
Epithelial Lung Cell Strains
and Cell Lines

Garry J. Smith
1. Introduction
Mesenchymal cells have been cultured successfully for many years
owing to the ease of their culturability. With the advent of more complex
growth media, and some changes to methods, it has become possible to
culture epithelial cell types from a wide variety of rodent and human tis-
sue sources successfully (1,2,3). We have used medium CMRL 1066 with
serum to establish and cultivate lung cell lines that are related to the type
II pneumocyte of lung alveolus (4-7). Other laboratories have success-
fully cultured different types of lung cells, including tracheal and bron-
chial cell types (1,2,3,8).
A critically important factor in satisfactory culture of any cell type is
the ability to observe markers that will identify the cell type of origin of the
culture. In some cell types, ultrastructural markers at the electron micro-
scope level can be very useful (for example, presence of microvilli and
properly structured desmosomes can indicate epithelial origin), and some
cells may exhibit specific structures (such as intracellular surfactant lamel-
lar bodies in type II pneumocytes, or Weibel-Palade bodies in endothelial

99
100 Smith

cells) that denote cell type of origin. However, ultrastructural markers can
be misleading, for example, desmosome-like bodies occur in a variety of
cultured cell types (1). Therefore, antibodies to specific ultra-structural
features can be more reliable. Identification by antibodies to keratinor pre-
keratins is an extremely valuable marker denoting epithelial origin, as
these intermediate filaments tend not to be expressed in culture bynonepi-
thelial cell types. Furthermore, it may be possible to prepare antisera to
cell-specific proteins (9,ZO). The major problem with identification of cells
in culture is the general observation that differentiated cell types do not
culture well. Secondly, because of the similarity of culturing procedures
and media formulations, cells from different tissue types tend to be placed
under the same selection pressures in culture. Therefore, cells of any tissue
type in culture tend to grow towards a similar end point with respect to
differentiated characteristics.
Cell culturing can be a particularly useful means for isolating purified
cell types for further study, provided that these limitations in ability to
culture cells of a desired type are taken into consideration. This present
chapter will use culturing of type II pneumocyte-related cells of the lung
alveolus as an example of epithelial culturing procedures. Several text-
books have very good overviews of cell culturing methods (11,12), We will
consider in this chapter the culturing of type II pneumocyte-related cells
from both normal lung and from chemically induced lung adenomas. This
chapter does not contain information on basic cell culture techniques, such
as media preparation, cell cloning, cryopreservation, cell passaging, pre-
vention of cell-line cross-contamination, laboratory design, and laboratory
safety, which are dealt with elsewhere in this volume.

2. Materials
The powdered medium CMRL 1066 is made up freshly or at most
within several days of use. Serum is kept frozen and added to the medium
as required together with antibiotics. The procedure in our laboratory is to
have the stocks of liquid medium, serum, antibiotics, and so on, made up
in a laminar flow hood as independent stocks that have been filtered
through 0.2 micron filters in order to eliminate prokaryotic contamina-
tion. Upon full formulation of the medium with serum and antibiotics, the
culture formulation is again filtered through 0.2 micron filters in order to
provide a second stage of sterilization. Enzymes such as trypsin are kept
as concentrated frozen stocks to be subsequently diluted with independ-
Lung Cell Strains and Cell Lines 101

ently sterilized phosphate buffered saline (II), and this formulation is sub-
sequently refiltered and kept at 4°C. Media constituents glutamine and
fungizone both have a life of only several days in unfrozen liquid formula-
tions.
Formulations for use in explanting and culture of type II pneumocyte
related cell lines from mouse lung are as follows:
1. Phosphate buffered saline (PBS): Made as a 10x concentrate without
calcium and magnesium consisting of (as w/v) 0.2%KCl, 0.2% KH,J?O~
8% NaCl, 1.15% Na,HPO,, pH 7.0-7.4 (with NaOH).
2. The complete formulation of culture medium, designated medium A,
consists of medium CMRL 1066 (seeAppendix), 10% (v/v> fetal calf
serum, 2.5 mg/mL Fungizone, 100 IU/mL penicillin, 100 mg/mL
streptomycin, and 2 mM L-glutamine.
3. Trypsin is used at 0.25% (w/v) in calcium and magnesium-free PBS.

3. Methods
The animals of choice for theseexperiments are specific-pathogen-free
animals. The initial dissection and explanting is performed under semi-
sterile conditions in a room that has preferably been presterilized by ultra-
violet light. The possibility of the operator contaminating the preparation
tissue material is minimized by the wearing of a mask or performance of
the explanting of tissue underneath a plastic screen standing in front of the
operator.
1. One adult female Balb / c mouse is killed in an ether chamber. The ven-
tral surface of the animal is swabbed well with alcohol, the neck region
dissected, and the trachea tied off tightly with fine cotton gauze in
order to prevent lung collapse upon opening the chest. The chest
region is then opened by a ventral incision from just above the dia-
phragm through the rib cage and up to the neck area. The rib cage on
either side is removed with scissors and forceps to expose the inflated
lungs.
2. An optional procedure at this stage is removal of blood cells by perfu-
sion of the lungs with medium A via the right ventricle of the heart.
However, this procedure does not enhance the culturing method. A
1 mL syringe is used to infuse the lungs gently via the trachea below
the thread tie with 1 mL of calcium and magnesium-free PBScontain-
ing 0.25% (v/v) trypsin. The trachea is then tied off below this needle
102 Smith

incision. This lung infusion procedure increases the yield of alveolar

cells in culture, but is optional for other epithelial cell types (I).
3. The lungs and heart are removed by dissecting through the trachea
with scissors above the tie and gently pulling upwards on the trachea,
while dissecting connective tissue away from the dorsal part of the
heart and lungs. The inflated lungs come free of the body wall, and
may be placed into a solution of medium A containing antibiotics and
Fungizone in order to wash off excess blood and to minimize con-
tamination from airborne prokaryotes.
4. The tissue is then taken to a laminar flow hood where it is put into a
glass Petri dish and a few drops of medium A are placed upon the
tissue. The lungs are dissected free of the other tissues and isolated in
the Petri dish.
5. The moist tissue is chopped into very fine pieces of maximum size 1
mm3 using two sharp, sterile scalpels drawn across the tissue in a
cross-chopping fashion. The mashed small tissue fragments may then
be transferred to a sterile, capped plastic centrifuge tube, and 4 mL of
calcium and magnesium-free PBS containing 0.25% (v/v> trypsin
added. The lung tissue is swirled to allow complete distribution
within the trypsin solution.
6. This tissue suspension is incubated at 37OC for 15 min in a water bath
with occasional gentle hand mixing. The lung tissue generally dis-
tributes between small tissue pieces floating on the top surface still
associated with air or small tissue pieces that sink in the solution.
7, Following the 37OCincubation, the tissue suspension is placed on ice
for several minutes to cool. Overtrypsinization can lead to cell lysis
and viscous DNA release. Centrifugation at 2000 rpm for 5 min in a
bench centrifuge results in a small pellet of tissue pieces that is re-
suspended in 15 mL of medium A. The upper floating fraction of lung,
which represents lung not in close contact with the trypsin solution
during infusion, is discarded.
8. The resuspended small tissue pieces are equally distributed into three
25 cm2 plastic culture flasks and the primary tissue explants incu-
bated at 37OC in a 5% (v/v> CO,-air atmosphere. The small size of the
chopped tissue fragments is crucial to the explant, since only these
small pieces are accessible to the medium and the atmosphere, and so
will remain viable.
9. At twice weekly intervals, the 25 cm2 flask is briefly placed on its side
at an angle so that nonattached tissue fragments may settle to the
Lung Cell Strains and Cell Lines 103

Fig. 1. A dark tissue explant with spindloid fibroblasticoutgrowth followed by cobble-

stone-like epithelioid cell outgrowth (arrowed). Six weeks after primary culture. Phase
contrast 165x.

bottom corner of the flask, half the medium A is removed with a

Pasteur pipet, and is replaced with fresh medium A. This is referred
to as a “twice weekly half-medium change.”
10. After 3-7 d, the small tissue pieces should attach to the plastic bottom
surface of the flasks, and fibroblastoid cells emerge onto the plastic
from the tissue piece followed by more epithelial-like cells (Fig. I). The
easily identified outgrowths reach a static phase or grow very slowly
after 1-2 wk in culture. The most desirable situation is when cob-
blestone-like epithelioid cells establish themselves in the explant and
can be easily distinguished from the spindle-shaped fibroblastoid
cells (Fig. 2). In the case of cells explanted from chemically induced
lung adenomas, an alternative primary culture behavior may occur.
Here the adenoma cells may appear as small, discrete foci of poly-
gonal cells emerging over wide areas of the plastic culture surface and
not deriving directly from small explanted tissue islands (Fig. 3).
11. In either the tissue focus or emerging island situation, the epithelioid
 Smith

Fig. 2. Mixed epithelioid (arrowed) and fibroblastoid cells early after cell outgrowth
in primary culture. Phase contrast 330x.

Fig. 3. Foci of multinuclear epithelioid cells appearing in primary culture of adenoma

tissue after two weeks. Phase contrast 330x.
Lung Cell Strains and Cell Lines 105

cells must be preserved free from overgrowth by fibroblastoid cells.

This may be efficiently achieved in a semis terile room by scraping the
cultures under an inverted microscope using a drawn-out, pointed
Pasteur pipet that has been sterilized by flaming. With practice, such
a fine glass probe can be used to detach encroaching fibroblas toid cells
without damaging the epithelioid cells nearby. The detached cells
may then be washed free of the epithelioid cells and removed from
culture with a Pasteur pipet, and the normal conditioned or half-
change medium of the cultures replaced onto the epithelioid cells.
Cultures are maintained on twice weekly half-medium changes.
12. The factors that will determine the success of the epithelioid cell cul-
ture are
a. the ability to identify discreet epithelioid colonies separated
from fibroblastoid cells;
b. the ability to prevent fibroblastoid overgrowth of cultures suc-
cessfully; and
c. the transition of the epithelioid cells to growth in culture.
13. Transition to growth in culture, which has occurred in some 20% of
culture experiments in our procedures, is observed as a proliferation
of the epithelioid cells a month or so after primary culture. It is a phen-
omenon that appears to be caused by minor rearrangement of the
mouse cell genome as the cells persist or slowly divide in culture. Only
with such transition to growth in culture can the epithelioid cells be
successfully established as growing cell lines and cell strains.
14. In the event that the cells do demonstrate continued growth in cul-
ture, the epithelioid cell colonies, when they have grown to some 50
cells or more and are showing continuousgrowth in culture, may be
selectively trypsinized away from other cell types. Having removed
the normal culture medium, a local application of 1 mL of 0.25% (v/
v) trypsin for one or two min using a Pasteur pipet lifts the cells off the
plastic. The cells, usually accompanied by some fibroblasts, may then
be transferred to a new culture flask and fed with normal conditioned
medium or half-change medium (Fig. 4). However, it is advisable to
attempt to maintain the cell density in culture at its highest practical
level during these procedures. This may be done by culturing the pas-
saged cells into small-well plastic culture dishes.
15. In the event that a primary culture remains static and epithelioid
transition to growth does not occur, trypsinization, after some 4 wk in
primary culture of the whole culture, and transfer to a new flask can
encourage epithelioid growth.
106 Smith

Fig. 4. A focus of epithelioid cells that have been selectively trypsinized from primary
culture.

16. The most desirable endpoint for passaged explants is where the epi-
thelioid cells continue to grow, with elimination of fibroblasts using a
glass rod as necessary, such that theepithelioid cells dominate the cul-
ture and approach confluence within several months. During this
phase, the epithelioid cells should maintain a healthy morphology,
with clearly polygonal cell shape, and clear nucleus and nuclei under
phase contrast microscopy, and a slow but definite growth rate in cul-
ture (Fig. 5). Tumor-derived cells generally exhibit a high nuclear to
cytoplasmic ratio (Fig. 3).
17. The next stage in cell culture is a verification of the derivation of the
growing cells. Some indication of the epithelial nature of the cells can
be obtained from overall morphology using the inverted light micro-
scope. Phase contrast microscopy should show polygonal cell types
and, in the case of nonneoplastic cells, a relatively low nuclear to cyto-
plasmic ratio. The cells should be quite visible, and show discreet
nuclei and vacuoles using phase contrast microscopy (Fig. 5). In gen-
eral, epithelial cell cultures from nonmalignant cells will grow in
closely associated islands of cells forming a polygonal cobblestone-
Lung Cell Strains and Cell Lines 107

Fig. 5. Proliferating culture of epithelioid cells that have transformed to growth in

culture and are approaching confluence. Phase contrast 330x.

like pattern. At the earliest time, as cell numbers increase, a sample of

cells should be taken for electron microscopy using a transmission
electron microscope. The electron microscopy should again indicate
polygonal cell structure on a trypsinized population with cells show-
ing microvilli or other epithelial-type ultrastructural marker (Fig. 6).
Particularly useful, especially on cells early in culture, is the ability to
see spot-desmosomes between cells that have not been trypsinized
apart. However, it is critical that the desmosome be of a correct
ultrastructural type (2) with a central lamella and the presence of lo-
nm filaments moving away on either side of the central lamella. One
of the best indications of epithelial nature is the ability to identify
prekeratin or keratin using specific antisera to this intermediate
filament protein family. Immunological identification (seeChapter
35) of keratin on cell pellets that have been sectioned in the usual
histopathology fashion or on cells that have been grown on glass
coverslips and permeabilized with 0.3% Triton X-100 is a good in-
dication of their epithelial nature (6,7). Finally, it may be possible to
identify cells as coming from a particular epithelial cell lineage by
108 Smith

micrographof
Fig. 6. Electron micrograph of epithelial cells in culture exhibiting surface microvilli
(closed arrows), osmiophilic lamellar bodies (open arrows), and polygonal cell shape
123lOx.
identification with cell-specific antisera. In the case of type II pneu-
mocyte-related cell strains cultured in our laboratory, we have iden-
ultrastructural observation of
tified them as epithelial based upon ultrastructural
microvilli and identification with an antiserum prekeratin Q&-7),
antiserum to prekeratin Q&-7),
and as type II pneumocyte-related by electron-microscopic obser-
vation of osmiophilic lamellar bodies and by identification with a
polyclonal antiserum that reacted with only type II pneumocytes and
pneumocyte-related cell types (4,ZO). It has been frequently observed
that some differentiated features of the primary culture, notably
ultrastructural lamellar bodies, are not well retained beyond passages
of around passage number 30.

4. Notes
1. An alternative culture procedure that has been found suitable for type
II pneumocyte-related cell strain culture, and may be more suitable for
Lung Cell Strains and Cell Lines 109

some epithelial cell types, is as follows: The minced tissue may be

directly diluted into 3 mL of the PBS trypsin solution for digestion and,
after digestion, the tissue solution gravity-sedimented for 1 min at
room temperature by standing in the capped, sterile centrifuge tube
on the bench. This allows large tissue pieces, which may subse-
quently necrose in culture, to go to the bottom of the tube and the
smaller tissue fragments, which do not sediment under gravity after
1 min, to be equally distributed between three 25-cm2 plastic culture
flasks each containing 4 mL of culture medium A. The floating lung
fragments again are discarded in this preparation procedure.
2. Prevention of overgrowth of cultures by fibroblasts can be one of the
most difficult procedures in establishing epithelial cell strains. Some
inhibitors can be used for prevention of fibroblast overgrowth, for
example cis-hydroxy-proline (13) or dexamethasone (14). However,
the results of using fibroblast inhibitors can be variable, and we have
found that mechanical elimination of the flat, stringy fibroblastoid
cells using sterile glass rods can be very effective provided that the
epithelioid cells transform to growth in culture.
3. At early stages of passaging from primary culture, it is advisable to
minimize the dilutionof cells and to passage them at l-2or at most 1-5
dilutions for rapidly growing cells. It is extremely important that the
cells, as they are passaged and with increasing passage number in
culture, are subjected to a constant and consistent culturing regime.
Therefore, it is of advantage to passage the cells at the same or similar
dilution ranges from passage to passage and, furthermore, never to
allow the cells to become too confluent prior to passaging. These
procedures minimize the possibility of karotypic changes in culture,
which may occur if cells are lef t at confluence, thereby minimizing loss
of differentiated features of a given cell strain.
4. The fundamental needs for setting up epithelial cell culture, and in this
case culture of type II pneumocytes, is a good laboratory setup in
which sterile conditions may be maintained together with thorough
and consistent culture protocols. Patience in waiting for cells to trans-
form to growth in culture, often required for nonmalignant cell types,
and the ability to keep strict and thorough cell records are very im-
portant. A good characterization of cells for their general morphol-
ogy, their ultrastructure by electron microscopy, and their specific
differentiated characteristics, such as expression of intermediate fila-
ment proteins or special differentiated cell functions, should be per-
110 Smith

formed early in the life of the culture and routinely thereafter during
the life of the cell lines. Surveillance preventing contamination by pro-
karyotes and for any possibility of cross-contamination of cell types is
most important in maintaining the integrity of the cell lines.
5. The main pointers that can prevent time wastage and ineffectual cell
culture are
a. do epithelioid cells appear in culture after l-2 wk of primary
tissue explant?
b. do fibroblasts overgrow the epithelioid cell types within a
month of culture?
c. do epithelioid cells transform to growth in culture either in
their own right after a month or so in culture or within a few
weeks of a trypsin passaging of the primary explant and
d.can epithelioid cells be confirmed as epithelial either ultra-
structurally or by expression of epithelial specific markers,
such as keratins, and can specific cell type markers be iden-
tified?

Acknowledgments
This work has been supported by The National Health and Medical
Research Council and The New South Wales State Cancer Council.

References
1. Franks,L.M.andWigley,C.B. (eds.)(1979)NeoplasficTrunsformutioninDilferentiuted
EpifheliaI Cell Systemsin Vitro (Academic Press,London).
2. Harris, C. C. (1987) Human tissuesand cells in carcinogenesis research. Cancer&s.
47,1-10.
3. Smith, G. J, and Stewart, B. W. (eds.) (1982) In Vitro EpitheZiuICeILDifferentiation and
Neoplusiu (TheAustralian CancerSot., Sydney).
4. Smith, G. J.,LeMesurier, S.M., DeMontfort, M. L., and Lykke, A. W. J. (1984) Estab-
lishment of epithelial cell strains from normal adult mouse lung resembling a ure-
thane-induced lung adenoma cell strain and a metastasing mouse lung carcinoma
cell line. Cell Biol. International Reportst&161-169.
5. Smith G. J.,Le Mesurier, S.M., DeMontfort, M. L., and Lykke, A. W. J. (1984) Devel-
opment and characterization of type II pneumocyte related cell lines from normal
mouse lung. PuthoIogy16,401-405.
6. Smith, G. J. and Lykke, A. W. J. (1985) Characterization of a neoplastic epithelial cell
strain derived by dexamethasone treatment of cultured normal mouse type II pneu-
mocytes. J. of Pathology147,165-172.
Lung Cell Strains and Cell Lines 111

7. Smith, G. J., Bennett, F. A., Steele,J. G., and Bentel, J. M. (3986) Onset of neoplastic
phenotype in an epithelial cell strain from adult BALB/c mouse lung alveolus. J,
Natl. Cancer Inst. 76‘73-79.
8. Nettesheim, P.,Terzaghi, M., and Klein-Szanto, A. J.P. (1982)Development andpro-
gression of neoplastic disease. Morphologic and cell culture studies with airway
epithelium, in Mechanisms of Chemical Carcinogenesis (Harris, C. C., ed.1, Liss, New
York, pp. 473-489.
9. Ward, J. M., Singh, G., Katyal, S.L., Anderson, L. M., and Kovatch, R. M. (1985) Im-
munocytochemical localization of the surfactant apoproteinand Claracell antigen in
chemically induced and naturally occurring pulmonary neoplasms of mice. Amer.
I: Pathol. 118,493-499.
10. Smith, G. J.,Kumar, R. K., Hristoforidis, C. P., and Lykke, A. W. J. (1985) Expression
of a type II pneumocyte-specific antigen by a cell strain from normal adult mouse
lung. Cell Biol. International Reports 9,1115-1122.
Il. Paul, J. (1975) CeZZand Tissue Culture, 5th Ed. (Churchill-Livingstone, London).
12. Kuchler, R. J. (1977) Biochemical Methods in Cell Culture and Virology. (Dowden,
Hutchinson and Ross,Stroudsburg, Pennsylvania).
23. Kao, W. W.-Y. and Prockop, D. J. (1977) Proline analogue removes fibroblasts from
cultured mixed cell populations. Nature 266,63-64.
14. Edwards, A. M. and Lucas, C. M. (1982) Gamma-glutamyltranspeptidase as a pre-
neoplastic marker in hepato carcinogenesis: Expression in hepatocytes isolated
from normal and carcinogen-treated rats, in In Vitro Epithelial Cell Differentiation and
Neoplasia (Smith, G. J. and Stewart, B.W., eds.) The Australian Cancer Sot., Sydney.
 Chapter 11

Culture of Human Brain

Tumors on an Extracellular
Matrix Derived from Bovine
Cornea1 Endothelial Cells
and Cultured Human Glioma
Cells

Manfred Westphal,
Marianne H&nsel,
Hildegard Nausch, Eva Rohde,
and Hans-Dietrich Herrmann
1. Introduction
Cellculture has becomeanintegralpart of thedailyroutineof most on-
cology laboratories. It has enabled researchers to investigate a wide range
of cellular parameters in a defined system in which the experimental con-
ditions can be controlled and repeated. Although the manufacturers of
tissue culture materials are continually improving their products, cell at-
tachment and initial survival of primary cultures from tumor cells are still
problems in many laboratories. Many different approaches have been
taken to circumvent this problem, and coating of tissue culture dishes with

113
114 Westphal et al.

attachment enhancers, such as polyamino acids (I), fibronectin (Z), laminin

(3), and collagen (4), has been found helpful. For a long time, it was known
that endothelial cells produce abasement membrane, and this led to the use
of bovine cornea1 basement membrane by Gospodarowicz et al. (5,6), in
their research into the phenomenon of regeneration and nonregeneration
of cornea1 endothelium in different species. The application of this bovine
cornea1 extracellular matrix (bECM) has sincebeen greatly expanded (5,6).
bECM has found broad approval, and has been used for mammary carci-
noma (7), urological tumors (a), and different kinds of pituitary adenomas
(9JO) as well as CNS tumors (II).
It is generally recognized that interaction of cells, normal or neo-
plastic, with ECMs may interfere with the cellular biology (12,13) and that
matrices from different sources may affect cells in a variety of ways. It
could be shown, for example, that an ECM derived from human arachnoid
cells induces differentiation in a glioma cell line (14). It should thus be
mentioned in this introduction that many different cells will be able to
produce matrices, including tumor cells themselves, and that the use of
ECMs in oncology will not only facilitate initiation of cell cultures, but is an
interesting research field of its own as will be illustrated by some obser-
vations obtained with an ECM derived from a cell line that was derived
from glioblastoma.

2. Materials
2.1. ECM from Bovine CorneaL Endothelium, bECM
1. Bovine eyes: following the instructions from protocols published
earlier and elsewhere (6,X), fresh bovine eyes have to be obtained
from the local slaughterhouse (seeNote 1).
2. Isolation of the cornea (seeNote 2): it requires
a. no. 1 cannulae
b. angled microscissors
c. foreceps
d. a grooved director
e. phosphate buffered saline (PBS).
All instruments should be sterile.
3. Cell culture medium: Dulbecco’s modification of Eagle’s medium
with high glucose content supplemented with 10% fetal calf serum,
glutamine, (1 mM), pyruvate (2 n&l), gentamycin (25 pg/mL), and
fungizone (2.5 pg/mL).
4. ECM production:
a. dextran (MW 40,000)
Human Brain Tumors on ECM 115

b. basic bovine fibroblast growth factor (FGF)

c. sterile ammonium hydroxide solution (20 mM)
d. sterile PBS
e. sterile PBS containing gentamycin (25 bg/mL).
5. FGF isolation (see16):
a. 4 kg of bovine brain (seeNote 3)
b. 8 L of 0.15M ammonium sulfate pH 5.6
c. 6N HCl
d. Ammonium sulfate
e. Dialysis tubing (MW cutoff approx. 10,000)
f. CM-Sephadex C-50
g. Buffers: O.lM sodium phosphate pH 6.0, O.lM sodium
phosphate pH 6.00,15M NaCl, O.lM sodium phosphate
pH 6.00,6M NaCl.
6. Tumor cell isolation (seeNote 4):
a. Hanks’ balanced salt solution, Ca*+and Mg2+ free
b. angled microscissors
c. 5 mL tissue culture pipets with smoothened openings
(best from Falcon or Costar)
d. Enzyme cocktail: Pronase (0.05%), Collagenase (0.03%),
DNase (O.Ol%), in Hanks’ filtered sterile buffer.

3. Methods
3.1. Preparation of bECM
1. Only eyes that are uninjured, and show no signs of inflammation or ul-
ceration should be used. They are taken with the left hand,rinsed thor-
oughly with alcohol from a spray bottle, and then held down firmly on
a paper towel that is also soaked in alcohol. The eye is then punctured
tangentially with a cannula, so the chamber water can drain and the
cornea can shrink after the pressure is taken off the eyeball. Starting
from the needle insertion, the cornea is carefully cut out without con-
taminating it with other tissues from within the eyeball (seeNote 2).
2. Once the cornea is removed,it is put external surface down onto a Petri
dish on which it will stick because of adhesive forces. It is then rinsed
with PBS. Then the exposed inner surface is scraped once with the
grooved director, and by this maneuver, sheets of endothelial cells
will detach. The grooved director is then dipped into a tissue culture
dish already containing the complete cell culture medium. Use 1
dish/eye. Process ten eyes in one session.
116 Westphal et al.

3. The dishes are then placed in an incubator and should not be disturbed
for 5 d. At this point, the first colonies will have formed. The differ-
ent stages in the production of bECM are illustrated in Fig. 1.
4. After the primary cultures have become about 50% confluent with col-
onies, these are detached with trypsin, which sometimes takes up to
20 min because the cells attach firmly to their support.
5. The total cells are next distributed to 25 dishes that have been pre-
coated with gelatin. To achieve this, 0.2% gelatin is dissolved in PBS
and autoclaved. It is advisable to filter this solution because it may
still contain granular matter that disturbs the cell culture. The IO-cm
diameter cell culture dishes are covered with 5 mL of the PBS-gelatin
and left in the refrigerator overnight. Prior to use, the excess PBS-
gelatin is aspirated. The advantage of this gelatin coating is that, at the
next step, the cells will be very easy to detach.
6. In the next step, the cells from the precoated dishes are passaged at the
point when they are just subconfluent. The dispersed cells harvested
from eight to ten lo-cm dishes are added to a 550-mL bottle of com-
plete medium, which in addition is supplemented with 5% dextran
(MW 40,000, dissolved by heating in a small volume of medium with-
out additives and sterile filtered). To this suspension, basic bovine
fibroblast growth factor (bFGF) is added (seeNote 5). From this sus-
pension, the cells are seeded onto the culture dishes to be covered with
ECM. Six mL are added to a T25 flask, 25 mL to a T75, and 1 mL/well
in a 24-mL multiwell tissue culture dish. At this density and in the
presence of bFGF, the cells should reach confluence within 2-3 days.
After confluence, the cells are left for one more week to ten d and then
they are lysed.
7. The cells are lysed by aspirating the culture media and replacing it
with distilled water containing ammonium hydroxide (20 mM). The
cells burst rapidly by osmotic shock, and within 3 min, the gelatinous
debris can be removed and rinsed off. The ECM is left behind and
should be washed at least once more with PBS. The ECM-coated
dishes can be stored at 4OCwhile being covered with PBS containing
gentamycin (25 pg/mL). At that time, the matrix can be seen with a
phase contrast microscope (Fig. 1, and seeNote 6).

3.2. ECM from a Human Brain Tumor CeZZ Line

WC2M)
Cells from a human glioblastoma cell line that is strongly positive for
cell surface fibronectin are used in early passages when the cells still grow
Human Brain Tumors on ECM 117

Fig. 1. Different stages in the production of ECM. (A) A sheet of endothelial cells as
it is scraped off the cornea. (B) A colony of endothelial cells after 6 d. (C) Subconfluent
layer of bovine cornea1 endothelial cellsin the stage of bECM production. (D) Appearance
of bECM in the phase contrast microscope after lysis of the cells.

in an orderly arrangement. The cells cannot be used for tECM production

after passage 20 (Fig. 2, andse.eNote 7). Cell culture conditions are the same
as for the preparation of ECM except that Earle’s modified MEM is used as
a basal medium.

3.3. Isolation of FGF

Of great importance to the production of any ECM appears to be the
addition of FGF to the cultures, since this speeds up the rate at which the
cultures become confluent and also influences the quality of the matrix.
The isolation of basic FGF (bFGF) has been described by Gospodarowicz,
(26) but is is not necessary for the production of ECM to purify bFGF to
homogeneity.
1. Eight frozen brains are homogenized in 8 L of 0.15M ammonium-
sulfate pH 5.6.
 118 Westphd et al.

Fig. 2. 7’he KM-producing gliom in two different stagesof culture. (A) In passage
15, in which they produce tECM and grow in streams of parallel arranged cells in mono-
layer. (B) At a later stage,passage 46, in which the cells grow on top of each other and
tECM is produced but comesoff the plastic support.

2. The homogenate is adjusted to pH 4.5 with 6N HCl and stirred for at

least 1 h at 4°C.
3. The homogenate is centrifuged at 20,000 x g, and the burgundy
colored supematants collected. CAUTION: If the color has turned
brown because the pH was too low, even if this was only for a short
moment, the procedure should be stopped and started anew because
bFGF is very intolerant to acidic treatment.
4. Ammonium-sulfate is now added slowly to the extract, with stirring,
to give a concentration of 200 g/L and the suspension again cenfri-
fuged at 20,000 x g after having been stirred for 60 min.
5. Again the supernatants are collected, and again ammonium-sulfate is
Human Brain Tumors on ECM 119

added to give a final concentration of 450 g/L. After centrifugation,

the supernatants are discarded and the pellets collected.
6. The collected pellets are dissolved in a small volume of water (approx.
250 mL), adjusted to pH 6.0 with formic acid, and dialyzed against
water. (Use dialysis tubing with mol wt cutoff between 10,000 and
12,000.)
7. The dialysate is then loaded onto a CM Sephadex C50 column (3.5 x 20
cm) that has been preequilibrated with O.lM sodium phosphate pH
6.0. The ionic strength of the dialysate should be equal to or less than
that of the running buffer.
8. The material retained on the column is eluted with a stepwise increase
of NaCl concentration in the sodium phosphate buffer, first to 0.15M
and then to 0.6M.
9. The material eluting at the high salt concentration is collected. It is less
than 0.1% pure, but this is sufficiently pure for use in the production
of ECM. It can be stored frozen as well as lyophylized until reconsti-
tution. The concentration at which the extract has to be used should
be tested in a bioassay on endothelial cells. It is variable between ex-
tractions.

3.4. Isolation and Culture of Brain Tumor Cells

Brain tumors for this study will be divided into meningeomas, gli-
omas, and other tumors (such as acoustic neuromas, pinealomas, pituitary
adenomas, meduloblastomas, or gangliogliomas). (Seealso notes on indi-
vidual tumor types.) First the cell isolation and cultivation is described.
1. Representative portions of tumors are obtained at surgery (seeNote8).
The selected tissue fragments are placed immediately into 15 mL of
sterile Hanks’ balanced salt solution without calcium and magnesium
(CMF). NOTE: In this solution, the material can be maintained at least
for 2 h at room temperature. A yellow decoloration of the Phenol red
indicator in the CMF around the tissue fragment will at that point con-
firm its metabolic activity and thereby its viability. Usually, 1 cm3is
plenty of material.
2. The tissue is freed from coagulated blood and minced with angled
scissors. The fragments are transferred into the enzyme cocktail of
pronase, collagenase, and DNase in Hanks’ buffer. After 20 min incu-
bation at 37OCand shaking of the incubation mixture, the fragments
are again agitated by repeated trituration. The undispersed fibrous
fragments should be allowed to settle and the turbid supernatant
transferred to a centrifuge tube.
120 Westphal et al.

3. Alternatively, mechanical dispersion can be used after mincing. The

fragments are agitated by repeated trituration until a turbid solution
is obtained. Allow large pieces or undispersed fibrous fragments to
settle, and remove the supernatant into a centrifuge tube.
4. The cells and small-cell aggregates obtained from the digestion proce-
dure or from the mechanical dispersion are pelleted at 80 x g for 10
min, and thereafter redispersed in cell culture medium containing
10% FCS. In the case of enzymatic dispersion, great care should be
taken in carefully removing the enzymes completely, because theleft-
over enzymes will digest the ECM. Otherwise, the cells can be washed
once in medium.
5. At this point, the suspension should be microscopically examined to
rule out the possibility that either, because of the nature of the selected
tissue or as a consequence of the isolation procedure, it consists of cell
debris or broken up fibers. (Trypan blue exclusion assay is optional at
this point, but with growing experience, the experimentor will be able
to do without.)
6. The suspension is then plated onto ECM-coated dishes and left to
attach for at least 6 h, but preferably overnight.
7. Cells are maintained at 37°C in a humidified atmosphere supple-
mented with 5% CO, for the meningeomas and 8% CO, for the
gliomas. Meningeomas are kept in Dulbecco’s Modified Eagle me-
dium (DME-H21) containing 10% fetal calf serum, 2 mM glutamine, 1
mM pyruvate, gentamycin (25 pg/mL), and amphotericin B (2.5 pg/
mL). Gliomas are maintained in MEM-Earle with the same supple-
ments. Splitting and passaging of the cells is performed with a 0.05%
[;i’v.) trypsin/0.04% (w/v) EDTA solution in PBS/Ca2+, Mg2+ free
.
8. Careful inspection of the cells on the next day is mandatory. In ex-
treme cases, the cultures are confluent, and the debris from red blood
cells and unattached fragments needs to be rigorously washed off. In
any case, it is advisable to change the medium as early as possible (see
Note 9).

4. Notes
4.1. Materials and Methods
1. Eyes can be conveniently transported in a plastic bag and do not need
special prerequisites for transportation, If they are obtained by
slaughterhouse workers at inconveniently early hours, they should be
Human Brain Tumors on ECM 121

kept in a refrigerator until the experimentor comes to pick them up.

Although cornea1 endothelial cells are rather sturdy, it is advisable to
use the eyes within 6-8 h after the animals have been killed.
2. The cornea is a very tough tissue, and tightly gripping forceps should
be used to lift the cornea after it is excised halfway. Also, the scissors
should be sharp.
3. After the brains are obtained from the slaughterhouse, they should
rapidly be frozen. If they are kept at -20°C, they should be put to 4°C
the evening before the isolation to turn them gradually into a waxy
consistency. It is necessary for the brains to be at least frozen once be-
fore they are used because that facilitates the homogenization.
4. For gliomas and meningeomas alike, the decision has to be made as to
whether the tissue is to be digested with enzymes or only mechani-
cally dispersed. Enzymatic dispersion has the advantage that a more
homogeneous suspension is obtained, which can immediately be
seeded for experiments in which comparable aliqots are to be used.
Purely mechanical dispersion has the advantage that only the loosely
connected tumor cells is obtained, vessels are left intact, and endo-
thelial contamination is minimized. Even from small fragments, only
those cells that will attach and migrate are mostly tumor cells.
IMPORTANT NOTE: If the decision is made to use mechanical dis-
persion by trituration with a tissue culture pipet, these pipets should
have a smoothened mouth. Some manufacturers have the tips cut, re-
sulting in a very rough mouth that will break up the cells. The decision
has to be made, however, on a case-to-case basis depending on the
softness of the tissue and the experimental intentions. Medulloblas-
tomas as well as pituitary adenomas and pineocytomas are usually
very soft and will readily disperse by repeated trituration through a 5-
mL tissue culture pipet alone. Acoustic neuromas are often very
tough and fibrous, so that enzymes help in releasing the tumor cells.
In any case, the tissue is first minced with scissors and then triturated.
Even in the face of abundant tumor material, one should refrain from
using too much tissue.
5. The addition of FGF is not quantified because the amount to be added
will depend on the preparation. The protein peak eluting from the
CM-Sephadex C-50 contains both acidic and basic FGF. The shape of
the peak will depend on the geometry of the column used and the ex-
act running conditions. Also, the final FGF concentration will depend
on the width of the peak and when it is decided to stop the collection.
An aliquot should be taken and tested at different dilutions to find out
the maximally effective dilution.
122 Westphal et al.

6. bECM will have a different appearance from batch to batch. It may

look homogeneous and bubbly as in Fig. 1, but it may also look more
like a webb or show pronounced ridges, depending on how long the
cornea1 endothelial cells had been propagated in culture, what the ini-
tial density was, and how long it took for the cells to become confluent.
7. It has to be noted, however, that the cells from tumors undergo
changes in vitro and may alter their growth pattern. In this cell line,
there was a gradual change in culture morphology towards a more
disorderly arrangement (Fig. 2). Before that event, the cells would lyse
just like the bovine cornea1 endothelial cells and leave a matrixbehind.
Thereafter, it appeared as if the adhesive forces between the cells were
stronger than the adhesion of the cells to their support. After the cells
had lysed and the tECM was to be separated from the cellular debris,
the whole matrix detached from the plastic in one piece.
8. It is advisable to obtain a preliminary histological diagnosis by frozen
section. Viable tissue is selected on the basis of surgical experience.
This means that, in those cases where the experimentor is not the
surgeon, a close collaboration with the surgeon is advisable. Frozen
sections from the immediate vicinity of the removed specimen are
helpful.
9. One should also note whether the ECM is intact or lysed. In 116 cases
of initiated gliomas, we noted that, in some of the initial cultures of
high-grade malignant tumor cells, a slow lysis of the matrix occured.
This could have been because of the release of proteolytic enzymes
after cell death or because of fibronectin degrading enzymes, which
are known to be present in membranes of transformed cells (17). In
these cases, the ECM was barely visible in the phase contrast micro-
scope 1 d after plating. Nevertheless, its beneficial effect was already
visible at that point.

4.2. Notes on the Applications of ECM

4.2.1. Meningeomas
Meningeomas are usually benign epithelial lesions that are regularly
homogeneous and of rather tough consistency. They are often highly vas-
cularized.
Meningeomas usually plate very well even without bECM. In our
hands, there were three cases from a total number of 60, however, that did
not plate on plastic but did so on ECM. Among them was the only men-
ingeoma in our series that was derived from a patient with v. Reckling-
hausen’s disease. In general, there was very little difference in the cellular
Human Brain Tumors on ECM 123

morphology between the coated and uncoated dishes in the primary cul-
ture of meningeomas. Nevertheless, the cultures seeded on ECM were
confluent several days prior to the cultures platedonuncoated dishes. This
is because of a better initial plating efficiency. Based on our experience,
there are two major applications for ECM in the culture of meningeomas:
First, those laboratories concerned with cytogenetic investigations of tu-
mor cells prefer to do karyotype analysis as early as possible. If the cells are
initiated on bECM, several days may be gained in reaching a point at which
experiments are possible. Second, the bECM is very useful in reducing the
requirements for serum and, therefore, can be used in those experiments
that require low serum or serum-free culture conditions.
4.2.2. Gliomas
Gliomas have to be divided into low-grade gliomas and malignant gli-
omas, and those cases of either that are operated upon as recurrencies after
prior surgery, radiation, or other treatment. Gliomas, especially the more
malignant lesions, are often very heterogeneous with numerous different
cell lines within one tumor (18). It has to be assumed that this heterogeneity
is the primary basis for the therapeutic difficulties that are intrinsic to gli-
oma management. Only the understanding of this heterogeneity and the
investigation of the different growth-control mechanisms present in one
glioma will lead to an advancement of specific therapies.
4.2.2.1. Pvimavy cultures. In contrast to meningeomas, we found great
differences in the behavior of primary glioma cultures on plastic and
bECM. Three recent examples from our series of 116 cases may illustrate
the different advantages, which are also documented in Fig. 3.
Case 1 is derived from a tumor that, in frozen section, seemed to be an
astrocytoma grade 2, but the definitive histological diagnosis classified it
as a mixed oligodendroglioma/astrocytoma B according to Ludwig. Equal
aliquots of mechanically dispersed tumor tissue were seeded onto either
plastic or bECM-coated flasks. There were clearly more single cells at-
tached on bECM. In addition, these cells had extensive communications
with each other by means of an elaborate fiber network. After several days
in culture, it was obvious, however, that there was very little mitotic
activity and the cells came almost to a rest. In this case, early experimen-
tation, i.e., assessment of cell lineage by immunocytochemistry with glial
cell markers or chromosome analysis while the cells were still proliferat-
ing, would have been impossible without bECM coating.
Case 2 is derived from a tumor in which histologically a glioblastoma
with beginning sarcomatous changes was diagnosed. This highly malig-
nant lesion did much better on bECM, allowing rapid expansion and early
124 Westphd et al.

Fig. 3. Phasecontrast micrographs from three different examples of cells after initiation
on bECM (upper row) and on plastic (bottom row). (A) Primary cultures from NCE-
G109x (mixed oligodendroglioma/astrocytoma grade B according to Ludwig). (B) Pri-
mary cultures from NCE-Clll, a glioblastoma with sarcomatous changes. (C) Primary
cultures from NCE-112, a gemistocytic astrocytoma grade 3.

experiments such as growth-factor receptor assays, immunostaining, and

karyotyping.
Case 3, which is morphologically distinct from the other two cases, is
derived from an astrocytoma in malignant transition. There is a good
amount of attachment also on plastic, but the amount of attached single
cells is much higher on bECM, thus permitting rapid change of medium,
for example, the day after plating. This increases the content of tumor cells
and minimizes the possibility of stromal contamination, which would be
derived from normal cells growing out of small clumps of undispersed
tissue.
4.2.2.2. Established cell Eines. If growth factors are to be investigated
under defined conditions or if conditioned media from cells with SUS-
petted autocrine growth factor secretion are to be collected, the useof ECM
greatly facilitates these experiments. Any kind of ECM, because of its fi-
bronectin content, lowers the requirements for serum in the culture media.
bECM influences the proliferation of established cell lines. However, the
effects of bECM canbebothpositive andnegativewithrespect to prolifera-
Human Brain Tumors on ECM 125

tion. Figure 4 describes two examples from an experiment in which cells

from established glioma cell lines were seeded onto uncoated and bECM-
coated multiwell dishes. One day after plating, the cells were changed to
serum-free conditions and counted. They were then counted on the days
indicated. It is obvious that, in one of the cases, the cells grew better on the
bECM and better on plastic in the other case. It has to be noted, however,
that, despite a better rate of proliferation in some cases, the cells look more
spherical, and it can be observed that, after longer times in culture, the cells
will detach and start floating. In some cases in which cells fromestablished
lines were used, the cells growing on bECM had a more differentiated ap-
pearance and grew more strictly in monolayer than those on plastic. The
comparative growth rates from such experiments indicate that there are
cell lines in which thereis rather a proliferation-inhibitingeffectexerted by
the bECM.
4.2.2.3. Applications of tECM. Production of an ECM by tumor cells is
rather the rule than the exception. In our laboratory, we were interested in
the effects of this tECM on primary cultures of gliomas. It was interesting
to note that, in contrast to the nongeometrically arranged bECM, there was
a profound effect of the tECM on the culture morphology (Fig. 5). Initial
plating in proliferation experiments as well as growth rates was also
affected (Fig. 6). In experiments in which the response of cultured glioma
cells to epidermal growth factor, platelet-derived growth factor, and bFGF
were evaluated, the quality of the responses were not different from
plastic, but the overall cell number was increased (Fig. 7).
For the reasons mentioned above, the experience with tECM is limited
to the use of few batches, but is included because it illustrates the broadness
of matrix biology. One aspect that is very well demonstrated in the com-
parison of the two ECMs is the drastically different geometrical arrange-
ment. Using immunostaining of bECM with an antiserum obtained by
immunizing rabbits against whole bECM, the disorderly structure can be
visualized. Using antisera against fibronectin, one can see the structure of
tECM that in itself has some kind of direction (Fig. 8).
4.2.3. Other Neurosurgically Obtained Tumors
4.2.3.2. Pituitary adenomas. The vast majority of these are soft and
should always be dispersed mechanically. Adenomas that are tough are
rare. Even with mechanical dispersion, fibroblast overgrowth will be a
problem after 3-6 wk of continuous culture because pituitary adenoma
cells do not proliferate. Cells usually attach within a few hours and can
then be used in endocrinological experiments in which test substances are
added. For this the medium needs to be completely changed for the hor-
126 Westphal et al.

I NE-GU

‘i.i,
NUgER OFlJ*-fS

Fig. 4. Two examples of established glioma cell lines showing the diversity of effects
when growing in serum-free medium on plastic (b and a) and on bECM (a and c). NCE-
G62 was derived from a glioblastoma in a child andNCFA84 from an adult glioblastoma.
The cell counts were obtained at the days indicated from quadruplicates detached from
multiwell plates with trypsin.

6- NCE-G22 6- NCE-GL7
5-
L-

1 I
1 2 3 4
NUMBER OF DAYS IN CULTURE

Fig. 5. Comparative growth curves of cells growing on tECM (closed squares) and
plastic (open squares). The cells were seeded at a density of 10,000 cells/well in tissue
culture multiwell dishes and counted on the indicated days after tryptic detachment.
Human Brain Tumors on ECM 127

Fig. 6. Phase contrast micrographs from primary cultures from two different gliomas
that were initiated on bECM, tECM, and untreated pIastic dishes. Upper row, NCE-G74B,
derived from an oligodendroglioma grade D; bottom row, cells from NCE-G70, a gemis-
trocytic astrocytoma grade 3.

mone assay. The adenoma cells are firmly attached and will not be re-
moved with the medium.
NOTE THE EXCEPTION: Adenomas derived from ACTH-secreting
adenomasinpatientswithCushing’sdiseaseattachmuchmoreslowlyand
sometimes it takes up to two days. The time it takes for the cells to attach
is correlated with the degree of the patient’s cortisol excess (9).
In addition to the rapid attachment, the most striking effect of bECM
in comparison to plastic is the rapid and obvious changes in cellular mor-
phology after exposure to hypothalamic releasing factors, such as GRF for
HGH-secreting cells and CRF or vasopressin for ACTH-secreting cells
WO).
4.2.3.2. Acoustic Neuromas. Acoustic neuromas are difficult to culture
even on ECM. If they are soft and cellular, they do attach well to plastic
support, too.
4.2.3.3. Pined Tumors. Pineal tumors are very heterogeneous. Those
that belong to the glial cell lineage behave as elaborated above. Two ger-
128 Westphal et al.

NCE- G66

tECM X x X AX

0 0 0
PLASTIC -0
CONTROL EGF FGF PDGF

Fig. 7. Comparison of the effects of maximally proliferation-stimulating doses of

epidermal growth factor (10 nM), platelet derived growth factor (35 ng/mL), and bFGF
(80 pg/mL), which were added every second day to serum-free cultures derived from a
human glioma on plastic and tECM.

minomas, however, showed very good attachment to bECM and did very
poorly on plastic. Morphologically, they had great resemblance to pitu-
itary adenomas.
4.2.3.4. Medulloblastomas: This particular type of tumor has been dif-
ficult to maintain in culture, even on bECM. In three cases in which the tis-
sue was very soft and only mechanically dispersed, there was little attach-
ment of thevery small tumor cells that presented as little clusters. Only one
case, which was a very malignant recurrence (presenting 6 wk after the
macroscopic total removal of the tumor), attached very well to bECM,
allowing immunocytochemical analysis, but it failed to continue to prolif-
erate in vitro.
4.2.3.5. Gangliogliomus: Of these very rare tumors, only one was tak-
en into culture, showing a remarkable heterogeneity. This was much more
visible on bECM. In this case, the influence of bECM on the expression of
cell-type specific markers with and without addition of B-Br-CAMP was
studied, showing that bECM did not change the immunostaining pattern.
Human Brain Tumors on ECM 129

Fig. 8. (A) Immunostaining of bECM with an&bECM antiserum and (B) immuno-
staining of tECM with antifibronectin antiserum. The staining was visualized by FITC
labeled second antibody.

4.3. General Notes

1. Cornea1 endothelial cells can be frozen and stored in liquid nitrogen.
They can be cultured for many in vitro passages and will continue to
produce matrix material. It has to be noted, however, that the bECM
gets thinner and less homogeneous with increasing age of the cornea1
endothelial cells. ECM can be stored at 4OCfor up to 1 yr. It can also
be dried after all salt has been washed off.
2. ECMs are different in their chemical composition, depending on the
tissue from which they are derived. One major aspect that has to be
taken into account if autocrine mechanisms of cell proliferation are in-
vestigated is the sequestration of growth factors such asbFGF into tie
130 Westphal et al.

ECM (29). They are then slowly released and act as an intrinsic mito-
gen. Furthermore, heparin binding growth factors appear to be stabi-
lized in the presence of ECM that contains heparin sulfate.

Acknowledgments
The authors are grateful to D. Gospodarowicz, in whoselaboratory the
introduction to ECM biology and the methodological training was ob-
tained. This study was supported by the Heinrich Bauer Stiftung fur Hirn-
tumorbiologie and the Deutsche Forschungsgemeinschaft. We express
our thanks to the departmental operating room staff for their collaboration
and appreciate the photographical work of S. Freist.

References
1. Bottenstein, J. E. and Sato, G. H. (1980) Fibronectin and polylysine requirement for
proliferation of neuroblastoma cells in defined medium. Exp. Cell Res. 129,361-366.
2. Terranova, V. I’., Aumalley, M., Sultan, L. H., Martin, G. R., and Kleinman, H. K.
(1986) Regulation of cell attachment and cell number by fibronectin and laminin. I.
Cell. Physiol. 127, 473-479.
3. Couchman, J.R.,Hook,M.,Rees, D.,andTimpl,R. (1983)Adhesion, growth,and ma-
trix production by fibroblasts on laminin substrates. J. Cell. Biol. 96,177-X%
4. Varani, J., Carey, T. E., Fligiel, S. E. G., McKeever, P. E., and Dixit, V. (1987) Tumor
type-specific differences in cell-substrate adhesion among human tumor cell lines.
lnt. J. Cancer 39,397-403.
5. Gospodarowicz, D., Vlodavsky, I., and Savion, N. (1981) The role of fibroblast
growth factor and the extracellular matrix in the control of proliferation and differ-
entiation of cornea1 endothelial cells. Vision Res. 21,87-103.
6. Gospodarowicz, D., Cohen, D., and Fujii, D. K. (1982) Regulation of cell growth by
the basal lamina and plasma factors: Relevance to embryonic control of cell prolifer-
ation and differentiation, in Cold Spring Harbor Conference on Cell Proliferation Vol. 9
(Cold Spring Harbor Laboratories, Cold Spring Harbor, New York), pp. 95-124.
7. Wichia, M. S., Lowrie, G., KohnE., Bagavandoss,P., and Mahn,T. (1982) Extracellu-
lar matrix promotes mammary epithelial growth and differentiation in vitro. Proc.
Natl. Acad. Sci. 79,3213-3217.
8. Pavelic, K., Bulbul, M. A., Slocum, H. K., Pavelic, Z. I’., Rustum,Y. M.,Niedbala, M.
J., and Bernacki, R. J. (1986) Growth of human urological tumors on extracellular Ma-
trix as a model for the in vitro cultivation of primary tumor explants. Cancer Res. 46,
3653-3662.
9. Westphal, M., Jaquet, P., and Wilson, C. B. (1986) Long-term culture of human corti-
cotropin-secreting adenomas on extracellular matrix and evaluation of serum-free
conditions. Acta Neuropath. 71,142-149.
10. Westphal, M., Hahn, H., and Liidecke, D. K. (1987) Culture of dispersed cells from
human pituitary adenomas from acromegalic patients on extracellular matrix. In
Growth Hormone, Growth Factors, and Akromegaly (Liidecke, D. K. and Tolis, eds.),
Raven Press, New York, pp. 125-133.
Human Brain Tumors on ECM 131

11. Westphal, M., Hansel, M., Brunken, M., Kijnig, A., Kdppen, J. A., and Herrman, H.
D. (1987) Initiation of primary cell cultures from human intracranial tumors on ex-
tracellular matrix from bovine cornea1 endothelial cells. Exp. Cell. Bid 55,152-163.
12. Ill, C. R. and Gospodarowicz, D. (1982) Factors involved in supporting the growth
and steroidogenic functions of bovine adrenal cortical cells maintained on extracel-
lular matrix and exposed to a serum-free medium. J. Cell. Physiol. 113,373-384.
13. Vlodavsky, I., Lui, G. M., and Gospodarowicz, D. J. (1980) Morphological appear-
ance, growth, behavior, and migratory activity of human tumor cells maintained on
extracellular matrix versus plastic. Cell 19,607-616.
14. Rutka, J.T. (1986)Effects of extracellular matrix proteinson the growth and differen-
tiation of an anaplastic glioma cell line. Can. J. Neurd. Sci. 13,301306.
15. Weiner, R. I., Bethea, C. L., Jaquet, I’., Ramsdell, J.S.,and Gospodarowicz, D. J. (1983)
Culture of dispersed anterior pituitary cells on extracellular matrix. Methods in En-
zymol. 103,287.
16. Gospodarowicz, D., Cheng, J., Lui, G. M., Baird, A., and Bohlen, P. (1984) Isolation
of brain fibroblast growth factor by heparin sepharose affinity chromatography:
Identity with pituitary fibroblast growth factor. Pm. NatI. Ad. Sci. 81,6963-6967.
17. Chen, J. -M. and Chen, W. T. (1987) Fibronectin degrading proteases from the mem-
branes of transformed cells. Cell 48,193-203.
18. Shapiro, J. R. (1986) Biology of gliomas: Chromosomes, growth factors, and onco-
genes. Seminars in Oncology 13,4-15.
19. Vlodavsky, I., Folkman, J.,Sullivan, R., Fridman, R., Ishai-Michaeli, R., Sasse,J.,and
Klagsbrun, M. (1987) Endothelial cell-derived basic fibroblast growth factor: Syn-
thesis and deposition into subendothelial extracellular matrix. Pm. Naf2. Acud. Sci.
84,2292-2296.
 Chapter 12

Embryonic
Rodent Brain Cells in Culture

Narayan R. Bhat
1. Introduction
Because of its cellular complexity and regional heterogeneity, the
mammalian central nervous system is not easily amenable for experi-
mental analysis. The study of the developing brain becomes even more
complicated because of the differential growth rates of different parts of
the brain. Primary culture techniques involving dissociation of discrete
regions of the developing brain into component cells offer an excellent
opportunity to study the regulation of growth and differentiation of neu-
ral cells and to investigate their biochemical, morphological, and physio-
logical behavior under well-defined conditions. Methods are available
now to isolate and grow individual classes of neural cells (i.e., neurons,
astrocytes, oligodendrocytes), thereby enabling one to study the cellular
behavior at the individual level and to uncover the nature of cell-cell
interactions that presumably govern cell differentiation.
The procedure widely used to prepare primary cultures of a mixed
neuronal-glial population involves dissociation of embryonic rat/mouse
brain. The cultures can be maintained either as suspension cultures or as
surface cultures. Both systems recapitulate normal developmental events,
and thus, have proved highly useful for the study of neuronal and glial
differentiation (1,Z). Described here is a version of the method to set up
surface-adhering primary cultures of embryonic rodent brain cells.

133
134 Bhat

2. Materials
1. Instruments: Surgicalinstruments such asmicroforceps (straight and
curved; sharp and blunt-pointed), dissecting scissors, nylon mesh (20
and 78 pm sieve size), glass pestle (Belco).
2. Animals: Pregnant (15 d) rats/mice.
3. Dulbecco’s Modified Eagle’s Medium (DMEM): To prepare DMEM,
dissolve the powdered medium, sodium bicarbonate (3.7g/L), glucose
(5g/L) an d an tbi io t ic-antimycotic mixture (10 mL/L) representing
100 U/mL penicillin, 100 pg/mL streptomycin, and 250 ng/mL am-
photericin B in an appropriate vol of distilled, deionized water. After
adjusting its pH to 7.4, the medium is filter-sterilized.

3. Method
Before the start of the experiment, make sure all the instruments,
nylon mesh, paper towels, and small pieces (-1 in*) of Whatman No. 1
paper are sterilized. Warm up the medium to room temperature.
Turn on the laminar flow hood and light the burner. Place the sterile
surgical instruments in a beaker containing 70% alcohol. Also, keep 100 x
20-mm dishes containing DMEM ready.
1. Removal of the embryos: Outside the hood, sacrifice the pregnant rats
by CO,-gassing. Wipe the abdominal area with 70% alcohol. Remove
the embryos by Caesarean section and place them in a culture dish
containing DMEM. The rest of the procedureis to be carried out under
the hood.
2. Dissection: Force out the fetuses from the sacs with a gentle cut with
the scissors. Individual embryos are wiped off by gently rolling on a
sterile paper towel. Dissect the head and place it upright in an empty
dish. Then, holding the front end down with a forceps, remove the
entire brain with another forceps through a midline incision in the
skull made with scissors. The brains thus taken out are collected in a
dish containing DMEM. The cerebral hemispheres may then be sepa-
rated from the rest of the brain with a cut across the mid brain.
3. Removal of meninges: This is done under a dissecting microscope
kept in the hood. Place the cerebral tissue on a piece of sterile paper
(Whatman No. 1) kept in a dish (60 x 15 mm) containing4 mL DMEM.
Looking through the microscope, one can visualize the blood vessels
embedded in the meningeal membrane covering the tissue. Remove
the meningeal membrane by probing through it with pointed forceps.
Rodent Brain Cells 135

If done correctly, the entire membrane should peel off in one piece.
4. Dissociation: Cerebra are then dissociated mechanically. The tissue
is first smoothed in a culture dish with the aid of a glass pestle, then
suspended in medium containing 15% FCS, and dispersed by passing
the suspension through a sterile pipet several times. Further dissocia-
tion is carried out by sequential passages through nylon meshes of
pore sizes 78 pm and 20 pm. The resulting suspension of dissociated
cells is centrifuged at a low speed (150 x g, 5 min) to obtain the pellet.
The pelleted cells are washed twice with the medium and suspended
at a density representing 2.5 x 106cells/mL. Final plating density: 3
x 105/cm2 surface area.
5. Maintenance of cultures: The initial change of medium is after 34 d
of culture (when most cells have attached to the surface of the dish/
flask) and twice a week thereafter. DMEM containing 10% FCS is used
for maintaining the cultures.

4. Notes
1. The dissociated embryonic brain cells have a tendency to reaggregate.
The aggregates of cells thus formed attach to the surface of the dish
and grow fibrous processes. By about a week in culture, an extensive
network of neuritic processes can be seen covering the surface of a
layer of proliferating astroglial precursors (Fig. 1). Cell proliferation,
mostly accounted for by glial cells, peaks around 6 d in culture (3).
This proliferative period is followed by cell differentiation, as indi-
cated by an increase in neuronal and glial properties (e.g., 4-7). The
time course of cell differentiation corresponds closely to the in vivo
pattern, and thus, the culture system serves as an excellent model of
the developing brain.
2. It should be pointed out that, under the culture conditions described
here, neurons start degenerating after the second week. Thus, if the
experiments call for prolonged neuronal survival, the following modi-
fications may be incorporated.
a. Use of antimitotic agents: An overgrowth of nonneuronal
(glial) cells in these cultures seems to be the main reason for
neuronal decay. In order to prevent this, investigators have
used antimitotic agents such as fluorodeoxyuridine and cytosine
arabinoside (8). Thus, for example, the cultures are set up as de-
scribed (the seeding cell density may be reduced to one-half).
The cultures grown for 6-8 d are treated with the inhibitor at a
concentration of 1 x 10-5M for 24 h.
136 Bhat

Fig. 1. Phase contrast micrograph of a 7-d-old primary culture established from 15-d
embryonic rat cerebral tissue.

b. It has been noticed that supplementation of the growth me-

dium with a combination of fetal calf serum and horse serum
results in a reduced @al growth, thereby promoting neuronal
survival. better still, Honegger et al. (9) have described a serum
free defined medium [based on the N-2 medium originally
described by Bottenstein and Sato (ZO)] that seems to support
the long-term growth of mixed neuronal-glial cell populations.
A similar growth medium has been used to study neuronal dif-
ferentiation in surface-adhering cultures of fetal mouse hypo-
thalamic cells (22).
3. Pettman et al. (22) have described a technique to grow pure neuronal
cultures from embryonic chick brain. In this procedure, dissociated
cells from 8-d embryonic brain are seeded onto polylysine-coated cul-
ture dishes. Glial cells fail to proliferate under these conditions for two
reasons: (a) the early embryonic age chosen when gliogenesis is mini-
mal and ( b) the growth of few glial cells present in the initial cell sus-
pension is inhibited by the polylysine substrate.
Alternatively, theproceduredescribed by Hansonet al. (13) separates
neurons and glial cells based on (a) the differential adhesion properties of
neurons and nonneuronal cells to a collagen substrate and (b) the capacity
Rodent Brain Cells 137

of neurons to form homotypic aggregates. Briefly, culture flasks are coated

with specially treated rat tail collagen. The cell suspension is plated on
collagen-coated flasks and placed in an incubator on an intermittantly
agitating platform. The nonneuronal cells progressively attach to the
substrate while the neurons form homotypic neuronal aggregates in the
supernatant. These aggregates are then triturated with a Pasteur pipet and
replated on a fresh flask coated with collagen. The original flasks are rinsed
free of the remaining neurons with fresh medium to obtain glial enriched
cultures.
The above procedures originally used to prepare culture of embryonic
chick brain cells can easily be adapted for embryonic rat/mouse brain cells.
It is a fact that the cellular composition of primary brain cell cultures varies
with the age of the donor tissue. Thus, cultures established from embry-
onic cortex generally results in a mixed neuronal-glial population, whereas
cultures prepared from neonates produce nearly pure glial populations.
The latter system can also be used as a source of pure cultures of astrocytes,
oligodendrocytes (14,15), and bipotential glial progenitors (16).

References
1, Seeds, N. W. (1973) Differentiation of aggregating brain cell cultures, in Tissue Cul-
fure of the Nervous System (Sato, G., ed.), Plenum Press, New York, London, pp. 35-
53.
2. Sensenbrenner, M. (1977) Dissociated Brain Cells in Primary Cultures, in Cell, Tissue,
and OrganCuIturesinNeurobioIogy(Fedoroff,S., andHertz,L.,eds.), AcademicPress,
New York, San Francisco, London, pp. 191-213.
3. Bhat, N. R., Shanker, G., and Peiringer, R. A. (1983) Cell proliferation ingrowingcul-
tures of dissociated embryonic mouse brain: Macromolecule and ornithine decarb-
oxylase synthesis and regulation by hormones and drugs. J. Neurosci. Res. 10,221-
230.
4. Yavin, A. and Yavin, E. (1977) Synaptogenesis and myelinogenesis in dissociated
cerebral cells from rat embryo on polylysine-coated surfaces. Exp. Bruin Res. 29,
137-147.
5. Bhat, N. R., Subbarao, G., and Pieringer, R. A. (1981) Investigations on myelination
in vitro: Regulation of sulfolipid synthesis by thyroid hormone in cultures of dis-
sociated brain cells from embryonic mice. J. Biol. Chem. 256,1167-1171.
6. Bhat, N. R., Shanker, G., and Peiringer, R. A. (1981) Investigations on myelination in
vitro: Regulation of 2’,3’-cyclic nucleotide 3’ phosphohydrolase by thryoid hormone
inculturesof dissociated braincellsfromembryonic mice. J. Neurochenz. 37,695701.
7. Bologa, L., Bisconte, J. C., Joubert, R., Marangos, P. J., Derbin, C., Rioux, F., and
Herschkowitz,N. (1982)Accelerateddifferentiationofoligodendrocytesinneuronal
rich embryonic mouse brain cell culture. Brain Res. 252,129-136.
8. Godfrey, E. W., Nelson, P. G., Schrier, B. K., Breuer, A. C., and Ransom, B. R. (1975)
Neurons from fetal rat brain in a new cell culture system: a multidisciplinary anal-
ysis. Bruin Res. 90,1-21.
138 Bhat

9. Honegger, I’., Lenoir, D., and Favrod, I’. (1979) Growth and differentiation of fetal
brain cells in a serum-free defined medium. Nature 282,305308.
20. Bottenstein, J. E. and Sato, G. H. (1979) Growth of a rat neuroblastoma cell line in a
serum-free supplemented medium. Proc. Natl. Acad. Sci USA 76,514517.
11. Faivre-Bauman, A., Rosenbaum, E., Puymirat, J., Grouselle, D., and Tixier-Vidal, A.
(1981) Differentiation of fetal mouse hypothalamic cells in serum-free medium. Dev.
Neurosci. 4,118-129.
12. Pettman, B., Louis, J. C., and Sensenbrenner,M. (1979) Morphologicaland biochem-
ical maturation of neurones cultured in the absence of glial cells. Nature (Lord .) 281,
378-380.
13. Hanson, G. R., Iversen, P. L., and Partlow, L. M. (1982) Preparation and partial char-
acterization of highly purified primary cultures of neuronal and non-neuronal
(glial) cells from embryonic chick cerebral hemispheres and several other regions of
the nervous system. Deu. Brain Res. 3,529~545.
14. McCarthy, D. K. and devellis, J. (1980) Preparation of separate astroglial and oligo-
dendroglial cell cultures from rat cerebral tissue. I. CeEZBio2.85,890-902.
15. Saneto, R. I’. and devellis, J. (1985) Characterization of cultured rat oligodendro-
cytes proliferating in a serum-free, chemically defined medium. Proc. Nafl. Acad. Sci.
USA 82,3509-3513.
26. Behar, T., McMorris, F. A., Novotny, E. A., Barker, J. L., and Dubois-Dalq, M. (1988)
Growth and differentiation properties of O-2A progenitors purified from rat cere-
bral hemispheres. J. Neurosci. Res. 21,168-180.
 Chapter 13

Human
Thyroid Epithelial Cells

D. W. Williams
and David Wynford-Thomas
1. Introduction
The thyroid gland contains two populations of epithelial cells,of quite
different embryological origin and function. Only the major component -
the follicular cells-will be considered here. The minor C cell population,
which forms only a few percent, can be ignored for the purposes of primary
culture.
The follicular cells are organized in the intact gland into discrete func-
tional units or follicles that consist of spheres lined by a single layer of
epithelium and filled with the secretory product of the lining cells-thyro-
globulin. The follicles are embedded in a vascular stroma.
As with other epithelial tissues, almost all published methods for
obtaining primary cultures have employed some form of proteolytic di-
gestion combined with mechanical disaggregation. Earlier workers used
trypsin in protocols similar to those used for embryo fibroblast preparation
(1,2,3). These yield mainly single-cell suspensions for which it is difficult
to obtain epithelial cells free of fibroblast contamination (4). A major im-

139
140 Williams and Wynford-Thomas

provement can be achieved by the use of collagenase (and/or dispase),

which unlike trypsin does not digest the intercellular junctional complexes
that hold together the apical poles of the follicular cells. As a result, pro-
vided that digestion and disaggregation are not excessive, the majority of
the follicular cells can be released in the form of follicles rather than single
cells (5), which can be separated from the single-cell fibroblast component
by differential sedimentation (6) or filtration (7). In addition, it has been
shown that, as expected, avoidance of trypsin improves the viability of the
resulting cultures and their sensitivity to hormones that act through cell-
surface receptors, notably thyroid stimulating hormone (TSH) (8). Finally,
this method of preparation allows for culture of follicular cells as intact
follicles in suspension as well as in monolayer. The follicular cell is highly
polarized both ultrastructurally and functionally (for example with re-
spect to hormone and growth factor receptor distribution), and it has been
shown that in monolayer its polarity is reversed, the apical rather than the
basal surface being in contact with the environment (9). This can result,
particularly in confluent monolayers, in a loss or diminution of many re-
sponses to extracellular agents, for example, the stimulation of adenylate
cyclase and iodide uptake by TSH (IO), and of proliferation by epidermal
growth factor (II). Culture in suspension as intact follicles avoids such
artifacts, since the normal polarity is maintained. Indeed, the thyroid of-
fers a rare opportunity to culture a glandular epithelium without disrupt-
ing its in vivo organizational unit, a condition that can never be entirely met
with branched ductular-acinar structures such as breast or pancreas.
All digestion protocols require that the tissue be first minced into
small enough pieces to permit rapid diffusion of enzyme to all areas.
Various strategies have been subsequently employed for digestion and
disruption. Most simply, the tissue fragments can be incubated for several
hours at 37OC and then disrupted in one step, for example by pipeting, fol-
lowed by harvesting of the released cells (12). This inevitably results, how-
ever, in some cells being exposed to enzyme for too long and hence to an
excessive proportion of single cells and poor viability. It is preferable to
fractionate the procedure by harvesting at multiple intervals during the
digestion (gentle mechanical disruption being applied throughout), so that
cells (in the form of follicles) are removed from the enzyme as soon as pos-
sible after their release. One variation of this approach that we have found
(13) to give reproducibly high yields of human follicles with minimal fibro-
blast contamination (~0.1%) is detailed here. Similar methods have been
applied to cultures of dog (5,14), sheep (E), pig (I 6), rat (6), and human (27)
thyroid, although not all including a fractionated harvesting protocol.
Thyroid Epithelial Cells 141

2. Materials
1. Hanks’ Balanced Salt Solution (HBSS): This is a simple inorganic salt
solution buffered for use in atmospheric CO,. We use the modified,
calcium- and magnesium-free formulation supplied by Flow Labora-
tories.
2. Collagenase: 200 U/mg. Store powder at 4OC.
3. Dispase: neutral protease from Bacillus polymyza, Grade II. Store
powder at 4°C.
4, Enzyme mixture: Dissolve 40 mg of collagenase plus 60 mg of dispase
in 60 mL of HBSS. Filter through 0.45 pm nitrocellulose filter to ster-
ilize. Prepare fresh each time; keep on ice until use.
5. Nylon mesh: 200 pm nominal pore size. Sterilize by autoclaving.
6. RPM1 1640 medium: The cells have also been grown successfully in
Dulbecco’s Modified Eagle’s Medium (DMEM). We routinely add
penicillin (50 U/mL) and streptomycin (50 pg/mL).
7. FCS (Fetal Calf Serum)
8. NBS (Newborn Bovine Serum)
9. Acridine orange and ethidium bromide: Prepare solution containing
O.lpg/mL of each in HBSS. Store in dark bottle at 4°C. These are
mutagenic and gloves should be worn when handling them.
10. Agar: high gel strength. Make 100 mL of a 2% solution in double-dis-
tilled water. Boil in a bottle with loosened lid to dissolve and sterilize.
Store at 4°C. Reboil before each use.
11. Trypsin/EDTA solution: 0.05% (w/v> trypsin plus 0.02% (w/v>
EDTA in Dulbecco’s calcium- and magnesium-free phosphate-buff-
ered saline. Dilute from 10x stock (stored at -20°C). Store working sol-
ution at 4OC (up to 1 mo).
12. Dimethylsulphoxide (DMSO): Tissue-culture tested product, sup-
plied sterile. Store at room temperature in dark. Store freezing mix-
ture (DMSO:NBS/ 1:4) at 4OCin dark.

3. Methods
The following procedure is summarized in Fig. 1.
1. Human thyroid tissue is best obtained as freshly excised surgical
material-most of our samples come from thyroid lobectomies per-
formed for removal of “cold” nodules (seeNote l), which usually pro-
vide at least 1 g of histologically normal gland. The tissue sample is
transported to the tissue culture laboratory in HBSS on ice (but speed
142 Williams and Wynford-Thomas

Follicles single
ceils & debris

Add to enzyme soiution

2oog
2 min

Remove
enzyme
solution
Resuspend
pellet in RPM

Allow to
sediment

Single ceils

debris foiiici& plus debris

Fig. 1. Protocol for preparation of human thyroid follicles.

Thyroid Epithelial Cells 143

at this stage is not particularly vital, since we have found no dele-

terious effect from storage at 4OCfor up to 1 h).
2. Prior to enzyme digestion, the tissue is first minced with sterile scalpel
blades as finely as possible (to around 2 mm cubes). The pieces are
then washed four times in 15 mL of HBSS at 4OCin a universal con-
tainer to remove as much blood as possible. (Each wash consists of
manual shaking followed by sedimentation under gravity- there is
no need to centrifuge.)
3. The HBSS is removed and replaced by 10 mL of enzyme mixture con-
sisting of collagenase (130 U/mL) and dispase (1 mg/mL) in HBSS.
The container is incubated in a static water bath at 37°C with intermit-
tent gentle agitation by manual shaking for 20 s every 15 min.
4. After the first hour of incubation, the first “fraction” is harvested (see
Note 2). The enzyme mixture (containing suspended follicles and
single cells) is removed (leaving the undigested tissue fragments at the
bottom of the universal container-see Note 3), and transferred to a
sterile 15 mL centrifuge tube. FCS is added to a final concentration of
0.5% to neutralize proteases, and the tube placed on ice. Fresh enzyme
mixture (prewarmed to 37OC) is added to the undigested tissue
fragments and the incubation continued.
5. After a further 30 min, the next fraction is harvested exactly as above.
The process is repeated until no more tissue remains (usually after
3 h).
6. While digestion is continuing, the fractions already collected are cen-
trifuged at 200g for 2 min in a swinging bucket rotor to recover the cells
and follicles. The supernatant (enzyme mixture plus FCS) is removed
and the pelleted cells thoroughly resuspended by pipeting in 10 mL
RPMI medium at 4°C. They are then allowed to resediment under
gravity for 0.5-l h (seeNote 4). The supernatant (containing single
cells) is then discarded and the sedimented follicles from all fractions
pooled in fresh RPMI.
7. After the final fraction has been processed, the pooled follicles are fil-
tered through a 200 pm nylon mesh to remove large debris such as
partially digested tissue fragments. Finally, the follicles are washed
twice by centrifugation (2OOg,2 min) and resuspended in l-4 mL of
RPMI.
8. To obtain a crude cell count and viability index, 15 PL of follicle sus-
pension is mixed with 15 PL of acridine orange/ethidium bromide
solution and viewed in a hemocytometer under phase contrast (for
cell/follicle counting > and UV light for assessment of viability (see
Note 5).
144 Williams and Wynford-Thomas

Fig. 2. Phase contrast of human thyroid follicular cells in monolayer culture, showing
an island of cells derived from a single follicle 3 d after attachment (350x).

9. Prior to plating out or freezing (seeNote 6), the follicles are first in-
cubated overnight (in a standard humidified 5% CO, atmosphere at
37OC)at a density of around lo5/cm2in RPMI (still serum-free) on Petri
dishes coated with 2% agar (seeNote 7) to prevent cell attachment.
This maneuver allows most of the follicles that will have been rup-
tured during extraction to reform and leads to death of most of the
small proportion of single cells (particularly fibroblasts) that may still
be present.
10. Monolayer Culture (Fig. 2): Follicles areplatedon standard tissue-cul-
ture grade plastic Petri dishes in RPMI medium supplemented with
10% FCS. Plating efficiency is poor at low density; densities are best
kept above 104/cm2. Initial doubling times average 2-3 d. At conflu-
ence, the cultures can be passaged (split ratio up to 1 in 8) by standard
trypsinization schedules using trypsin/EDTA solution (see Materi-
als). As expected for a primary culture, senescence occurs after a finite
number of divisions-in 10% FCS between 4 and 8. The cells flatten,
mitoses are no longer observed, and after several weeks, the culture
eventually dies. Escape from senescence has never been observed in
these human cultures.
Thyroid Epithelial Cells 145

Fig. 3. Phase contrast of thyroid follicles in suspension culture (350x).

The cells retain the tissuqxxific control of growth by thyroid-

stimulating hormone (TSl3, which can be demonstrated by stimula-
tion of PI-II-thymidine labeling index in serum-free medium by TSH,
in the presence of insulin-like growth factor-l (IGF-1) (18).
11. Suspension (Follicle) Culture (Figs. 3 and4): Cells (follicles) are plated
at 5 x W-5 x 105/mL in RPMI supplemented with the desired serum
concentraton. The cells will survive even in serum-free medium for
many days (-95% viability after 1 wk). In 10% FCS, proliferation is
very limited in comparison to monolayer cultures, [3Hl-thymidine up-
take falling rapidly during the first few days. This is probably partly
the result of a simple physical restriction on cell spreading in the fol-
licle, but there is evidence also that the loss of proliferative capacity
correlates with the development of follicular inversion, which occurs
over the first week of culture (seeNote 8). In the absence of serum, in-
version is delayed, and clear proliferative responses to addition of
pure growth factors (e.g., TSH plus IGF-1) can be observed for at least
a week. This culture system was the first to demonstrate the growth-
stimulatory effect of TSH on human follicular cells in vitro (23).
146 Williams and Wynford-Thomas

Fig. 4. Electron micrograph of follicles in suspension. The follicles were collected by

centrifugation, fixed in 1% glutaraldehyde, and embedded in L.RWhite resin. Sections
were stained with uranyl acetate/lead citrate. O3OOOx).

4. Notes
1. Cold nodules are most often a benign follicular adenoma; the sur-
rounding unaffected tissue is usually clearly distinguishable macro-
scopically from the tumor, but the normality of the samples used for
Thyroid Epithelial Cells 147

tissue culture is checked retrospectively by histological examination

of tissue sections. Laryngectomies provide an alternative source of
surgical samples. Some workers have successfully employed post-
mortem material, which is of course available in much larger quantity
w.
2. Very little release of cells occurs during the first hour of enzymic di-
gestion. Most follicles are recovered in the next four successive 30-
min incubations. The whole procedure usually lasts 3-4 h.
3. The undigested tissue pieces sink rapidly compared to the released
follicles, so that on standing after a brief shake the fragments will all
have settled in less than 1 min, leaving the follicles still in suspension.
4. At step 6, the tube is allowed to stand for a much longer period than
in Note 3 (longer than 30 min). In this time, the follicles will sediment
even under unit gravity in a medium of sp. gr. 1, permitting adequate
separation from the single cells (including fibroblasts), which remain
in suspension. This is made possible by the large size of human fol-
licles; with rat follicles, for example, a Percoll separation medium is
needed (6).
5. Viable cells exclude the ethidium bromide and are stained green by
the a&dine orange. Dead cells take up the ethidium bromide and con-
sequently show an orange fluorescence. The quantitation of cell num-
ber is subject to observer error, because of the difficulty of distinguish-
ing individual cells in some follicles. This has not proven a problem
provided absolute figures are not important and that all observations
for a given experiment are made by the same operator. If higher ac-
curacy is needed, a follicle sample can be treated with trypsin to pro-
duce a single-cell preparation.
6. Freezing: The follicles are first incubated for 24 h in suspension to al-
low reformation of closed follicles, since this has been found to im-
prove subsequent viability after thawing. The follicles are harvested
from the agar-coated dish by washing with RPMI medium into a 15-
mL centrifuge tube, and then centrifuged at 200g for 2 min. The pellet
is resuspended in ice-cold RPM1 containing 10% NBS at a maximum
cell concentration of lO’/mL. An equal vol of a 1:4 (v/v> mixture of
DMSO and NBS is added, and the cells thoroughly dispersed by gentle
pipeting. The suspension is aliquoted into freezing ampules and al-
lowed to cool slowly to 70°C inside a tightly closed polystyrene box.
The next day, the ampules are transferred to liquid nitrogen for pro-
longed storage. We have routinely obtained viability of >80% after 3
yr in store.
7. The agar is prepared by boiling a 2% solution in double-distilled
148 Williams and Wynford-Thomas

water. The hot solution is dispensed rapidly to avoid cooling, using

a disposable plastic pipet. An excess is applied (around 5 mL for a 9-
cm Petri dish) and allowed to cool for 1 min. Most of the agar is then
removed, leaving a thin coating.
8. Follicle inversion is a well-recognized process in serum containing
media (20). The mechanism is unclear, but the result is an inside-out
structure in which the polarity of the cell is reversed, the microvillous
border coming to lie on the outside in contact with the medium, ac-
companied by corresponding changes in position of the Golgi appara-
tus, nucleus, and intercellular junctions. Perhaps as a result of inap-
propriate fluid and electrolyte transfer, the lumen of the follicle
becomes greatly distended, and the lining epithelium very attenu-
ated. Such cells show a marked alteration in both functional and pro-
liferative responses (13,21).

Acknowledgments
We are grateful to the Cancer Research Campaign and to the Welsh
Scheme for Development of Health and Social Research for grant support.

References
1. Pastan,I. (1961)Certain functions of isolated thyroid cells.Endocrinology 68,924-931.
2. Tong, W., Kerkof, P. R., and Chaikoff, I. L. (1962) Iodine metabolism in dispersed
thyroid cells obtained by trypsinization of sheep thyroid glands. Biochim.Biophys.
Ada. 60,1-19.
3. Fayet, G., Pacheco,H., and Tixier, R. (19701Sur la reassociation in vitro des cellules
isolees de thyroide de port et la biosynthese de la thyroglobuline. 1. Conditions
pour I’induction des reassociations cellulaires par la thyreostimuline. Bull. Sot.
Chim. Biol. 52,299-306.
4. Murphy, A., Mothersill, C., O’Connor, M. K., Malone, J.F.,Cullen, M. J.,and Taaffe,
J. K. (1983) An investigation of the optimum culture conditions for a differentiated
culture of sheep thyroid cells. Ada Endocrinol. 104,431-436.
5. Rapoport, B. (1976) Dog thyroid cells in monolayer tissue culture: Adenosine 3, S-
cyclic monophosphate response to thyrotropic hormone. Endocrinology 98,1189-
1197.
6. Smith, I’., Williams, E. D., and Wynford-Thomas, D. (1987) In vitro demonstration
of a TSH-specific growth desensitizing mechanism in rat thyroid epithelium. Mol.
Cell. Endocrinol. 51,51-58.
7. Nitsch, L. and Wollman, S. H. (1980) Suspension culture of separated follicles
consisting of differentiated thyroid epithelial cells. Proc.Nat. Ad. Sci.USA 77,472-
476.
8. Stockle,G., Wahl, R., and Seif, F. J.(1981) Micromethod of human thyrocyte cultures
for detection of thyroid-stimulating antibodies and thyrotrophin. Ada Endocrino-
logica 97,369-375.
Thyroid Epithelial Cells 149

9. Chambard, M., Gabrion, J., and Mauchamp, J. (1981) Influence of collagen gel on the
orientation of epithelial cell polarity: Follicle formation from isolated thyroid cells
and from preformed monolayers. J. Cell. Bid. 91,157-X6.
10. Chambard, M., Verrier, B., Gabrion, J., and Mauchamp, J. (1983) Polarization of
thyroid cells in culture: Evidence for the basolateral localization of the iodide pump
and of the thyroid-stimulating hormone receptor-adenyl cyclase complex. J. Cell.
Bid. 96,1172-1177.
11. Westermark, K., Westermark, B., Karlsson, F. A., and Ericson, L. E. (1986) Location
of epidermal growth factor receptors on porcine thyroid follicle cells and receptor
regulation by thyrotropin. EndocrinoZogy118,1040-1046.
12. Stringer, B. M. J., Wynford-Thomas, D., and Williams, E. D. (1985) In vitro evidence
for an intracellular mechanism limiting the thyroid follicular cell growth response
to thyrotropin. EndocrinoIogy 116,611-615.
13. Williams, D. W., Wynford-Thomas, D., and Williams, E. D. (1987) Control of human
thyroid follicular cell proliferation in suspension and monolayer culture. Mol. CeZZ.
Endocrinol. 51,33-40.
14. Roger, P. P., Hotimsky, A., Moreau, C., and Dumont, J. E. (1982) Stimulation by
thyrotropin, cholera toxin and dibutyryl cyclic AMP of the multiplication of differ-
entiated thyroid cells in vitro. MoZ. Cell. Endocrinol. 26,165-176.
15. Westennark, K. and Westermark, B. (1982) Mitogenic effect of epidermal growth
factor on sheep thyroid cells in culture. Exp CeZlRes.138,47-55.
26. Westermark,K., Karlsson, F. A.,and Westermark,B. (1983) Epidermal growth factor
modulates thyroid growth and function in culture. Endocrinology112,1680-1686.
17. Davies, T. F., Platzer, M., Schwartz, A., and Friedman, E. (1983) Functionality of
thyroid-stimulating antibodies assessed by cryopreserved human thyroid cell bio-
assay. 1. CZin. Endocrinol. Metab. 57,1021-1027.
18. Williams, D. W., Wynford-Thomas, D., and Williams, E. D. (1987) Human thyroid
adenomas show escape from IGF-1 dependence for growth. Annales d’EndocrinoZo-
gie 48,82A.
19. Roger, P. I’. and Taton, M. (1987) TSH is a direct growth factor for normal human thy-
rocytes. Annales d’EndocrinoZogie48,158A.
20. Nitsch, L. and Wollman, S. H. (1980) Ultrastructureof intermediate stagesin polarity
of thyroid epithelium in follicles in suspension culture. 1. Cell BioZ.86,875-880.
21. Gartner, R., Greil, W., Stubner, D., Permanetter, W., Horn, K., and Pickardt, C. R.
(1985) Preparation of porcine thyroid follicles with preserved polarity: functional
and morphological properties in comparison to inside-out follicles. Mol. Cell Endo-
crinol. 40,9-16.
 Chapter 14

Preparation of Isolated
Rat Liver Hepatocytes

Bjern Quistorfg John Dich,

and Niels Grunnet
1. Introduction
The bulk volume (about 85%) of the mammalian liver parenchyma is
contributed by the hepatocytes, whereas at least four other types of cells
constitute the remainder (1). Procedures for the isolation of all five cell
types have been described, although not from the same liver. However, the
isolation of hepatocytes has clearly been the most widely used prepara-
tion and has proved extremely valuable for a wide variety of experiments,
spanning fields like pharmacokinetics, drug metabolism, and metabolic
functions of the liver.
Historically two classes of methods have been used for the prepar-
ation of the isolated hepatocytes, either mechanical/chemical methods or
enzymatic methods. The enzymatic techniques in which collagenase
perfusion is applied as the principle of disintegration of the liver as first
described by Berry and Friend (2) is completely dominating the field now
and forms the basis of the method described in detail in this chapter. A
number of symposia proceedings and reviews on the preparation and use
of isolated hepatocytes have appeared (3-B).

151
152 Quistorfi Dich, and Grunnet

An added complication in the use of isolated hepatocytes is the

demonstration in recent years of functional metabolic differences of hep-
atocytes of periportal and perivenous origin (for review, see 7). This
chapter does not take such differences into account; however, a following
chapter describes a method by which preparation of isolated hepatocytes,
enriched in either periportal or perivenous cells, may be accomplished
(this book, Chapter 16).
2. Materials
1. Adult Wistar rats weighing about 200 g were housed at 21OC with
alternating 12-h cycles of light (7 AM - 7 PM) and darkness.
2. Solution I: 118.0 mM NaCl, 4.7 mM KCl, 1.2 mM KH.J?O, and 25 mM
NaHCO,.
3. Solution II: As solution I plus 3 mM CaCl,, 1.2 mM MgSO,, and
collagenase (0.2-0.4 mg/mL).
4. Solution III: As Solution II without collagenase. In order to avoid
micro-floral contamination, all solutions were prepared from indiv-
idual stock solutions 5 times more concentrated. Solutions were equi-
librated with oxygen/carbon dioxide (19:1, vol/vol) for 1 h before use
and kept in a thermostated waterbath at 39OCduring the cell isolation
procedure (Fig. 1). Because of heat loss in the tubing system, the per-
fusate temperature at the liver outlet is 34-35°C.
5. Trypan blue solution: 0.5 g Trypan blue in 100 mL of buffer contain-
ing 140 mM NaCl, 4.7 mM KCl, 1.2 mM KHJ?O, 0.6 mM MgSO,, and
10 mM HEPES, pH 7.4. The solution is stirred for 2 h and filtered.
Aliquots can be preserved at -18OC.
6. Water bath: A conventional thermostated shaking water bath with an
accuracy of & 0.2”C and a stroke amplitude of about 2 cm is used.
7. Oxygenation: It is sufficient to bubble solutions with mounted Pas-
teur pipets connected to the oxygen/carbon dioxide tank (19:1, vol/
vol).
8. A two-channel roller pump (e.g., LKB 2115, LKB, Bromma, Sweden)
is preferable. The pump should be able to give a variable flow rate of
20-40 mL/min.
9. Bubble trap: Since infusion of only small amounts of air is deleterious
to the perfusion, it is highly advisable to mount a simple bubble trap
on the infusion line.
10. Cannula: A glass cannula or a double cannula (19-gage x l-3/4 in
[needle], 16-gage x l-l /4 in [catheter]) can be used. The latter is easier
to handle by unexperienced persons.
Rat Liver Hepatocytes 153

BT

V. V.
PORTA CAVA

L -------w

Fig. 1. Perfusion setup for preparation of isolated rat liver hepatocytes. The three
solutions I, II, and III are kept in a thermostated water bath. Pl, P2 is a two-channel rol-
ler pump. BT bubble-trap. B is a beaker for collection of effluent for recirculation during
collagenase perfusion with perfusate II.

11. Filtration device: A simple but efficient filtration device consisting of

a 250-mL beaker, a Biichner funnel rubber gasket, and a nylon net (100
mesh [160 pm]) is shown in Fig. 2.
12. Centrifuge: A conventional laboratory centrifuge equipped with a
swing-out rotor can be used, provided that low g-values can be
obtained reproducibly.
13. Tubing: Silastic tubing 4 mm/2 mm o/i diameter was used. For
cannulation of the superior caval vein, a 2-mm/1.5-mm o/i diameter
polyethylene catheter is used.
154 Quistorff, Did, and Grunnet

lib MESH
NYLON NET

RUBBER
HEPATOCYTE GASKET
>
SUSPENSION u

250 ml BEAKER

Fig. 2. Simple filtration device. The rubber gasket with the nylon net is pushed toward
the bottom before the cell suspension is loaded on the net. The filtration is carried out by
moving the gasket up and down a couple of times, if necessary,with addition of extra
buffer.
14. Counting chamber: A conventional counting chamber with avolume
of 0.1 PL.
15. Collagenase (type II) is used.

3. Methods
The essential features of the collagenase cell isolation procedure from
rat liver involve five steps as listed in Table 1. Quantitative details of the
procedure described below and in Table 1 refer to a rat weighing 200 g. The
procedure may, of course, be scaled to larger or smaller rats as required.
1. The liver of the anesthetized rat is perfused in situ via a cannulation
of the portal vein. The abdomen is opened widely by a long midline
incision and two perpendicular cuts at the level of the kidneys. The
guts are moved to the left, out of the abdominal cavity, giving free
access to the portal vein and the inf. v. cava.
Rat Liver Hepatocytes 155

Table 1
Schematic Presentation of the Preparation
of Isolated Rat Liver Hepatocytes

Perfusion/ Flow Time Buffer/

Phase incubation mL/min min comment

Phase 1 Calcium re- 35-40 8-12 Solution I

moval Flow-
through
perfusion

Phase 2 Collagenase 20-30 15-25 Solution II

perfusion Recircula-
tion

Phase 3 Rinse-cycle 35-40 2-4 Solution III

perfusion Flow-
through
perfusion
Phase 4 Disintegration - 10 Incubation
of the liver, with shaking
followed by (60-70
incubation strokes/
mid in solu-
tion III

Phase 5 Filtration and - Centrifugation

cell separation 3x5og
x2min

2. Prior to cannulation, ligatures are placed loosely around the v. porta

(ca. 5 mm from the bifurcation to the different lobes), and around the
v. cava inf. above the right kidney.
3. As soon as the cannula is in place in the portal vein and retrograde
bleeding is observed, the ligature is tightened, the perfusion is started
with perfusate I (seeFig. I), and the inf. v. cava is cut open below the
right kidney. The chest cavity is cut widely open, the v. cava sup. can-
nulated through the left atrium, and the ligature around v. cava inf. in
the abdominal cavity is tightened.
4. Establishing the flow-through perfusion to this point usually takes a
trained technician about 44 min. The liver must be homogenously
 Quistorff, Dich, and Grunnet

bleached and absolutely without spots of unperfused areas. If that is

not the case, the liver should be discarded. An additional interval of
8-12 min of perfusion with the calcium-free buffer is required to
accomplish efficient calcium removal. Inefficient calcium removal
will decrease cell yield drastically, since the cells will not separate
properly.
5. The perfusate is switched to perfusate II (collagenase), and recircu-
lation is established (seeFig. 1). The total perfusate volume at this time
amounts to approximately 150 mL.
6. The collagenase perfusion is usually continued for 15-25 min de-
pending upon the appearance of the liver. With experience, it is
possible to judge quite accurately from the appearance of the liver
surface the degree of disintegration. One “semi-objective” indicator
of sufficient collagenase perfusion is when a slight impression made
(with the handle of a pair of tweezers) on the liver surface tends to
remain. During collagenase perfusion, the perfusate starts to leak out
through the liver surface, and in most experiments, 75-100% of the
perfusate will be lost by the time when the collagenase treatment is
sufficient. In a few experiments, it is necessary to prepare additional
collagenase perfusate in order to complete the treatment.
7. The liver is now flushed for 3 min with perfusate III in a flow-through
perfusion.
8. The perfusion is stopped, and the liver cut out and gently dispersed in
a Petri dish with two forks in perfusate III.
9. The dispersed liver is incubated in a conical 1-L flask with a O,/CO,
atmosphere (95/5 ~01%) at 37OCfor 10 min with gentle shaking (60-70
strokes/min).
10. The next step is filtering of the disintegrated liver through a nylon net.
This may be accomplished as shown in Fig. 2. One should not try to
force the liver cell suspension through the filter by mashing with a
spoon or a spatula. Instead, the filter on the rubber ring may be moved
up and down a couple of times and some extra medium may be added
(see Table 1). A successful cell preparation at this stage is usually
characterized by a remainder on the filter that constitutes mostly the
fibrous vascular tree of the liver.
11. The final step consists of a separation of hepatocytes from cell debris
and other smaller cells of the liver. This is accomplished by three low-
speed centrifugations (2 min at 508) in two 30-mL centrifuge tubes.
The hepatocytes will sediment as a dark brown layer. Only these cells
Rat Liver Hepatocytes 157

are useful, whereas debris and other cells, which will remain in the
supernatant, are discarded. The centrifugation is repeated with two
resuspensions of the cell pellet; the last time cells are pooled in a
graduated tube in a medium to which is usually added l-3% bovine
serum albumin. After the last centrifugation, the cells are suspended
in the final buffer to be used in the particular experiment. The cell
concentration here ranges from 2-10% in most experiments (seeNotes
l-6).

4. Notes
1. With the procedure described above one routinely obtains a yield of
2-3 mL of packed cells corresponding to 2-4 x lo8 cells, which is about
50% of the total hepatocyte population of the liver.
2. The most frequently used test of cell viability is Trypan blue staining.
The test is carried out immediately after the final centrifugation by
mixing 50 PL of cell suspension with 50 PL Trypan blue solution. The
cell suspension is allowed to stand for 1 min prior to a lo-fold dilution
with 0.9% NaCl, followed by examination in an hemocytometer (vol
0.1 PL). Several factors are observed in the light microscope at a
magnification power of 40:
a. The number of cells/O.1 PL
b. The number of cells smaller than 10 pm
c. The appearance of the cells, rounded and with or without blebs
d.The number of hepatocytes that do not have blue stained
nucleus. This number in percentages is referred to as the
percentage of viable cells. More than 200 cells should be
counted.
Usually >90% of the cells obtained by the procedure described above
exclude Trypan blue and show normal morphology, i.e., rounded
cells without blebs of a size of 15-25 ym. Less than 5% nonhepato-
cytes are found in the preparation.
The criterion for positive Trypan blue staining is a visible blue
nucleus. Unfortunately, there is no correlation between the per-
centage of stained cells and the metabolic capability of the isolated
hepatocytes (9). Therefore, a number of other viability criteria have
been suggested. Among these, the content of ATP is probably the
easiest. ATI? should be > 2 ~mol/108 cells. Also the capacity for
glucose synthesis, the leak of cytosolic enzymes (LDH, ALAT), or the
 Quistorff, Dich, and Grunnet

response in terms of oxygen consumption upon incubation with

various substrates may be used (3,7-g). In the practical experiment,
however, it is difficult to perform an extensive testing of the isolated
cells prior to the experiment itself, since the cells tend to deteriorate in
terms of metabolic capacity over about 2-3 h. In terms of the viability
test, the bottom line is “that anyone who proposes to work on isolated
hepatocytes should satisfy himself that he can reproduce the max-
imum rates reported in the literature before studying new aspects.
Merely looking at cells- at their shape and at dye exclusion-is not
good enough. Nor is reproducibility of data, because this does not
reveal systematic errors...” as stated by Krebs et al. (9).
3. Inclusion of hyaluronidase in the collagenase-containing perfusate
(2), addition of EDTA to the calcium-free perfusate (IO), the use of
buffers other than Krebs-Henseleit bicarbonate, e.g., Hanks’ solu-
tions (12), higher concentrations of collagenase, inclusion of eryth-
rocytes (12), and ex situ perfusion in thermostated perfusion cabinet
are variations of the described technique, which have been used
successfully, but without significant improvement over the simpler
technique described here. Likewise, oxygenators are superfluous as
long as the perfusion solutions are bubbled efficiently with 95% OJ
5% CO, (vol/vol>. Thus, it appears that many details of the cell
preparation procedure are uncritical and only a few essential factors
may be identified. In our opinion, these are the quality of the liver
perfusion, efficient depletion of calcium, and the presence of calcium
in the collagenase solution.
Some authors recommend inclusion of DNase in the buffer used for
incubation of the digested liver (phase 4, Table I) to prevent clumping
of the cells (23). Normally, however, clumping of the cells is not a
problem.
For the purpose of comparison, it may be desirable to obtain a liver
sample before the preparation of isolated cells. After the perfusion has
been established, the small caudate lobe (next to the right kidney) can
be ligated and cut off without affecting the remaining part of the
procedure (24).
4. Isolated hepatocytes are usually incubated in conical flasks at 37OC at
a concentration of 0.5-10 x lo6 cells/mL, corresponding to approx-
imately 4-70 mg wet wt./mL (15). Shaking of the flasks, 60-80
strokes/min, is necessary. With a gas phase of 95% 0,/5% CO,, the
depth of the suspension should not under these conditions exceed 6
mm in order to ensure proper oxygenation. Flasks equipped with a
Rat Liver Hepatocytes 159

center well improve stirring of the cell suspension. Albumin (l-3%)

is often included in the final suspension medium and in the incuba-
tion medium. Unless fatty acids are included in the medium, there ap-
pears to be no beneficial effect of albumin on metabolic rates or on cell
survival. The incubation time should be kept below 2 h. Isolated hepa-
tocytes may be used for primary cultures (this book, Chapter 15).
5. Isolated hepatocytes in suspension are able to carry out most intra-
cellular processes at rates close to those attainable in the perfused
liver. In some respects, however, isolated hepatocytes appear less
competent. This is the casefor glycogen synthesis, which is nonexist-
ent in isolated hepatocytes unless special conditions prevail, i.e. high
glucose concentrations (>15 mM) or lactate (5 mM), pyruvate (1 mM>
and glutamine (10 mM> (26,27). Isolated hepatocytes are in a state of
marked, negative nitrogen balance, which may be alleviated by the
addition of amino acids to the incubation medium (18).
During preparation, the hepatocytes are depleted of low mol wt in-
termediates (3). Consequently, they are in an oxidized state (a state 2)
(29) withsubstratesupplybeingrate-limiting for therespiratory chain
and synthetic processes. Addition of lactate (5 mM) plus pyruvate (I
mM> appears adequate to establish a physiological NAD-redox level
in the cytosolic and mitochondrial cell compartment and to supply
precursors for, e.g., gluconeogenesis and fatty acid synthesis. Exoge-
nous fatty acids results in a rather reduced NAD-redox state, unless
lactate and pyruvate are added (20). Secretory processes, e.g., the se-
cretion of albumin, occur at a rate comparable to that in perfused liver,
but lower than in vivo (21).
6. The enzyme assays and metabolite assays employed in the evalua-
tion of cell viability and functional capacity may be used directly asde-
scribed, e.g., in the handbooks (22,23).

References
2. Greengaard, O., Fedex-man, M., and Knox, W. E. (1972) Cytomorphometry of devel-
oping rat liver and its application to enzyme differentiation. J Cell Biol. 52,261-X2.
2. Berry, N. M. and Friend, D. S. (1969) High yield preparation of isolated rat liver pa-
renchymal cells. 1. Cell Bid. 43,506-520.
3. Krebs, H. A., Cornell, N. W., Lund, P., and Hems, R. (1974) Isolated liver cells as ex-
perimentalmaterial,inRegulationofHepaticMefuboZism. (Lundquist,F.andTygstrup,
N., eds.), Munksgaard, Copenhagen, pp. 726-750.
4. Seglen, P. 0. (1976) Preparation of isolated rat liver cells, in Methods in Cell Biology,
vol. 19 (Prescott, D. M., ed.), Academic, New York, pp. 29-83.
160 Quistorfi Dich, and Grunnet

5. Tager, J. M.,SGlling,H. D., and Williamson, J. R. (eds.) (1976) UseoflsoIated Liver Cells
and Kidney Tubules in Metabolic Studies (Elsevier, Amsterdam).
6. Seglen, P. 0. (1979) Disaggregation and separation of rat liver cells, in Cell Pop&z-
fions (Reid, E., ed.), Wiley, New York, pp. 25-46.
7. Thurman, R. G., Kauffman, F. C., and Jungermann, K. (eds.) (1986) Regulation of
hepatic metabolism, in Infer- and Infru-Cellular Compartmenfafion (Plenum, New
York).
8. Harris, R. A. and Cornell, N. W., eds. (1983) Isolation, Characterization, and Use of
H~tocytes (Elsevier Biomedical, Amsterdam).
9. Krebs, H. A., Lund, I’., and Edwards, M. (1979) Criteria of metabolic competence of
isolated hepatocytes, in Cell PopuIafions (Reid, E., ed.), Wiley, New York, pp. l-6.
10. Seglen, P. 0. (1972) Preparation of rat liver cells. I. Effect of calcium on enzymatic
dispersion of isolated perfused liver. Exp. Cell. Res. 74,450-454.
II. Berg, T. and Msrland, J. (1975) Induction of tryptophan oxygenase by dexametha-
sone in isolated hepatocytes. Dependence oncomposition of medium and pH. Bio-
chim. Biophys. Acfu. 392,233-241.
12. Quistorff, B., Bondesen, S., and Grunnet, N. (1973) Preparation and biochemical
characterization of parenchymal cells from rat liver. Biochim. Biophys. Acfu. 320,
503-516.
13. Bellemann, P., Gebhardt, R., and Mecke, D. (1977) An improved method for the iso-
lation of hepatocytes from liver slices. And. Biochem. 81,408-415.
24. Quistorff, B. (1986) Gluconeogenesisinperiportal and perivenoushepatocytesof rat
liver, isolated by a new high-yield digitonin/collagenase perfusion technique. Bio-
them. J 229,221-226.
15. Jurin, R. R. and McCune, S. A. (1985) Effect of cell density on metabolism in isolated
rat hepatocytes. 7. Cell. Physiol. 123,442448.
16. Katz, J., Golden, S., and Wals, P. (1976) Stimulation of hepatic glycogen synthesis by
amino acids. PYOC.Nufl. Acud. Sci. 73‘3433-3437.
17. Boyd, M. E., Albright, E. B., Foster, D. W., and McGarry, J. D. (1981) In vitro rever-
sal of the fasting state of liver metabolism in the rat. 1. Clin. Invest. 68,142-152.
18. Seglen, P. O., Gordon, P. B., and Poli, A. (1980) Amino acid inhibition of autophagic/
lysosomal pathway of protein degradation in isolated rat hepatocytes. Biochim. Bio-
phys. Acfu. 630,103-118.
29. Chance, B. and Williams, G. R. (1955) Respiratory enzymes in oxidative phospho-
rylation. III. The steady state. 1. Biol. Chem. 217,409-427.
20. Berry, N. M., Grivell, A. R., and Wallace, P. G. (1980) Energy dependent regulation
of the steady state concentration of the components of the lactate dehydrogenase re-
action in the liver. FEBS Left. 119,317-322.
21. Dich, J. A. and Glud, C. N. (1976) Effect of glucagon on cyclic AMP, albumin metab-
olism and incorporation of l4c-leucine into proteins in isolated parenchymal rat liver
cells. Acfu. Physiol. Scud. 97,457469.
22. Bergmeyer, H. U. (1974)Mefhods of Enzymatic Analysis. 2nd English Ed. (Academic,
New York).
23. Lowry, 0. H. and Passonneau, J. V. (1972) A Flexible System of Enzymatic Analysis
(Academic, New York).
 Chapter 15

Primary Cultures
of Rat Hepatocytes

John Dich and Niels Grunnet

1. Introduction
With the advent in 1969 of the collagenase-perfusion technique for the
high-yield preparation of isolated, differentiated hepatocytes (I), easy es-
tablishment of primary cultures of hepatocytes was made possible. Since
then, this experimental system has been increasingly used in many re-
search fields.
Hepatocytes are anchorage-dependent cells, and cultures can be es-
tablished as confluent, nonproliferating monolayers. This experimental
system is superior to other in vitro liver preparations because of a longevity
of several weeks, as compared to a few hours for perfused liver or suspen-
sions of isolated hepatocytes, making studies of cellular processes that
occur within a time scale of hours or days feasible. Another advantage of
primary hepatocyte culture, as compared to freshly isolated suspensions,
may be that deficiencies of the latter, resulting from damage of the cell
membrane by the collagenase treatment, are resolved during the early
stages of culture (2-4).
Examples of topics that have been studied in primary cultures of hepa-
tocytes are DNA-synthesis (5,6), transcriptional and translational events
including hormone effects (7-g), the regulation of hepatic metabolism,
(ZUJZ), secretory processes, (12-14), metabolism and effects of xenobiotics

161
162 Dich and Grunnet

such as ethanol and drugs (25-17), and hepatotoxicity, including carcino-

genic and mutagenic effects (18). Hepatocyte cultures have also been used
as an activating system in the Salmanellu mutagenicity test (19).
One important consideration when using primary cultures of hepato-
cytes is the phenotypic stability of the cells. Early work with hepatocyte
cultures demonstrated changes in characteristic functions of differentiated
hepatocytes, such as declining activity of liver-specific enzymes (24) and
cytochrome P-450 content (3,15,20,22), increased production of a-fetopro-
tein (22), and enhanced response to @-adrenergic agents (4,23,24). Al-
though much work has been devoted to the establishment of phenotypic
stable cultures, many reports appear without adequate characterization of
the culture system actually used. This aspect will be discussed in more de-
tail in section 4.
In the following sections, a standard procedure for the preparation of
primary monolayer cultures of hepatocytes from adult rats will be de-
scribed, together with some methods for the preparation of samples for
analysis. Methods for preparation of hepatocyte cultures from neonatal
and postnatal rats and from other species, and conditions for proliferating
hepatocyte cultures will not be described in detail, but will be discussed
briefly in section 4.

2. Materials
1. Adult Wistar rats, weighing 200-230 g and starved for 16 h. Animals
are housed at 21°C with alternating 12-h cycles of light (HAM-~IJM) and
darkness.
2. A solution of collagen prepared from rat tails. Six to seven tails are im-
mersed in ethanol (70%) to ensure sterility. Under sterile conditions,
the tails are skinned, and the four ligaments are pulled off and trans-
ferred to a sterile flask with 1 L of acetic acid (0.1%). After stirring for
48 h at 4OC,the suspension is centrifuged (10,OOOgfor 30 min) and the
supernatant is decanted. The collagen solution is kept at 4OCand can
be used for at least 1mo. Coating of Petri dishes is normally performed
the day before isolation of the cells. A collagen solution of 1.0-1.5 mL
is evenly distributed on the dish. Immediately before plating, the dish
is tilted, the solution is carefully sucked off, and the dish is used with-
out further treatment.
3. The composition of the standard hepatocyte culture medium is given
in Table 1. Stock solutions, 50x concentrated, of solutions 1,3a, 3b, 3c,
and 4 can be prepared in water. Preparation of stock solution of folic
Cultures of Rat Hepatocytes 163

Table 1
Hepatocyte Culture Medium, HCM, Standard
Vitamins and glutathione w mg/L
L-Ascorbic acid 99.4 17.5
D-Biotin 4.1 1.0
Choline. HCl 1790.4 250.0
Inositol 10.0 1.8
Nicotinamide 8.2 1.0
Panthothenate, Ca 2.1 1.0
Pyridoxine. HCl 4.9 1.0
Thiamine. HCl 29.6 10.0
Vitamin B,, 0.15 0.2
Glutathione, reduced 48.8 15.0
Folic acid and riboflavin
Folk acid 2.3 1.0
Riboflavin 2.7 1.0
Inorganic salts mM mg/L
CaC& 2H,O 0.82 120.0
MgSO,. 7I-JO 1.99 490.0
NaCl 72.33 4225.2
KC1 2.01 150.0
Na,HPO,. 2H,O 2.11 376.0
wpo4 0.59 80.0
NaHCO, 25.00 2100.0
Carbohydrates
D-Glucose, H,O
Amino acids
L-Alanine 1.80 160.4
L-Asparagine 1.50 225.2
L-Arginine 3.00 522.6
L-Aspartic acid 0.46 61.2
L-Cysteine 0.57 69.1
L-Glutamic acid 0.80 117.7
Glycine 4.00 300.3
L-Histidine 0.51 79.2
L-Isoleucine 0.81 106.3
L-Leucine 1.50 196.8
L-Lysine, HCl 2.45 447.4
(continued)
164 Dich and Grunnet

Table 1 (continued)
Amino Acids mg/L
L-Methionine 0.50 74.6
L-Phenylalanine 1.50 247.8
L-Proline 3.00 345.3
L-Serine 1.00 105.1
L-Threonine 0.99 117.9
L-Tyrosine 1.20 217.4
L-Tryptophan 0.69 140.9
L-Valine 1.00 117.2
L-Glutamine 4.00 584.5
Antibiotics, pH-indicator
Penicillin G 60.0
Streptomycin sulfate 100.0
Phenol red 10.0
acid and riboflavin requires the addition of NaOH. Preparation of
stock solution of amino acids (minus glutamine) requires addition of
150-mL cont. HCl. Portions of 20 mL are stored at -18OC and can be
used for at least half a year.
Preparation of 1L of HCM:
a. Dissolve 20 mL each of solutions 1,2, and 5 in 500 mL of water.
Adjust pH to 7.4 with NaOH (about 25 mmol).
b. Add 20 mL of solutions 3a, 3b, 3c, and 4.
c. Add NaHCO,, L-glutamine, phenol red, and antibiotics as in-
dicated in Table 1.
d. The solution is bubbled with OJCO, (19:1, vol/vol) for 30 min
and pH is adjusted to pH 7.4 if necessary.
e. Add water to 1 L.
4. Stock solutions of oleate (250 mM), palmitate (175 mM), plus linoleate
(75 mM) are prepared by dissolving the fatty acids in dimethyl sulfox-
ide. Appropriate volumes are filled in ampules, sealed under N2, and
stored at -18OC.
5. HCM with albumin and fatty acids: Prepare 1 L of HCM as described
above, and then heat the medium to 37OC. Add, with stirring, 10 g of
fatty acid-free bovine albumin, obtained commercially or prepared
according to ref. 25. Complete solubilization takes 1-2 h. To the warm
albumin-containing medium add slowly, with stirring, 1.0 mL of the
stock solution of fatty acids. Continue stirring at 37OC for 2-3 h and
afterwards at 4OCuntil the next day. Centrifuge the medium in sterile
tubes (10,OOOgfor 30 min).
Cultures of Rat Hepatocytes 165

Other hydrophobic compounds may be solubilized in dimethyl sul-

foxide. Care should be taken to avoid high concentrations of dimethyl
sulfoxide in the medium, as concentrations of l-2% have been re-
ported to have their own effects on cultures of hepatocytes (26,27).
6. Pentothal-Natrium: 0.5 g is dissolved in 10 mL of sterile water before
use.
7. Hibitane: A 0.5% solution in water is prepared before use.
8. Iodide solution containing 50 mg of iodine and 35 mg of potassium
iodide/ml of ethanol.
9. Trypan blue: A solution is prepared as described in Chapter 14 (28).
Portions are stored at -18OC.
10. Penicillin and streptomycin.
Il. Dexamethasone is obtained as Decadron (4 mg/mL) from Merck,
Sharp & Dome, Haarlem, Nederlands and kept at 4°C.
12. Crystalline insulin and glucagon. Solutions are prepared in 40 mM
phosphate buffer, pH7.4 with 1% human albumin after dissolving the
desired amounts of insulin and glucagon in a small amount of diluted
HCl and NaOH, respectively. Aliquots can be stored at -18OC for at
least 1 yr. The water used for preparation of all solutions should be of
high quality as, e.g., triple distilled water.
13. During the isolation of the hepatocytes, sterile technique is necessary.
Glassware, instruments for surgery, tubings, and so on, should be
autoclaved. Alternatively, washing with ethanol (70%) is advisable.
All solutions and media are filtered through 0.45 pm filters. Before
use, media are brought to room temperature and bubbled with sterile
OJCO, (19:1, vol/vol) if pH is too high. After isolation of the cells, all
work is performed in a laminar flow bench.

3. Methods
1. Rats,usually starved for 16h, areanesthesizedwithPentothal-Natrium
(10 mg/lOO g rat), administered intraperitoneally. Care is taken to
avoid noise and harmful distress. The rat is immobilized on the dorsal
side. Before surgical procedures, the ventral side of the animal is
washed thoroughly with Hibitane solution. A midline incision through
the skin of the abdominal cavity is performed. The skin is pulled aside
and the underlying tissue is washed with iodine solution before the
abdominal cavity is opened. Cannulation of portal and caval veins,
perfusion of the liver, incubation, and filtration of dispersed cells are
performed as described in Chapter 14.
 Dich and Grunnet

2. After filtration of the dispersed liver, the cells obtained are transferred
to two 50-mL centrifuge tubes with caps, and centrifuged at ZOg,, for
1 min.
3. The supernatant is discharged, and the cell pellets are gently sus-
pended in a few mL of HCM. Additional medium is added to bring
the volume in the tubes to about 30 mL, and the cells are thoroughly
mixed by inverting the tubes several times. The cells are centrifuged
as described above.
4. The cell pellets are resuspended in a few mL of HCM, and the suspen-
sion is transferred to a graduated 15-mL centrifuge tube. Sufficient
medium is added to bring the volume to about 12 mL, and the cells are
mixed again before centrifugation at 30gm, for 2 min. The amount of
packed cells varies from preparation to preparation and is between 2
and 5 mL.
5. Immediately after centrifugation, the packed cells are suspended and
diluted to 20x the packed volume with HCM.
6. To 0.1 mL of the diluted cell suspension, 0.1 mL of Trypan blue so-
lution and 0.8 mL of Krebs-Henseleit buffer are added. After mixing,
an aliquot is immediately transferred to a counting chamber, and the
number of stained and unstained cells counted. Only cell prepara-
tions in which more than 85% of the cells exclude vital dye should be
used for culturing. Normally, the hepatocytes from starved rats at this
stage of preparation appear spherical with a distinct cell border in the
light microscope.
7. The stock solution of cells is diluted with HCM and horse serum. Dex-
amethasone and insulinare added to final concentrations of 10% (vol/
vol), 1 w and 0.1 @I, respectively. The diluted cell suspension
should contain 0.55 million cells /mL. Normally, the yield from a 200
g starved rat is between 250 and 350 million hepatocytes, which is suf-
ficient for 100-150 culture dishes (60 mm).
8. Cell suspension of 4.5 or 3.0 mL is pipeted into 60- and 35-mm Petri
dishes, respectively, equivalent to application of 120,000 cells/cm2. If
multidishes with 24 chambers are used, the cell suspension is further
diluted to O.Emillion cells/mL. To each chamber is added 1.5 mL cell
suspension, equivalent to 120,000 cells/cm2. Since hepatocytes sedi-
ment rapidly, it is necessary to swirl the suspended cells frequently
during pipeting.
9. After plating, the cells are incubated at 37OC with atmospheric air/
CO, (19:1, vol/vol> for 3 h.
10. At that time, the medium is sucked off in order to eliminate loose cells.
3.0 or 2.0 mL HCM with albumin. fattv acids. dexamethasone (0.1
Cultures of Rat Hepatocytes 167

or lw), insulin (0.1 u.M or 0.01 @I), and glucagon (0.1 nM) are added
to 60- and 35-mm Petri dishes, respectively. Tomultidishes are added
0.5 mL of the same medium to each well. The cultures are normally
changed every second day. With the medium used, the interval be-
tween changes can be prolonged to 3 d. When medium is changed, the
dishes are tilted slightly. The medium is sucked off by using, e.g., a
sterile Pasteur pipete connected to a suction device. Replacement of
medium should be done cautiously to avoid mechanical damage of
the monolayer.
11. Preparation of samples for analyses:
a. Dishes to be used for analyses are initially placed on ice. The
medium is collected and stored at -18OC or -70°C until analysis.
The dishes are afterward tilted and residual medium is sucked
off.
b. For determination of enzyme activities and DNA, the cells are
homogenized in 1.5 mL (60-mm dish) glycyl-glycine buffer
(glycyl-glycine, 25 mM, KCl, 150 mM, MgSO, 5 mM, EDTA, 5
mM, dithiothreitol, 1 mA4, and defatted albumin, 0.2%,pH7.5).
Homogenization by ultrasonication (40 W for 10 s) is per-
formed in the dishes without previous scraping. Unless analy-
sis is performed immediately, homogenates should be frozen
at -7OOC.
c. For determination of the content of cytochromeP-450, medium
is removed as described and 0.8 mL (60-mm Petri dish) phos-
phate buffer (phosphate, O.lM, EDTA, 1 mM, dithiothreitol, 1
mM, glycerol, 20%, and Lubrol PX, 2%) is added to the dishes.
The cells are scraped with a rubber policeman, transferred to
vials, and frozen at -70°C.
d. For determination of metabolites, 0.2 mL perchloric acid (70%)
is added to a 60-mm dish with 3 mL medium. The dishes are
swirled and allowed to stand on ice for 10 min. All material is
transferred to a centrifuge tube and centrifuged. An aliquot of
the supernatant is neutralized and stored at -7OOCuntil analysis.

4. Notes
1. Results should preferably be related to DNA, and the DNAcontent/
culture dish should be reported. The common practice of using pro-
tein content as a measure of cell number may lead toerroneousresul ts,
since the protein content/cell may vary with culture conditions. Fur-
168 Dich and Grunnet

thermore, accurate determination of the cellular protein content is in-

compatible with albumin- and serum-containing media.
2. Cell death may be evaluated by determination of the DNA-content of
the monolayer, since dead cells detach from the substratum and are re-
moved by medium changes. Cell disintegration (cell death or mem-
brane damage) may be quantitated by determination of enzyme leak-
age to the culture medium. Lactate dehydrogenase is well suited for
this purpose because of its high activity and cytosolic localization, but
other enzymes may also beused, e.g., transaminases and the liver-spe-
cific argininosuccinate lyase. Uptake of Trypan blue by cells in culture
is a poor criterion for cell survival, since dead cells detach from the
substratum. Light microscopic observations are difficult to quan-
titate. Counting of cells is possible but cumbersome.
3. Several parameters can and should be applied to evaluate the meta-
bolic competence and the maintenance of hepatocyte-specific proper-
ties during the culture period. It is worth mentioning that the majority
of hepatocyte cultureexperiments have been carried out for a few days
only, and that the choice of parameters of course has to depend on the
purpose of the actual experiments. Some useful parameters are de-
scribed in the following:
a. Hepatocyte-specific enzymes (seealso Chapter 14, this volume).
Glucokinase, pyruvate kinase (L-form), and urea-cycle en-
zymes are hepatocyte specific and should ideally be main-
tained at least at the same activity/cell as in vivo.
b. ATP-level. The level of ATP reflects cell integrity and the meta-
bolic condition of the cells, and ought to be no less than 2.5
pmol/g cells (29,30), corresponding to 1 pmol ATP/mg DNA.
c. Rate of metabolic pathways. Gluconeogenesis (10,21,32,32),
glycogen synthesis (33,34), urea synthesis (35,36), fatty acid ox-
idation (16), and esterification (12,16,37), ethanol metabolism
(38,39,40), and protein synthesis (4) are examples of integrated
metabolic pathways that require the interaction of intracellular
compartments, and that have been shown to proceed at accep-
table rates in hepatocyte cultures.
d. Rate of secretory processes. Hepatocytes in vivo secrete anum-
ber of proteins, which are not produced by other cell types.
Examples are albumin, acute-phase proteins, and some lipo-
proteins. Albumin secretion is often used as a criterion for the
maintenance of specific characteristics. It is, however, not the
most sensitive liver-specific parameter, since other proper ties
Cultures of Rat Hepatocytes 169

are lost before albumin secretion starts to decrease (36). Cul-

tured hepatocytes produce and secrete bile acids (23), however,
there are some indications that these are secreted at a much
lower rate than in vivo as the cultures age, and that primary cul-
tures of hepatocytes may represent a sort of cholestatic state
(40.
e. Response to hormones. Primary cultures of hepatocytes have
been shown to respond to glucocorticoids (4,11,36,40,42,43), in-
sulin (4,36,40,44), glucagon (12,35,36), a-adrenergic agonists
(4,23), triiodothyronine (45,46), and growth hormone (45,47-50),
indicating that receptors for these hormones are present and
that a functional receptor-coupling exist in the cells. Converse-
ly, the estrogen receptor has been reported almost to disappear
within 24 h of culture (49).
4. Hepatocytes in primary culture survive better, when plated on some
sort of substratum rather than directly on tissue culture polystyrene
plastic. On tissue culture plastic, the cells do not spread, and the typi-
cal monolayer of polygonal cells with sharply delineated cell borders
is not established (52). Furthermore, the cells detach rapidly from
naked plastic (52). Primaria plastic (Falcon, Becton Dickinson) has
been claimed suitable for hepatocyte cultures without further treat-
ment (22). Satisfactory results have been obtained with plastic or glass
surfaces coated with rat tail collagen (36,52), with floating collagen
gels (53) or with collagen-gels supported by nylon mesh (25). Pure
substances, e.g., Con A, collagen I, collagen IV, fibronectin, and lam-
inin have also been successfully used as substrata (9,51,52,54,55), al-
though they in some respects, e.g., transcription of liver-specific
genes, appear inferior to rat tail collagen (9). Extracts of extracellular
components, so-called biomatrix, have been reported to have advan-
tages over other substrata (52,56); however, little comparative work
has been carried out. Cocultures of hepatocytes and a rat liver epithe-
lial cell line (57), the latter of which may be considered as a special case
of substratum for the hepatocytes, appear to be superior to hepato-
cytes cultured on plain plastic. No systematic comparison of such coc-
ultures and hepatocyte cultures on collagen-coated plastic has been
reported.
5. Commercially available tissue culture media (seeAppendix) may be
used for hepatocyte cultures, the most commonly used being Leibo-
vitz L15, Dulbecco’s Modified Eagle’s Medium, Williams E, Way-
mouth 752/l, RPM1 1640, and Ham F-12, or mixtures thereof. A num-
170 Dich and Grunnet

ber of modifications havebeen claimed to improve the performance of

hepatocyte cultures. In the following, the various media components
will be discussed shortly:
a. Serum. l-10% is beneficial during the initial adhesion period of
Z-4 h (58). At later time-points, serum addition has been re-
ported to improve cell survival (59) but also to increase the level
of liver-unspecific mRNA’s (60). Satisfactory culture condi-
tions can be established without serum addition, and since ser-
um is not chemically defined, it should be avoided, except for
the initial adhesion period.
b. Substrates. A high concentration of glucose (25 mM) has pro-
ven superior to a lower concentration (5 mM) as judged by the
activity of alcohol dehydrogenase (40). By the same criterion,
high amino acid concentrations (Table 1) are superior to lower
concentrations (40). This effect of amino acids may be related
to the depletion of several amino acids from the usual standard
media within 24 h (24,55,61). A high concentration of amino
acids may enable medium change every second or third day,
only. Substitution of arginine by ornithine may retard growth
of nonparenchymal cells without affecting hepatocytes (62).
However, growth of nonparenchymal cells is normally not a
problem.
c. Fatty acids are not an essential component of media for hepato-
cyte cultures, at least not for short time cultures. They are, how-
ever, a quantitatively important substrate for the liver cells in
vivo, and may be added to culture media as an albumin-fatty
acid complex to attain physiological concentrations.
d. Micronutrients. Although a requirement for metal ions, vita-
mins, or other usual media components (inositol, choline, pu-
rine, and pyrimidine bases) has not been documented for hepa-
tocyte cultures, these substances are often included as a matter
of precaution. Ascorbate (63) and selenium (64) have been re-
ported to improve the maintenance of cytochrome P-450.
e. Hormones. There is a general agreement that glucocorticoids
significantly improve survival and performance of primary
hepatocyte cultures. Dexamethasone is normally preferred
over naturally occurring glucocorticoids because of its resis-
tance to degradation (65). Insulin is also in many respects bene-
ficial to the cultures (4,7,12,14,36,44), and the combined action
of glucocorticoid and insulin is necessary for the expression of
some hepatocyte-specific functions (7,11,50,66). It is therefore
Cultures of Rat Hepatocytes 171

advisable to add both hormones to cultures of hepatocytes. In-

sulin at a concentration of 10-8M is rapidly degraded by hepa-
tocytes with a half-life of about 8-16 h in primary cultures
(unpublished results, 67). A general effect of other hormones
on confluent hepatocyte cultures has not been reported, al-
though they respond to several hormones (seeNote 3).
f. Other medium additions. Addition of 2% dimethyl sulfoxide
to the medium has been reported to maintain albumin secre-
tion in cultures for up to 40 d (26). Addition of glycosamino-
glycans or proteoglycans to the medium has been shown to
enhance the expression of liver-specific genes and to suppress
the expression of tissue-unspecific genes (9). A number of me-
dium additions have been recommended for maintaining the
content of cytochrome P-450 in hepatocyte cultures, e.g., nico-
tinamide (68) metyrapone (69), heme (64), ascorbate (63), selen-
ium (64), or high concentrations of dexamethasone (3). How-
ever, addition of these substances only partially prevents the
decrease of the content of cytochrome P-450.
6. Hepatocytes prepared from fetal rat liver divide in primary culture,
respond to several hormones including androgens and estrogens, and
mature during culture to express adult hepatocyte characteristics (70).
Hepatocytes from adult rat livers proliferate in primary culture, if the
cells are plated at a low density (50,000 cells/cm*), and if the medium
contains glucocorticoid, insulin, and epidermal growth factor (5,60).
Once confluency is reached, the cells stop dividing. Expression of
hepatocyte characteristics in this preparation appears inversely pro-
portional to the rate of cell division (71).
7. The preparation of primary cultures has been described using livers
from a number of other species including humans (72,73), dogs (73),
monkeys (73), mice (74), hamsters (74), guinea pigs (73,74), rabbits
(73,74), chicken embryos (75), and fish (76).

References
2. Berry, M. N. and Friend, D. S.(1969) High-yield preparation of isolated rat liver pa-
renchymal cells. J Cell Bid. 43,506-520.
2. Bissell, D. M., Hammaker, L. E., and Meyer, U. A. (1973) Parenchymal cells from
adult rat liver in nonproliferating monolayer culture. J Cell Bid. 59,722-734.
3. B&sell, D. M. and Guzelian, P.S.(1980) Phenotypic stability of adult rat hepatocytes
in primary monolayer culture. Ann. N.Y. Acud. Sci. 349,85-98.
4. Ichihara, A.,Nakamura, T., and Tanaka, K. (1982)Use of heptocytes in primary cul-
ture for biochemical studies on liver functions. MoI. Cell. Biochem. 43,145-X0.
172 Dich and Grunnet

5. Nakamura, T. and Ichihara, A. (1985) Control of growth and expression of differen-

tial functions of mature hepatocytes in primary culture. Cell Structure and Function
lO, l-16.
6. Tomura, A., Sawada, N., Sattler, G. L., Kleinman, H. K., and Pitot, H. C. (1987) The
control of DNA synthesis in primary cultures of hepatocytes from adult and young
rats: Interaction of extracellular matrix components, epidermal growth factor, and
the cell cycle. J. CeII. Physiol. 130,221-227.
7. Jefferson, D. M., Reid, L. M., Giambrone, M.-A., Schafritz, D. A., and Zern, M. A.
(1985) Effects of dexamethasone on albumin and collagen expression in primary cul-
tures of adult rat hepatocytes. HepatoIogy 5,14-20.
8. Steward, A. R., Wrighton, S. A., Pasco, D. S., Fagan, J. B., Li, D., and Guzelian, P. S.
(1985) Synthesis and degradation of 3-methylcholanthrene-inducible cytochrome
P-450 and their mRNAs in primary monolayer cultures of adult rat hepatocytes.
Arch. Biochem. Biophys. 241,494-508.
9. Fujita, M., Spray, D. C., Choi, H., Saez, J. C., Watanabe, T., Rosenberg, L. C., Hertz-
berg, E. L., and Reid, L. M. (1987) Glycosaminoglycans and proteoglycans induce
gap function expression and restore transcription of tissue-specific mRNAs in pri-
mary liver culture. Hqxltology 7,1S-9s.
10. Probst, I., Schwartz, P., and Jungermann, K. (1982) Induction in primary culture of
“gluconeogenic” and “glycolytic” hepatocytes resembling periportal and perivenous
cells. Eur. 1. Biochem. 126,271-278.
21, Kletzien, R. F., Stumpo, D. J., Kelley, D. S., and Todderud, C. G. (1983) Primary cul-
tures of hepatocytes as a model system for studies on the chronic effects of hormones
on hepatic carbohydrate metabolism, in Isolation, Characterization, and Use of
Hep&cytes (Harris, R. A. and Cornell, N. W., eds.), Elsevier, New York, pp. 77-86.
12, Mangiapane, E. H. and Brindley, D. N. (1986) Effects of dexamethasone and insulin
on the synthesis of triacylglycerols and phosphatidylcholine and the secretion of
very-low-density lipoproteins and lysophosphatidylcholine by monolayer cultures
of rat hepatocytes. Biochem. ].233,151-160.
13. Gebhardt, R. (1986) Use of isolated and cultured hepatocytes in studies of bile forma-
tion, in Isolated and Cultured Hep~focytes (Guillouzo, A. and Guguen-Guillouzo, C.,
eds.), John Libbey, London and Paris, pp. 353-376.
14. Hutson, S. M., Stinton-Fischer, C., Shiman, R., and Jefferson, L. S. (1987) Regulation
of albumin synthesis by hormones and amino acids in primary cultures of rat hep-
atocytes. Am. J Physiol. 252, E291-E298.
25. Sirica,A.E.andPitot,H.C.(1980)Drugmetabolismandeffectsofcarcinogensincul-
tured hepatic cells. Phurmucol. Reu. 31,205-228.
16. Grunnet, N., Kondrup, J., and Dich, J. (1985) Effect of ethanol on lipid metabolism
in cultured hepatocytes. Biochem. J. 228,673-681.
27. Guillouzo, A. (1986) Use of isolated and cultured hepatocytes for xenobiotic
metabolism and cytotoxicity studies, inIsolated and Culfured Heputocyfes (Guillouzo,
A. and Guguen-Guillouzo, C., eds.), John Libbey, London and Paris, pp. 313-332.
18. Michalopoulos, G. K., Strom, S. C., and Jirtle, R. L. (1986) Use of hepatocytes for
studies of mutagenesis and carcinogenesis, in IsoZuted and Cultured Hepatocytes
(Guillouzo, A. and Guguen-Guillouzo, C., eds.), John Libbey, London and Paris, pp.
333-352.
19. Gayda, D. P. and Pa&a, M. W. (1980) Rat hepatocytesinprimary monolayer culture:
A tool for investigating the regulation of carcinogen metabolism, in Microsomes,
Drug Oxidations, and Chemical Carcinogenesk (Coon, M. J., Conney, A. H., Estabrook,
Cultures of Rat Hepatocytes 173

R. W., Gelboin,H. V., Gilette, J. R., and O’Brien, P. J., eds.), Academic,New York, pp.
1165-1168.
20. Guzelian, P. S., Bissell, D. M., and Meyer, U. A. (1977) Drug metabolism in adult rat
hepatocytes in primary culture. Gust~oente~oZogy 72,1232-1239.
21. Fry, J. R. and Bridges, J. W. (1979) Use of primary hepatocyte cultures in biochemi-
cal toxicology. Rev. Biochem. Toxicol. 1,201-247.
22. Sirica, A. E., Richards, W., Tsukada,Y.,Sattler, C. A.,and Pitot,H. C. (1979) Fetal phe-
notypic expression by adult rat hepatocytes on collagen gel/nylon meshes. Proc.
Natl. Acad. Sci. 76,283-287.
23. Tsujimoto, A., Tsujimoto, G., Azhar, S., and Hoffman, B. 8. (1986) Altered respon-
siveness to a- and j3-adreno-receptor stimulation in hepatocytes cultured in defined
medium. Biochem. Phamacol. 35,1400-1404.
24. Refsnes, M. Sandnes, D., and Christoffersen, T. (1987) The relationship between j3-
adrenoreceptor regulation and beta-adrenergic responsiveness in hepatocytes. Eur.
I. B&hem. 163,457-466.
25. Chen, R. F. (1967) Removal of fatty acids from serumalbuminby charcoal treatment.
1. Bid. Chem. 242,173-181.
26. Isom, H. C., Secott, T., Georgoff, I., Woodworth, C., and Mummaw, J. (1985) Main-
tenance of differentiated rat hepatocytes in primary culture. Proc. Natl. Acud. Sci.
USA 82,3252-3256.
27. Baribault, H. and Marceau,N. (1986) Dexamethasone and dimethylsulfoxide as dis-
tinct regulators of growth and differentiation of cultured suckling rat hepatocytes.
I. Cell. Physiol. 129,77&k
28. Seglen, P. 0. (1972) Preparation of liver cells. ExptI. CeZI Res. 74,450X%.
29. Krebs, H. A., Cornell, N. W., Lund, I’., and Hems, R. (1974) Isolated liver cells as ex-
perimental material, in Regulation of HepdicMetuboZism (Lundquist, F. and Tygstrup,
N., eds.), Munksgaard, Copenhagen, pp. 726-750.
30. Cornell, N. W. (1983) Evaluation of hepatocyte quality: Cell integrity and metabolic
rates, in Isolation, Characterization, and Use of Heptocytes (Harris, R. A. and Cornell,
N. W., eds.), Elsevier, New York, pp. 11-20.
31. Quistorff, B., Dich, J., and Grunnet, N. (19861 Periportal andperivenous hepatocytes
retain their zonal characteristics in primary culture. Biochem. Biuphys. Res. Commun.
139,1055-1061.
32. Fleig,W. E.,Noether-Fleig,G.,Roeben,H.,andDitschuneit,H. (1984)Hormonalreg-
ulation of key gluconeogenic enzymes and glucose release in cultured hepatocytes:
Effects of dexamethasone and gastrointestinal hormones in glucagon action. Arch.
Biochem. Biophys. 229,368-378.
33. Dich, J., Bro, B., Grunnet, N., Jensen, F., and Kondrup, J. (1983) Accumulation of tri-
acylglycerol in cultured hepatocytes is increased by ethanol and by insulin and dex-
amethasone. Biochem. J 212,617-623.
34. Fleig, W. E., Enderle, D., Steudter, S., Niither-Fleig, G., and Ditschuneit, H. (1987)
Regulation of basal and insulin-stimulated glycogen synthesis in cultured hepato-
cytes. J Biol. Chem. 262,1155-1160.
35. Gebhardt, R., Bellemann, P., and Mecke, D. (1978) Metabolic and enzymatic charac-
teristics of adult rat liver parenchymal cells in non-proliferating primary monolayer
cultures. Exp. Cell Res. 112,4311141.
36. Dich, J., Vind, C., and Grunnet, N. (1987) Long-term culture of hepatocytes. Effect
of hormones on enzyme activities and metabolic capacity. Heptology 8,3945.
37. Chambaz, J., Guillouzo, A., Cardot, P., Pepin, D., and Bereziat, G. (19861 Essential
174 Dich and Grunnet

fatty acid uptake and esterification in primary culture of rat hepatocytes. Biochim.
Biophys. A& 878,310-319.
38. Dich, J., Vind, C., and Grunnet, N. (1987) Long-term culture of hepatocytes. Ethanol
oxidation and effect of ethanol on enzyme activities and albumin secretion. Alcohol
and Alcoholism Suppl. 1,271-275.
39. Wood, C. K. and Lamb, R. G. (1979)Theeffect of ethanol onglycerolipid biosynthesis
by primary monolayer cultures of adult rat hepatocytes. Biochim. Biophys. A&. 572,
121-131.
40. Grunnet, N., Vind, C, and Dich, J. (1989) Maintenance of alcohol dehydrogenase
activity in long-term culture of hepatocytes from female rat. Alcohol. Clin. Exp. Res.
13, in press.
41. Buscher, H.-P., Friedrich, H., Thorn, H., Schneider, S., Gerok, W., and Kurz, G. (1985)
Hepatocyte primary cultures represent a cholestatic state. Eur. Assoc. for the Study
of the Liver, abstract no. 51.
42. Miyamaga, O., Evans, C., and Cottam, G. L. (1983) The effect of dexamethasone on
pyruvate kinase activity in primary cultures of hepatocytes. Biochim. Biophys. Acfu.
758,42-48.
43. Marceau, N., Baribault, H., and Leroux-Nicollet, I. (1984) Dexamethasone can mod-
ulate the synthesis and organization of cytokeratins in cultured differentiating hep-
atocytes. Can. J. Biochem. Cell Biol. 63,448-457.
44. Katz, N. R., Nauck, M. A., and Wilson, P. T. (1979) Induction of glucokinase by in-
sulin under the permissive action of dexamethasone in primary rat hepatocyte cul-
tures. Biochem. Biophys. Res. Commun. SS,23-29.
45. Gebhardt, R. and Mecke, D. (1979) The role of growth hormone, dexamethasone and
triiodothyronine in the regulation of glutamine synthetase in primary cultures of rat
hepatocytes. Eur. J. Biochem. 100,519-525.
46. Wilson, E. J. and McMurray, W. C. (1983) Effects of hormones on the maintenance
and mitochondrial functions of rat hepatocytes cultured in serum-free medium.
Can. J. Biochem. Cell Biol. 61,636-643.
47. Schaffer, W. T. (1985) Effects of growth hormone on lipogenic enzyme activities in
cultured rat hepatocytes. Am. J. Physiol. 248, E719-E725.
48. Mezey, E., Potter, J. J., and Rhodes, D. L. (1986) Effect of growth hormone on alcohol
dehydrogenase activity in hepatocyte culture. Hepatology 6,1386-1390.
49. Crabb, D. W. and Roepke, J. (1987) Loss of growth hormone-dependent characteris-
tics of rat hepatocytes in culture. In Vitro 23,303-307.
50. Grunnet, N., Vind, C., and Dich, J. (1988) Synergistic effect of insulin and glucocor-
ticoid on albumin secretion in cultured hepatocytes. Acta Physiol. Scund. 133,593,
594.
51. Gjessing, R. and Seglen, P. 0. (1980) Adsorption, simple binding and complex bind-
ing of rat hepatocytes to various in vitro substrata. Exp. Cell Res. 129,239-249.
52. Reid, L. M.,Narita,M., Fujita,M.,Murray,Z.,Liverpool, C.,andRosenberg,L. (1986)
Matrix and hormonal regulation of differentiation in liver cultures, in IsoZafed and
Cultured Hepufocyfes (Guillouzo, A. and Guguen-Guillouzo, C., eds.), John Libbey,
London and Paris, pp. 225-258.
53. Michalopoulos, G. and Pitot, H. C. (1975) Primary cultureof parenchymal liver cells
on collagen membranes. Exp. Cell. Res. 94,70-78.
54. Bissell, D. M., Stamatoglou, S. C., Nermut, M. V., and Hughes, R. C. (1986) Interac-
tion of rat hepatocytes with type IV collagen, fibronectin, and laminin matrices.
Distinct matrix-controlled modes of attachment and spreading. Eur. J. Cell Biol. 40,
72- 78.
Cultures of Rat Hepatocytes 175

55. Sawada, N., Tomomura, A., Sattler, C. A., Sattler, G. L., Kleinman, H. K., and Pitot,
H. C. (1987) Effects of extracellular matrix components on the growth and differenti-
ation of cultured rat hepatocytes. In Vitro 23,267-273.
56. Rojkind, M., Gatmaitan, Z., Mackensen, S.,Giambrone, M.-A., Ponce, P., and Reid,
L. M. (1980) Connective tissue biomatrix: Its isolation and utilization for long-term
cultures of normal rat hepatocytes. J. Cell Bid. 87,255-263.
57. Guguen-Guillouzo, C. and Guillouzo, A. (1983) Modulation of functional activities
in cultured rat hepatocytes, Mol. CeZZ.,BioZ. 53/54,3!?-56.
58. Rubin, K,, Oldberg, A,, Hook, M., and Obrink, B. (1978) Adhesion of rat hepatocytes
to collagen. Exp. Cell Res.117,165-177.
59. Williams, G, M., Bermudez, E.,and Scaramuzzino, D. (1977) Rat hepatocyte primary
cell cultures. III. Improved dissociation and attachment techniques and the en-
hancement of survival by culture medium. In Vitro 13,809-817.
60. Enat, R.,Jefferson, D. M., Ruiz-Opaza, N., Gatmaitan, Z., Leinwand, L. A., and Reid,
L, M. (1984) Heptocyteproliferationintifro: Itsdependence on theuseof serum-free
hormonally defined medium and substrata of extracellular matrix. Proc.N&l. Ad.
Sci. USA 81,1411-1415.
61. Schwarze, P, E., Solheim, A. E., and Seglen, P. 0. (1982) Amino acid and energy re-
quirements for rat hepatocytes in primary culture. In Vitro 18,43-54.
62. Acosta, D., Anufuro, D. C., and Smith, R. V. (1978) Primary monolayer cultures of
postnatal rat liver cells withextended differentiated functions. In Vitro 14,428-436.
63. Bissell, D. M. and Guzelian, P, S. (1979) Ascorbic acid deficiency and cytochrome
P-450 in adult rat hepatocytes in primary culture. Arch. Biochern.Biophys. 192,
569-576.
64. Engelmann, G. L., Richardson, A. G., and Fierer, J. A. (1985) Maintenance and in-
ductionofcytochromeP-45Oinculturedrathepatocytes. Arch. Biochem.Biophys.238,
359-367.
65. Lovell-Smith, C, J. and Garcia-Webb, P. (1986) Glucocorticoids and the isolated rat
hepatocyte. Biochem.Biophys.Res.Commun. 135,160-165.
66. Christoffersen,T.,Refsnes, M., Brsnstad,G.O., Ostby,E., Huse, J.,Haffner, F.,Sand,
T.-E., Hunt, N. H., and Sonne, 0. (1984) Changes in hormone responsiveness and
cyclicAMPmetabolisminrathepatocytesduringprimarycultureandeffectsofsup-
plementing the medium with insulin and dexamethasone. Eur. I+Biochem.138,217-
226.
67. Fleig, W. E., N&her-Fleig, G., Stendter, S., Enderle, D., and Ditschuneit, H. (1986)
Effect of down-regulation and return of insulin receptors on glycogen synthesis in
cultured hepatocytes. Biochim.Biophys.Actu 888,191-198.
68. Paine, A. J.,Williams, L. T., and Legg, R. F. (1979) Apparent maintenance of cyto-
chrome P-450 by nicotinamide in primary cultures of rat hepatocytes. Life Sci. 24,
2185-2192.
69. Paine, A, J.,Villa, P.,and Hockin, L. J. (1980)Evidence that ligand formation is a me-
chanism underlying the maintenance of cytochrome P-450 in rat liver cell culture.
Biochem.J, 188,937-939.
70. Kremers, P. (1986) Drug metabolism in cultured fetal hepatocytes, Is&fed and Cul-
tured Hepafocyfes (Guillouzo, A. and Guguen-Guillouzo, C., eds.), John Libbey,
London and Paris, pp. 285312.
71. Nakamura, T., Yoshimoto, K., Nakayama, Y., Tomita, Y., and Ichihara, A. (1983)
Reciprocal modulation of growth and differentiated functions of mature rat hepa-
tocytes in primary culture by cell-cell contact and cell membranes. Proc.NufZ.Ad.
Sci. USA N&7229-7233.
176 Dich and Grunnet

72. Clement, B., Guguen-Guillouzo, C., Campion, J.-P., Glaise, D., Bourel, M., and
Guillouzo, A. (1984) Long-term co-cultures of adult human hepatocytes with rat
liver epithelial cells: Modulationof albumin secretion and accumulation of extracel-
lular material. Heputology4,373-380.
73. Byard, J. L., Reese,J.A., and Knadle, S.A. (1983) Isolation and culture of hepatocytes
from liver biopsies, in Isolation, Churucterizdion, and Use ofHeputocytes (Harris, R. A.
and Cornell, N. W., eds.), Elsevier, New York, pp. 69-76.
74. Maslamsky, C. J. and Williams, G. M. (19831Isolation and culture of primary hep-
atocytes derived from rat, hamster, guinea pig, and rabbit, in Isolation, Churucteriza-
tion, and Use of Hqutocytes (Harris, R. A. and Cornell, N. W., eds.), Elsevier, New
York, pp. 87-92.
75. Sinclair, J. F., Smith, L., Bement, W. J., Sinclair, P. R., and Bonkowsky, H. L. (1982)
Increases in cytochrome P-450 in cultured hepatocytes mediated by 3- and 4-carbon
alcohols. Biochem. Pharmucol. 31,2811-2815.
76. Campbell, J.W., Aster,P. L., Casey,C. A., and Vorhaben, J.E. (1983) Preparation and
use of fish hepatocytes, i&o&ion, Characterization, and Useofkkputocytes (Harris, R.
A. and Cornell, N. W., eds.), Elsevier, New York, pp. 31-40.
 Chapter 16

Preparation of Isolated
Periportal or Perivenous
Hepatocytes from Rat Liver

Bjern Quistorff
1. Introduction
There are a number of functional, metabolic differences between peri-
portal and perivenous hepatocytes in the mammalian liver resulting from
zonal differences in the activity of several enzymes, and possibly from
morphological differences as well (1). Examples of key enzymes where the
periportal/perivenous activity ratio is >1 are glucose-6-phosphatase (Z),
phosphoenolpyruvate carboxykinase (3), alanine aminotransferase (4),
ornithine carbamoyltransferase (5), and carbamoylphosphate synthetase
(6), whereas enzymes such as glucokinase (7), P-450-dependent hydroxyl-
ation reactions (8), and glutamine synthetase (9) show the opposite zona-
tion. Recent studies on lipogenic enzymes (acetyl-CoA carboxylase, citrate
lyase, and fatty acid synthase) show that, although there is a periportal pre-
dominance in both the fed and fasted male rat, the ca. threefold enzyme in-
duction observed after a fasting-refeeding transition takes place almost
entirely in the perivenous zone (IO).
It is clear that this metabolic heterogeneity that develops a few weeks
after birth is closely related to the microcirculatory pattern of the liver (1).
However, the signals responsible for the heterogenous expression of the

177
178 Quistorff

hepatocte genome as a function of location in the microcirculation are un-

known at present, although the gradients of oxygen (II) and hormones (12)
are implicated. The physiological significance of zonation is largely un-
known. In case of ammonia fixation it has been suggested that urea syn-
thesis located in the inlet part of the microcirculation serves as the main
process of ammonia fixation, whereas the high affinity system of gluta-
mine synthesis located at the very outlet of the liver is considered the back-
up system for ammonia removal (13).
Thus, it is of considerable interest to study isolated periportal and
perivenous hepatocytes separately. Several attempts have been made to
separate isolated hepatocytes according to various physical and chemical
characteristics, but generally with limited success. Recently, however, it
was observed that digitonin perfusion allowed selective destruction of the
periportal or perivenous part of the microcirculation (24), and by applying
this principle in combination with traditional collagenase cell isolation (see
Chapter 14), a method was developed for high-yield preparation of iso-
lated hepatocytes enriched in either periportal or perivenous cells (15,16).
Applying this technique, periportal predominance of the pathways of
gluconeogenesis (16,17) and urea synthesis (17,M) in freshly islolated cells
as well as in primary culture has now been demonstrated. A number of re-
view papers on the topic of metabolic zonation of the liver have appeared
(1,19,20,21).

2. Materials
1. Rats weighing 170-190 g. Animals are anesthetized with pentobarb-
ital ip; 400 pL, 50 mg/kg.
2. Perfusion system: Switching between forward and reverse perfusion
as well as switching between different perfusates may be done man-
ually. We find it advantageous, however, to use a valve system as the
one shown schematically in Fig. 1, e.g., Hamilton miniature valves
(Hamilton Bonaduz AG, CH-7402 Switzerland) type HV 86729 for the
switching of low direction and two type HV 86727 valves for perfusate
switching.
Alternatively miniature pinch-solenoid valves may be used. Our
current perfusion system is based on such 3-way valves (model S305-
01 EG; Sirai Inc., Italy). The perfusion system is convenient to operate
and ensures rapid and precise switching with avery small dead-space
in the tubing system. Furthermore, the perfusion system is readily
computer controlled, e.g., for more complicated perfusion sequences
(22). The system is shown schematically in Fig. 2.
Isolated Hepatocytes 179

Fig. 1. Schematic perfusion setup for preparation of isolated periportal and perivenous
hepatocytes. The liver is connected to the perfusion system via the portal vein (PP) and
the v. cava sup. (PV). A 90° turn of valve A switches the flow direction through the liver
from P+ C to C+P. A 90’ turn of valve B switches the perfusate from I to II.

3. Solution I: 118 mM NaCl, 4.7 mM KCl, 1.2 mM KH,PO, 1.2 mM

MgSO, and 25 mM NaHCO,.
4. Solution II: As solution I plus 3 mM CaCl, and 0.2-0.4 mg/mL colla-
genase.
5. Solution III: As solution I plus 1.2 mM CaCl,.
6. Solution IV: As solution I plus 4 mg/mL digitonin.
All solutions (I-IV) are kept in a thermostated water bath at 39OC
during the cell isolation procedure, constantly equilibrated with oxy-
gen/carbon dioxide (19:1, vol/vol), except for solution IV, which is
gassed only until the addition of the digitonin, a few minutes before
use.
7. Equipment as described in Chapter 14, except that two roller type
pumps (LKB MultiPerpex 2115, Bromma, Sweden) are used. A bubble
trap is mounted on the infusion line between valve A and B, Fig. 1.
8. Digitonin: Commercial preparation of digitonin vary considerably in
terms of purity and in terms of water solubility. We usually use the
SIGMA digitonin. In order to obtain good water solubility, the follow-
ing purification steps are necessary. Five g of digitonin is dissolved in
150 mL boiling methanol (65OC). This solution is cooled to 45OC and
180 Quistorff

RI R2

Pl Pl

8j ~
I ! I I i I

V‘1
4 4 v2

PD

v7

Fig. 2. Schematic perfusion setup based on three-way micro-pinch-solenoid valves.

The liver is connected to the perfusion system as in Fig. 1. The valves V5 and V6 are op-
erated synchroneously, switching the perfusion direction through the liver. 01 and 02 are
two Clarck type of oxygen electrodes mounted on the infusion line and on the out flow
line from the liver, respectively. PD is a pressure difference gage. V7 regulates whether
the eluate from the liver is collected or discarded. Vl-V4 regulate which of four different
perfusates are being infused. When one of these valves is not selected, the perfusate is
automaticallly recycled to the proper reservoir, Rl-R4. N-P4 are different decks on the
multiperpex pump or, alternatively, different pumps, e.g., used for fast switching of flow
rate (27,221.

precipitated by addition of 450 mL room temperature diethyl ether.

The suspension is filtered on a Buckner funnel. After three washings
with diethyl ether, the remnant is collected from the funnel, resus-
pended in water, and freeze-dried for 2-3 d in order to remove all di-
ethyl ether. The freeze-dried digitonin is weighed and redissolved in
water (5 mL) in portions of 150 mg digitonin, which is the amount used
for each experiment. After centrifugation (20000 g x min), the super-
Isolated Hepatocytes 181

natant is freeze-dried overnight and kept in the powder form in the

same tube until use.
Note that digitonin is toxic and easily absorbed through the skin.
Handling should be done carefully, wearing gloves, and weighing of
the fluffy, freeze-dried powder should be done in a hood.
9. Collagenase (type II) is used.

3. Method
The principle of the technique is to destroy selectively cells of either
the periportal or perivenous part of the liver microcirculation with digito-
nin and then perform a standard collagenase cell isolation on the remain-
ing part. Thus, only periportal or perivenous hepatocytes may be prepared
from one liver. Table 1 lists the essential steps of the procedure.
1. A method for establishing the initial perfusion of the liver with Krebs-
Henseleit buffer (solution III) is as described in Chapter 14. After 5-10
min of perfusion, the small caudate liver lobe is routinely ligated, and
cut off (approximately 0.3 g of tissue) and homogenized in ice-cold
buffer (solution III with 5 mM dithiotreitol). This sample (the start
biopsy) is used for analysis of marker enzymes in order to quantitiate
the zonal purity of the prepared cells (seeNote 2).
2. The selective destruction of perivenous cells in order to isolate peri-
portal cells is described first (see also Table 1). Flow direction is
switched to cava+ porta, and flow rate is decreased to 7 ml/min.
3. Thirty seconds later, the digitonin containing perfusate (solution IV)
is switched on and the liver surface is observed. Because of complex-
ing with cholesterol, the digitonin will progressively destroy the plas-
ma membrane of the cells along the path of the microcirculation.
Experience has shown that under the present conditions it is possible
to achieve a largely synchroneous destruction of the sinusoids, which
advance slowly along the path of the perfusate, 50-loo-fold slower
than the linear flow rate (14). This is demonstrated very clearly on the
liver surface by the development over 30-50 s of a pattern of pale,
white spots in the perivenous areas (the dark-brown spots on the nor-
mal liver). The goal is to continue the digitonin perfusion until a reg-
ularly scattered perivenous decoloration pattern on the liver surface
is obtained as demonstrated in Fig. 3B. Under the present conditions,
this involves 3045 s of perfusion.
4. The perfusion is then switched to the calcium-free perfusate (solution
I), and a few seconds later (when the digitonin perfusate in the tubing
182 Quistorff

Table 1
Essential Steps in the Preparation of Periportal or Perivenous Hepatocytes
Flow rate,
mL/min
flow Time
direction min Samples Buffer
1. Initial 30 lo-15 Biopsy Solution III
perfusion P+C
2. Selective destruction 7 0.5-l .o Solution IV
of PV cells C+P
3. Calcium 25 8-12 Eluate Solution I
removal’ P-C first
20 s
_------------------
2a.Selective destruction 7 0.5-l .o Solution IV
of PV cells P+C
3a.Calcium 25 8-12 Eluate Solution I
removal C+,P first
20 s
-------------------

4. Collagenase 25 10-20 Solution II

perfusion P+C (recircu-
lation)
5. Rinse 25 l-2 Solution III
cycle P+C
6. Incubation 10 Solution III
7. Cell Solution III
separation,
washing
‘Steps 2 and 3 are replaced by steps 2a and 3a in order to prepare perivenous cells in-
stead of periportal cells.

between the liver and valve B, Fig. 1, is cleared), the flow direction is
switched to porta+ cava and flow rate increased to ca. 25 mL/min.
Initial eluate (15 s) during this interval is collected in 200 PL 40 mM di-
thiotreitol for marker enzyme analysis (seeNote 2).
5. The procedure for selective destruction of periportal cells in order to
prepare perivenous cells is as described above, except that flow direc-
tions were opposite with periportal infusion of digitonin at a rate of
Isolated Hepatocytes 183

Fig. 3. Zonqxcific decolorization of the liver after digitonin perfusion. A: Peripor-

tal decoloration pattern, obtained after45 s of digitonin perfusion 4 mg/mL, 7 mL/min.
B: The complementary perivenous pattern obtained after 45 s of digitonin perfusion 4
mg/mL, 5.0 mL/min.

only 5 mL/min (see Table 1). The decoloration pattern obtained is

complementary to the pattern described above, now with pale, peri-
portal spots rapidly developing into a network-like pattern, (seeFig.
3A).
6. The switch to the calcium-free buffer (solution I) initiates the stan-
dard hepatocyte preparation, which is described in detail in Chapter
14.
184 Quistorff

4. Notes
1. Cell yields are usually higher for periportal than for perivenous cells,
i.e., 1.3-2.0 x lo6 and 0.8-1.2 x lo6 cells, respectively (X,16). It is essen-
tial to keep the flow rate low during digitonin perfusion. With a flow
rate of 20 mL/min, the cell yield was ca. fivefold lower, even though
the amount of digitonin infused under these conditions was the same
as with the low flow (26). It may improve the yield of cells to include
deoxyribonuclease 20 pg/mL during the incubation of the isolated
cells, step 6 in scheme 1 (15).
In terms of viability, it is our experience that both periportal and
perivenous cell preparations show the same viability of 85-95% un-
stained cells as judged by Trypan blue staining, which is not different
from control cell preparations. Occasionally preparations of low vi-
ability occur. We recommend that such preparations be discarded.
Alternatively, some separation of stained and unstained cells may be
obtained (at the expense of cell yield) by metrisamide gradient centri-
fugation (15). (For other viability tests and metabolic competence
cells, see Chapter 15.)
2. The purity of the cell preparation in terms of contamination of peri-
portal cells to a preparation of perivenous cells and vice versa is deter-
mined by the “zonal specificity” of the digitonin destruction. It is clear
from the pattern of marker enzymes eluted during continuous digi-
tonin perfusion that ideal conditions of equal destruction of all sinu-
soids do not prevail (14), except for the initial eluation, which has a
very high zonal specificity [< 1% contamination (22)]. Thus, a quanti-
tative evaluation of the zonal purity of the final cell preparation is
required. Figure 4 illustrates the problem schematically in a simpli-
fied model of the microcirculatory unit of the liver. Using this model,
which assumes only two populations of hepatocytes, a quantitative
evaluation may be attempted as explained below, based on the activ-
ity of alanine aminotransferase (ALAT), which is known to have a
periportal/perivenous activity gradient of approximately IO-fold
(42).
In cell preparation obtained according to the method described
above, we have measurements of ALAT specific activity:
a. in the start biopsy,
b. in the eluate obtained after digitonin perfusion and
c. in the final cell suspension.
Isolated Hepatocytes 185

Control conditions PP- Digitonin pulse

A B
I I I

PP
Schematic
microcirculatory
unit of liver
PV

PV-Digttonin pulse

BIOPSY. PV- CELLS : PP- CELLS :

6P*6V 1 3P+6V I
TP*;v ---~P+.$v
12= 9

Fig. 4. Model of the liver microcirculation: effect of digitonin. The model assumes only
two populations of hepatocytes, periportal (PI3 and perivenous (PV), each of which is
homogenous with respect to the parameter in question, but not necessarily of the same
size. The figure shows two examples of selectivedestruction with a pulse of digitonin. B
shows periportal destruction, which, with ensuing collagenase cell isolation will give a
cell preparation enriched in perivenous cells by 67% (see numeric example below figure).
Conversly, C shows the case of perivenous cell destruction.

Assuming the two-population model of Fig. 4, the volume fraction

v of periportal cells in the intact liver may be calculated to 0.52 (23).
This value for v may now be substituted in the equation
Abi = (v)App + (1-v)Apv 0)
where Abi is the specific activity of ALAT in the start biopsy, expres-
sed as units/mg of cytosolic protein (see23), while App and Apv is the
specific activity in periportal and perivenous cells respectively (ex-
pressed as units/mg of cytosolic protein). Since in one particular ex-
periment either App or Apv is known (measured in the initial eluate
collected during digitonin perfusion), the other may be calculated.
The values obtained for App and Apv may now be substituted in the
equation
 Quistorff

Ace11 = 0) App + (l-f, Apv (2)

where Ace11 is the specific activity in the actual preparation of peri-
portal or perivenous cells, f the fraction of periportal cells, and con-
versely (l-f> the fraction of perivenous cells.
This relation thus allows calculation of the purity of the cell prepar-
ation (for l-f> under the assumption of a two-population model for the
distribution of the enzyme in question. Typical results using this cal-
culationshowcontaminationofperiportalcellpreparationwith25-35%
perivenous cells and vice versa 20-30% contamination of perivenous
cells with periportal cells. It has proven difficult to improve these fig-
ures significantly. More extensive digitonin destruction does not
seem to be the way to go, since in our experience this only decreases
cell yield without increasing purity significantly.

References
2. Thurman, R. G., Kaufman, F. C., and Jungermann, K. (eds.) (1986) Regulation of
Hqafic Metabolism. Inter- and Infra Cellular Comparfmentation (Plenum, New York).
2. Katz, N., Teutsch, H. F., Sasse,D., and Jungermann, K. (1977) Heterogeneous dis-
tribution of glucose-6-phosphatase in micro dissectedperiportal and perivenous rat
liver tissue. FEBS Left. 76,226-230.
3. Guder, W. G. and Schmidt, U. (1976) Liver cell heterogeneity. The distribution of
pyruvate kinase and phosphoenolpyruvate carboxykinase in the lobule of fed and
starved rats. Hoppe-Seylers Z. Physiol. Chem. 357,1793-1800.
4, Welsh, F. A. (1972) Changes in distribution of enzymes within the liver lobule dur-
ing adaptive increases. 1. Hisfochem. Cytochem. 20,107-111.
5, Mizutani, A. (1968) Cytochemical demonstrationof ornithine carbamoyltransferase
activity in liver mitochondria of rat and mouse. 1. Hisfochem. Cyfochem. 16,172- 180.
6. Gaasbeek-Janzen,J. W., Lamers. W. H., Morman, A. F., de Graf, A., Los, J. A., and
Charles, R. (1984) Immunohistochemical localization of carbamoylphosphate syn-
thetas@in adult rat liver. J. Histochem. Cytochem. 32,557-564.
7. Trus, M., Zawalich, J.,Gaynor, D., and Matschinsky, F. M. (19801 Hexokinase and
glucokinase distribution in the liver lobule. 1.Histochem. Cyfochem. 28,579581.
8. Smith, M. T. and Wills, E. D. (19811 Effects of dietary lipid and phenobarbitone on
the distribution of cytochrome P-450 in the liver studied by quantitative histochem-
istry. FEBS Left. 127,33-36.
9. Gebhardt, R. and Mecke, D. (1983) Heterogeneous distribution of glutamine syn-
thase among rat liver parenchymal cells in situ and in primary culture. Embo 1.2,
567-570.
10. Evans, J. L., Quistorff, B.,and Witters, L. A. (1988) Zonation of hepatic lipogenic en-
zymes identified by dual-digitonin-pulse perfusion. Biochem, J., in press.
Il. Wolfe,D. and Jungermann,K. (1985) Long-termeffectsof physiologicaloxygencon-
centrations on glycolysis and gluconeogenesis in hepatocyte culture. Eur. J. Bio-
them. 151,299-303.
Isolated Hepatocytes 187

12. Probst, I. and Jungermann, K. (1983) The glucagon-insulin antagonism and glu-
cagon-dexamethasone synergism in the induction of PEP-carboxy kinase in cul-
tured rat hepatocytes. Hoppe-Seylers Z. Physiol. Chem. 364,1739-1746.
23. H&rssinger, D. (1983) Hepatocyte heterogeneity in glutamine and ammonia me-
tabolism and the role of an intracellular glutamine cycle during ureagenesis in
perfused rat liver. Eur. J Biochem. 133,418-422.
24. Quistorff, B.,Grunnet, N., and Cornell, N. W. (1985)Digitonin perfusion of rat liver.
A new approach in the study of intra-acinar and intra-cellular compartmentation in
the liver. Biochem. J. 226,289-297.
35. Lindros,K. 0. and Penttila, K. E. (1985) Digitonin-collagenaseperfusion for efficient
separation of periportal or perivenous hepatocytes. Biochem. J 228,575-560.
16. Quistorff, B. (1985)Gluconeogenesis in periportal and perivenous hepatocytesof rat
liver, isolated by a new high-yield digitonin/collagenase perfusion technique.
Biochem. J 229,221-226.
17. Quistorff, B., Dich, J., and Grunnet, N. (1986) Periportal and perivenous hep-
atocytes retain their zonal characteristics in primary culture. Biochem. Biophys. Res.
Commun. 139,1055-1061.
18. P&so,R. A., Penttila, K. E., Suolinna, E. M., and Lindros, K. E. (1986) Urea synthesis
in freshly isolated and in cultured periportal and perivenous hepatocytes. Biochem.
J. 239,263-267.
19. Jungermann, K. and Katz, N. (1982) Functional hepatocellular heterogeneity. Hep-
atology 5385-395.
20. Jungermann, K. (1986)Functional heterogeneity of periportal and perivenous hepa-
tocytes. Enzyme 35,161-180.
21. Gumucio, J.J. and Chianale, J. (1988) Liver cell heterogeneity and liver function, in
The Liver, Biology and Pathology (Arias, I. M., Jakoby, W. B.,Popper, H., andschachter,
D., eds.), Raven, New York, pp. 931-948.
22. Quistorff, B. and Grunnet, N. (1987) Dualdigitonin-pulse perfusion. Concurrent
sampling or periportal and perivenous cytosol of rat liver for determination of me-
tabolites and enzyme activities. Biochem. J 243,87-95.
23. Quistorff, B. (1987) Digitonin perfusion in the study of metablic zonation of the rat
liver: Potassium as an intracellular concentration reference. Biochem. Sac. Transact.
15,361363.
 Chapter 17

Primary Kidney Cells

Mary Taub
1. Introduction
Hormonally defined serum-free media havebeen developed for growth
and functional studies with kidney epithelial cell cultures. Not only can
several established kidney epithelial cell lines (MDCK and LLC-PKl) be
grown in a serum-free environment (1,2), but in addition, primary cul-
tures of kidney epithelial cells can also be grown serum free (l-4). Invest-
igations with primary kidney cell cultures are advantageous for several
reasons. First of all, kidney cells can be grown in vitro from the animal of
choice. Thus, the results of tissue culture studies can be more closely cor-
related with animal studies. Secondly, new tissue culture systems can be
developed that more closely resemble kidney cells in vivo than those
obtained with established kidney cell lines.
This report describes the use of serum-free medium to grow primary
rabbit kidney proximal tubule cell cultures that express proximal tubule
functions. Rabbit kidney proximal tubules are first purified from the renal
cortex by a modification (4-6) of the method of Brendel and Meezan (7,B).
The purified proximal tubules are then put into tissue culture dishes
containing serum-free medium supplemented with insulin, transferrin,
and hydrocortisone. Within the first day of culture, the tubules attach to
the dish, and subsequently, epithelial cells grow out from the tubules.
After one week confluent monolayers are obtained that express a number
of proximal tubule functions (Table 1, Fig. 1). These monolayer cultures
can be used for a wide range of purposes, including viral production,

189
 Taub

Table 1
Properties of Primary Proximal Tubule Cell Cultures
Morphology (4,9,2(I)
Domes
Form polarized monolayers
Adjacent cells form tight junctions
Brush border (although not as elaborated as in vivo)
Transport properties
Na+-dependent sugar transport (4,9)
Na+-dependent phosphate transport (6)
Probenacid-sensitive p-aminohipurate transport (21)
Amiloride-sensitive sodium transport (12)
Hormone responses
Parathyroid hormone-sensitive CAMP production (4)
Enzymes; metabolic properties
Leucine aminopeptidase (4)
Alkaline phosphatase (4)
y glutamyl transpeptidase (4)
Glutathione S-transferase (16)
Angiotensin converting enzyme (10)
Phosphoenolpyruvate carboxykinase (13)
Aerobic metabolism (9)
Hexose monophosphate shunt (9)
Growth properties
Cell growth in serum-free medium (4)
Growth in response to insulin, transferrin, and hydrocortisone (4)
Growth in response to laminin, collagen, and fibronectin (14)
Can undergo two passages

transfection and immortalization, biochemical and physiologic studies, as

well as molecular biologic studies concerning control of expression of
differentiated function.

2. Materials
1. The basal medium used in these studies is a 50/50 mixture of Dul-
becco’s Modified Eagle’s Medium (with 4.5 g/L D-glucose and L-glut-
amine, and without Na+ pyruvate or Na+ bicarbonate), and Ham’s F12
Primary Kidney Cells I91

Fig. 1. Culture of primary proximal kidney tubule cells.

medium supplemented with 15 mM HEPES buffer (pH 7.4) and sodi-

um bicarbonate (brought to a final concentration of 20 mM) (DME/
F12). The medium used for tissue culture studies generally is not
supplemented with penicillin and streptomycin (unless specifically
indicated), since we have observed that theinclusion of these two anti-
biotics in the medium was inhibitory to primary proximal tubule cell
growth. The water to be used for medium preparation is purified first
with a Millipore Reverse Osmosis System, and then with a Milli-Q re-
agent grade water system as previously described (4,5). The medium
is sterilized by pumping it through a 0.22 J.&I Millistak filter.
2. Sterile stock solutions of 5 mg/mL bovine insulin in O.OlN HCl, 5 mg/
mL human transferrin in 30, and 103M hydrocortisone in 100% etha-
nol are used. Insulin and transferrin stock solutions are sterilized
using a 100 mL Nalgene filter apparatus (0.22 micron pore size), and
distributed into sterile plastic tubes in I-mL aliquots. Insulin and hy-
drocortisone are kept at 4°C for up to 2mo, whereas transferrin is kept
frozen and thawed no more than four times.
3. Sterile 0.25% trypsin 1 mM EDTA in Phosphate Buffered Saline (PBS)
(Grand Island Biological Corporation) is used in the trypsinization of
proximal tubule cell cultures.
4. During the perfusion procedure, a 0.5% iron oxide solution (wt/vol)
is used. The iron oxide solution is prepared as described by Cook and
192

Pickering (14). Sodium hydroxide (2.6 g) and (20 g) potassium nitrate

are dissolved in 100 mL of oxygen-saturated water. Ferrous sulfate (9
g) is dissolved in 100 mL of oxygen-saturated water. These two loo-
mL solutions are mixed, and the mixture is boiled for 20 min. The re-
sulting precipitate of magnetic iron oxide is washed 5-10 times. In
each wash, the precipitate is resuspended in water, and is then
attracted to the bottom of the wash flask with a strong magnet. The
wash water is then decanted. The iron oxide is resuspended in several
liters of 0.9% NaCl. The concentrated iron oxide solution is then
autoclaved for future use. Immediately prior to use, the iron oxide
solution is diluted in PBS.
5. Collagenase IV. Prior to its routine use, a sample of a particular lot of
collagenase is first tested to evaluate if it may have deleterious effects
on the outgrowth of cells from the proximal tubules.
6. A stock solution of 10 mg/mL soybean trypsin inhibitor in PBS is filter
sterilized and stored frozen for future use.
7. Kidneys are obtained from male New Zealand white rabbits (2-2.5
kg). The kidney is prepared and sliced for tissue culture using a sterile
50-mL beaker, a sterile curved nose scissors, 2-3 sterile forceps, and
2-3 sterile glass loo-mm diameter Petri dishes. Heavy black suture
thread and a Kelly hemostat (5-l/2 in straight end) are used in
suturing the renal artery over a sterile 20-gage needle (blunt ended,
using a file). The needle is attached to a 50-mL glass syringe contain-
ing perfusate.
8. A sterile 15 mL Dounce tissue homogenizer (Type A pestle) is used for
homogenization of the renal cortical slices. The preparation of proxi-
mal tubules for tissue culture also entails the use of Nylon nitrex
screening fabric both 253 m and 85 m (TETCO, Inc., Depew,NY 14043),
a lOOO-mL beaker, a metal spatula, and magnetic stirring bars. The
nylon mesh is held in place in embroidery hoops and sterilized either
in a container filled with 95% ethanol or in an autoclave.

3. Method
1. Insulin (5 pg/mL), transferrin (5 pg/mL), and hydrocortisone (5 x
lOaM) are added to the culture medium (DME/F12) on the day of
preparation of proximal tubules for tissue culture.
2. A rabbit is killed by cervical dislocation. The kidneys are removed (re-
nal artery and vein intact) using sterile scissors, and placed in a sterile
50-mL beaker with ice-cold DME/F12 containing 192 IU/mL penicil-
Primary Kidney Cells

lin and 200 pg/mL streptomycin. The kidney is kept ice cold through-
out the procedure.
3. The kidney is perfused as follows. The kidney is placed in a 1OOLmm
diameter glass Petri dish containing ice-cold DME/Fl2 with penicil-
lin and streptomycin, and washed (using thesame medium). A sterile
needle is inserted into the renal artery8 which is sutured. The kidney
is thenperfused,firstwith DME/F12 (toremoveblood), andthenwith
iron oxide (the kidney turns gray-black in color).
4. The renal capsule is removed using sterile forceps, and the kidney is
immediately transferred into another sterile 100-mm glass Petri dish
containing ice-cold DME/F12. (After this point, greater care must be
taken with regard to sterility.) Renal cortex slices are then removed
from the kidney using a sterile, curved nosed scissors.
5. The nephron segments are separated using the nylon mesh sieves.
Toward this end, the 85 pm sieve is placed directlyon top of a sterile
l,OOO-mL beaker; then the 253-p sieve is placed over the&S-p sieve.
The sieves are washed with DME/Fl2 containing 192 IU/mL penicil-
lin and 200 pg/mL streptomycin. The tubules and glomeruli are re-
moved from the top of the 85-pm sieve using a sterile metal spatula
and transferred into a sterile 50-mL plastic tube containingDME$F12.
6. In order to remove the glomeruli in the tubule suspension, a sterile
magnetic stir bar is placed into the tube. Glomeruli (covered with iron
oxide) adhere to the stir bar. The stir bar is carefully removed. This
process is repeated.
7. First soybean trypsin inhibitor, and then collagenase are added to the
tubule suspension (the final concentration of both reagents is 0.050
mg/mL). The tubules are incubated with the collagenase for 2 min at
23OC. In order to stop the collagenase treatment, the tube containing
the tubules is placed in a desktop centrifuge and spun for 5 min. The
tubules are resuspended in DME/F12 and again washed by certtrifu-
gation. Empirically, we have observed that the use of collagenase is
required if the outgrowth of epithelial cells from the tubules is to occur
in vitro.
8. After collagenase treatment, the tubules are suspended in DME’/‘F12
supplemented with 5 pg/mL insulin, 5 l.tg/mL transferrin, and 5 x
lOaM hydrocortisone but lacking antibiotics. We have used up to 400
mL of the medium to suspend proximal tubules obtained from a single
rabbit kidney. The tubule suspension is then inoculated into 35-mm
diameter tissue culture dishes (2 ml/dish).
9. The culture dishes are placed in a37’C incubator with a humidified 5%
CO,/95% air environment.
194 Taub

10. The medium is changed the day after plating the tubules, and routine
ly every 3-4 d thereafter. Confluent monolayers can be obtained after
5-6 d in culture.
11. Primary proximal tubule monolayers on plastic dishes can readily be
subcultured by trypsin treatment on a 1:2 basis. Confluent monolay-
ers will again be obtained if care is taken to minimize the trypsiniza-
tion period. Trypsin action is stopped by the addition of an equimolar
concentration of soybean trypsininhibitor. The cells are then removed
from the dish, suspended in a tube containing DME/F12, and spun in
a desktop centrifuge at 400 rpm. The cells are resuspended in DME/
F12 containing the three growth supplements, and inoculated into
plastic dishes. The proximal tubule cells can also achieve confluency
following a second subculturing into plastic dishes.

4. Notes
If the cell cultures do not grow to confluency, a number of problems
may have occurred.
1. Enough tubules must be added to the culture dishes to obtain conflu-
ent monolayers (0.5 mg protein/ml tubule suspension). The tubules
may easily be lost if they are not carefully harvested at the end of the
sieving procedure.
2. Cell cultures obtained from kidneys of young adults (as opposed to
older animals) are the most successful.
3. A third point of concern is the tissue culture medium. The purity of
the water is critical in defined medium studies. Loss of purity because
of contamination from a dirty pH probe for example may result in me-
dium that does not support cell growth. Our laboratory determines
pH using samples of the medium rather than placing the probe in the
medium to be used for tissue culture. In addition, a set of glassware
is used in medium preparation that is specifically designated for that
purpose.
4. Another point of caution with regard to medium is the hormone
supplements. Improper preparation or storage of the growth supple-
ments may be deleterious to cell growth. The growth stimulatory
effect of insulin may be lost if the stock solution is frozen. Further-
more, the medium supplements should be added to the medium im-
mediately prior to its use for tissue culture, since (unlike the case with
serum-supplemented medium) their stability in the medium is uncer-
tain.
Primary Kidney Cells 195

5. Care should be taken that the incubator hold a constant 37°C temper-
ature and a 5% CO,/95% air humidified environment. Animal cell
growth in the absence of serum is more sensitive to shifts in tempera-
ture and changes in the medium pH than with serum. The addition of
HEPES buffer to the medium alleviates this latter problem. Finally,
the primary rabbit kidney proximal tubule cells are less adherent to
plastic dishes than many other cell types. Thus, the cells may detach
during their manipulation for cell growth studies or transport studies.
This problem may be alleviated by growing the cultures on tissue
culture dishes coated with laminin, or with laminin and type IV col-
lagen (25). Proximal tubule cell cultures on plastic cell culture dishes
can readily be passaged once. Additional subculturings can be ob-
tained, but confluent monolayers are obtained with greater difficulty.
However, proximal tubule cell cultures grow more rapidly, achieve a
higher saturation density, and a higher passage number when pas-
saged into laminin-coated dishes.

References
1. Taub, M., Chuman, L., Saier, M. H., and Sato, G. (1979) Growth of Madin Darby ca-
nine kidney epithelial cell (MDCK) line in hormone-supplemented serum-free med-
ium. Proc. Natl. Acad. Sci. USA, 76,3338-3342.
2. Chuman, L., Fine, L. G., Cohen, A. I., and Saier, M. H. (1982) Continuous growth of
proximal tubular epithelial cells in hormone-supplemented serum-free medium. J,
Cell. Biol. 94,506-510.
3. Taub, M. and Sato, G. (1979) Growth of functional primary cultures of kidney epi-
thelial cells in defined medium. J. Cell. Physiol. 105,369-378.
4. Chung, S. D., Alavi, N., Livingston, D., Hiller, S.,and Taub, M. (1982) Characteriza-
tion of primary rabbit kidney cultures that express proximal tubule functions in a
hormonally defined medium. J. Cell. Biol. 95,118-126.
5. Taub, M. (1985) Primary culture of proximal tubule cells in defined medium J. Tis-
sue Cult. Methods 9,67-72.
6. Waqar, M. A., Seto,J.,Chung, S.D., Hiller-Grohol, S.,andTaub, M. (1985) Phosphate
uptake by primary renal proximal tubule cells grown in hormonally defined medi-
um. J. Cell. Physiol. X4,411-423.
7. Brendel, K. and Meezen, E. (1975) Isolation and properties of a pure preparation of
proximal kidney tubules obtained without collagenase treatment. Fed. Proc. 34,803.
8. Meezan, E. K., Brendel, J., Ulreich, J., and Carlson, E. C. (1973) Properties of a pure
metabolicallyactiveglomerularpreparationfromratkidneys. I. Isolation. J.Pham-
acol. Exp. Ther. 187,332-341.
9. Sakhrani, L. M., Badie-Dezfooly, B., Trizna, W., Mikhail, N., Lowe, A. C., Taub, M.,
and Fine, L. G. (1987) Transport and metabolism of glucose by renal proximal tubu-
lar cells in primary culture. Amer. 1. PhysioI. 246, F757-F764.
10. Matsuo, S.,Fukatsu, A., Taub, M. L., Caldwell, P. R. B., Brentjens, J. R., and Andres,
G. (1987) Nephrotoxic glomerulonepritis induced in the rabbit by antiendothelial
antibodies. j. C&z. Invest. 79,1798-1811.
196 Taub

11. Yang, I. S., Gokiinger,J M, Hong, S. K, and Taub, M. (1988) Preparation of basolat-
era1 membranes that transport paminohippurate from primary cultures of rabbit
kidney proximal tubule cells J. CeU. Physiol. 135,481-487.
12. Fine, I.. GandSakhmni, L. M. (19861 Proximal tubular cells in primary culture. Min-
eral Eke. Met& 12,X-57.
Wang, Y. and Taub, M. (1987) Unpublished observations.
ii: Cook, W. F. and Picketi~ G. W. (1958) A rapid method for separating glomeruli
from rabbit kidney. N&E ¶@, 1103,1104.
15. Taub, M. and Wang, Y. (1987) Control of rabbit kidney proximal tubule cell growth
andfunction~~~~u~~matrixcomponents,inBiologyofG~owfhFucto~,Trien-
nisi Symposium, program and abstracts, U. Toronto, Ontario, Canada.
26. Aleo, M. D, Taub# &!I. L, O&song J R., Nickerson, P. A., and Kostyniak, P. J. (1987)
Primary culturesof rabbit renal proximal tubule cells as an in vitro model of nephro-
titi&yz &e&of 2 mercurials, in In Viti Totiolagy: Approaches to Validation, vol.
5 G3Aibew Aim M.., ed3, Liebert, New York, pp. 211-225.
 Chapter 18

Adipocytes

Shinobu Gamou, Yoshiko Shim&u,

and Nobuyoshi Shim&u
1. Introduction
There are several preadipocyte cell lines reported, but in this chapter,
we will describe mainly the 3T3-Ll cells, which are best characterized and
widely used for molecular biological studies. The 3T3-Ll cells were
clonally isolated by Green and Kehinde from 3T3-Swiss albino. When
these cells are growing exponentially, they appear indistinguishable from
their parental Swiss/3T3 cells. 3T3-Ll cells, however, undergo a differ-
entiation to adipocytes when they enter a confluent and contact-inhibited
resting stage (2). Many clones isolated from the original 3T3 stock are able
to convert to adipocytes, in most cases, with a much lower frequency than
that of Ll cells. Subclones can be generated by serial cloning, which
differentiate to adipocytes at a high frequency. 3T3-F422A is such a clone
isolated from the same 3T3-Swiss albino stock culture as Ll cells (2).
A number of morphological, biochemical, and genetic changes occur
during the differentiation of 3T3-Ll cells to adipocytes. Ll cells start to
accumulate triglyceride in their cytoplasm, which can be seen as Oil Red 0
staining droplets as they change from a spindle-like fibroblastic cell to a
spherical cell (3). Thus, the differentiation can be easily identified under
the microscope.
The differentiation is also accompanied by a large increase in the
enzyme activities involved in de ~OVOfatty acid and triglyceride synthesis
197
198 Gamou, Shimizu, and Shimizu

such as glycerol-3-phosphate dehydrogenase (4). The differentiated cells

acquire a sensitivity to physiological concentrations of insulin via an
increased number of insulin receptors that have a higher binding capacity
than those in preadipocyte cells (5).
Although the differentiation of Ll cells can occur spontaneously over
a period of 2-4 wk following confluency, the differentiation can be syn-
chronously induced at a high frequency by treating growth-arrested Ll
cells with dexamethasone (DEX) and l-methyl-3-isobutylxanthine (MIX)
for 2 d, followed by incubation in normal medium for 2-5 d (6). The
combinations of DEX and MIX with insulin and biotin can accelerate the
differentiation.
Growth arrest at a specific stage of the cell cycle plays an important
role in the induction of the adipocyte differentiation, and various factors
appear to be involved in this process (7). The relationship between cell
division in growth-arrested Ll cells and their initiation of differentiation
can be closely examined under serum-free, hormone-supplemented con-
ditions (8).
We can observe the complex differentiation phenomenon under the
microscope within a week by rather simple induction procedures. Thus,
Ll cells’ differentiation is useful as a model of in vivo differentiation. The
differentiated adipocytes are also useful as a model hormone-responsive
cell sys tern.

2. Materials
2.1. Culture Medium
1. Medium: DMEM/FCSlO. Dulbecco’s Modified Eagle’s Medium with
a high concentration of glucose supplemented with fetal calf serum
(FCS) at 10% and kanamycin (100 pg/mL). FCS concentration can be
reduced to 5% by replacing with heat-inactivated calf serum.
2. PBS(-): Ca2+-and Mg2+-free Dulbecco’s phosphate buffered saline
(PBS; 2.7 mM KCl, 1.5 mM KH.J?O, 8.1 mA4 Na,HPO,, 137 mM NaCl).
3. Trypsin/EDTA solution: 0.1% trypsin and 0.01% EDTA in PBS(-).

2.2. Inducers for Adipocyte Differentiation

1. Dexamethasone (DEX) 1,000 x stock solution: 0.25 mM (98 pg/mL).
Dissolve in DMSO and store at -2OOC in small aliquots.
2. l-methyl-3-isobutylxanthine (MIX) 1,000 x stock solution: 0.5M (111
mg/mL). Dissolve in DMSO and store at -2OOC in small aliquots.
Adipocytes 199

3. Insulin 1,000x stock solution: 10 mg/mL. Dissolve in O.OlNHCl, ster-

ilize by filtration through a sterile Millipore filter (0.22 pm> and store
at -2OOC.
4. Biotin 1,000 x stock solution: 100 pg/mL. Dissolve in distilled water,
sterilize by filtration through a sterile Millipore filter (0.22 pm) and
store at -2OOC.

2.3. Oil Red 0 Staining

1. PBS(+): Dulbecco’s PBS with Ca2+(0.9 mM) and Mg2’ (0.5 mM).
2. 10% Formalin/PBS(+): Mixture of 10 mL formalin solution and 90 mL
of PBS(+).
3. Oil Red 0 solution: Dissolve Oil Red 0 (700 mg) in isopropanol(200
mL), stir at 4OCovernight, and then filter through Whatman 3 MM
paper. Mix 3 volumes of Oil Red 0 in isopropanol with 2 volumes of
distilled water, allow to stand at 4°C overnight and filter through
Whatman 3 MM. Store at 4OC.
4. 50% (v/v) glycerol solution.
5. Mayer’s Hematoxylin solution.
2.4. Detection of Glycerot3-Phosphate
Dehydrogenase Activity
1. Extraction buffer: 0.5% lYiton X-100,10 mM Tris-HCl, pH 7.5,l mM
MgCl,, 20 mM KCl, 0.2 mM Dithiothreitol, 10% Glycerol.
2. Reaction mixture: 125 mM Triethanolamine-HCl, pH 7.5, 2.5 mM
EDTA, 0.16 mMNADH, 0.36 mM Dihydroxyacetonephosphate,0.125
mM P-mercaptoethanol.
3. Glycerol-3-phosphate dehydrogenase from rabbit muscle.
2.5. Insulin Binding Assay
1. Insulin binding buffer: 120 mM NaCl, 1.2 mM MgSO,, 5 mM KCl, 10
mM Glucose, 1 mM EDTA, 15 mM Sodium acetate, 50 mM Hepes-
NaOH, pH 7.8,1% Bovine serum albumin. Store at -2OOC.
2. Cold insulin solution: Dilute the insulin stock solution described
above with the insulin binding buffer at a final concentration of 1 pg/
mL and 10 pg/mL. Make fresh before use.
3. ?insulin: Insulin iodination can be carried out as described in
Chapter 36, Vol. 1 of this series. Some technical skill is required to
iodinate insulin at a high specific activity without compromising its
biological activity. Thus, lZI-insulin obtained from commercial sources
200 Gamou, Shim& and Shimizu

may be more convenient. Approximately 50,000 cpm/mL (-1 ng/

mL) of y251-insulin in insulin binding buffer is used for binding assay.

3. Methods
3.1. Maintenance of 3T3-Ll Cells
1. Ll cells can be purchased from American Type Culture Collection
(12301 Parklawn Drive, Rockville, MD 20852-1776, USA). Ll cells are
routinely maintained in DMEM/FCSlO at 37°C in 5% CO, and 100%
humidity. Ll cells actively grow with a population doubling time of
20-30 h before becoming contact-inhibited at a saturation density of
-5 x lo5 tells/3-cm dish.
2. The morphology of the parental 3T3-Swiss albino is quite uniform and
shows typical 3T3 shape at confluency. Ll cells, however, show rather
heterogeneous morphology, occasionally overlapping each other at
confluency, but not piling up like transformed cells.
3. Contact-inhibited Ll cells will differentiate into the adipocytes spon-
taneously after 2-3 w when refed regularly.
4. Ll cells are easily detached after washing 2-3 times with PBS(-) and
treating with trypsin/EDTA at 37°C. Transfer cells every4-5 d by 1:20
dilution to maintain the preadipocyte state.

3.2. Induction of AcZipocyte Differentiation

1. The induction medium contains DEX (0.25 w), MIX (0.5 mM), in-
sulin (10 pg/mL) and biotin (100 ng/mL) in DMEM/FCSlO. Add l/
1000 vol of DEX and MIX stock solution into DMEM/FCSlO, mix
immediately and vigorously, and then add insulin and biotin at the
indicated concentration. Prepare the induction medium fresh before
use.
2. Replace the medium with induction medium when Ll cells reach
confluency, and incubate for 48 h at 37OC.
3. During the 48-h induction incubation, Ll cells start growing to ap-
proximately double in cell number, and the medium becomes slightly
viscous because of triglyceride secretion.
4. Replace the induction medium with DMEM/FCSlO containing bio-
tin (100 ng/mL) and insulin (10 pg/mL), and incubate for 5 d to allow
differentiation into mature adipocytes.
Adipocytes 201

3.3. Oil Red 0 Staining to Identifjt

the Morphological Differentiation
1. After the five days of incubation, the medium becomes viscous and
spherical adipocytes are seen under the microscope (Fig. 1). The adi-
pocytes often appear in clusters of varying sizes.
2. Wash the differentiated cultures twice with PBS(+), and then fix with
10% formalin/PBS(+) for 30 min. Stain the fixed culture dishes with
Oil Red 0 solution for 10 min, wash under tap water for a few min-
utes, and then cover with 50% glycerol. The adipocyte colonies can be
readily identified.
3. For microscopic observation or photography, the cells can be stained
with Mayer’s Hematoxylin for 2-3 min, washed for about 10 min
under tap water, and covered with 50% glycerol.
4. To determine the differentiation rate, wash the adipocytes and un-
differentiated cells 2-3 times with PBS(-) and detach them by trypsin/
EDTA treatment at 37OC.
5. Resuspend the cells with an appropriate volume of DMEM/FCSlO,
and count the number of adipocyte and undifferentiated cells in the
hemocytometer. The adipocytes are identified by the presence of
large lipid droplets,
6. When the lipid droplets are not large enough to identify under the
microscope, add half a volume of Oil Red 0 solution to the cell sus-
pension and incubate for a minute at room temperature. The adi-
pocytes can be easily identified by the presence of red droplets in their
cytoplasm.

3.4. Detection of Glycerol-S-Phosphate Dehydrogenase

Activity as a Biochemical Differentiation Marker
1. Ll cells grown in 6-cm dish are washed twice with ice-cold PBS(-) and
scraped with a rubber policeman. Transfer the Ll cells to a 1.5-mL
centrifuge tube, and wash the cells twice with ice-cold PBS(-) by brief
centrifugation.
2. Add 100 PLof ice-cold extraction buffer, and incubate for 30 min at 4OC
with gentle and continuous agitation. Centrifuge at 16,000 rpm for 60
min at 4°C in a Sorval SS34 rotor with a 1.5-mL tube adapter.
3. The resulting supernatant is kept frozen at -7OOC in small aliquots
until use. The amount of protein is determined by Lowry’s method as
described in Chapter 1, Vol. 1 of this series.
202 Gamou, Shimizu, and Shimizu

Fig. 1. Morphology of 3T3-Ll cells. (A) 3T3-Ll cells at high density. 09 3T3-Ll cells after
adipocyte differentiation. The cells were treated for 2 d with inducers (DEX, MDC, and
insulin) and allowed to differentiate for 5 d.

4. Put 0.4 mL of reaction mixture into a narrow width cuvette and set it
in a spectrophotometer.
5. Dilute the cell lysate to 100 PL (W-200 pg protein) with the extraction
buffer. Add 100 PL of diluted cell extract into the cuvette and mix rap-
idly with the reaction mixture to initiate the reaction.
Adipocytes

6. Record the absorbance at 340 nm 30 s after initiation and at 30-s inter-

vals. The absorbance will decrease.
7. The reaction is linear with time (for several minutes) and cell extract
concentration. Calculate the difference in absorbance/min/pg pro-
tein.
8. To determine the absolute enzyme activity, purified glycerol-3-phos-
phate dehydrogenase from rabbit muscle (Sigma) can be used as a
standard.
9. The enzyme activity in adipocytes increased lO-loo-fold over the
activity in preadipocytes. The increase in enzyme activity can be
detected just after the induction incubation.

3.5. Insulin Binding Assay

I. Ll cells are grown in three 3-cm dishes. One of these is used to
determine cell number and differentiation rate as described above.
2. Place the remaining two dishes on ice, and wash twice with ice-cold
insulin-binding buffer.
3. Prepare 3 mL of *251-insulin solution, and save 0.5 mL to determine
input radioactivity (cpm/mL) with a gamma counter (50,000 cpm/
mL, -1 ng/mL).
4. Add 1 mL of 1251-insulinsolution to one of the dishes, and incubate at
15OCfor 90 min with gentle agitation.
5. Wash the cells four times with ice-cold insulin binding buffer, and
then dissolve in 1 mL of 1NNaOH by incubation at 37°C for 1 h. Cell-
asso-ciated radioactivity (TB) is determined with a gamma counter.
6. To determine nonspecific binding (NSB), the remaining dish is pre-
incubated with cold-insulin solution containing 1 pg/mL insulin at
4°C for 30 min and then incubated with 1251-insulinsolution contain-
ing one-tenth the volume of the 10 pg/mL cold-insulin solution (final
insulin cont. 1 pg/mL) at 15OCfor 90 min. Cell-associated radio-
activity is determined as above. NSB is usually one-tenth to one-
fourth of TB,
7. Insulin specific binding (SB) is calculated as follow: SB = (TB-NSB)/
input/lo” cells x 100%. Specific binding in the undifferentiated Ll
culture ranges from l-3% of input/l06 cells, whereas that in Ll
adipocytes is 5-15% depending on the differentiation rate. Insulin
binding capacity gradually increases after the induction incubation.
8. Duplicate or triplicate dishes of Ll cells are preferable for the TB and
NSB determinations.
204 Gamou, Shimizu, and Shimizu

4. Notes
1. Although the differentiation potential of Ll cells is heritable, they
have a tendency to lose this capacity (2,2). The cells that have differ-
entiated at 90% frequency may lose half of their differentiation ability
after lo-20 transfers. Thus, a series of experiments must be carried out
using cells with same differentiation potential. Numerous frozen
stocks should be stored as early as possible.
2. If all of the frozen cell stocks with high differentiation potential have
been depleted, subclones may be isolated using the same method by
which the original Ll cells were established (1,2). Isolate lo-20 areas
in which the cells have small lipid droplets in their cytoplasm at
confluency by the standard cloning method. Measure the differ-
entiation ability and subclone if necessary.
3. The timing for induction is an important factor in order to achieve a
high differentiation rate. Sparse cultures can be induced, resulting in
a high frequency of adipocyte differentiation. Induction during late
confluency appears to be less effective in achieving the high differ-
entiation rate than that induced during early confluency. Ll cells must
traverse the cell cycle once before expressing the adipocyte pheno-
type (8).
4. Incubation of more than five days for the expression of the adipocyte
phenotype does not yield higher differentiation rates; however, the
adipocytes become more fatted and anchorage independent with
increased incubation time.
5. The differentiation rate may be measured quantitatively from the
fixed, Oil Red 0 stained culture. Elute Oil Red 0 from the stained
culture with isopropanol and measure the absorbance at 510 nm.
Cultures containing no adipocytes occasionally bind some dye and
give a high background. This, however, does not interfere the quan-
tification of the differentiation rate.
6. When the cell lysates are analyzed by two-dimensional electropho-
resis as described in Chapter 10, Vol. 1 of this series, an alteration in the
biosynthesis of more than 60 species of differentiation-specific pro-
teins is detected. This includes a decrease in the biosynthesic rates of
p and y actins as well as a and p tubulins by 90% that reflect the changes
in cell shape. These observations have been confirmed using cDNA
probes (9).
7. The biochemical nature of the reaction of glycerol-3-phosphate de-
hydrogenase is: Dihydroxyaceton phosphate + NADI&&lycerol-3-
phosphate + NAD+. The decrease in the absorbance at 340 nm is the
result of the conversion of NADH to NAD’.
Adipocytes

8. Since glycerol-3-phosphate dehydrogenase activity is very low in pre-

differentiated Ll cells, the cell extract without dilution or the extract
from more than two dishes may be required to measure the activity
accurately.
9. An automated spectrophotometer with a recorder is preferable, but
not necessary, for the measurement of the enzyme activity because the
reaction will be linear 2-3 min after initiation.
10. cDNAs for glycerol-3-phosphate dehydrogenase and other major
differentiation-related proteins have been cloned from cDNA librar-
ies constructed from the differentiated 3T3-F422A cells (10) and Ll
cells (11).
11. The increase in insulin binding capacity during adipocyte differ-
entiation involves an increase in both receptor number and affinity.
Thus, Scatchard plot analysis must be used to measure changes in
insulin binding capacity during the differentiation process. Prepare
1251-insulin solutions with different specific activities (50,000 cpm/mL
of *251-insulin with 1,2,4,10,20,40, and 100 ng/mL of cold-insulin, and
1,2,4,10,20,40, and 100 pg/mL of cold-insulin for nonspecific bind-
ing using duplicate or triplicate dishes for each specific activity). The
1251-insulinbinding assay is carried out as described above (3.5.). Cal-
culate how much insulin is bound as follows: Sp. AC. = Input cpm/
[Input insulin concentration]. [Bound] = SB/Sp. AC. [Free] = [Input
insulin concentration] - [Bound]. Plot [Bound]/ [Free] on the ordinate
and [Bound] on the abscissa. The negative reciprocal of the slope
indicates the affinity of the insulin receptor for insulin (Kd: dissocia-
tion constant). The intercept on the abscissa indicates the number for
insulin receptor. Both predifferentiated Ll cells and adipocytes yield
curvilinear plots, suggesting a heterogeneity in receptor affinity types.
12. Serum-free hormone-supplemented medium has been developed for
3T3-Ll cells (8,X?). The basal medium is a 3:l mixture of DMEM and
Ham’s F12 supplemented with several amino acids, vitamins, and
trace elements. Hormone-supplemented medium contains the fol-
lowing components in the basal medium: transferrin (2 pg/mL),
insulin (10 pg/mL), epidermal growth factor (20 ng/mL), biotin (100
ng/mL), bovine serum albumin (O.l%, Fr. V), and ethanolamine (0.1
mM). Under these conditions, the cells’ commitment to adipocyte
differentiation is separated from the expression of the adipocyte
phenotype.
13. There are several drugs that can inhibit the differentiation through
several distinct mechanisms. Nicotinamide possibly inhibits the
differentiation through the inhibition of poly (ADP-ribose) synthe-
tase (23). Dihydroteleocidin B (DHTB), a potent tumor promoter,
206 Gamou, Shimizu, and Shimizu
almost completely inhibits the differentiation whether added before,
during, or after the inducer treatment. Similar inhibition is observed
by 12-0-tetradecanoylphorbol-13-acetate (TPA), but with over 90%
less efficiency (14). Thus, these inhibitions are useful for dissecting the
complex differentiation process.
14. Using insulin-diphtheria toxin A fragment conjugate, variant cells
with altered insulin receptors have been isolated from Ll cells (25).
Also, DHTB-nonresponsive variants have been isolated by mitotic
shake-off method (16). Using these variants, the mechanism of
mitogenic signal transduction is being investigated.
15. Proto-oncogene c-fos expression in Ll cells and their derived adi-
pocyte have been studied in response to a variety of growth-pro-
moting agents (17). The intracellular location of protein kinase C has
been also studied in the Ll cell during growth and after exposure to
phorbol esters (18). Vesicles containing insulin-responsive glucose
transporters have been isolated from Ll cells (29).
16. 3T3-F442A cells isolated from 3T3-Swiss albino are able to differ-
entiate into adipocytes with high frequency by growth hormone
treatment (20).

References
1. Green, H. and Kehinde, 0. (1974) Sublines of mouse 3T3 cells that accumulate lipid.
Cell 1,113-l 16.
2. Green, H. and Kehinde, 0. (1976) Spontaneous heritable changes leading to in-
creased adipose conversion in 3T3 cells. Cell 7,205-113.
3. Novikoff, A. B., Novikoff, I’. M., Rosen, 0. M., and Rubin, C. S. (1980) Organelle
relationships in cultured 3T3-Ll preadipocytes. J. Cell Bid. 87,180-196.
4. Wise, L. S. and Green, H. (1979) Participation of one isozyme of cytosolic glycer-
ophosphate dehydrogenase in the adipose conversion of 3T3 cells. J. Biol. Chem. 254,
273-275.
5. Rosen, 0. M., Smith, C. J.,Fung, C., and Rubin, C. S. (2978) Development of hormone
receptors and hormone responsiveness in vitro: Effect of prolonged insulin treat-
ment on hexose uptake in 3T3-Ll adipocytes. J. Bid. Chem.253,7579-7583.
6. Rubin, C. S., Hirsch, A., Fung, C., and Rosen, 0. M. (1978) Development of hormone
receptors and hormonal responsiveness in vitro: Insulin receptors and insulin sen-
sitivity in the preadipocyte and adipocyte forms of 3T3-Ll cells. J. Biol Chem. 253,
7570-7578.
7. Scott, R. E., Florine, D. L., Wille, J. J., Jr., and Yun, K. (1982) Coupling of growth ar-
rest and differentiation at a distinct state in the G, phase of the cell cycle: GD. Proc.
Natl. Ad. Sci. USA 79,845-849.
8. Gamou, S. and Shimizu, N. (1986) Adipocyte differentiation of 3T3-Ll cells in
serum-free hormone-supplemented media: Effects of insulin and dihydrotele-
ocidin B. Cell Struct. Fun& 11,21-30.
Adipocytes

9. Spiegelman, B. M. and Farmer, S. R. (1982) Decreases in tubulin and actin gene

expression prior to morphological differentiation of 3T3 adipocytes. Cell 29,53-60.
10. Spiegelman, B. M., Frank, M., and Green, H. (1983) Molecular cloning of mRNA
from 3T3 adipocytes: Regulation of mRNA content for glycerophosphate dehy-
drogenase and other differentiation-dependent proteins during adipocyte dev-
elopment. J. BioI. Chern.258,10083-10089.
Il. Bernlohr, D. A., Angus, C. W., Lane, M. D., Bolanowski, M. A., and Kelly, T. J., Jr.
(1984) Expression of specific mRNAs during adipose differentiation: Identification
of an mRNA encoding a homologue of myelin P2 protein. Proc.N&Z. Acud. Sci. USA
81,5468-5472.
12. Shimizu, N. (1984) Use of hormone-toxin conjugates and serum-free media for the
isolation and study of cell variants in hormone responses,inMethods for Serum-Free
Culture ofEpitheliu1 and FibrobZusticCelZs.(Barnes, D. W., Sirubasku, D. A., and Sato,
G. H., eds.), Alan R. Liss, New York, pp. 233-248.
13. Lewis, J. E., Shimizu, Y., and Shimizu, N. (1982) Nicotinamide inhibits adipocyte
differentiation of 3T3-Ll cells. FEBS Lett. 146,37-41.
14. Shimizu, Y., Shimizu, N., Fujiki, H., and Sugimura, T. (1983) Distinct inhibitory
effects of dihydroteleocidin B and the phorbol ester tumor promoters on the
adipocyte differentiation of 3T3-Ll cells. Cancer Res. 43,49744979.
15. Shimizu, N., Miskimins, W. K., Gamou, S., and Shin&u, Y. (1982) A genetic
approach to the mechanisms of insulin action: Use of the insulin-Diphtheria toxin
fragment A conjugates for isolation of genetic variants. Cold Spring Harbor Confer-
ences on Cell Proliferation 9,397-402.
16. Shimizu, Y., Fujiki, H., Sugimura, T., and Shimizu, N. (1986)Mouse 3T3-Ll cell var-
iants unable to respond to mitogenic stimulation of dihydroteleocidin B: Genetic
evidence for the synergism of tumor promoters with growth factors. CancerRes. 46,
4027-4031.
17. Stumpo, D. J. and Blackshear, P. J. (1986) Insulin and growth factor effects on c-fos
expression in normal and protein kinase C-deficient 3T3-Ll fibroblasts and
adipocytes. Proc. Nutl. Acud. Sci. USA 83,9453-9457.
18. Halsey, 0. L., Girard, I’. R., Kuo, J, F., and Blacksher, P. J. (1987) Protein kinase C in
fibroblasts: Characteristics of its intracellular location during growth and after
exposure to phorbol esters and other mitogens. J. Biol. Chem. 262,2234-2243.
19. Biber, J. W. and Lienhard, G. E. (1986) Isolation of vesicles containing insulin re-
sponsive, intracellular glucose transporters from 3T3-Ll adipocytes. 1. Bid. Chem.
261,16180-16184.
20. Morikawa, M., Nixon, T., and Green, H. (1982) Growth hormone and the adipose
conversion of 3T3 cells. Cell 29,783-789.
 Chapter 19

Tissue Culture
of Skeletal Muscle

Terry A. Partridge
1. Introduction
In its simple forms, no special difficulty attaches to the tissue culture
of skeletal muscle. Indeed, it is one of the easiest tissues to culture in large
amounts because the starting material, skeletal muscle, is plentiful and
readily obtainable from a wide variety of species, including humans.
Moreover, the stem cells from which muscle develops in tissue culture
seem to be very resilient; they will survive anoxic conditions within muscle
at temperatures between 4 and 37OC, for up to a few days, and when
obtained in suspension, are easy to freeze (with 10% DMSO as a cryopro-
tectant) and store in liquid nitrogen.
Mature muscle fibers, being sensitive to physical damage and anoxia,
cannot be maintained in good functional condition out of the body for more
than a few hours. In tissue culture, therefore, muscle fibers must be grown
afresh from their mononucleate precursor cells (mpc). These latter cells
may readily be obtained in large numbers from growing or regenerating
muscle and, in smaller but still usable numbers from mature normal
muscle. It is of biological interest, but not of great practical value, that the
thymus is also a source of such cells (1).

209
210 Partridge

Here, the method we use to grow skeletal muscle from neonatal

mouse and a method for partially separating myogenic from nonmyoge
nit cells from this species are described in detail (2). The disaggregation
method gives satisfactory results with fetal human (15 wk) and neonatal
rabbit muscle. A simple method we have used for growing avian muscle
in culture is also described, and the use of some mouse muscle cell lines is
commented upon. Complications, problems, and refinements are indi-
cated in the Notes.

2. Materials
1. All glassware is washed in 1.5% RBS(Chemical Concentrates), rinsed
ten times in tap water and twice in double-glass distilled water, and
then air-dried in a warm oven.
2. Muscle cells are usually grown in vented polystyrene tissue culture
Petri dishes, 35 or 60 mm diameter. Where we need to make per-
manent microscopic preparations, immunostained preparations, or
microscopic observations at high resolution on the cultures, we grow
the cells on a sterile coverslip, placed in the Petri dish prior to addition
of the cell suspension and removed when it is required. We routinely
coat tissue culture surfaces with sterile rat-tail collagen in distilled
water and allow it to air dry. This is not strictly necessary for the mixed
cell cultures described here.
3. For filtration of the single-cell suspension, three layers of 45 pm pore
size nylon cloth are placed in sterile Swinnex filter holders.
4. Phosphate Buffered Saline (PBS): 8.0 g NaCl, 0.2 g KCl, 1.12 g
Na,Hl?O,, and 0.2 g KHJ?O, in 1 L distilled water, adjusted to pH 7.2
with NaOH. The solution is dispensed in lOO-mL aliquots and steri-
lized by autoclaving.
5. Antibiotic solution: 100 mL sterile PBS, 10 mL Fungizone (am-
photercin B, 250 pg/mL), and 2 mL penicillin/streptomycin (I? 5000
IU/mL, S 5000 yg/mL). Can be stored at -2OOC.
6. Hepes-buffered calcium and magnesium-free Hanks’ balanced salt
solution, pH7.4 (Ca2+-and Me-free BSS). Purchased as a 10 x solution
and diluted l/10 with double-glass distilled water. This solution is
buffered by the addition of 1M Hepes solution to give a final concen-
tration of 25 mM Hepes, sterilized by filtration through a 0.45 pm
Millipore filter and stored at -2OOC.
7. Trypsin stock solution: Trypsin (1%) (1:250; Difco) in Ca2+-and Mg2+-
free BSS. Trypsin and BSSare mixed for 2 h at room temperature and
filtered through a Whatman No. 1 filter paper. The pH is adjusted to
Tissue Culture of Skeletal Muscle 211

7.4 with lNNaOH, sterilized by filtration through a 0.45-pm Millipore

filter, and stored at -20°C in 20-mL aliquots.
8. Pangestin stock solution: Pangestin (1:75; Difco) mixed in a 0.5% ratio
(w/v) in Ca2+-and Mg2+-free BSSfor 2 h at 37OC,centrifuged at 26008
for 15 min at 4OC,and filtered through a Whatman No. 1 filter paper.
The pH is adjusted to 7.4 with 1NNaOH. The solution is sterilized by
filtration through a 0.45-pm Millipore filter and stored in 20-mL
aliquots at -2OOC.
9. Enzyme solution: On the day prior to use, 60 mL of trypsin stock
solution, 20 mL pangestin stock solution, and 20 mL Ca2+-and Mg2+-
free BSSare mixed under aseptic conditions. The solution is dispensed
in 5-mL aliquots in sterile universal bottles, stored at 4OCovernight
and brought to room temperature immediately before use.
10. Inhibition medium: 100 mL Hanks’ balanced salt solution + 20 mL
Fetal Calf Serum (FCS) and 2 mL penicillin/streptomycin (I?5000 IU/
mL, S 5000 pg/mL). Stored at 4OCuntil use.
11, Culture media: Basic medium constituents: Eagle’s Minimum Essen-
tial Medium (EMEM), Dulbecco’s Minimum Essential Medium
(DMEM), Fetal Calf Serum (FCS), Horse Serum (HS), Chick Embryo
Extract (CEE), Penicillin and Streptomycin stock solutions, and Fungi-
zone solution are all obtainable from a number of commercial sources
(seeNote 1).
We make our own CEE by a standard method (3). Fertile chicken
eggs, 10-12 d incubation, are held blunt end uppermost, swabbed
with 70% ethanol, and given a sharp tap, delivered horizontally, with
a sterile scalpel handle at about the lower edge of the air space (this can
be seen by holding the egg against a bright light). The cap of the shell
and adherent shell membranes are then carefully lifted off with sterile
coarse forceps. One prong of a second pair of sterile coarse forceps is
plunged through the embryonic membranes and beneath the neck of
the embryonic chick, which can then be lifted carefully from the egg
and transferred to a sterile beaker. When all the embryos have been
removed, they are washed free of all traces of blood and yolk with
several changes of BSS.These embryos are then crudely homogenized
by removing the plunger from a 20-mL syringe, loading the embryos,
a few at a time, into the barrel, carefully reinserting the plunger, and
pushing it home to express the contents, via the needle fitting, into
sterile plastic universal containers. To the mashed embryo in each
container an equal quantity of Hanks’ BSSis added, and the two are
mixed thoroughly by stirring with a sterile glass rod. After 30 min
incubation at room temperature, the containers are centrifuged at
212 Partridge

20008 for 20 min and the supernatant CEE is stored in small aliquots
at -20°C. On thawing, the CEE is recentrifuged before use.
12. Media for mouse primary cultures: 100 mL Dulbecco’s Minimum
Essential Medium (DMEM), 10 mL FCS, 2 mL CEE.
13. Cell lines and their growth media:
a. GB-1: For growth: 100 mL DMEM, 10 mL FCS. For differen-
tiation: 100 mL DMEM, 2% HS.
b. C2: For growth: 100 mL DMEM, 20 mL FCS. For differen-
tiation: 100 mL DMEM, 2% H S.
c. MM14: 50 mL DMEM, 50 mL Ham’s FlO, 15 mL HS, 3 mL CEE.
d. Human primary cultures: 100 mL DMEM, 20 mL FCS.
e. Avian cells: Following the methods described by Konigsberg
(4), we have used “high growth” media consisting of: For
chick, 79 mL Eagle’s MEM (with Earle’s salts), 15 mL HS, 5 mL
CEE, 1 mL Penicillin/Streptomycin, 0.25 mL Fungizone. For
Quail, 74mL Eagle’s MEM (with Earle’s salts), 15 mL HS, 10 mL
CEE, 1 mL Penicillin/Streptomycin, 0.25 mL Fungizone. Pri-
mary cultures, given one change of “high growth” medium on
the day after they were set up, and left in this medium, initially
proliferate and subsequently, as they exhaust the nutrients,
differentiate to form large numbers of myotubes. However,
Konigsberg (4) recommends reducing the concentration of
CEE to produce a “low growth” medium that, on addition to
the culture, limits proliferation and encourages differentiation.
14. Culture substrata:
a. Collagen: This is made by a standard method (3). Rat tails are
sterilized by steeping them in 95% ethanol for a few hours.
Each tail is then broken into short segments, starting at the tip.
To do this, the tail is grasped crossways at the distal end with
two pairs of sterile Spencer-Wells forceps, held side by side, the
tip is snapped off by a sharp rotation of the more distal of the
two forceps, and carefully drawn away to pull out the long
white tendons that insert into the tip of the tail. These are then
cut off with a pair of sterile scissors and allowed to fall into a
beaker containing sterile 1% acetic acid in distilled water (150
mL/ tail). This process is repeated, working progressively
toward the base of the tail. When all of the accessible tendon
has been removed, the beaker is covered to maintain sterility
and kept for a few days at 4OC, after which the solution is
centrifuged for 1 h at 25008. The supernatant is then dialyzed,
in autoclave-sterilized dialysis tubing, against daily changes of
distilled water until it begins to become viscous. This solution
Tissue Culture of Skeletal Muscle 213

is removed, after sterilizing theoutside of the dialysis tube with

a wash of 70% ethanol, and can be stored for several months in
aliquots under sterile conditions at 4OC. Should the solution
become too gelatinous, its viscosity can be reduced by addi-
tion of a little sterile dilute acetic acid. A drop of collagen
solution is placed on the culture surface (glass or plastic),
spread thinly with a glass spreader (made by using a fine flame
to seal the tip of a Pasteur pipet and to bend the terminal cm or
so at right angles to themain stem), and left to air-dry in asterile
flow hood. Such collagen coated surfaces can be stored dry for
several weeks.
b. Gelatin: A 0.01% solution of gelatin (Porcine skin, Type I) in
distilled water is distributed into I-2-mL aliquots and ster-
ilized by autoclaving. A drop of the solution is placed on the
culture surface, spread, and air-dried as for collagen (above).

3. Methods
3.1. Tissue Culture of Cells
thorn Neonatal Mouse Muscle
1. Newborn mice, (l-2 d, seeNote 2) are killed by decapitation, the tail
and paws are cut off, and the body washed in 70% ethanol, then
soaked in antibiotic solution. By slitting the skin along the length of
the spine, it can be peeled off in one operation by means of two pairs
of coarse forceps: one pair is used to grasp the mouse by the exposed
end of the spine at the neck and the second pair to pull the free edge
of the skin at the sternum back toward the tail.
2. The skinned carcass is then placed in a sterile Petri dish where as much
as possible of the skeletal muscle is removed from the limbs and trunk,
with the aid of watchmakers forceps and microscissors, and placed in
a second sterile Petri dish where it is kept damp with a drop of BSS.
Care must be taken not to include fragments of bone or of the brown
fat that is present in large amounts in these young animals, especially
between the shoulder blades. It is also important, particularly if the
animal has suckled, not to puncture the gut. Depending on the age
and size of the mice and the skill of the operator, approximately 0.5 g
of muscle can be obtained in this way from 3-5 mice in about 2 h: this
is the optimal amount to be taken through the remainder of the
procedure by one person.
3. Themuscle is minced into small fragments with a pair of sharp, curved
fine scissors (if well dampened with saline, the fragments tend to be
kept together by surface tension) and divided into ten equal portions,
214 Partridge

each of which is placed in 5 mL of enzyme solution in a sterile plastic

universal bottle.
4. These are placed in a 37°C waterbath and shaken at 2 Hz for 10 min.
Each aliquot is gently sucked into and expelled from a Pasteur pipet
continuously for one min to disperse the cells. Remaining muscle
fragments are allowed to settle, and the supernatant containing the
released cells is decanted into 5 mL of inhibition medium.
5. Fresh enzyme solution (5 mL) is then added to the remaining muscle
fragments in each universal bottle and the above disaggregation
procedure repeated.
6, The inhibition medium containing all the released cells from both in-
cubations is poured into a large, sterile plastic syringe fitted with the
Millepore filter holder containing the 45-pm nylon cloth, and slowly
expressed, thus removing any remaining clumps of cells and muscle
fragments to give a suspension consisting almost entirely of separated
mononucleate cells.
7. This is centrifuged at 3508 for 10 min at 4OC, and the resultant cell
pellets are resuspended in a known volume of growth medium.
8. A 100~PLsample of this suspension is added to 50 u.Lof a 0.5% solution
of Trypan Blue in PBS, and a count of viable cells (i.e., those that ex-
clude the dye) is made in a hemocytometer.
9. For culture in 35-mm plastic Petri dishes, cells are suspended in
culture medium at concentrations of between lo4 - 2 x 105cells/mL,
and plated out at 2 mL of cell suspension/dish. It is common to coat
dishes with collagen or gelatin. This is important when cloning
myogenic cells, but with cultures prepared as above, we can detect no
advantage in doing so.
10. The Petri dishes are placed in a 37OCincubator in an atmosphere of
10% CO, and examined daily under an inverted microscope fitted
with long-working distance phase contrast objectives. By 5-7 d,
numerous myotubes will have begun to form (Fig. 1, seeNote 3). If
longer term cultures are required, the medium must be replaced by
one containing less FCS and no CEE, so as to discourage overgrowth
of the culture by nonmyogenic cells.
3.2. Partial Separation
of Myogenic and Nonmyogenic Cells
on Discontinuous PercoZZ Gradients
Cell suspensions obtained by enzymatic disaggregation of neonatal
mouse muscle contain a majority of nonmyogenic cells that in long-term
cultures, tend to overgrow the developing muscle fibers. The method
Tissue Culture of Skeletal Muscle 215

Fig. 1. 7-d primary culture of cells obtained by enzymic disaggregation of neonatal

mouse muscle, stained with hematoxylin and eosin, 100x. Note small multinucleated
myotubes and the numerous unfused cells. Of the latter, some of the small spindle-
shaped cells are probably myoblasts.

described below, modified from that described by Turner (5) for the sep-
aration of chick muscle cells on Ficoll gradients, can be used to obtain a
degree of separation of myogenic from nonmyogenic populations in
mouse muscle cell suspensions (seeNote 3 and 4).
1. Percoll dilutions of 26% and 34% are made up in sterile HEPES-
buffered Ca2+- M g%?ee BSS, one day before use, stored at 4OC
overnight, and brought to room temperature immediately before use.
2. Cells obtained from the disaggregation procedure are suspended in
Ca2+- Mg!+-free BSS and placed in 5-mL aliquots in sterile Universal
bottles.
3. The 26% Percoll solution is taken up into a syringe equipped with a 21-
gage needle, and layered beneath the cell suspension by placing the tip
of the needle on the bottom of the Universal bottle and gently expel-
ling 5 mL of the solution. In the same way, 5 mL of the 34% Percoll
solution is layered beneath the 26% layer.
4. The gradients are then centrifuged at 3508 for 15-20 min at 15OC. Cells
that accumulate as cloudy layers at the interfaces and the cell pellet at
the tip of the bottle are then carefully removed with a Pasteur pipet.
5. Each fraction is placed in 5 mL of inhibition medium, pelleted at 3508,
and suspended in growth medium. Samples of each fraction are
216 Partridge

Fig. 2. Cells obtained from the pellet beneath 34% Percoll after separation of cells
obtained from neonatal mouse muscle on a discontinuous density gradient, stained with
hematoxylin and eosin. (a) After4 d in culture. Note the high proportion of small strongly
stained rounded or bipolar cells, characterized by a high nuclear/cytoplasmic ratio, 250x.
(b) Cells fusing after 7 d in culture. Note the conspicuous myotubes. In addition, the
majority of nonfused cells have a myoblastic appearance, 100x.

counted, and the remainder plated out in 35-mm plastic Petri dishes
and cultured as before. The pellet from the tip of the tube contains a
high proportion of small rounded or bipolar spindle-shaped baso-
philic cells (Fig. 2a).
6. After 5-7 d in culture, large numbers of myotubes can be seen to form
by fusion of such cells together (Fig. 2b). Cells obtained from the 26/
34% Percoll interface do not include the small basophilic type and, on
culture, do not produce myotubes (Fig. 3a,b). Very few cells can be
harvested from the BSS/26% Percoll interface, but they do include
some small basophilic cells and do form a few myotubes when
cultured. How the low buoyant density and high buoyant density
myogenic cells differ from one another is uncertain; they may re-
present cells in G, and G, stages, respectively, of the cell cycle.

3.3. Tissue Culture of Avian Muscle

Chick and quail embryonic muscles are among the easiest to prepare
and culture. No sterilization of the carcass is required. At 9 d for the quail
and 11 d for the chick, the muscles can be disaggregated enzymically as
described for the mouse, but are also sufficiently soft that the mpc can be
released by mechanical disruption of the tissue. Further advantages of the
avian muscle culture systems, especially quail, are that a greater pro-
portion of the cells differentiate into myotubes and that there is less exper-
iment to experiment variation than with mouse muscle. For both chick and
Tissue Culture of Skeletal Muscle 217

Fig. 3. Cells obtained from the 26/34% Percoll interface after separation of cells ob-
tained from neonatal mouse muscle on a discontinuous density gradient, stained with
hematoxylin and eosin. (a) After 4 d in culture. The majority of cells are large and of
fibroblastic appearance, haying a low nuclear/cytoplasmic ratio and pale-staining cyto-
plasm, 250x. (b) Cells after 7 d in culture. Note absence of myotubes. Only the occasional
cell has the bipolar spindle-shape characteristic of myoblasts, 100x.

quail muscle, we have found the following method, based on that de-
scribed by Caplan (6) and by Konigsberg (4) to be a very quick and sat-
isfactory means of obtaining large numbers of mpc.
1. Eggs of the correct incubation age are swabbed with 70% ethanol and
opened at the blunt end. The method normally recommended for
doing this is to cut around the end of the egg with sterile scissors, but
this generates egg shell dust, which may fall onto the embryo. We find
it more satisfactory to use the method described for the preparation of
CEE (above): breaching the shell with a blow of a sterilized scalpel
handle, lifting off the cap of the egg shell, still adherent to the un-
derlying membrane, with sterile coarse forceps, then using a second
pair of forceps to lift out the embryo by the neck, without squeezing
it, and transfer it to a sterile Petri dish.
2. The skin overlying the breas t muscle is dissected away, and the muscle
is carefully removed and placed in a drop of sterile PBS in a second
Petri dish.
3. The pooled muscle is then finely minced with curved scissors (as for
the mouse), and the fragments are suspended in sterile growth medi-
um; approximately 0.5 ml/quail breast, 1 ml/chick breast.
4. This suspension is distributed in 2-mL aliquots into large (20-25 mL)
conical centrifuge tubes, and each tube is mixed for 30 s on a labora-
tory vortex mixer at high speed.
218 Partridge

5. After allowing the large fragments to settle, the supernatant is drawn

off. The remaining fragments can be subjected to a second agitation
in growth medium.
6. The combined supernatants are then filtered through3 layersof ZO-pm
pore nitex mesh mounted in a Millepore filter holder and samples
counted for viable cells as with the mouse cell preparations.
7. Depending on the desired timing and degree of myoblast fusion, the
cells can then be seeded out in growth medium at 104-106cells/35mm
Petri dish. Cell debris and nonviable cells initially contaminate
cultures prepared in this way, but they can be removed by giving the
cultures a brief wash and change of medium on the following day.

3.4. Cell Lines

It is possible to trypsinize primary cultures of muscle cells, using
standard tissue-culture methods (see Chapter 1, this volume), and to
propagate them for several generations. With an increasing number of
passages, there is a tendency to lose the myogenic phenotype. This loss of
myogenicity can be offset by seeding the cells at low densities, such that
clones develop from individual cells, and subcloning from the nonfused
cells present in myogenic clones (7). Stable continuous myogenic lines
sometimes appear after several such serial passages. Continuous lines are
available from the rat (L6, L8) and mouse (MM14, G8-1, C2), but not from
human or avian origins. These lines provide the most experimentally
reliable material; in most of them, the majority of cells fuse into myotubes
in the right conditions and they present little difficulty in tissue culture. Of
the lines of which we have experience, the C2 line (8) is the most satis-
factory, being almost compulsively myogenic. It should be kept in mind,
however, that these cells are to some extent neoplastic (all lines we have
tested form tumors as well as muscle when implanted into nude mice) and
should not be accepted uncritically as models of normal myogenesis.

4. Notes
1. Culture media: Sera and commercially available Chick Embryo
extract are extremely variable and must be batch tested. As a general
rule, all procedures, especially those concerning biologically derived
components, must be rigidly standardized.
2. Ageof muscle: In general, the older the starting muscle, the longer the
disaggregation period required, and the lower the yield of mpc. We
obtain lo’-lo8 total cells/g from neonatal muscle; 1-5 x lo5 cells/g
from muscle of 6-7 wk-old mice. This problem can be partially over-
Tissue Culture of Skeletal Muscle 219

come by causing the muscle to start to regenerate in vivo prior to

taking it for extraction of mpc: this has been done by ligaturing the
vascular supply of the muscle, by injecting myotoxic agents such as
marcain or notexin, or as we have sometimes done with good results,
removing, mincing, and reimplanting the muscle as an autograft. This
latter procedure, applied to 7-wk mouse muscle, 3-5 d after grafting
yields some lo8 cells/g. About half of these cells are macrophages, but
these can be removed by virtue of their rapid adherence to glass (see
below), and the yield of mpc is still much improved.
3. Methods for enhancing myogenicity of cell preparations: Apart from
the discontinuous density gradient centrifugation method described
above for obtaining mouse cell populations enriched and impov-
erished in myogenic cells, a number of other methods have been
employed for this purpose. Below are two of the more commonly
used, and what looks to be a promising new method.
a. Methods based on differences in the rate at which myogenic
and nonmyogenic cells adhere to a solid substratum were
originally developed on chick cells where they are reported to
be very successful. They involve plating out the mixed cell
suspension in a Petri dish for a short time (30-60 min), during
which the majority of fibroblasts adhere to the plastic: the
supernatant, now enriched in myogenic cells, is then removed
and replated in another Petri dish. Such methods have also
been applied to mouse skeletal muscle cells, apparently with
success (e.g., 9,10), but except possibly for the removal of
macrophages, we have not found them to be at all effective with
muscle cells from this species.
b. A very similar principle is the differential rate of de-adhesion
of myogenic and nonmyogenic cells from a sustratum on mild
trypsinization: the supernatant of a brief trypsinization is
enriched in myogenic cells (10).
c. Antigenic differences between myogenic and nonmyogenic
cells offer one of the most promising routes to mass separation
of the two populations. A monoclonal antibody to human
myogenic cells, 5.1 Hll, has been used to separate large num-
bers of myogenic cells on the fluorescence-activated cell sorter
(II). In the chick, a similar monoclonal antibody has been
shown to cause rounding and detachment of myogenic cells
from the substratum, while leaving nonmyogenic cells attached
(12).
4. Enhancement of differentiation: In vivo, motor innervation and the
220 Partridge

physiological workload greatly influence the state of differentiation of

skeletal muscle fibers. Addition of extracts of peripheral nerve has
been reported to produce better differentiated and longer-lived cul-
tures than media lacking such extracts (23). By far the most effective
in vitro model of skeletal muscle, however, is the nerve muscle
coculture system developed by Peterson and Crain (14). In such
cultures, slices of fetal mouse spinal chord are placed in culture
together with fragments of skeletal muscle. Outgrowths from the
spinal chord of nonneural cells, when they contact the muscle frag-
ment, stimulate the proliferation and early differentiation of myo-
genie cells within it. Subsequently, axonal outgrowths of nerve cells
make synaptic contacts with the developing muscle fibers, causing
them to twitch and to express many of the phenotypic characteristics
of normal skeletal muscle. The chief disadvantages of this method of
culture are that some experience and a meticulous technique are
required to set up and maintain the cultures, and that the amount of
well-differentiated muscle tissue formed within them is too small to
be of use for many biochemical techniques.
5. Problems: Two major problems of tissue culture in general seem to be
exemplified in extreme forms by skeletal muscle in tissue culture.
a. It is difficult to set up tissue cultures in which a high proportion
of cells become differentiated and form muscle fibers. The
principal reason for this is the difficulty of ridding cultures of
cells with no myogenic potential. However, even with cloned
myogenic cells, not all, and sometimes only a minor pro-
portion, develop into muscle cells.
b. The degree of differentiation attained by muscle fibers grown
in tissue culture falls far short of what is found in vivo. The
innervated cocultures described above are the most effective
attempt so far to remedy this.
6. Two practical difficulties we have encountered are:
a. Poor reproducibility. Even with standardized techniques and
batched media components, variation between individual
primary culture preparations is often too great to permit
comparisons to be made between them. It is safest, therefore,
to study the effects of variables within particular preparations.
Cultures of muscle of the Japanese quail seem less susceptible
to this problem.
b. Long-term cultures. As the contractile myofibrils assemble
within developing muscle fibers, these cells respond to minor
disturbances, such as illumination under a microscope, change
Tissue Culture of Skeletal Muscle 221

of medium, or addition of more cells, by contracting violently.

This may cause irreversible detachment of the entire mono-
layer of cells from their substratum.
In general, it is best to use the simplest culture system for skeletal
muscle that will critically test the hypothesis under consideration. If it is
necessary to use one of the more complicated methods, it can save much
time and effort to acquire experience of it in a laboratory where it is used
routinely, because written accounts of a technique do not record all of the
subliminal practices that are frequently crucial to its success.

Acknowledgments
I wish to thank Drs. J. Morgan and D. Watt for their advice and criti-
cism. The work on mouse muscle culture is supported by the Muscular
Dystrophy Group of Great Britain and Northern Ireland.

References
1. Wekerle, H., Paterson, B., Ketelson, U.-P., and Feldman, M. (1975) Striated muscle
fibers differentiate in monolayer cultures of adult thymus reticulum. Nature 256,
493,494.
2. Watt, D. J., Lambert, K., Morgan, J. E., Partridge, T. A., and Sloper, J. C. (1982)
Incorporation of donor muscle precursor cells into an area of muscle regeneration
in the host mouse. J. Neural. Sci. 57,319-331.
3. Paul, J. (1975) Cell and Tissue CuIture (Churchill Livingstone, Edinburgh).
4. Konigsberg, I. R. (1979) Skeletal myoblasts in culture. Methods in Enzymology LVIII,
51 l-527.
5. Turner, D. C. (1978) Differentiation in cultures derived from embryonic chicken
muscle. The post-mitotic fusion-capable myoblast as a distinct cell type. Differ-
entiation 10,81-93.
6. Caplan, A. I. (1976) A simplified procedure for preparing myogenic cells for culture.
J. Embyol. Exp. Morphol. 36,175-181.
7. Hauschka, S. D., Linkhart, T. A., Clegg, C., and Merrill, G. (1979) Clonal studies of
human and mouse muscle, in Muscle Regeneration (Mauro, A., ed.), Raven, New
York, pp. 311322.
8. Yaffe, D. and Saxel, 0. (1977) Serial passaging and differentiation of myogenic cells
isolated from dystrophic mouse muscle. Nature 270,725-727.
9. Cossu, G., Pacifici, M., Marino, M., Zani, B., Coletta, M., and Molinaro, M. (1981)
Developmental changes in glycopeptides synthesized by normal and dystrophic
satellite cells in culture. Exp. Cell. Res. 132,349357.
10. Kiihl, U., Timpl, R., and Von Der Mark, K. (1982) Synthesis of type IV collagen and
laminin in cultures of skeletal muscle cells and their assembly on the surface of
myotubes. Dev. Sol. 93,344-354.
Il. Webster, C., Pavlath, G. K., Walsh, F. S., Parks, D. R., and Blau, H. M. (1988) Isola-
tion of human myoblasts with the fluorescence activated cell sorter. Exp. Cell Res.
174,252-265.
222 Partridge

22. Sasse,J., Horwitz, A., Pacifici, M., and Holtzer, H. (1984) Separation of precursor
myogenic and chondrogentc cells in early limb bud mesenchyme by a monoclonal
antibody. J. Cell Bid. 99,1856-X366.
13. Oh, T. H. and Markelonis, G. J. (1979) Neurotrophic effects of a protein fraction
isolated from peripheral nerves on skeletal muscle in culture, inMuscle Regeneration
(Mauro, A., ed.), Raven, New York, pp. 417-427.
24. Peterson, E. R. and Cram, S. M. (1979) Maturation of human muscle after innerva-
tion by fetal mouse spinal chord explants in long term cultures, inMuscIe Regenem-
tion (Mauro, A., ed.), Raven, New York, pp. 429-441.
 Chapter 20

Derivation and Maintenance

of Embryonic Stem Cell
Cultures

Elizabeth J. Robertson
1. Introduction
The ability to derive permanent tissue culture lines (I) of pluripo-
tential stem cell lines (ES cells) from mouse embryos has provided a
valuable model system for fundamental research into the cellular differ-
entiation processes occuring in the normal embryo. Perhaps the most
attractive feature of embryonic stem cell lines is that they can be mani-
pulated to differentiate into a diversity of cell types either directly in the
tissue culture dish or within the context of the normal developing embryo
following their return to the embryonic environment.
ES cells can be maintained in the undifferentiated state for periods of
several months without loss of their developmental capacity. When ES
cells are reintroduced into host carrier embryos by the technique of
blastocyst injection (2), they resume a normal regulated pattern of pro-
liferation and differentiation to form chimeric conceptuses. ES cells are
unrestricted in their pattern of functional differentiation contributing
extensively both to the somatic and germ cell lineages of adult chimeras (3).
Transmission of ES cells in the germ cells provides a route to deriving
transgenic mouse strains, i.e., animals that harbor exogenous introduced
DNA sequences. ES cells offer several attractive advantages over the
223
224 Robertson

alternative techniques that are currently available for the production of

transgenic mice, such as direct injection of the one-cell mouse egg or
retroviral infection of embryos (reviewed 4) as the cells are accessible for
experimental manipulation and screening in culture prior to their incor-
poration into the animal (5-7).
This chapter describes the technique for the derivation of ES cells from
cultured blastocyst-stage embryos and the protocols for the routine cul-
ture of established ES cell lines. For techniques involved in the production
and analysis of ES chimeras, the reader is referred elsewhere (2).

2. Materials
1. Complete growth medium: Dulbecco’s Modified Eagle’s Medium
(high glucose formulation) plus 10% (v/v) fetal calf serum, 10% (v/v
newborn calf serum, seeNote l), 5 x 1094 Z-mercaptoethanol, anti-
biotics (50 U/mL penicillin, 50 pg/mL streptomycin), and 1% non-
essential amino acids.
2. Trypsin-EDTA; 0.25% (w/v> powdered porcine trypsin in 0.04%
(w/v) EDTA in Phosphate buffered saline (pH 7.6).
3. Phosphate buffered saline (calcium and magnesium free): 0.17M
NaCl, 3.4 mM KCl, 4 mM Na,HPO,*12H,O, 2.4 mM KHJYO,.
4. Mitomycin C medium; Dulbecco’s Modified Eagle’s Medium sup-
plemented with 10% (v/v) newborn calf serum and 10 pg/mL mito-
mycin C (prepare freshly at each use).
5. Lightweight paraffin oil.

3. Methods

3.1. Isolation of ES CeZZLines (See Note 2)

3.1.1. Preparation of Embryonic Outgrowths
1. Prepare feeder layers in a 24-well tissue culture plate. Seed 5-7 x lo*
mitomycin C-treated ST0 fibroblasts in 1 mL complete growth me-
dium/well. Feeder layers should be prepared 1 d prior to collecting
embryos.
2. Set up timed matings using the mouse strain of choice. The blastocyst
stage of embryonic development is reached 3.5 d after fertilization of
the oocyte. Inspect females daily for copulation plugs (day of plug is
day 0). On day 3, kill the females by cervical dislocation, swab the
abdomen with liberal quantities of 70% ethanol, and dissect out
Embryonic Stem Cell Cultures 225

individual uterine horns (cut below the uterotubal junction and above
the cervix). Remove the horns into a sterile Petri dish. Flush the
uterine lumen by introducing a 26-gage needle attached to a l-mL
syringe filled with complete growth medium (seeNotes 3 and 4).
3. Remove the dish to a dissecting binocular microscope and locate the
blastocysts. Using a finely drawn out Pasteur pipet attached to a
mouth-controlled tube, collect the blastocysts. Transfer the embryos
into a small drop of complete growth medium (overlay the drop with
lightweight paraffin oil to prevent evaporation).
4. Transfer the embryos individually into the feeder wells taking care to
place the embryo in the center of the well (seeNote 5). Return the plate
to the incubator and culture undisturbed for 4 d. After this time,
inspect the embryos. The blastocysts will have attached by the out-
growth of the trophectoderm cells. The inner cell mass (ICM) of the
embryo will have proliferated to form a small discrete mass of cells
(Fig. 1).

3.1.2. Disaggregation of Inner Cell Mass (ICM) Clumps

1. Prepare fresh feeder layers in a 24-well plate (as above).
2. Aspirate the medium from the wells containing the embryos and add
1 mL of PBS.
3. Prepare a tissue culture dish containing an array of small drops of
trypsin-EDTA solution (approx. 100 pL). Overlay the drops with
lightweight paraffin oil.
4. Place the 24-well plate under a dissecting microscope, and using a
finely drawn out Pasteur pipet, carefully dislodge the ICM clumps
from the trophectoderm cells. Transfer each ICM into a separate
trypsin drop. Return the dish to the incubator and leave for 5 min.
5. Pull out a very fine diameter capillary in a bunsen flame from a Past-
eur pipet. Attach the pipet to a mouth-controlled tube. Using gentle
suction, suck a small quantity of complete growth medium into the
pipet.
6. Return the dish of trypsin drops to the binocular microscope. Dis-
sociate each ICM as follows. Blow a small quantity of medium from
the pipet into the trypsin drop (to neutralize the trypsin action). Us-
ing mouth control, draw the ICM clump in and out of the end of the
pipet (the end of the pipet should be approximately a quarter the
diameter of the ICM clump). Fragment the ICM into several small
cellular clumps, and transfer the fragments into a fresh feeder well.
Repeat for all embryos, and return the 24-well plate to the incubator.
226 Robertson

Fig. 1 Attachment and growth of a single blastocyst on a gelatinized tissue culture

surface. (a) 1.5-d culture, blastocyst attaching by trophoblast cell outgrowth; (b) 2.5-d
culture, ICM component is revealed by the continued spreading of the trophoblast cells;
(c) 4-d culture, cells of the ICM proliferating rapidly; (d) 5-d culture, ICM has formed a
mass of cells.

3.1.3. Identification and Expansion of ES Cell Colonies

1. Incubate the ICM-derived fragments for 4 d. During this period, the
fragments will attach and proliferate to form small primary colonies
of cells.
Embryonic Stem Cell Cultures 227

Fig. 2 Morphologies of primary colonies that result from the disaggregated ICM. (a)
patch of “spread” trophoblast cells; (b) fibroblast-like cells; (c) epithelioid-like cells; (d)
epithelioid cells, single colony of cells at high power; (e) appearance at high power of
trophoblast-like cells before cell flattening has occurred; (fl colony of stem cells.

2. Inspect each well under high-power phase contrast (e.g., 200x). With
reference to Fig. 2, classify the individual primary colonies according
to morphology. A variety of differentiated phenotypes will be noted.
 .:d -

Fig. 3 Progressive alteration to the morphology of a putative stem cell colony over a 3-
d culture period. (a) When first located 2 d after disaggregation of the ICM, the constit-
uent cells superficially resemble stem cells. (b) Following a further 3 d culture, the colony
has spread and flattened to form a patch of giant trophoblast-like cells.

Mark potential ES cell colonies by circling the posi tion of the colony on
the underside of the well with a fine-tipped, indelible marking pen.
Return the plate to the incubator. Inspect the marked colonies daily
for the following 3 d. Eliminate colonies that differentiate to form
large flat cells (seeFig. 3). Genuine colonies of ES cells will increase in
size in the absence of overt differentiation (seeFigure 4, seeNote 6).
Embryonic Stem Cell Cultures 229

Fig. 4 Progressive growth of a colony of stem cells. (a) d 3; (b) d 4; (c) d 5. Note that
the colony increases in size in the absence of overt differentiation. The cells pile up and
maintain a discrete colony on top of the feeder cells. The constituent cells adhere tightly
to one another making it difficult to see individual cells. This growth pattern is charac-
teristic of stem cells, although the colonies may appear slightly flatter and less regularly
shaped than the colony illustrated here.

3. Using the following procedure, subculture colonies of stem cells

approximately 6-7 d after the primary disaggregation event. To avoid
the possibility of cross-contamination with nonstem cells, each puta-
tive ES cell colony should be dissociated individually.
230 Robertson

4. Prepare the feeder layers in l-cm wells. Aspirate the well containing
primary colonies and add 1 mL of PBS.
5. Using a drawn out Pasteur pipet dislodge the marked colony, trans-
fer it to a small drop of trypsin-EDTA, and incubate for 2-3 min in the
incubator.
6. Using a finely drawn out Pasteur pipet dissociate the colony into a
single-cell suspension. This should be achieved easily, since stem cells
are more readily dissociated than the original ICM-derived clump.
7. Transfer the cell suspension to a l-cm feeder well containing 1 mL of
complete growth medium and incubate.
8. Inspect the wells after 2 d for the presence of stem cells. These are
readily identifiable as having a very characteristic morphology (see
Fig. 5).
9. After 2 d of further culture, passage the wells containing ES cell
colonies. Wash the well with 1 mL of PBS and trypsinize the cells by
the addition of 100 PL trypsin-EDTA solution. Allow a 3-4 min in-
cubation in trypsin. Then using a Pasteur pipet and bulb, add 100 PL
complete growth medium to each well. Vigorously disperse the cells
to ensure that a single-cell suspension is obtained.
10. Seed the single-cell suspension from each well into a 3-cm feeder plate
containing 2 mL of complete growth medium.
11. Refeed the plates daily. After 3 d, sufficient stem cells should be
present to necessitate subculture. Stem cells can also be frozen down
at this stage (seeNote 7).

3.2. Routine Culture of ES Cell Lines

It is recommended that ES cells be cultured exclusively on feeder
layers. However, for some experimental procedures, it is desireable to
omit feeder cells. This can be achieved by culturing the cells on gelatinized
plates in complete growth medium supplemented with 50% BRL con-
ditioned medium (8).
Stem cells grow rapidly and divide approximately every 15-18 h. The
ES cell line should be maintained at relatively high densities to ensure
maximal growth rate in order to minimize spontaneous differentiation.
1. Cultures of ES cells should be inspected and refed daily. The cultures
should be passaged when the stem cells have reached an approx-
imately 80% confluency.
2. Refeed the cultures 2-3 h prior to subculture to maximize the cell
viability.
Embryonic Stem Cell Cultures 231

Fig. 5 Examples of typical areas of cells found within a subconfluent culture of es-
tablished stem cells, to illustrate the cellular morphology exhibited by pluripotential cells.
The culture was seeded initially as single cells. These divide and the sister cells stay to-
gether to form small nests. With time the nests increase in size and merge to form a con-
fluent layer.
3. Aspirate the growthmedium andwashwith2-3mLof PBS- Add1 mL
of trypsin-EDTA solution, and return the plates to the incubator for
4-5 min.
232 Robertson

4. Gently rock the plates to dislodge the cells. Add 1 mL of complete

growth medium, and using a Pasteur pipet and bulb, dissociate the
cells to a single-cell suspension.
5. Centrifuge the cell suspension at 1000 rpm for 5 min. Aspirate the
medium and resuspend the cells in 5 mL of complete growth medi-
um. Determine the cell density, and replate the cells into fresh feeder
plates at a density of IO6 cells/&cm plate.
6. Refeed andinspect the plates daily. An 80% confluent dishof cells will
be obtained after 34 d of culture, and at this time, a 6-cm plate will
typically yield 2 x 10’ cells.
4. Notes
1. Both fetal serum and newborn calf serum are used as medium sup-
plements. For the successful growth of most tissue culture cells, ser-
um quality is critical, and this is especially true for blastocysts and
embryonic cells. The use of unsuitable serum is undoubtedly one of
the reasons for a failure to isolate stem cells from embryos. All sera
should be tested for their ability to support the growth of pluripo-
tential stem cells: serum testing is laborious, and the best strategy is
to simultaneously request the maximum number of available sam-
ples from all the commercial suppliers. The samples can then be tested
in parallel and a bulk order placed for the most suitable sera. Serum
can be stored for periods of up to 2 yr at -2OOC. The most sensitive test
of the ability of a given serum to support the growth of stem cells is by
a comparison of the plating efficiency of single cells from either an
established embryo-derived stem cell line or, alternatively, feeder-
dependent EC cell line (e.g., PSA-4). The stem cells are coplated with
feeder ST0 cells. Control samples of fetal newborn calf sera are in-
cluded in the test. Only sera that are comparable to, or better than, the
control sera should be bought. The technique is to set up duplicate
plates containing a suspension of mitomycin C-treated ST0 fibro-
blast cells and a small number of s tern cells in serum-free medium. The
plates are placed in sets with each set being allocated a specific serum
batch. Serum is added to a final concentration of 10%. To a single plate
in each set, serum is added to a final concentration of 30%. At this
concentration, any toxic component will readily be detected. The
plates are incubated for 7-10 d, and the stem cell colonies stained and
counted. A good quality serum gives a plating efficiency of 20% or
higher. Toxicity effects will be evidenced by a lower plating effic-
iency and/or the formation of smaller colonies at the 30% concen-
tration Serum toxicity may be the result in part of high levels of
Embryonic Stem Cell Cultures 233

complement. Heat treating serum (56OC for 30 min) to inactivate

complement may dramatically decrease toxicity.
2. Karyotype analysis procedures should be used to ascertain the chro-
mosome complement and sex chromosome constitution of a cell line
as soon as possible after it has been established. The majority of ES cell
lines derived, to date, have a normal euploid chromosome com-
plement when established cultures are analyzed. However, it is
inevitable that cell populations will drift away from the normal
genotype and subsets of aneuploid cells will be selected over a period
of continuous culture. For this reason, it is important to routinely
check the chromosome complement of stem cells in culture. Simple
chromosome counts are used to determine the model number, and
estimate the range and variability of the population. Ideally, this
should be complemented by G-banding analysis, which enables the
exact chromosomal constitution to be determined.
3. The starting material of choice for the isolation of stem cells is the
blastocyst stage embryo. Since the time course of mouse embryo-
genesis is well documented, and the time of fertilization can be readily
determined, the recommended method for obtaining the maximum
number of fully viable embryos is to leave the embryos to develop
within the reproductive tract until they have reached the appropriate
blastocyst stage. Blastocysts are amenable to culture in standard
tissue culture media and do not require special requirements. This is
not the case for embryos at preblas tocys t stages that have to be grown
in a simpler serum-free medium. The mouse strain from which a cell
line is to be made is another important consideration. The majority of
in-bred strains will, for example, yield relatively few embryos (normal-
ly between five and eight). If approaching this technique for the first
time, it is recommended that more fecund outbred stocks of mice be
used to generate embryos. Outbred mouse stacks are readily available
from commercial breeders, and matings provide large numbers of
embryos on which to practice the various techniques before turning to
the particular strain, or mutations, of interest. The rate of recovery of
stem cell lines is rarely higher than 30%, and may be lower than 10%.
Therefore, a large percentage of the embryos are wasted. If the mouse
strain of interest is available in restricted numbers, or if the desired
genotype is carried by only a fraction of the embryos from a given
mating, attempts to isolate a stem cell line can be frustrating.
4. Blastocysts that have undergone a period of implantational delay
prior to flushing may be more suitable material from which to recover
stem cells; the success rate can be significantly improved by the use of
 Robertson

delayed blastocysts. Implantational delay is brought about by alter-

ing the hormonalstatus of thepregnantfemale. The sourceof estrogen
production is removed by surgical ovariectomy 2.5 d postcoitum.
Lowering the level of estrogen, together with the postoperative ad-
ministration of a synthetic progesterone, prevents the embryos from
implanting. The embryos develop normally to the blastocyst stage,
but since the tissues of the uterine wall have been artificially rendered
nonreceptive, the embryos arrest and remain free floating within the
uterine lumen. In the delayed state, the embryos remain viable for up
to 10 d. For the purposes of isolating stem cells, a 4-5-d delay period
is normally used. Embryos are flushed from the uterine horns and
treated exactly as described for normal blastocysts.
5. There are a variety of technical considerations to be made regarding
the best strategy for the culture of blastocysts. These are as follows:
a. Embryos collected on a single day can be kept together and
cultured as one group, or split up into individual embryo
cultures. For most purposes, it is recommended that embryos
be cultured singly. This has the advantage of avoiding the
necessity of recloning the stem cell cultures obtained, since a
cell line derived from a single embryo is effectively a clonal
population in terms of genotype and sex chromosome com-
plement. An additional advantage is that the monitoring and
recording of embryos on an individual basis provides infor-
mation about the overall viability of cultures and the effi-
ciency of recovering stem cells. The major disadvantage to
this approach is that it takes considerably more time and
materials to maintain large numbers of individual cultures.
b. Blastocysts can either be grown in drop cultures under a layer
of liquid paraffin oil or in larger tissue culture dishes. The
advantage of drop cultures is that embryos are easily located
for observation because they are physically confined to a
small area. In our experience, however, embryos grown in
drop cultures for the necessary 5-6 d fare less well. This is
probably because of depletion of an essential growth require-
ment in the medium, or more likely results from the gradual
accumulation into the medium of traces of toxic residues from
the overlying oil. The most satisfactory method is to use small
tissue culture wells (10 mm) that hold about 1 mL of medium.
c. Embryos may either be cultured directly on tissue culture
plastic or on feeder layers. The use of feeder layers during
initial culture appears to maximize the viability of blastocyst
Embryonic Stem Cell Cultures

cultures. The stage at which an individual ICM-derived com-

ponent is selected for disaggregation is fairly critical. Figure
1 shows examples of the morphologies of blastocyst out-
growths. Under the conditions of culture described above,
embryos will normally attain a sui tablemorphology following
a 5-d culture period. A degree of variability in the size of the
ICM-derived cell clump between embryos of the same batch
is a common observation. As a result, it may be necessary to
disaggregate ICM clumps over a 2-d period according to the
morphology of individual cultures. It is important to note that
there is by no means an absolute correlation between the
phenotype of the outgrowth and the successful isolation of a
stem cell line; stem cells can be ob tained from both vigorously
growing ICMs and from small outgrowths. In our experience,
however, ICM clumps in which there is extensive endoderm
formation, or a relatively rapid progression to an overtly
multilayered egg cylinderlike structure, tend to have a re-
duced chance of retaining pluripotential cells.
6. It is important to be able to accurately identify colonies of pluri-
potential stem cells. Positive identification is difficult unless one is
familiar with the appropriate cellular morphology. Pluripotential
cells are typically small, have a large nucleus and minimal cytoplasm,
and the nuclei contain one or more prominent dark nucleoli struc-
tures, The cells pack tightly together in small nests in which it is
difficult to discern the individual component cells. Individual cells
aremost easily seen at the edges of the colony. A helpful exercise, prior
to attempting to isolate stem cells from embryos, is to plate out single
cells from a feeder-dependent Embryonal Carcinoma cell line onto a
feeder plate (e.g., 5 x lo2 cells on a IO-cm plate). Carefully observe the
growth of the resulting colonies over a period of 5-8 d.
7. It is recommended that embryo-derived stem cell lines should be
grown exclusively on feeder layers. This, together with careful cul-
ture in high-quality medium, acts to prolong the embryonic pheno-
type of the cell line specifically in terms of maintaining a high differ-
entiation ability and euploid chromosome complement. As with all
permanently established cell lines, long-term tissue culture will, how-
ever, select for abnormal cells within the population. For this reason,
it is advised that many samples be frozen as soon as possible after
founding a specific cell line. Cultures can be replaced as necessary or
recloned to establish euploid cultures. Recloning can be achieved by
picking single cells into feeder wells. The resulting colonies of cells are
236 Robertson

reexpanded by the method described. For experimental procedures in

which it is desirable to minimize the number of contaminating fibro-
blasts from the feeder layer, it is possible to passage the stem cells once
or twice on gelatinized plates. If cultures are kept under these con-
ditions for longer periods of time, extensive cell death occurs.

References
2. Evans, M. J. and Kaufman, M. H. (1981) Establishment in culture of pluripotential
cells from mouse embryos. Nature 292,154.
2. Bradley, A. (1987) Production and analysis of chimeric mice, in Terufocarcinomas and
Embryonic Stem Cells: A Prucficd Approach (Robertson, E. J., ed.), IRL Press Ltd.,
Oxford, p. 113.
3. Bradley, A., Evans, M., Kaufman, M. H., and Robertson, E. (1984) Formation of germ
line chimeras from embryo-derived teratocarcinoma cell lines. Nature 309,255.
4. Wagner, E. F. and Stewart, C. L. (1987) Integration and expression of genes intro-
duced into mouse embryos, in Experimental Approaches to Mammalian Emb yonic De-
velopment, (Rossant, J. and Pedersen, R. A., eds.), Cambridge University Press, UK,
p. 509.
5. Robertson, E., Bradley, A., Kuehn, M.,and Evans, M. (1986) Germ-line transmission
of genes introduced into cultured pluripotential cells by retroviral vector. Nature
323,154.
6. Kuehn, M., Bradley, A., Robertson, E., and Evans, M. (1987) A potential animal
model for Lesch-Nyhan syndrome through introduction of I-IPRT mutations into
mice. Nature 326,292.
7. Hooper, M., Hardy, K., Handyside, A., Hunter, S., and Monk, M. (1987) HPRT-
deficient (Lesch-Nyhan) mouse embryos derived from germ line colonization by
cultured cells. Nature 326,292.
8. Smith, A. and Hooper, M. (1987) Buffalo Rat Liver cells produce a diffusible activity
which inhibits the differentiation of Murine Embryonal Carcinoma and Embryonic
Stem Cells. Develop. Biol. 121,l.
 Chapter 21

Defining Hormone
and Matrix Requirements
for Differentiated Epithelia

Lola M. Reid
1. Introduction
The culture of differentiated cells requires conditions that acknowl-
edge the complicated cell-cell interactions that both occur in vivo and are
responsible for affecting and maintaining the differentiated states of cells.
In brief, one must use conditions that mimic the epithelial-mesenchymal
relationship that is universal and constitutes theorganizational basis for all
metazoan tissues. This relationship is sustained by a set of soluble signals
(autocrine, paracrine, and endocrine) and by a set of insoluble signals (the
extracellular matrix). Since this is a technical and methodological article,
neither the scientific evidence for the importance of the epithelial-mesen-
chymal relationship nor the evidence forming the basis for the culture
conditions will be described. Recent reviews have discussed this back-
ground in considerable detail (I-12).

1.1. Classical Cell Culture Conditions

As background and for the sake of comparison to the new forms of cul-
ture, the response of cells under classical cell culture conditions will be
noted. For a detailed description and discussion of classical cell culture
methods, see any of a number of recently published books (13-25, and
Chapter 1 of this volume).

237
238 Reid

All methods of preparing cells for culture start with the disruption of
the tissue and its dispersal into chunks or single-cell suspensions. In clas-
sical cell culture, the dispersed tissue or cells are plated onto an inflexible,
charged plastic substratum, and suspended in or covered with a liquid me-
dium. The plastic substratum, in the form of culture dishes, consists of
polystyrene dishes or containers exposed to ionizing plasma gas that pro-
duces a negatively charged surface that will permit the cells to adhere by
charge binding. The cells adhere and spread via the negative charges, and
then are fed with a liquid basal medium that typically consists of a specific
mixture of salts, amino acids, trace elements, carbohydrates, and so on, and
that is supplemented from 5 to 25% with a biological fluid, usually serum.
The basal media that are commonly used (e.g., RPMI, DME, BME,
Waymouth’s) wereoriginally developed for cultures of fibroblasts (23-U).
Although most of the constitutents in these basal media are requirements
for both epithelial cells and fibroblasts, some aspects of the basal media
must be redefined for some epithelial cell types (26-18). Two classic exam-
ples are trace elements and calcium levels. Highly differentiated epithelia
require various trace elements or other factors that act as cofactors for the
enzymes associated with their tissue-specific functions (16-22). The cal-
cium levels in many of the commercially available media are above 1 mM,
concentrations that are permissive for growth of fibroblasts, but that in re-
cent years have been shown to be inhibitory to most epithelial cell types
(16-22). This problem is exacerbated by culturing the cells in serum-sup-
plemented media, since serum also contributes significantly to calcium
levels. Most normal epithelial cells can grow in calcium concentrations of
approximately 0.4 mM (19-22). An excellent review of these aspects of ba-
sal media construction has been written by Ham and McKeehan (22).
Someinvestigators utilize serum autologous to the cell types to be cul-
tured. However, it is more common that the serum derives from animals
that are routinely slaughtered for commercial usage, such as cows, horses,
sheep, or pigs (13-25). Fibroblasts do well in serum-supplemented media
(SSM).” By contrast, epithelial cell types cultured in SSM dedifferentiate
rapidly and then die, usually within a week (23,24).
Tumor cell lines derived from epithelia are resistant to some of the ad-
verse effects of serum, and will grow and even express some of their tissue-
specific functions in SSM (25). Still, if these cell lines are switched into
serum-free, defined conditions, their differentiated functions are mark-
edly improved (26).

‘SSM, serum supplemented medium; HDM, serum-free, hormonally defined medium;

EGF, epidermal growth factor; GAGS, glycosaminoglycans, PGs, proteoglycans.
Requirements for Differentiated Epithelia 239

1.2. Realistic Expectations for Cultures

of Distinct Classes of Epithelial Cells
In order to culture epithelial cells, one must define the soluble signals
(hormones, growth factors) and the matrix signals needed to sustain the
cells either in a growth state or in a differentiated state. Using the defined
conditions results in cells that can approximate the behavior of their in vivo
counterparts. However, even with optimal defined conditions, one must
be realistic in one’s expectations for cultures of particular cell types. The
expectations differ with several general classes of cells.
Three classesof cells have been described in adult tissues that require
distinct conditions, and require distinct expectations and assumptions for
what the cells will do in culture. All involve stem cells and the lineage of
cells derived from them, and generally support the original hypothesis of
Pierce and Potter that all adult tissues have stem cells that can become the
targets for carcinogenic events (27). A schematic diagram of these classes
is given in Fig. 1. These classes are:
1. Stem cells that give rise to constantly regenerating tissues-eg., epi-
dermal cells (skin), bone marrow stem cells (hemopoietic cells), in-
testinal stem cells (2832). The stem cells require attachment to a form
of extracellular matrix, and the presence of specific hormones and
growth factors for viability and for sustained growth potential. The
adherent cells grow, but do not express their tissue-specific functions.
Detachment from the extracellular matrix substratum (mimicking the
loss of contact with the stromal layer) results in terminal differentia-
tion occurring in parallel with the loss of viability and restriction of
growth potential. The kinetics in the terminal differentiation process
and the loss of viability and growth potential vary with the cell type
(28-32). Cell cultures of this class of stem cells can be routinely cloned
and subcultured. However, cultures of the differentiated derivatives
are inherently short-lived; the cultures survive for a matter of days to,
at most, a few weeks. Obviously, the life span of the differentiated de-
rivatives is dictated by the inherent life span of the cell in vivo. It is this
in vivo life span that sets the limits for what is possible in culture.
2, Stem cells that give rise to quiescent tissues having some regenerative
potential-e.g., liver, lung, pancreas, prostate, mammary glands. Re-
cent studies (33-38) have provided strong evidence that all of these tis-
sues contain small numbers of stem cells that give rise to a cell lineage
gradually developing into a cell type forming the bulk of the cells in
the quiescent tissues. The cell type predominating in the quiescent tis-
sue is stabilized into a G, state (or a very long G, state) by extracellu-
240 Reid

Figure 1. Classes of Epithelia

Class 1: Tissues that are undergoing rapid renewal: skin, intestine,

haemopoietic cells

STEM CELLS --) --) + --) TERMINALLY DIFFERENTIATED

CELLS

Intermediates in the differentiation pathway

Class 2: Tissues that exist for long periods in a “quiescent” state but retain
some regenerative potential: liver, prostate, pancreas, lung

STEM CELLS + + + INTERMEDIATE CELL(S) STABILIZED

(the number BY MATRIX IN QUIESCENT TISSUE
of stem cells
declines with
age) L

1
~
1
Regeneration Signals (limited cell division)

TERMINALLY DIFFERENTIATED CELLS

Class 3: Tissues in which there is little to no regeneration possible: mus-

cle, neurons in the CNS

STEM CELLS + + -+ MATRIX STABILIZED TERMINALLY

DIFFERENTIATED CELLS
Requirements for Differentiated Epithelia 241

lar matrix and hormonal conditions. When regenerative conditions

occur, the signals, again both hormones and matrix, trigger these
quiescent cells to undergo a few rounds of division, typically 2-4 divi-
sions, accompanied by a terminal differentiation process. The signals
for regeneration also recruit the stem cells to send more cells through
the cell lineage. Generally speaking, fetal tissues have greater num-
bers of these stem cells than do adult tissues, and the older the animal,
the fewer the stem cells in these tissues. However, studies to date have
not defined well either the kinetics of aging of the stem cells or the real
replicative ability of these cells for any tissue in this class.
Defined culture conditions for the predominant cell in the tissue can
result either in amitotic cultures that are fully differentiated and able
to be maintained for a long-term (months) (39), or in cultures that can
undergo 2-4 divisions and then go into a state of growth arrest (40). In
this state, they can survive for months. However, no conditions are
known that will enable any of the predominant cell types from these
tissues to be routinely subcultured or cloned. It is hypothesized that
cultures of the stem cells from these tissues will be able to be cloned or
subcultured. Although a number of investigators are making consid-
erable progress in culturing stem cells (41,42), the conditions for rou-
tinely maintaining the stem cell populations have not yet been fully
defined for any stem cell from this class of tissues.
3. Stem cells that give rise to cells stabilized in the terminally differenti-
ated state-e.g., CNS neurons and muscle cells. Muscle and brain do
not regenerate, or do so in a most limited fashion (43,44). The tissues
are assumed, in adult state, to consist of terminally differentiated cells
stabilized by the matrix and hormones. Whether or not there are stem
cells in the adult muscle and brain tissues is unknown. All of the
known matrix and hormonal conditions for cultures of these tissues
result in cell populations that survive and function, but do not grow.

1.3. Defining the&equirements

for a Specific Epithelial Cell 5?)pe
1.3.1. Soluble Signals:
Autocrine, Paracrine, and Endocrine Factors
An approach to defining the soluble signals from cell-cell interactions
has been to replace the serum supplements in medium with known and
purified hormones and growth factors (16-18). This results in the develop-
ment of serum-free, hormonally defined media (HDM). Table 1 provides
some representative defined media for several epithelial cell types. There
242 Reid

Table 1
Constituents of “Hormonally Defined Media” for Several Representative
Epithelial Cell9
Proximal kidney tubule cells
ms~lm (5 &mL), transferrin (5 &mL), hydrocortisone (5 x WW), T3
(5 x l(PM), prostaglandin El (25 ng/mL), selenium (5 x 10m8M)(4%
Mammary epithdial ceh
Dexamethasone (10-8M), epidermal growth factor (10 ng/mL), insulin
(s-10 pg/mL),transferrin (5-10 pg/mL), prolactm (1 pg/mL), T, (IO-‘OM),
17P-estradiol (10-8M), linoleic acid/E&I (5 kg/mL) (50)
SertoIi cells
Epidermal growth factor (1 ng/mL), insulin (5 pg/mL), transferAn
(5 pg/mL), growth hormone (0.1 pg/mL), follicle stimulating hormone
(0.5 ng/mL,), hydrocortisone (10-8M), adrenocorticotropm (5 mU/mL),
vitamin A (50 ng/mL), and vitamin E (0.001%) (18)
*For recipes of other epithelial cell types (or for fibroblasts) seeMethods in Cell Biology,
Barnes and Sato (27) and Mather (18).

are several books and review articles giving recipes for defined media for
other cell types (26-18). Use of HDM results in selection of the epithelial
cell type of interest from primary cultures containing multiple cell types,
Almost all of the published HDM are optimized for cell growth. To observe
optimal expression of differentiated functions, the HDM must be re-
tailored (22). Each tissue-specific function requires a discrete set of
hormones and growth factors, often at concentrations that differ from
those required for cell growth. Furthermore, some of the hormones
conducive to growth can markedly inhibit tissue-specific functions. Thus,
a rule of thumb is to develop an HDM for growth of cells and then retailor
this medium for the conditions needed for differentiation of those cells (22).
The development of an HDM for a cell type begins by defining the
hormonal and growth factor requirements for growth of a minimally de-
viant neoplastic cell line (16-18). Since a cell line is already adapted to cul-
ture, it can readily be used in clonal growth assays to define growth re-
quirements under serum-free conditions. The HDM for the tumor cells
usually contains a subset of the requirements and is, therefore, a starting
point in defining the requirements for the normal cellular counterparts to
these tumor cells (45-48; seeTable 2 for examples).
Using the minimally deviant tumor cell line, one determines its clonal
growth efficiency (percent of cell colonies that grow at low seeding den-
Requirements for Differentiated Epithelia 243

Table 2
Comparison of the Composition of a Hormonally Defined Medium
for a Normal CelI (Hepatocytes)
vs Its Neoplastic Counterpart (Hepatomasp
Hepatocytes Hepatomas

EGF 50 ng/mL Not required

Insulin 1-5 pg/mL 5-10 pg/mL (or more if
insulin-degrading en-
zymes are produced by
the tumor cells)
Glucagon 10 pg/mL 10 pg/mL
Glucocorticoidb 10-8M 10”M
Transferrin Not required 5-10 pg/mL (depends on
whether the hepatoma
cells have, and to what
degree, the capacity to pro-
duce transfer&$
Linoleic acid’ 5 pg/mL 10 pg/mL
HlzL’ n. t. 10 pg/mL
Triiodothyronine Not required 1 x 10-9M
Prolactin 20 mU/mL 2 mU/mL
Growth hormone 10 pU/mL 10 pU/mL
Trace elements
zinc 1 x lO-‘OM 1 x lO-‘OM
Copper 1 x 10-7M 1 x 10-7M
Selenium 3 x lo-‘OM 3 x lo-‘OM
“These culture conditions (21,12,45,48)permit clonal growth of many hepatoma cell
lines (human, rat, and mouse cell lines have been tried). However, normal hepatocytes
will grow at high densities (>105cells/60mm dishes), but not at clonal seeding densities
(10*-l@ cells/6Omm dishes) when maintained in the above hormonally defined me-
dium. Furthermore, unless protease inhibitors and/or specific matrix substrata are used,
the cells must be seeded in a mixture of 5-10% serum plus the hormones for a few hours
(4-6 h) and then transferred into the serum-free, hormonally defined medium.
bGlucocorticoid can be hydrocortisone or dexamethasone, depending on the species.
‘Linoleic acid must be present in combination with fatty acid free bovine serum albu-
min (Pentax, Inc.) at a l/l molar ratio. If added alone, it is quite toxic to the cells. Some
species of hepatomas (e.g., the human hepatoma cell line, HepG2) have more complex
lipid requirements, especially those cell lines that are poor in making cholesterol or one
or another of the lipoproteins present in high density lipoprotein (HDL). These cells must
be given a more complex lipid source such as HDL.
244 Reid

sities, such as 100-1000 cells/bO-mm dish) on tissue culture plastic plates,

and under serum-supplemented medium (SSM) conditions or under se-
rum-free medium (SFM) conditions. This will establish the positive and
negative control conditions.
Since most differentiated epithelial cells are quite dependent for sur-
vival on substrata of extracellular matrix, one should determine the cells’
clonal growth efficiency in SSM and SFM using a simple collagen substra-
tum (45). Usually, the cells will attach and survive on collagen substrata
and in SFM for a week or more. Thus, one can use this condition as a base
in which to test the growth effects of soluble signals, one by one. For epi-
thelial cell types, type IV collagen is preferred. However, some epithelial
cell types will also accept type 1 collagen. Since type 1 collagen substrata
are easier and cheaper to prepare, one can try to use them if the cells permit.
The clonal growth efficiency of the minimally deviant tumor cells under
these various conditions establishes the critical controls:
1. Negative control-a condition (cells on tissue culture plastic and in
SFM) in which the cells have minimal survival and growth. This con-
trol will determine the extent to which the components in the basal
medium are able to facilitate cell survival and growth. For most epi-
thelia, the SFM condition permits no survival at all.
2. Survival state-a condition (collagen substratum + SFM) in which the
cells can survive but show limited growth or no growth.
3. Growthstate/positivecontrols-acondition (SSM +/-collagen)under
which the cells will show the highest clonal growth efficiency. It is
likely that the minimally deviant tumor cells will grow best on colla-
gen and in SSM.
In all future experiments, replicate plates of cells are plated under the
experimental conditions and under the control conditions, respectively.
To test for hormone or growth factor requirements, the cells are plated on
collagen substrata and in SSM, and then switched after they have attached
(3-4 h) to SFM. The cells are seeded at the lowest density at which they will
attach and survive when on collagen and in SFM.
Usually, purified hormones and growth factors are prepared indivi-
dually (see Compendium 1) and aliquoted as 1000 x stocks. The initial
concentration tried is based on any information available from in vivo or
in vitro studies on the relevance of the factor to the cell type of interest. The
factor is added to cultures that are under survival conditions: SFM and on
a collagen substratum. There should be 3-4 plates for each control or
experimental condition resulting in a total of 12-16 control plates plus
3-4 x the number of test conditions. Medium changes are done every
Requirements for Differentiated Epithelia 245

3-4 d. The cells are allowed to grow for 1-2 wk. The number of cell colonies
(number to attach and survive) and the diameter of the cell colonies (epi-
thelial cells) or cell density (fibroblasts) are measured. Any hormones or
growth factors that increase the number of cell colonies or the colony
diameter are studied further. Dosage studies are done, again using clonal
growth assays, over a 2-3 log range to determine the optimal concentration
of the factor in its effects on growth.
Once active factors are determined, they form the basis for a new con-
trol condition: SFM, collagen substratum, plus any biologically active fac-
tors at their optimal concentrations. This condition is then used as the base
condition to screen again for hormones and growth factors. All the factors
are tested again, since a factor found negative in the first screen can prove
positive in subsequent screens because of synergies with another soluble
signal. Any new growth factor requirements that are identified are again
assessed for their optimal concentrations. The cycle of:
1. Screening for factors on cells under the minimal survival or under the
defined “base” conditions
2. Determining the active factors
3. Determining their optimal concentrations and
4. Adding them at that concentration to the “base” condition for the next
cycle
is repeated until no new factors emerge in the screens. Once the active fac-
tors are identified, the optimal concentrations of each one are reassessed in
the completed defined medium.
A serum-free HDM has been made for many cell lines with the known
repertoire of hormones and growth factors. However, if a specific cell line
still grows better in SSM than in the HDM, then one must purify from
serum or screen tissues for unidentified factors (52).
One uses the HDM developed for the minimally deviant tumor cells
as a starting condition for their normal cellular counterparts. There is one
important caveat: normal cells will not survive at the cell densities that tu-
mor cells do. Empirically, one develops an HDM for the normal cells with
a starting seeding density of approximately 105/60-mm dish or higher. By
contrast, most tumor cells will survive seeding densities of 102-1@/60-mm
dish.
1.4. General Rules That Have Emerged
porn Studies Developing Hormonally Defined Media
From the studies in which HDM were developed for a variety of cell
types, several rules have been realized:
246 Reid

1. Since the primary determinant of cell attachment and survival is a ma-

trix substratum, development of an HDM for a cell line or for normal
cells is more rapidly accomplished with cells cultured on matrix sub-
strata permissive for growth.
2. All cells require transferrin, an iron-carrying protein. Iron is a require-
ment for the cells’ polymerases.
3. All cells require insulin or an insulin-like growth factor (insulin, IGFI,
IGFII, NGF).
4. All cells require an antioxidant when they are plated in culture. A
cheap antioxidant that is very potent is selenium.
5. All cells require a lipid source. For many cell types, linoleic acid
(bound to bovine serum albumin as a carrier) is sufficient as a primary
lipid source. For other cell types, more complex lipids, such as high
density lipoprotein (HDL), must be given.
6. If cells are on tissue culture plastic, most cells attach better and survive
longer, if a glucocorticoid (dexamethasone or hydrocortisone) is
given. The glucocorticoid is not usually required if the cells are on col-
lagenous or matrix substrata.
7. The hormone requirements for the minimally deviant tumor cells are
a subset of the requirements for their normal cellular counterparts
(45-47,53). EGF is a common growth factor requirement of normal
epithelial cells and not of their neoplastic derivatives. Platelet-derived
growth factor (PDGF) is a common requirement that is normal for
stromal cells but not for their neoplastic derivatives. Inboth cases,the
autonomy of the growth factor is the result of constitutive secretion of
the factor or activation of its hormonal pathway intracellularly be-
cause of an overexpression or a mutated oncogene (53).
8. Tumor cells will surive much lower seeding densities than the normal
cells.

1.5. Influence of HDM on Epithelial Cells in Culture

HDM has been found to select for parenchymal cells even when the
cells are on tissue culture plastic (11,12,16-19,24,45,48). This results, within
a few days, in cultures that are predominantly the cell type for which the
HDM was developed. However, if the cultures are plated onto tissue cul-
ture plastic and in HDM, the life span of the primary cultures has been
found to be approximately 1 wk, at which time, the cells peel off the plates
in sheets. Achievement of longer culture life spans has been found tobe de-
pendent upon using collagenous or matrix substrata in combination with
HDM (11,12,39,48).
Requirements for Differentiated Epithelia 247

1.6. Insoluble Signals: Extracellular Matrix

For many years, it has been apparent that extracellular matrix is neces-
sary, especially for normal cells to survive and function (12,12,17,28). In re-
cent years, the knowledge of matrix chemistry and biology and of its com-
plexity has become better understood (54-65). It is now recognized that,
in order to optimize cell attachment and survival, and to enable the cells to
optimally respond to hormones and growth factors, it is essential for the
cells to be plated onto substrata of extracellular matrix (21,22,26,66-70).
The extracellular matrix is a complex mixture of molecules between and
around cells made insoluble by crosslinking. Excellent reviews are avail-
able on its known chemistry and functions (55-65). It is important to note
that there are many types of matrices (see Tables 3 and 4) with distinct
chemical composition. Each cell type secretes and is associated with a
specific type of matrix. Furthermore, the matrix chemistry changes when
the cell is growing vs quiescent or when it is in some pathological condition
(11,22,55-57,71-73). Thus, if you want the cell type of interest to mimic its
in vivo counterpart in a specific physiological state, you must identify the
matrix chemistry associated with the cell in that physiological state. One
is helped by the understanding that there is a paradigm to how all types of
extracellular matrix are made, and by the numerous available studies
identifying the matrix chemistry in various tissues.

1.7. The Paradigm in the Construction

of Extracellular Matrix
All cells produce an extracellular matrix, and the extracellular matrix
in between any given set of cells contains components derived from all the
cell types in contact with the matrix. There are many chemically distinct
types of extracellular matrices produced by cells (55-65). Although no type
of matrix has been studied sufficiently to know all of its components, nev-
ertheless, all matrices consist of collagen scaffoldings (65) to which cells are
bound by adhesion proteins (60-64). In association with the collagens,
adhesion proteins, and the cell surface are polymers of sulfated amino
sugars, glycosaminoglycans (GAGS) that are bound to a protein core to
form proteoglycans (PGs) (58,59).
The chemically distinct extracellular matrices are readily grouped
into types dictated by the type of collagen used as the scaffolding (65,67).
Each collagen type is associated with specific adhesion proteins and
proteoglycans that are coordinately synthesized and assembled with the
collagen. For example, a “type I” matrix consists of type I collagen, fibro-
248 Reid

Table 3
Representative Matrix Types Produced by Cells In Vivo
Collagen Associated Associated cell Surface Cell(s)
type anchorage protein proteoglycan(s) receptor producing

Type * Fibronectin Dermatan sulfate, Integrin All physiological

chondroitin states: fibro-
sulfate blasts
Type II Fibronectin Chondroitin Integrin Chondrocytes
sulfate (cartilage)
Type III Fibronectin Heparan sulfate, Integrin Hepatocytes in
heparin quiescent state;
fibroblasts in
association with
epithelia
Type IV Laminin Heparan sulfate, Laminin All epithelial
heparin receptor cells; endothe-
lial cells; regen-
erating hepato-
cytes
Type V Fibronectin Heparan sulfate Integrin Hepatocytes in
quiescent state
Type VI Fibronectin Heparan sulfate, Integrin Hepatocytes in
heparin quiescent state

Table 4
Families of Matrix with No Known Collagen Scaffolding
Anchorage Associated Cell surface Tissue localization
protein(s) proteoglycams) receptor (cells producing)
Adhesion Heparan sulfate, ?? Central nervous system
molecules heparin (neurons)
(N-CAMs)
Fibronectin Heparan sulfate, Integrin Bone marrow, lymphatic
heparin (GP/GP) tissue (platelets)
Fibronectin Heparan sulfate, Integrin Bone marrow, lymphatic
heparin (VLA-l,VLA-2) tissue (lymphocytes:
(VLA-3,VLA-4) other hemopoietic cells)
Fibronectin Heparan sulfate, Integrin (Myeloid cell lines)
heparin (VLA-5)
Complement Heparan sulfate, Integrin Bone marrow, lymphatic
C3bi heparin (Mac-l) tissue, (monocytes,
granulocytes, large
granular lymphocytes)
Requirements for Differentiated Epithelia 249

nectin (plus any other anchorage proteins), and chondroitin sulfate proteo-
glycan or dermatan sulfate proteoglycan. By contrast, a “type IV” matrix
consists of type IV collagen, laminin, and heparan sulfate proteoglycan.
There are two tissues for which collagen scaffoldings have not been
identified. Hemopoietic cells produce an extracellular matrix that includes
a collagen scaffolding when they are part of the bone marrow. During their
differentiation into the various hemopoietic cells in the blood, the collagen
synthesis is lost. However, they retain the capacity to produce other matrix
components: the anchorage proteins and the proteoglycans. Increasingly,
these matrix molecules are being found to be of importance to the functions
of these cell types (31). Similarly, embryonic neurons of thecentral nervous
system produce a matrix that includes a collagen scaffolding (7,74). How-
ever, differentiation into adult, mature neurons results in the loss of colla-
gen synthesis. However, these cells are associated, both intracellularly and
extracellularly, with proteoglycans and glycosaminoglycans and extracel-
lularly with adhesion molecules (e.g., the N-CAMS). The proteoglycans
have been found to be important for regulating various aspects of neuronal
functions (7).
1.8. Rules for Cultures on Matrix Substrata
1. Cells attach in seconds to minutes on an appropriate matrix substra-
tum. Therefore, when plating cells on any matrix substrata, add the
cell in sufficient volume of medium to allow equal distribution over
the plates.
2. Since matrix is a mixture of proteins and carbohydrates, use DNA or
RNA stains rather than protein or carbohydrate stains for assessing
plating efficiency or clonal growth efficiency of cells on matrix.
3. Cells on complex matrices (matrigel, biomatrix) do not detach readily.
To do growth curves of cells on complex matrices, you may have to
scrape the plates, isolate the DNA, and determine DNA content.
4. Antibodies (and many other reagents) may nonspecifically stick to
matrices. For antibody staining of cells on matrices, expose the cells
to a control antiserum first and then to the specific antiserum.
5. Cells can achieve a very high density if cultured on matrix substrata,
especially if proteoglycans or glycosaminoglycans are used. For ex-
ample, typical saturation densities for normal cells on a loo-mm tissue
culture dish are 7-10 x 106;on appropriate matrix substrata, they can
be greater than lOa cells/lOO-mm dish.
6. Extraction of RNA or DNA from cell cultures on collagens, proteogly-
cans, or complex matrices should avoid first detaching the cells enzy-
matically unless (1) the enzymes are guaranteed to be free of nucleases
250 Reid

and (2) detachment does not result in alteration of the RNA levels.
Rather, use guanidine or guanidinum solutions on the cultures and
extract everything on the plate, both cells + matrix. Then use a purifi-
cation protocol that will accommodate the increase in proteins and
carbohydrates from the matrix (e.g., centrifuging through cesium
chloride).

2. Methods
2.1. Protocols for Culturing Epithelial Cells
2.1.1. In a Growth State with Completely Defined
Conditions
2.1.1.1. Normal Cells
1. Isolate the cells by standard protocols for the cell type you are using
(seeChapter 1, this volume).
2. Use a basal medium that has a low calcium concentration, approxi-
mately 0.04 mM.
3. Plate the cells at densities of 3-4 x W/60 mm or at least l@/lOO mm.
Use dishes coated with type IV collagen, a medium that contains both
the hormones and growth factors used in the HDM, and some serum
(2-10%) to permit them to attach and to inactivate any enzymes that
might have been used in isolating the cells. You can avoid the serum
altogether if you use laminin with the collagen and if you use no en-
zymes in isolating the cells (or have effective ways of inactivating
those enzymes without providing a condition that is toxic to cells).
4. Incubate the cells for 4 h at 37OC or until they are firmly attached.
5. Rinse the plates, removing debris and floating cells, and feed with the
serum-free HDM. Change the medium every day, or at the very least,
every other day (prepare the HDM fresh).
6. The cells will grow and will survive up to a month under these condi-
tions.
7. If the cells are stem cells from class I (seeFig. l), you can expect to be
able to subculture and clone the cells under these conditions. If the
cells are from class II, you can expect several rounds of division and
then growth arrest. You will not be able to clone or subculture them,
or you will have limited ability to do so. Cells from class III will not
grow under these conditions, but may survive for some weeks.
2.1.1.2 Tumor Cells
1. Tumor cells are plated onto tissue culture plastic (for more anaplastic
Requirements for Differentiated Epithelia 251

tumor cells) or on type IV collagen (for more differentiated tumor

cells) at densities of lo4 cells/60 or lo5 cells/lOO-mm dishes and into
HDM/SSM.
2. After 4 h, the cultures are rinsed and switched into the HDM designed
for the tumor cells.
3. The cells will survive and grow under these conditions for 7-10 d.
4. The tumor cells can be subcultured and cloned under the defined con-
ditions.
5. To date, no HDM plus matrix condition is sufficiently defined to per-
mit indefinite maintenance of stock flasks of cells.
2.1.2. In a Fully Differentiated State
with Completely Defined Conditions
2.1.2.1. Normal Cells
1. Isolate the cells as a single-cell suspension by standard methods.
2. Plate the cells at high density,6-8 x 106/100-mm dish, into HDM/SSM
and onto one of the following substrata: type IV collagen mixed with
laminin for most epithelial and endothelial cells or a fibrillar collagen,
ideally type III collagen, mixed with fibronectin for hepatocytes.
3. After 4 h (or whatever length of time it takes to allow the cells to at-
tach), rinse the cells and give them the serum-free HDM retailored to
optimize tissue-specific gene expression. Tailor the medium (as de-
scribed earlier) to suit whichever genes are to be optimally expressed.
4. The medium fed to the cells after 4 h (not the plating medium) must be
supplemented with an appropriate species of glycosaminoglycan (at
10-50 pg/mL) or ideally the corresponding proteoglycan (at l-5 pg/
mL). One can mix the GAG or proteoglycan in with the mixture of
collagen and adhesion protein; however, there may be some problems
in attachment of the cells. For epithelial and endothelial cells, the gly-
cosaminoglycan/proteoglycan will be a species of heparin or ahighly
sulfated species of heparan sulfate proteoglycan. For cartilage cells,
the glycosaminoglycan/proteoglycan will be chondroitin sulfate or
chondroitin sulfate proteoglycan.
5. Note that the cells will not grow, but will survive. They will usually
be three-dimensional and highly differentiated.
6. The cultures will survive for perhaps 1 mo and will retain most of their
differentiated functions. If you are able to add the proteoglycan, you
will have normal transcription rates for most of the tissue-specific
genes. If you use the glycosaminoglycan, you will also seenormal or
near normal transcription rates of some tissue-specific genes, but not
252 Reid

as many genes or to the same extent as that seen with the proteogly-
can. The proteoglycans and glycosaminoglycans can be tissue-spe-
cific. So, ideally use the proteoglycan or GAG from the tissue being
cultured. With the limited availability of GAGS (and even more so for
proteoglycans) you will probably have to screen the available ones
and use the one that is most active on your cells.
2.1.2.2. Tumor Cells
1. Prepare stock flasks of tumor cells as single-cell suspensions.
2. Plate the cells at high density (6-S x lo6 cells/lOO-mm dish) under the
same conditions as those used for the normal cells.
3. The cells will go into growth arrest and, simultaneously, will differen-
tiate to the extent that they are capable of differentiating. Note that, if
a tumor cell line is completely silent for a gene, the known culture con-
ditions for differentiation have yet tobe shown to activate such a gene
(45). However, if the tumor cells express even tiny amounts of the
gene product, the cells can usually be made to express very high levels
of that product. Thus far, this effect has been found to be predomi-
nantly by posttranscriptional mechanisms such as mRNA stability.
However, some limited improvements in transcriptional signals for
tissue-specific mRNAs have been observed.
4. The cells will stay in a state of growth arrest for a matter of days (if they
are anaplastic tumors) to a few weeks (if they are minimally deviant
tumor cells). The well-known ability of tumor cells to secrete enzymes
that degrade matrix and/or to overexpress oncogene products (usu-
ally products that are the same as a hormone active on the cell or prod-
ucts that activate the hormonal pathways intracellularly) ultimately
overrides the growth-arrested/differentiated state caused by the de-
fined matrix and hormonal conditions. The tumor cells will begin to
grow in foci, similar to transformation foci, piling up on top of one
another. The cells in these foci will be much less differentiated than the
growth-arrested cells. If you remove the growing foci of cells and
plate them on fresh matrix plates and into fresh medium, they will
again go into growth arrest and into a differentiated state.

2.2. Culturing Cells in Defined Medium Conditions

but with Substrata of Tissue Extracts Enriched
in Extracellular Matrix
One can use the tailored HDM in combination with a substratum of tis-
sue extract enriched in extracellular matrix. This will replace the defined
Requirements for Diferentiated Epithelia 253

collagen, adhesion protein, and proteoglycan combination. There are

several published protocols for preparation of such extracts.
1. Matrigel-a urea extract (75) from EHS tumors, a transplantable
mouse embryonal carcinoma that constitutively produces basement
membrane components (76). It has remarkable effects on the differen-
tiation of cells (77-79). Commercially, it is sold at great cost. However,
you can readily (and much more cheaply) prepare it yourself from the
tumor. Tumor stocks are maintained either in C57BL/6 mice or in
athymic nude mice (seethe protocol for matrigel).
2. Biomatrix-a tissue extract prepared by extracting the tissue with
NaCl solutions, nucleases, and detergents (26,39). The extract can be
made from any tissue. It is the only matrix that has been shown to
result in very stable, long-term, quiescent cultures of normal epithe-
lial cells (39). It is partially depleted of proteoglycans and glycosami-
noglycans, which are restored during the first few days of cell culture
(restored by the cells). However, if you need maximal differentiation
within a few days of plating the cells, you must add the proteoglycans
or glycosaminoglycans to the medium (seethe protocol for biomatrix).
3. Matrix from amnions-amnions are extracted with dilute alkali or
detergent and rinsed (80). The matrix on the side of the epithelial
surface is a type IV matrix. On the other side (the stromal surface) is
a type III matrix. You can choose which surface is needed for your cell
type. Studies using this matrix indicate that a given cell type will
respond dramatically differently if plated on one side vs the other (82).
4. ECM (extracellular matrix)-cell cultures are extracted with dilute
alkali or detergent and rinsed (82). The matrix chemistry depends on
the cell type used in culture and the conditions under which the cells
were grown. Cells attach readily, grow well, and show improved dif-
ferentiation (albeit not to the same extent on matrigel or on long-term
cultures as on biomatrix) (83).

2.3. Protocols for Preparing and Storing Hormones

and Extracellular Matrix Components
2.3.1. Methods of Preparation and Storage
Protocols are described in Compendia 1 and 2 for sources and prepa-
ration of commercially available hormones and matrix components. They
can be used for all cell types, For the specific requirements for each cell
type, consult the literature. Since the costs of the hormones and other
factors vary with time, they will not be given.
254 Reid

2.3.2. Preparation of l)pe I and lV Collagens

Protocols for the preparation of two collagen types, type I and IV, will
be given. For other collagen types, see the methods of Miller (65).
2.3.2.1. Protocol for Preparation of Qpe I Collagen Substrata
[modified from protocols published by Pitot and associates (54)]
1. Materials
a. Several rat tails (preferably about 250 g body wt).
b. One large hemostat or common pliers.
c. One medium pair of scissors.
d. One medium forceps.
e. Several loo-mm Petri dishes, or 4-6 in. weighing boats.
f. One filtration apparatus consisting of a Buchner funnel linked
with two layers of cheesecloth, set in a 2-L sidearm flask.
Autoclave entire apparatus.
g. One large-mouthed bottle (such as a roller culture bottle) ster-
ilized with a magnetic stir bar.
h. Acetic acid (l/1000 v/v) solution.
2. Procedures
a. Freeze tails after cutting from rats and keep frozen until use.
b. When ready to use, wash the tails slightly in 70% ethanol.
c. After tails are thawed, skin them.
d. Grasp the base of the tail in one hand, and with the hemostat
or pliers in the other hand, grasp the tip of the tail, break at a
joint, and slowly pull out the tendons. Place them on a paper
towel to dry.
e. Repeat step 4 one or two joints at a time until no more tendons
can be pulled out.
f. The tendons dry rapidly. When dry, cut off the fleshy verte-
brae and cut the remaining fibers into 5-4 cm lengths.
g. Then with scrubbed hands, rub bundles of fibers between
thumb and forefinger to break into individual fibers. Spread
in dishes or boats and place under ultraviolet light for 24 h.
h, Using the ratio 1 g fibers/300 mL of acetic acid solution, deter-
mine a proper amount of solution, and add this to a sterile
bottle together with a magnetic stir bar.
i. With a flamed forcep, put the sterilized fibers into the bottle
with the acetic acid solution.
j. Stir moderately for 48 h or more.
k. Filter the collagen solution through the cheesecloth using
Requirements for Differentiated Epithelia 255

house line vacuum. This process gets rid of any remaining

undissolved fibers.
1. Using sterile technique, pour into bottles andlabel as Solution
A.
2.3.2.2. Preparation ofPlates Coated with Type I Collagen. Gels are
made by combining two solutions: Solution A: collagen in 1 /lOOO acetic
acid (prepared earlier); Solution B: 2 parts serum-free medium (10x) to 1
part 0.34N NaOH (with DME, use 6x strength).
1. Set out on trays the desired number of 60-mm (or loo-mm) Falcon
plastic tissue culture dishes Put 1.7 mL of Solution A into each dish
with a 10 mL pipet.
2. Then add 0.4 mL of Solution B to each dish, a few dishes at a time (with
a 2-mL pipet), and swirl until the color (dark pink or lavender) is con-
sistent throughout the dish.
3. The collagen solution gels in a matter of seconds (usually less than 1
min). (If 150-mm dishes are used, add 1.3 mL (2.6 mL) Solution B to
5 mL (10 mL) of Solution A. If 150-mm dishes are used, add 3.0 mL
Solution B to 11.5 mL of Solution A.
4. After all gels are made, add 3 mL (7 mL for large dishes) of phosphate
buffered saline at pH 6.8 to each dish. This is usually allowed to sit
overnight, although this is not absolutely necessary.
2.3.3. Type IV Collagen
Based on the protocol by Kleinman and associates (75), which in turn
is based on the original procedures by Miller (65).
2.3.3.1. EHS Tumors (Starting Material)
1. EHS tumor cells are grown as transplantable tumors in C57B1/6 mice
or in athymic nude mice of any strain. The EHS tumor is known for
its unusual property of constitutively secreting a matrix with a chem-
istry appropriate for basal lamina (type IV matrix). It is now the pri-
mary source for most basal lamina components. The protocol given is
the one used with transplantable tumors maintained in athymic nude
mice. Tumors should be grown subcutaneously on both sides of the
animal and allowed to grow to l-3 cm in diameter. For a starting
preparation of EHS tumor-derived type IV collagen, use 250-350 g of
tumor. Since each tumor removed from the mice weighs from l-2 g
(bigger tumors become necrotic), the number of mice needed, carry-
ing tumors on both sides, will be a minimum of 125 and, on average,
about 200. From approximately 300 g of tumor, you can get approxi-
256 Reid

mately 20 mg of pure type IV collagen. To get such yields, the animals

must be treated with lathyrogen, a reagent that will block cross-link-
ing of the matrix components. The lathyrogen typically used is p-
aminoproprionitrile (BAPN). For reasons not yet fully understood,
BAPN alone results in rather limited lathyritic condition with this
tumor. Kleinman and associates have found that other pharmacologi-
cal treatment of the mice, in addition to the BAPN, is necessary. Their
method is: mix 100 g of BAPN, 2 g of Iproniozid (Sigma), 0.2 g of par-
gyline (Sigma) into 300 mL of distilled water. Sterilize by filtration.
Then mix the solution with 5 kg of autoclaved chow. Let it sit for 15
min. Serve it cold.
2. The mice should be killed, and the tumors dissected free under ster-
ile conditions. The tumors should then be placed immediately into a
cellector (Bellco Glass, Inc.) with a 20 mesh (860 @4) grid and the
tumors pressed through the grid into ice-cold, sterile, serum-free me-
dium (e.g., RPM1 1640) containing antibiotics. This step will eliminate
the capsule that surrounds the tumor; the capsulecontains type1 colla-
gen and fibrin.
3. Collect the tumor cells (each one is surrounded by a thick band of basal
lamina); put into a 50-mL, sterile centrifuge tube and centrifuge for 5
min at 1500 rpm. Obviously, if you have dissected free many tumors,
you can pool them into a large container and centrifuge all of them
together.
4. Throw away the supernatant, and use the pellet for preparing type IV
collagen. You can store the pellets in liquid nitrogen or in a deep
freezer at -7OOCuntil you are ready to extract and purify type IV col-
lagen or any other basal lamina component.
2.3.3.2. Extraction and Purification of Type IVCollagen from EHS
Tumors. All solutions and centrifugations should be at 4OCunless other-
wise indicated.
1. Buffer A g/L
3.4M NaCl 198.7
0.05M Tris-HCl, pH 7.4 6.0
2 mM N-ethyl maleimide (NEM) 0.25
8mM EDTA 2.69
2. Thaw the EHS tumor cells (or if you have just isolated them, use them
immediately) and put them in ice-cold Buffer A at a ratio of 1 mL of
buffer/g of tumor.
3. Homogenize with a polytron homogenizer for about 2 min and spin
Requirements for Differentiated Epithelia 257

at 12,000 rpm for 30 min. Throw away supernatant. (This eliminates

blood from the tumors.)
4. Homogenize the pellet in Buffer A with the same vol of buffer used in
step2,and spin again at 12,000 rpm for 30 min. Again, throw away the
supernatant.
5. Resuspend pellets in
Buffer B R/2 L
0.5M NaCl 58.4
0.05M Tris-HCl, pH 7.4 12.1
2mMNEM 0.5
8 mM EDTA 5.4
Use 3 mL of Buffer B/g of tumor wt (original tumor wt in step 1) and
stir overnight. Spin at 12,000 rpm for 30 min. Save supernatants for
laminin preparation; save pellets for type IV collagen.
6. Resuspend pellets in Buffer B again. Stir for several hours, and then
spin at 12,000 rpm for 30 min. Again save supernatants for laminin
preparation and pellets for type IV collagen. (Supernatants can be
frozen at -7OOC until ready for purification of laminin.)
7. Extract pellets with Buffer C.
Buffer C g/L
2.OM GuHCl (Guanidine hydrochloride) 190
0.05M Tris-HCl, pH 7.4 6
2mMNEM 0.25
8 mM EDTA 2.69
Buffer C should be used at ratio of 2 mL/g of tumor (original tumor
wt). Stir overnight. Spin at 12,000 rpm for 30 min. Save both superna-
tants and pellets.
8. Reextract pellet again in Buffer C for several hours. Spin at 12,000 rpm
for 30 min. Pool supernatants from first and second extractions. Pool
pellets from first and second extraction. Steps 7 and 8 result in extrac-
tion of the type IV collagen, but further extractions in increasingly
harsher conditions increases the yields. Freeze supernatants at -70°C
until ready for them (supernatants: go to step 10).
9. Suspend pellet (from second extraction with Buffer C) in Buffer D
Buffer D g/L
2.OM GuHCl 190
0.05M Tris-HCl, pH 7.4 6
2mMNEM 0.25
258 Reid

8 mM EDTA 2.69
2 mM Dithiothreitol (D’IT) 0.31
Use 1.5-2 mL of Buffer D/g of tumor (original tumor wt-step 1) and
stir overnight. Centrifuge at 12,000rpm for 30 min. Save supernatants
and pellets. Extract pellet again overnight with Buffer D, centrifuge
and save supernatant and pellet. Extract pellet a third time over up to
48 h, centrifuge and save supernatant. Extract pellet a fourth time.
Then discard pellet. Pool supernatants from all extractions with
Buffer D. (You will treat the supernatants from extraction with Buffer
D the same as those resulting from extraction with Buffer C. However,
keep those from Buffer C and Buffer D separate.) Supernatants: go to
step 10.
10. Dialyze supernatants (separately) resulting from extraction with Buf-
fer C or D against Buffer E (ratio of at least 1:lO v/v>. Change Buffer
E several times. Rotate bags.
Buffer E g/l0 L
1.7M NaCl 993.5
0.05M Tris-HCl, pH 7.4 60.5
2mMDTT 3.1
1mM EDTA 3.4
0.1 mM Phenylmethylsulfonylfluoride (PMSF) 10 mL of stock
solution (100 mM in 100% ethanol)
11. Spin at 12,000 rpm for 90 min to collect the precipitates from the dial-
ysis. Combine pelleted precipitates from extractions with Buffers C
and D.
12. Resuspend pelleted, precipitated material in Buffer D, about 1/lO the
original volume, and stir overnight in cold room to get the precipitated
material into solution.
13. Spin down any undissolved pellet (usually small) at 16,000 rpm for 30
min. Throw away pellet. Save the supernatant.
14. Determine protein concentration using Bradford’s reagent and run a
5-16% reducing gradient SDS gel to check purity.
15. Store supernatant in this solution at 4°C (do not freeze).
16. To make gels, dialyze supernatant from 13 vs Buffer F at 4OC. Use l/
10 ratio (v/v). Change at least twice and rotate bags.
Buffer F g/2 L
4M Urea 480
0.25M NaCl 29.2
O.,O5MTris-HCl, pH 8.6 12
Requirements for Differentiated Epithelia

2mM DTT 0.32

0.1 mM PMSF 2mL
NOTE: If the resulting gels are not washed thoroughly, they are very
toxic to cells; we have obtained more consistent results by dialyzing
the supernatant against any basal medium, such as RPM1 1640 or
DME/12.
17. Determine protein concentration, as above (step 14).
18. Remove collagen solution from dialysis bags and add to plates to give
thin coating on culture plates. Put plates at room temperature (or at
37OC). Gel will form within an hour. Sterilize gels on the plates using
gamma irradiaton (10,000 rads).
19. Rinse plates repeatedly with sterile, double-distilled water to elimi-
nate the salts, urea, and so on.
20. Culture the cells on the plates as usual.
2.3.4. Tissue Extracts Enriched in Extracellular Matrix
2.3.4.1. Matrigel. Matrigel _is.~~
prepared
~. from EHS tumors according
to the protocol of Kleinman et al. (75,78).
1. Preparation of EHS tumors (starting material).
a. EHS tumor cells are grown as transplantable tumors in
C57Bl/6 mice or in athymic nude mice of any strain (seesec-
tion 2.3.3.1.).
b. The mice should be killed and the tumors dissected free under
sterile conditions. The tumors should then be placed imme-
diately into a cellector (Bellco Glass, Inc.) with a 20 mesh (860
l..tM>grid and the tumors pressed through the grid into ice-
cold, sterile, serum-free medium (e.g., RPM1 1640) containing
antibiotics. This step will eliminate the capsule that sur-
rounds the tumor; the capsule contains type I collagen and
fibrin.
c. Collect the pooled tumor cells (eachone is surrounded by a
thick band of basement membrane); put into a 50-mL, sterile
centrifuge tube and centrifuge for 5 min at 1500 rpm.
d. Throw away the supernatant and use the pellet for preparing
the matrigel. You can store the pellets in liquid nitrogen or in
a deep freezer at -7OOCuntil you are ready to extract and pre-
pare the matrigel. For the preparation of matrigel from EHS
tumors, all solutions and centrifugations should be at 4OC
unless otherwise indicated.
2. Preparation of matrigel from the tumors
260 Reid

a. Thaw the EHS tumor cells (or if you have just isolated them,
use them immediately) and put them in ice-cold Buffer A (see
extraction of Type IV collagen) at a ratio of 10 mL of buffer/
g of tumor.
b. Homogenize with a polytron homogenizer for about 2 min
and spin at 12,000 rpm for 30 min. Throw away supernatant.
(This eliminates blood from the tumors).
c. Suspend the pellets in Buffer A and stir for 1 h at 4°C.
d. Centrifuge for 20 min at 12,000 rpm at 4OC, and again save the
pellets.
e. Extract the pellets overnight with 2 M urea in 0.5M Tris-HCl,
pH 7.4.
f. Centrifuge at 10,000 rpm at 4°C for 30 min and save superna-
tant.
g. Extract pellet again, centrifuge, and pool supernatant with
that from step 6.
h. Dialyze the urea extract against 0.15M NaCl in 0.05M Tris-
HCI, pH 7.4.
i. Centrifuge at 15,000 rpm in a Sorvall for 20 min to remove any
insoluble material.
j. Store aliquoted supernatant at -20°C.
3. To coat dishes, slides, or any substratum
a. Add the viscous solution (4OC)onto plates or dishes. Use ap-
proximately 1 mL/60-mm dish. The thickness of the gel is
dictated by the length of time you want to culture the cells. For
unknown reasons, the gels with cells on them gradually thin
over time. Once the gel is gone, the cultures die rapidly.
b. Bring the plates to room temperature.
c. Sterilize the dishes with 10,000 rads of gamma irradiation.
The dishes can be stored at 4OCuntil used.
d. Once the matrigel is on the plates, you can store the plates at
4OCuntil use.
2.3.4.2. Biomatrix
1. Isolation of biomatrix: biomatrix can be prepared from any tissue,
from any animal, and from tissue that is normal or diseased. This
protocol is based on the procedures of Reid and associates (26,39).
a. Mince tissue and homogenize in a Waring Blender or with a
Polytron homogenizer using 10 vol (to 1 vol of mince) of
Requirements for Differentiated Epithelia 261

“Buffer G “: 34MNaCl+ 0.1 mg/mL soybean trypsin + anti-

biotics (e.g., penicillin/streptomycin) at concentrations stan-
dard for cell culture (all at 4OC; seeAppendix, this vol.). Ho-
mogenize thoroughly with 10-15 s pulses.
b. Centrifuge for 20 min at 10,000 rpm at 4OC in a preparative
centrifuge.
c. Save the pellets and resuspend in Buffer G. Stir for 1 h at 4°C.
d. Centrifuge for 20 min at 10,000 rpm at 4OCand again save the
pellets.
e. Repeat the extractions of the pellets with Buffer G until the
supernatants after centrifugation are clear and are negative
for proteins by Lowry or Bradford assays or by optical density
at 280 nm.
f. Rinse pellets twice, each time for 1/2-h, in serum-free basal
medium (e.g., RPM1 1640) containing the soybean trypsin in-
hibitor at 4OC.
g. Suspend pellets in a serum-free basal medium (must contain
calcium and appropriate salts for the nucleases; e.g., RPM1
1640) containing nucleases: For each 10 mL vol of pellet, sus-
pend it in 100 mL of solution containing 1.0 mg DNase and 5
mg RNase. Stir at 37OC for 1 h.
h. Centrifuge for 20 min at 10,000 rpm at 4°C and save the pellet.
i. Repeat steps 7 and 8 until the supernatants (after centrifuga-
tion) give spectrophotometric readings of less than 0.1 at 260
nm.
j. Suspend pellets in 1% sodium deoxycholate in basal medium
for 1 h at room temperature. You must prepare the detergent
as a 10x stock in distilled water first-it does not dissolve
easily; then, disperse it into any basal medium-e.g., RPM1
1640, DME, F12.
k. Centrifuge samples as above and save the pellets.
1. Rinse the pellets (now considered ‘biomatrix”) three times in
1/2-h rinses with serum-free basal medium (e.g., RPM1 1640).
m. Suspend the pellets in serum-free basal medium (v/v of l/5).
n. Sterilize with 10,000 rads of gamma irradiation.
o. Store at -7OOCfor long-term storage or at 4°C for samples to be
used within a month. (Sterilization is imperative for all sam-
ples stored at 4°C.)
262 Reid

2. To coat dishes, slides or any substratum

a. Freeze matrix pellets at liquid nitrogen temperature.
b. Pulverize matrix pellets-with a SpexMill Freezer Mill (Me-
tuchen, New Jersey) into powder, keeping the sample at
liquid nitrogen temperature.
c. When the sample is completely pulverized, allow it to come to
room temperature. It will become like paint. Use a simple,
Camel hair brush (small artist’s brush) to paint the pulverized
biomatrix onto dishes, slides, and so on.
d. Sterilize the dishes with 10,000 rads of gamma irradiation.
The plates, dishes, and so on, can be stored at 4OCuntil used.
Just prior to use, soak the plates with serum-free basal me-
dium to be used with the cells (e.g., RPM1 1640) all at 37°C in
a regular CO, incubator.

Acknowledgments
I wish to thank Elaine Halay for helping to collect the information on
the commercial sources and preparations of the hormones and matrix mol-
ecules. I also wish to thank Rosina Passela for excellent secretarial assist-
ance, and Isabel Zvibel, Elaine Halay, and Maria Lourdes Ponce for helping
to edit and proofread the manuscript. This research was supported by a
grant from the American Cancer Society (BC-439) and by grants from the
National Institutes of Health (CA30117, P3O-CA13330, AM17702). Lola
Reid receives salary support through a Career Development Award (NIH
CA00783).

References
I. Cunha, G. R. (1976) Epithelial-stromal interactions in development of the urogenital
tract. Int. Rev. CytoZ. 47,137-194.
2. Yamada, K. M., ed. (19831 Molecular Mechanisms, Cell Interactions and Deueloprnent
(Wiley Interscience, NY).
3. Protero, J. (1980) Control of stem cell proliferation: A density-dependent commit-
ment model. J. Theor. Bicl. 84,725-736.
4. Reid, L. M. and Jefferson, D. M. (1984) Culturing hepatocytes and other differenti-
ated cells. Hqxltology 4,548-559.
5. Fleischmajer, R. and Billingham, R. E., eds. (1968) Epitheliul-Mesenchymalln teructions
(Williams and Wilkins, Baltimore).
6. Stoker, M. and Gherardi, E. (1987) Factors affecting epithelial interactions. Ciba
Found. Symp. IU, 217-239.
7. Edelman, G. M. (1987) Epigenetic rules for expression of cell adhesion molecules
during morphogenesis. Cibu Found. Symp. 125,192-216.
Requirements for Differentiated Epithelia

8. Clayton, D. F., Harrelson, A. L., and Darnell, J.E.,Jr. (1985lDependenceof liver-spe-

cific transcription on tissue organization. Mo2.and Cell. Biol. 5,2623-2632.
9. Guillouzo, A. and Guguen-Guillouzo, C. (1986) Isolated and Cultured Hep&cytes.
(John Libbey, Eurotext Ltd. INSERM, Paris).
10. Fraslin, J. M., Kneip, B.,Vaulont, S.,Glaise, D., Munnich, A., and Guguen-Guillouzo,
C. (1985) Dependence of hepatocyte-specific gene expression on cell-cell interac-
tions in primary culture. EMBO ].4,2487-2491.
11. Reid, L. M. and Jefferson, D. (1984) Cell culture studies using extractsof extracellular
matrix to study growth anddifferentiationinmamma lian cells. Mammalian CeuCul-
ture (Mather, J.,ed.), Plenum, NY, pp. 239-280.
12. Reid, L. M., Abreu, S., and Montgomery, K. (1988) ExtracelIular matrix and hor-
monal regulation of gene expression. Liver BioZogyandPathology(Arias, I. M., Jakoby,
W. B., Pepper, H., Schachter, D., and Shafritz, D. A., eds.), Plenum, NY, chapter39,
pp. 717-737.
23. Freshney, R. I., ed. (1985) Animal Cell Culture: A Practical Approach ORI..,NY).
14. Paul, J. (1975) Cell and Tissue Culture (Churchill Livingstone, NY).
25. Jacoby, W. B. and Pastan, I. (1979) Cell culture. Methods in Enzymology, vol. 58.
16. Barnes, D. and Sato,G. (1980)Serum-freeculture: aunifyingapproach. CeIf 22,649-
655.
17. Barnes, D. and Sato, G. (19841Methods for serum-free culture of cells. Cell Culture
Methodsfor Molecular and Cellular Biology ~01s.l-&(Liss, NY).
28. Mather, J.,ed. (1984) Mammalian Cell Culture (Plenum, NY).
29. Boyce, S.T. and Ham, R. G. (1983) Calcium regulated differentiation of normal hu-
man epidennal keratinocytes in chemicalIy defined clonal culture and serum-free
serial culture. J. Invest. Dermafol. 81,33s-44s.
20. Eckl,P.M.,Whitcomb,W.R.,Michalopoulos,G.,andJirtle,R.L.(1987)EffectsofEGF
and calcium on adult parenchymal hepatocyte ProIiferation. 1. CeII Pht@. 132,
363-366.
21. Stanley, J. R. and Yuspa, S.H. (1983) Specific epidermal protein markers modulated
during calcium-induced terminal differentiation. J. Cell Biol. 96,1809-1814.
22. Ham, R. and McKeehan, W. (1979) Media and growth requirements. Methods in
Enzymology55,44-93.
23. Clay ton, D. F.and Darnell, J.E.,Jr. (1983)Changesinliver-specificcompared to com-
mon gene transcription during primary culture of mouse hepatocytes. Mol. andCell.
Biol. 3,1552-1561.
24. Jefferson, D. M., Clayton, D. F., Darnell, J. E., Jr., and Reid, L. M. (1984) Posttran-
scriptional modulation of gene expression in cultured rat hepatocytes. Mol. Cell.
Biol. 4,1929-1934.
25. Clayton, D. F., Weiss,M., and Darnell, J. E., Jr. (1985) Liver-specific RNA metabo-
lism in hepatoma cells: Variationsin transcription ratesand mRNAlevels. Mol. Cell.
Biol. $2633-2641.
26. Muschel, R., Khoury, G., and Reid, L. M. (1986) Regulation of insulin mRNA abun-
dance and adenylation: Dependence on hormones and matrix substrata. Mol. Cell.
Biol. 6,337-341.
27. Pierce, G. B.,Arechaga, J.,and Wells, R. S.(1986) Embryonic control of cancer. Pmg.
Clin. Biol. Res.226,67-77.
28. Potter, C. S.,Chadwick, C., Ijiri, K., Tsubouchi, S.,and Hanson, W. R.(19841There-
cruitabiltiy and cell cycle state of intestinal stem cells. J. Cell Cloting 2,12t%I40.
29. Dudouet, B., Robine, S., Huet, C., Sahuquillo, M., Blair, L., Coudrier, E., and Lou-
264 Reid

vard, D. (1987) Changes in villin synthesis and subcellular distribution during intes-
tinal differentiation of HT29-18 clones. J. Cell Biol. 105,359-369.
30. Dexter, T. M., Whetton, A. D., Spooncer,E., Heyworth, C., and Simmons,P. (1985)
The role of stromal cells and growth factors in hemopoiesis and modulation of their
effects by the src oncogene. J. Cell Sci.3,83-95.
32. Gallagher, J. T., Spooncer, E.,and Dexter, T. M. (19831Role of the cellular matrix in
hemopoiesis. I. Synthesis of glycosaminolglycans by mouse bone marrow cell cul-
tures. J. Cell Sci. 63,155-171.
32. Barrandon, Y. and Green, H. (1985) Cell sizeas a determinant of the clone-forming
ability of human keratinocytes. Pm. Natl. Acud. Sci. USA 82,5390-5394.
33. Haynes,N. T., Braun, L.,Yaswen, P.,Brooks, M., and Fausto,N. (1984) Isozyme pro-
files of oval cells, parenchymal cells and biliary cells isolated by centrifugal elutria-
tion from normal and preneoplastic livers. CancerRes.44,332-338.
34. Hixson, D. C. and Allison, J. P. (1985) Monoclonal antibodies recognizing oval cells
induced in the liver of rats by N-2 fluorenylacetamide or ethionine in a choline de-
ficient diet. CancerRes.45,3750-3760.
35. Tatematsu, M., Kaku, T., Ekem, J. K., and Farber, E. (1984) Studies on the prolifer-
ation and fate of oval cells in the liver of rats treated with Zacetylaminofluorene and
partial hepatectomy. Amer. J. Puthol. 114,418430.
36. Hanahan, D. (1985)Heritable formation of pancreatic beta-cell tumors in transgenic
mice expressing recombinant insulin/simian virus 40 oncogens. Nature 315, 115-
122.
37. Power, R. F., Holm, R., Bishop, A. E., Varndell, I. M., Alpert, S.,Hanahan, D., and
Polak, J. M. (1987) Transgenic mouse model: A new approach for the investigation
of endocrine pancreatic B-cell growth. Gut. 28,121-129.
38. Cunha, G. R. and Donjacour, A. (1987) Strom&epithelial interactions in normal and
abnormal prostatic development. Prog. Clin. Biol. Res. 239,251-272.
39. Rojkind, M., Gatmaitan, Z., Mackensen, S.,Giambrone, M. -A., Ponce, P., and Reid,
L. M. (1980) Connective tissue biomatrix: Its isolation and utilization for long-term
cultures of normal rat hepatocytes. J. Cell Biol. 87,255-263.
40. Michapoulos, G., Cianciulli, H. D., Novotny, A. R., Kligerman, A. D., Strom, S. C.,
and Jirtle, R. L. (1982)Liver regeneration studies with rat hepatocytes in primary cul-
ture. Cancer Res. 42,4673-4682.
41. Germain, L., Noel, M., Gourdeau, H., and Marceau, N. (1988) Promotion of growth
and differentiation of ductular oval cells in primary cultures. Cancer Res. 48,
368-378.
42. Hammond, S. L., Ham, R. G., and Stampfer, M. R. (1984) Serum-free growth of
human mammary epithelial cells: rapid clonal growth in defined medium and
extended serial passagewith pituitary extract. Proc. Nutl. Acud. Sci. USA 81,5435-
5439.
43. Morrison, R. S.,Sharma, A., devellis, J., and Bradshaw, R. A. (19861Basic fibroblast
growth factor supports the survival of cerebral cortical neurons in primary culture.
Pm. Nutl. Acud. Sci. USA83,7537-7541.
44. Dodson, M. V., Allen, R. E.,and Hossner, K. L. (1985) Ovine somatomnedin, multi-
plication stimulating activity, and insulin promote skeletal muscle satelite cell pro-
liferation in vitro. Endocrinology 117,2357-2363.
45. Gatmaitan, Z., Jefferson, D., Ruiz-Opazo, N., Leinwand, L., and Reid, L. M. (1983)
Regulation of growth and differentiation of a rat hepatoma cell line by the synergis-
tic interactions of hormones and collagenous substrata. I. Cell. Biol. 97,1179-l 190.
46. Reid, L. M.,Stiles, C., Saier, M., Jr., and Rindler, M. (1979) Growthof nontumorigenic
Requirements for Differentiated Epithelia

cells in millipore diffusion chambers implanted in mice: Implications for in viva

growth regulation. Cancer Res. 20,1467-1473.
47. Cherington, P. V., Smith, B. L., and Pardee, A. B. (1980) Loss of epidermal growth
factor requirement and malignant transformation. Proc. N&l. Acud. Sci. USA 76,
3937-3941.
48. Enat,R., Jefferson, D. M., Ruiz-Opazo,N., Gatmaitan, Z., Leinwand, L. A., and Reid,
L. M. (1984) Hepatocyte proliferation in vitro: Its dependence on the use of serum-
free hormonally defined medium and substrata of extracellular matrix. Pm. Nufl.
Acud. Sci. USA 81,1411-1415.
49. Taub, M. (1984) Growth of primary and established kidney cell cultures in serum-
free media. Cell Culture Methods for Molecular and Cell Biology, vol. 3 (Barnes, D., Sir-
basku, D., and Sato, G., eds.) Liss, NY, pp. 3-24.
50. Lippman, M. E. (1984) Definition of hormones and growth factors required for op-
timal proliferation and expression of phenotypic responses in human breast cancer
cells. Ceil Culture Methods for Moleculur and Cell Biology, vol. 2 (Barnes, D., Sirbasku,
D., and Sato, G., eds.) Liss, NY, pp. 183-200.
51. Deleted in proof.
52. Hoshi, H. and McKeehan, W. L. (1984) Brain and liver cell-derived factors are re-
quired for growth of human endothelial cells in serum-free culture. Proc. N&Z. Acud.
Sci. USA 81,6413-6417.
53. Waterfield, M. D. (1983) Platelet-derived growth factor is structurally related to the
putative transforming protein ~28’~ of simian sarcoma virus. Nature 304,35-39.
54. Michalopoulos, G. and Pitot, H. C. (1975) Primary culture of parenchymal liver cells
on collagen membranes. Exp. Cell Res. 94,70-78.
55. Piez, K. A. and Reddi, A. H., eds. (1984) Extracellular Matrix Biochemistry (Elsevier,
NW.
56. Ruohslahti, E. and Pierschbacher, M. D. (1988) Molecular basis of cell-extracellular
matrix interactions. Liver Biology and Puthobiology, 2nd Ed. (Arias, I. M., Jakoby, W.
B., Popper, H., Schachter, D., and Shafritz, D. A., eds.) Raven, NY, pp. 739-751.
57. Mecham, R. I’. ed. (1986) Biology of Exfrucellulur Matrix, vol. 1, Academic, NY.
58. Evered, D., Hascall, V. C., and Whelan, J. (eds.) (1986) Functions of the Proteoglycuns
(Ciba Foundation Symposium, Wiley, NY).
59. Fransson, L. (1987) Structure and functionof cell-associated proteoglycans. TIBS 12,
406-411.
60. Hynes, R. 0. (1987) Integrins: A family of cell surface receptors. CeZI 48,549-554.
61. Ruoshlahti, L. E. and Pierschbacher, M. D. (1986) Arg-Gly-Asp: A versatile cell rec-
ognition signal. Cell 44,517,518.
62. Buck, C. A. and Horwitz, A. F. (1987) Cell surface receptors for extracellular matrix
molecules. Ann. Rev. Cell Biol. 3,179-205.
63. Hynes, R. 0 (1986) Fibronectins Sci. Amer. 254,42-51.
64. von-der-Mark, -K., and Kuhl, U. (1985) Laminin and its receptor. Biochim Biophys.
Actu. 823,147-X0.
65. Miller, E. J. (1986) Collagen types: Structure, distribution and functions, in CoElagen:
Biochemistry, vol. 1 (CRC Press), pp. 134-156.
66. Martin, G. R. and Kleinman, H. K. (1981) Extracellular matrix gives new life to cell
culture. Hepufology 1,264-266.
67. Yamada, K. M. (1983) Cell surface interactions with extracellular materials. Ann.
Rev. Biochem. 52,761-799.
68. Narita, M., Jefferson, D. M., Miller, E. J., Clayton, D. F., Rosenberg, L. C., and Reid,
L. M. (1985) Hormonal and matrix regulation of differentiation in primary liver cul-
266 Reid

tures, in Gmwth and Diffkentiution of Cells in Defined Environments (Murakami, H.,

Yamane, I., Hayashi, I., Mather, J. P., Barnes, D. B.,and Sato, G. H., eds.) Springer
Verlag, NY, pp. 89-96.
69. Spray, D. C., Fujita, M.,Saez,J.C., Choi, H., Watanabe, T., Hertzberg, E.,Rosenberg,
L. C., and Reid, L. M. (1987) Proteoglycans and glycosaminoglycans induce gap
junction synthesis in primary liver cultures. 1. Cell BioI. 105,541-551.
70. Fujita, M, Spray, D. C., Choi, H., Saez,J. C., Watanabe, T., Rosenberg, L. C., Hertz-
berg, E. L., and Reid, L. M. (1987) Glycosaminoglycans and proteoglycans induce
gap junction expression and restore transcription of tissue-specific mRNAs in pri-
mary liver cultures. Hepatology1, Is-9s.
71. Sell, S.and Ruoslahti, E. (1982) Expression of fibronectin and laminin in the rat liver
after partial hepatectomy,during carcinogenesis,and in transplantable hepatocellular
carcinomas. J. Natl. Can. Inst. 69,1005-1014.
72. Carlsson, R., Engvall, E., Freeman, A., and Ruoshlahti, E. (1981) Laminin and fibro-
nectin in cell adhesion: Enhanced adhesion of cells form regenerating liver to lam-
inin. Proc. Natl. Acad. Sci. USA 78,2403-2406.
73. Fedarko, N. S. and Conrad, H. E. (1986) A unique heparan sulfate in the nuclei of
hepatocytes; structural changes with the growth state of the cells. J. Cell Biol. 102,
587-599.
74. Chuong, C. M., Crossin, K. L., and Edelman, G. M. (1987) Sequential expression and
differential function of multiple adhesion molecules during the formation of cere-
bellar cortical layers. J. Cell BioZ. 104,331~342.
75. Kleinman, H. K., McGarvey, M. L., Hassell, J. R., and Martin, G. R. (1983) Formation
of a supramolecular complex is involved in the reconstitution of basement mem-
brane components. Biochemistry 22,4969-4974.
76. Orkin, R. W., Gehron, P.,McGoodwin, E. B.,Martin, G. R.,Valentine, T., and Swarm,
R. (1977) A murine tumor producing a matrix of basement membrane. J. Exp. Med.
145,204-220.
77. Hadley, M. A., Byers, S. W., Suarez-Quian, C. A., Kleinman, H. K., and Dym, M.
(1985) Extracellular matrix regulates Sertoli cell differentiation, testicular cord for-
mation, and germ cell development in vitro. J. CeZZBiol. 101,1511-1522.
78. Kleinman, H. K., McGarvey, M. L., Hassell, J. R., Star, V. L., Cannon, F. B., Laurie, G.
W, and Martin, G. R. (1986) Basement membrane complexes with biological activ-
ity. Biochemistry 25,312-318.
79. Martin, G. R., Kleinman, H. K., Terranova, V. P., Ledbetter, S., and Hassell, J. R.
(1984) The regulation of basement membrane formation cell-matrix interactions by
defined supramolecular complexes. Ciba Found. Symp. 108,97-212.
80. Liotta, L. A., Lee, C. W., and Morakis, D. J. (1981) New method for preparing large
surfaces of intact basement membrane for tumor invasion studies. CancerLeft. 11,
141-152.
81, Madri, J. A. and Williams, S.K. (1983) Capillary endothelial cells cultures: Pheno-
typic modulation by matix components. 1. Cell Biol. 97,153-X5.
82. Gospodarowicz, D. (1984) Preparation of extracellular matrices produced by cul-
tured bovine cornea1 endothelial cells and PF-HR-9 endodermal cells: their use in
cell culture. CeZZCultureMethods for Molecular and CeZZBioZogy (Barnes, D., Sirbasku,
D., Sato, G., eds.) Liss, NY, vol. 1, pp. 275-293.
83. Vlodavsky, I., Levi, A., Lax, I., Fuko, Z., and Schlessinger, J. (1983) Induction of cell
attachment and morphological differentiation in a pheochromocytoma cell line and
embryonal sensory cells by the extracellular matrix. Dev. BioZ. 93,285-300.
Requirements for Differentiated Epithelia 267

Compendia of Commercially Available

Hormones and Matrix Components
Sterilization Protocols: There are four protocols listed in these com-
pendia and used for either hormones or matrices. In the description of
preparation of the individual hormones, growth factors, or matrix mole-
cules, the sterilization protocol is indicated by a number correlating with
the following:
1. Sterilize by filtration through a 0.22~pm filter.
2. Sterilize by filtration through a filter with low protein binding prop-
erties (e.g., Millex-GV).
3. Sterilize by gamma irradiation, 10,000 rads.
4. Sterile as prepared in the reconstituted form.
5. For protoglycans and glycosaminoglycans, sterility is achieved by
ethanol precipitation. Prepare samples in sterile 0.15M sodium ace-
tate in a sterile tube. For best results, stock solutions should be at
concentrations of 10-20 mg/mL. However, lower concentrations will
also precipitate, but there will be some loss in yields. Add 100% etha-
nol at a ratio of 21 (v/v>, which results in precipitation of the pro-
teoglycan or glycosaminoglycan. Leave the sample at 4°C overnight.
Centrifuge sample at 5000 rpm. Wash pellet 2x with 100% ethanol to
eliminate the sodium acetate. Allow the pellet to air dry under vac-
uum (or more slowly in air) under sterile conditions. Weigh the
sample, again under sterile conditions. Dissolve the sample at the
desired concentration in sterile aqueous solutions.

Compendium 1
Commercially Available Hormones and Growth Factors
ACTH (adrenocorticotropic hormone)
Sources: Armour Pharmaceuticals, NM, Sigma, Calbiochem, Serva, ICN, Bio-
medicals, Inc.
Storage: Freezer at -20°C
Reconstitution: Dissolve in PBS
Sterilization protocol: 2
Typical concentration: 5 pU/mL
Ceruloplasmin
Sources: Sigma, Calbiochem Serva
Reconstitution: Dilute in PBS
Storage: Freezer at -20°C
(continued)
268 Reid

Compendium 1 (confinued)
Sterilization protocol: 2
Typical concentration: 0.05 U/mL
Cl apric sulfate
Sources: Sigma, AESAR (Johnson Matthey), Serva
Storage: Dry-room temperature
Solution: Freezer at -20°C
Sterilization protocol: 1
Typical concentration: 10-6-10-2M
DHT-Dihydrotestosterone
Sources: Sigma, Collaborative Research
Storage: Dry-room temperature
Solution: Refrigerator (4’0
Reconstitution: Seeprotocol for glucocorticoids
Sterilization protocol : 4
Typical concentration: 10-7-10-1*M
EGF (Epidermal Growth Factor)
Sources: Collaborative Research, Boehringer Mannheim Biochemicals, Gibco,
Calbiochem, Chemicon, BTI, IBT, Serva, Upstate Biotechnology, Inc., ICN
Reconstitution: Bring vialed product to room temperature. Add 1 mL of ster-
ile, distilled H,O and mix.
Solution: Contains 100 yg/mL EGF in 0.55M NaCl
Storage: Lyophilized EGF should be stored at 4”C, reconstituted products
store at -20°C (stable for 3 mo)
Sterilization protocol : 4
Note: Do not freeze and thaw more than 4 times; therefore, aliquot accord-
w$Y
Estradiol-17B
Sources: Sigma, Calbiochem, BTI
Storage: Dry-room temperature or refrigerator
Solution: Room temperature or refrigerator
Reconstitution: Seeprotocol for glucocorticoids
Sterilization protocol : 4
Typical concentration: 10-7-10-10M
Endothelial cell growth factor
Sources: Gibco, Chemicon, Serva, ICN, Collaborative Research, Sigma
Storage: Lyophilized powder (refrigerator 4”C), solubilized (freezer -20°C)
Reconstitution: Bring to room temperature and add l-5 mL of sterile serum-
free media. Aliquot in plastic tubes.
Solution (commercial preparations): may contain streptomycin sulfate and so-
dium chloride
Sterilization protocol : 2
Typical concentration: 50-100 Fg/mL
(continued)
Requirements for Differentiated Epithelia 269

Compendium 1 (continued)
FGF (Fibroblast Growth Factor)
Sources: Collaborative Research, Boehringer Mannheim Biochemicals, R & D
Systems, Inc., Gibco, Calbiochem, Chemicon, BTI, Serva, ICN
Reconstitution: Bring vialed product to room temperature. Add 1 mL sterile,
distilled H,O and mix
Solution (commercial preparations): Contains 10 pg/mL FGF, 100 pg/mL al-
bumin in 0.006M NaCl, 0.002M Na,HPO,
Sterilization protocol : 2
Storage: Lyophilized 4”C, reconstituted product should be stored at -20°C
Typical concentration: 0.2-2 ng/mL
Note: Do not freeze and thaw more than 4 times
FSH (Follicle Stimulating Hormone)
Sources: Sigma, Calbiochem, Serva
Storage: Dry-refrigerator
Solution: Freezer at -20°C
Reconstitution: In protein containing media, e.g., 1% bovine serum albumin
solution in PBS
Sterilization protocol : 2
Typical concentration: 0.4 -10 pg/mL
GHL (Glycyl-histidyl-lysine)
Sources: Sigma, Calbiochem, BTI, Serva, ICN
Storage: Lyophilized 4”C, Solubilized (-20°C)
Reconstitution: Dissolve in PBS or tissue culture medium
Sterilization protocol : 2
Typical concentration: 10-200 ng/mL
Glucagon
Sources: Sigma, Calbiochem, Serva
Storage: Dry-refrigerator
Solution: Freezer at -20°C
Reconstitution: Dissolve in basic H,O (approximately pH 9) at low concentra-
tion
Sterilization protocol : 2
Typical concentration: 0.5 pg/mL
Glucocorticoids (hydrocortisone, dexamethasone)
Sources: Sigma, Calbiochem
Reconstitution: Make a solution of 10w3M in 95% ethanol, place in air-tight con-
tainer (a glass scintillation vial is best), and mark the level of the liquid. No
evaporation should occur. Store at room temperature. Make a fresh dilution
to 10m4M in 95% ETOH every 3 wk. This is a 1000x solution and is the im-
mediate stock. It can be added to media in the bottle of media itself or to the
individual plate. When adding to plate, add after the media has been placed
and before the cells are added. If cells are already attached to the plate, then
270 Reid

Compendium 1 (continued)

tilt the plate to increase the depth of the media and drop the stock solution
onto the top of the deepest part of the liquid. In this way, no cells are dam-
aged by the ethanol. A fresh stock of 1WM should be made up every 3-4 mo.
Storage: Dry-room temperature; solution-refrigerator (4OC)
Sterilization protocol: 4
Typical concentration: 104-10-7M
High Density Lipoprotein (HDL)
Source: Sigma
Reconstitution: Dissolve in aqueous solutions
Sterlization: 2
Storage: Refrigerator
Typical concentration: 10 pg/mL
Insulin 500 mg/bottle
Sources: Sigma, Boehringer Mannheim Biochemicals, Gibco, Calbiochem, BTI
Serva, ICN, Collaborative Research
Reconstitution: Make a solution of 2 mg/mL in O.OlM HCl. Store in refriger-
ator (up to 6 wk). Sterilize by filtration through a swinney filter that mini-
mizes protein sticking. Just before use, dilute to 500pg/mL (100x) with PBS.
If any precipitate occurs, discard
Sterilization protocol: 2
Typical concentration: l-5 pg/mL
IGF I (Insulin-like Growth Factor or Somatomedin C)
Sources: Sigma, Boehringer Mannheim Biochemicals, Iemcera, Chemicon,
Incell, Amgen
Storage: Freezer at -20°C
Reconstitution: Make up in buffer at pH 7.6 (stable for 4-6 mo if frozen)
Sterilization protocol: 2
Typical concentration: l-50 ng/mL
IGF II
Source: Bachem
Storage: Freezer at -20°C
Reconstitution: Dissolve in a aqueous buffer at pH 8.0
Sterilization: 2
Typical concentration: 5 ng/mL
Linoleic acid sodium salt
Sources: Sigma, Serva, Collaborative Research
Storage: Freezer at -20°C
Reconstitution: Dissolve 1 g of linoleic acid in a solution containing 1% bovine
serum albumin (fattv acid free)
(continued)
Requirements for Differentiated Epithelia 271

Compendium 1 (continued)
Sterilization protocol: 2 (necessary because of the BSA)
Typical concentration: 5-10 &mL
LRF or LHRF (Luteinizing Hormone Releasing Factor)
Sources: Calbiochem, Boehringer Mannheim Biochemicals, Serva, Collabor-
ative Research, Accurate, Chemicon
Storage: Freezer at -20°C
Reconstitution: Dissolve in PBS
Sterilization protocol: 2
Typical concentration: 10 ng/mL
LH (Luteinizing Hormone)
Sources: NIH, Calbiochem
Storage: Freezer at -20°C
Reconstitution: Dissolve in PBS
Sterilization protocol: 2
Typical concentration: l-5 pg/mL
MSA (Multiplication Stimulating Activity)
(an extract containing both insulin-like growth factor I and II)
Sources: Collaborative Research, Sigma, Serva
Storage: Lyophilizedmaterial-refrigerator (4°C);reconstituted-freezer-20°C
Reconstitution: Bring to room temperature; add 1 mL of 1M acetic acid or ster-
ile basal (serum-free) medium containing 0.25-l mg of bovine serum albu-
min
Sterilization protocol: 4 (if commercial preparation); 2 (if lyophilized mater-
ial not sterile)
Typical concentration: 8-250 ng/mL
NGF (Nerve Growth Factor) 10 pg/mL
Sources: Collaborative Research, Boehringer Mannheim Biochemicals, Cal-
biochem, Chemicon, Biomedical Technology, Inc., Serva, ICN, Sigma
Storage: 4°C for lyophilized NGF and -2OOCfor the solubilized form
Reconstitution: Dissolve the vialed product with 1 mL H,O. Sterilize by filtra-
tion
Sterilization protocol: 2
Typical concentration: 100 ng/mL
Note: NGF absorbs to glass and plastic surfaces. Preparing the NGF in a BSA
solution of at least 1 mg/mL will minimize nonspecific absorption. Do not
freeze and thaw more than 4 times
PDGF ( Platelet Derived Growth Factor)
Sources: Imcera, Gibco, Boehringer Mannheim Biochemicals, R & D Systems,
Chemicon, PDGF, Inc., ICN, Collaborative Research, Sigma
(continued)
272 Reid

Compendium 1 (continued)
Storage: Lyophilized (4°C); solubilized (-20°C)
Reconstitution: Dilute in serum-free culture medium
Sterilization protocol: 2
Typical concentration: 2-8 ng/mL
P ‘ogesterone 5 g/bottle
Sources: Sigma, Calbiochem, Serva, Collaborative Research
Storage: Room temperature or refrigerator
Reconstitution: Seeprotocol for glucocorticoids
Sterilization protocol: 4
Typical concentration: 10-7-10-10M
Prolactin (Luteotropic Hormone)
Sources: Sigma, Calbiochem, Accurate
Storage: Freezer at -20°C
Solution: Refrigerate
Reconstitution: Dissolve in 100% ethanol. Use glass scintillation vial if pos-
sible (it seals very well and prevents evaporation)
Sterilization protocol: 2
Typical concentration: 10 ng/mL
PTH (Parathyroid Hormone)
Sources: Calbiochem, Collaborative Research, York Biological International
Storage: Freezer at -20°C
Reconstitution: Dissolve in PBS
Sterilization protocol: 2
Typical concentration: 0.5 ng/mL
Selenous Acid
Sources: AESAR (Johnson Matthey Chemicals Ltd., England), BTI
Storage: Dry-room temperature;
Solution: Refrigerate (4°C)
Reconstitution: Dissolve in sterile, distilled water
Sterilization protocol: 1
Typical concentration: 10-8-10-7M
Somatostatin (GH Releasing Inhibiting Factor) 0.25 mg/bottle
Sources: Sigma, Boehringer Mannheim Biochemcical, Calbiochem, Serva,
Collaborative Research
Storage: Freezer at -20°C
Reconstitution: Dissolve in PBS
Sterilization protocol: 2
Typical concentration: 10 ng/mL
Somatotropin (human growth hormones)
Sources: Sigma, Calbiochem, Chemicon, Boehringer Mannheim
(continued)
Requirements for Differentiated Epithelia 273

Compendium 1 (continued)
Storage: Dry-refrigerate (4°C)
Solution: Freezer (-20°C)
Reconstitution: Put in PBS and titrate with O.OlN HCl to dissolve
Sterilization protocol: 2
Typical concentration: ng/mL
T3 (Triiodo-Thyronine) 100 mg/bottle
Sources: Sigma, Calbiochem, Collaborative Research
Storage: Freezer at -20°C (up to 6 mo)
Reconstitution: Make up a 10-2-104M solution in 0.02N NaOH
Sterilization protocol: 2
Typical concentration: 10-g-lO-*lM
Transferrin 100 mg/bottle
Sources: Sigma, Boehringer Mannheim Biochemicals, Gibco, Calbiochem,
Chemicon, BTI, Serva, ICN, Collaborative Research
Reconstitution: Make solution of 5 mg/mL in PBS. Store this in 0.5-l mL ali-
quots in freezer. When needed, dilute to 500 pg/mL (100x). Do not thaw and
freeze more than 4 times
Sterilization protocol: 2
Typical concentration: 5 pg/mL
Zinc sulfate
Sources: AESAR (Johnson Matthey Chemicals Ltd., England), Serva, Sigma
Storage: Dry-room temperature
Solution: Refrigerate
Reconstitution: Make solution in sterile, distilled water
Sterilization protocol: 1
Typical concentration: 10-g-lO-loM
274 Reid

Compendium 2:
Commercially Available Matrix Components
Adhesion proteins
Fibronectin (human)
Sources: Collaborative, Gibco, Upstate Biotechnology, Accurate, Biomedical
Technologies
Storage: Lyophilized-refrigerator (4OC)
Solution: Freezer (-20°C)
Reconstitution: Allow to come to room temperature, add sterile, distilled
water and allow to solubilize for 30 min. Use plastic pipet to transfer to plas-
tic tubes for storage or to plastic plates
Sterilization protocol: 4
Typical concentration: 1-2 pg/cm2
Laminin (mouse)
Sources: Gibco, Iemcera, Serva, ICN, Biomedical, Sigma, Collaborative Re-
search
Storage: Freezer at -20°C
Reconstitution: Thaw the solution; make a lo-20 pg/mL dilution in sterile cul-
ture medium. Add to plates at 1-2 pg/cm2. Allow to attach to plates for 45
min at room temperatue. Then add cells
Sterilization protocol: 4
Typical concentration: 1-2 pg/cm2
Human Vitronectin
Source: Iemcera
Storage: Freezer at -20°C
Reconstitution: As for fibronectin
Sterilization protocol: 4
Typical concentration: l-2 pg/cm2
Collagens
Type I Collagen (rat tail; bovine achilles tendon)
Sources: BTI, Serva, ICN, Collagen Corp., Gibco, Seikagaku America, Inc.
Storage: Refrigerator (4°C)
Preparation for plates: Seeprotocol section 2.3.2.
Sterilization protocol: 3
Typical concentration: 1 mg/60-mm plate for gels
Type IV Collagen-Mouse, human placenta
Sources: Gibco, Collaborative, Seikagaku America, Inc.
Storage: Refrigerator (4OC)
Preparation for plates: Seeprotocol section 2.3.2.
Sterilization protocol: 3
Requirements for Differentiated Epithelia 275

Compendium 2 (continued)
Typical concentration: 10-100 pg/60-mm dish
Note: amount required can be quite variable. A thin coating can be achieved
with 10-20 pg/plate. However, for some cells, thicker coatings are neces-
sary. Increase the thickness of the layer as desired
Glycosaminoglycans
Heparins-bovine, porcine, ovine; from intestine or lung
Sources: Made by Hepar Industries (Madison, Wisconsin) and distributed by
various companies including Sigma, Serva, Accurate Chemical
Storage: Room temperature for powder
Solution: Refrigerator (4°C)
Reconstitution: PBS
Sterilization protocol: 5
Typical concentration: 20-50 pg/mL

Chondroitin Sulfates-whale, shark; from cartilage

Sources: ICN Biologicals, Serva, Sigma
Storage: Refrigeration for solution, room temperature for powder
Reconstitution: PBS
Sterilization protocol: 5
Typical concentration: 20-50 pg/mL
Dermatan Sulfates-bovine, porcine; from cartilage
Sources: ICN Biologicals, Serva, Sigma
Storage: Refrigeration for solution, room temperature for powder
Reconstitution: PBS
Sterilization protocol: 5
Typical concentration: 20-50 pg/mL
Heparan Sulfates-bovine, kidney and intestine
Sources: ICN Biologicals, Serva, Sigma
Storage: Refrigeration for solution, room temperature for powder
Reconstitution: PBS
Sterilization protocol: 5
Typical concentration: 20-50 pg/mL
Hyaluronates -human, bovine, porcine; from umbilical cords
Sources: ICN Biologicals, Serva, Sigma
Storage: Refrigeration for solution, room temperature for powder
Reconstitution: PBS
Sterilization protocol: 5
Typical concentration: 20-50 pg/mL
(cmti?iuf?d)
276 Reid

Compendium 2 (continued)
Matrigel
Sources: Serva, Collaborative Research
Storage: Freezer at -20°C; just prior to usage, thaw and maintain at 4°C.
Preparation for plates: add viscous solution (4°C) to plates. Bring to 37°C. At
this temperature gels will form.
Sterilization Protocol (seeTable 4 for detail): 4 (if bought); 3 (if made yourself).
ECM
Sources: International Biotechnology, Accurate Chemicals
Storage: Refrigerate (4°C).
Preparation for plates: sold as coated plates.
Sterilization Protocol (Table 4): 4 (if bought); 3 (if made yourself). Matrigel is
prepared from EHS tumors according to the protocol of Kleinman et al.
(75,78X
 Chapter 22

Culturing Primitive
Hemopoietic Cells

Long-Term Mouse Marrow

Cultures and the Establishment
of Factor-Dependent (FDCP-Mix)
Hemopoietic Cell Lines

Elaine Spooncer
and T. Michael Dexter
1. Introduction
The maintenance of normal primitive hemopoietic cell types in culture
for long periods can be achieved either by culture of hemopoietic cells in
association with bone marrow-derived stromal cells, as in long-term cul-
tures, or by establishing factor-dependent primitive hemopoietic cell lines
(FDCP-Mix cell lines). The long-term culture technique and a reproducible
method for establishing FDCP-Mix cell lines will be described here (1,Z).
In the stromal cell-dependent long-term cultures a wide range of
hemopoietic cell types are maintained for several months, including the
stem cells (as recognized by the CFU-S assay (3) and the ability of the cells

277
278 Spooncer and Dexter

to rescue potentially lethally irradiated mice), mature cell types of the

myeloid lineage, and all the committed progenitor cell types. Cloned
factor-dependent cell lines can be derived from long-term cultures and the
method described here produces cells that maintain a homogeneous
primitive nature under certain culture conditions and yet can be induced
to mature in response to normal regulatory influences. Clearly, long-term
in vitro growth of primitive hemopoietic cells is valuable for studying the
maintenance of homeostasis in the hemopoietic system in which both
stromal cells and hemopoietic growth factors (e.g., interleukin 3, IL3) are
thought to play important roles. These cultures are also useful for inves-
tigating the biochemistry of the response of hemopoietic cells to biological
regulators, the molecular biology of the nature of stem cells and for testing
the response of hemopoietic cells to drugs, chemicals, and other insults.

2. Materials
1. Fischer’s medium supplemented with sodium bicarbonate (1.32g/L),
penicillin (500 p/m), streptomycin (50 pg/mL) and L-glutamine (7
mL/L of 200 mM). Store at 4OCand use within 2 wk after the addition
of the L-glutamine.
2. Horse serum: each batch must be pretested for its quality to support
long-term bone marrow cultures. We have used batches from North-
umbria Biologicals Ltd. and Flow Laboratories. Store at -2OOC until
use and then keep the aliquot in use at 4OCfor up to 10 d.
3. Hydrocortisone sodium succinate. Make a stock solution in Fischer’s
medium and sterilize through an 0.22 p filter. For use dilute to lOAM
and store at -20°C.
4. Complete growth medium made up in 100 mL aliquots containing
20-25% v/v horse serum (depending on batch quality) and 1% v/v
lOAM hydrocortisone stock (final concentration WM) in Fischer’s
medium.
5. Donor mice, 8 wk or older.

2.1. Additional Materials for Establishing CZonaZ,

Factor-Dependent (FDCP-Mix) Cell Lines
3I. NIH-3T3 producer cell line for the 2-1 recombinant syc virus grown as
a Moloney murine leukemia virus pseudotype (41, [SK (MoMuLV)
producer cells]. The SYC(MoMuLV) has been categorized in this insti-
tute to be used under the good microbiological practice Code of
Practice by the local Advisory Committee on Genetic Manipulation.
Culturing Primitive Hemopoietic Cells

2. Dulbecco’s Modification of Eagle’s Medium supplemented with 10%

v/v newborn calf serum (DMEM 10% NBCS): culture medium for SIC
(MoMuLV) producer cells.
3. Radiation source to irradiate producer cell cultures at a dose to kill all
producer cells (for example, 20 Gy delivered at 84 Gy/min by linear
accelerator).
4. Source of IL3: filtered, conditioned medium (CM) from Wehi-3b-
leukemic cell line established at the Walter and Eliza Hall Institute,
cultures chemically purified IL3 from Wehi-3b CM, or recombinant
IL3.
5. Semisolid media for plating cells in the granulocyte-macrophage col-
ony-forming cell type assay, e.g., Noble Agar final concentration 0.3%
or methyl cellulose. For more details about working with semisolid
media seeChapters 26-28 in this volume.

3. Methods
3.1. Long-Term Mouse Bone Marrow Cultures
1. Assemble the tissue culture equipment and reagents in a microbiolog-
ical safety cabinet. Prepare the complete growth medium, 10 mL per
long-term culture, in 100 mL aliquots.
2. Kill donor mice by cervical dislocation. One femur (approximately 1.5
-2 x 10’ nucleated cells) is used for each long-term culture. Remove
both femora from each donor, removing as much muscle tissue as pos-
sible and cutting first below the knee joint and then at the hip joint.
Place the bones in a Petri dish containing a small volume of Fischer’s
medium on ice. If mice are in short supply, then tibiae can also be re-
moved. Use two tibiae to make one long-term culture. The time be-
tween killing donor mice and flushing the cells into growth medium
should not exceed 15-20 min. Therefore, the speed at which you can
work dictates how many donor mice to kill at once. A competent oper-
ator should be able to work in batches of five mice (i.e., 10 cultures).
3. Take the femora to the microbiological safety cabinet. Using sterile
gauze swabs, clean off any remaining muscle tissue from each bone.
Carefully cut off the ends of the femora and avoid splintering the bone
(Fig. 1). Insert the tip of a 21 gage needle and 1 mL syringe into the knee
end of the bone. For an average sized mouse (approximately 25 g) the
needle tip should fit firmly into the bone cavity. If the donor mice are
very large or small a slightly larger or smaller needle may be required
to give a good fit. Hold the syringe and needle so the bone is just below
280 Spooncer and Dexter

1 OOml. growth medium

Femur

33°C 5% CO, in air 10 ml marrow suspension

Fig. 1. Preparationof long-term bone marrow culture.

the level of the medium and flush medium through the bone 5-10
times until the bone shaft is empty of cells and looks white. Repeat
with nine more bones into the same 100 mL aliquot of medium. It is
not necessary to make a single cell suspension, but do ensure that
whole marrow plugs are broken up into fragments. Dispense 10 mL
aliquots of the marrow cell suspension into ten tissue culture flasks.
Connect a plugged sterile pipet to the 5% CO, in air supply line and
turn on the gas at a low pressure. Gas the cultures by inserting the tip
of the pipet into the air space of the flask for a few seconds, cap the flask
firmly and place the cultures horizontally in a 33°C incubator (prefer-
ably nonhumidified).
4. The cultures are routinely fed at weekly intervals by removing half the
growth medium (5 mL) and replacing it with fresh medium. Gently
agitate the flasks to suspend non-adherent cells uniformly in the
growth medium and carefully remove 5 mL, taking care not to disturb
the adherent cells with the tip of the pipet. Tip the flask sideways and
add 5 mL fresh medium down the side of the flask so it does not flow
directly onto the adherent cells. Gas the flask gently and replace the
top. The feeding routine can be altered if required and the cultures will
tolerate a complete media change weekly or more frequent feeding.
Culturing Primitive Hemopoieti s 281

Fig. 2. (a) Inverted phase-contrast micrograph of hemopoietic long-term culture: “f ar-

rows a small focus of hemopoietic cells; “F” arrows an area of fat cells. Note the hetero-
geneous arrangement of the adherent cell layer and the association of the hemopoietic
cells (small round cells) with the fat cellsand underlying stromal cells. (b) Inverted phase-
contrast micrograph of long-term culture 15 wk after infection withsrc (MoMuLV). There
is an absence of fat cells and extremely reduced numbers of hemopoietic cells. The highly
organized structure of the adherent layer is lost and the predominant cell type is a macro-
phage, many of which have abnormal morphology.

Cells harvested during feeding comprise CFU-S, progenitor cells, and

mature cells.
5. During the first 1-3 wk of culture a highly organized adherent layer
develops on the base of the flask containing stromal cells of bone mar-
row origin. The establishment of the adherent layer is essential to the
subsequent onset of hemopoietic activity. The development of the ad-
herent layer can be observed using inverted phase-contrast or bright
field microscopy. The major cell types in this cellular environment are
macrophages, fat cells, and the large, well spread “blanket cells” that
can only be easily observed in the electron microscope. Within the
adherent layer, foci of hemopoietic cells develop and grow (Fig. 2a).
These are known as “cobblestone” areas and from these foci hemo-
poietic cells are released into the growth medium. Hemopoietic cells
retained within the adherent layer are the most primitive population
in the culture. They include a high concentration of CFU-S and these
282 Spooncer and Dexter

CFU-S have higher self-renewal capacity than that of the CFU-S

released into the growth medium. The cells released into the growth
medium include the bulk of the mature granulocytes, GM-CFC and
CFU-S with lower self-renewal capacity (as measured by serial trans-
plantation). Between 34 wk after establishing a long-term culture, it
will have achieved a relatively steady-state of hemopoietic cell pro-
duction. This level of hemopoietic activity should persist for at least
another 8 wk and often longer. Signs that the hemopoietic activity is
declining are a predominance of macrophages in the non-adherent
cell population and a decline of hemopoietic cell foci in the adherent
layer. Table 1 describes some of the characteris tics of active long-term
cultures and of nonhemopoietic long-term cultures.

3.2. Establishing CZoned Factor-Dependent

(FDCP-Mix) Cell Lines
1. Establish long-term bone marrow cultures. When active hemopoiesis
is established (after 4-6 wk) infect the cultures with thesrc (MoMuLV)
as described below in steps 3-6 (5).
2. The NIH-3T3 (SYC)MoMuLV producer cells are cultured in DMEM
10% NBCS at 37OCin 5% CO, in air. Subculture twice weekly at a di-
lution of approximately 1:3 using 0.25% trypsin to detach the cells,
according to standard procedure for the culture of adherent cell lines.
3. The day before infecting long-term cultures with src (MoMuLV) feed
the long-term cultures by removing 5 mL of growth medium and re-
placing with 5 mL fresh medium.
4. Prepare NIH-3T3 SYC(MoMuLV) cultures so they will be subconfluent
on the day of infection. The day before infection remove all the growth
medium and replace with half the usual volume of fresh growth
medium.
5. On the day of infection irradiate the producer cell cultures with a dose
of irradiation (for example, 20 Gy delivered at 84 Gy/min) sufficient
to kill all the cells. Harvest the growth medium from the cultures and
centrifuge at lOOO-15OOgon a bench centrifuge for 5 min. Remove all
growth medium from the long-term cultures and add 24 mL of the
supernatant growth medium from the irradiated producer cells. This
should contain 2 x 104-2 x lo5 focus-forming units of sarcoma virus
when titered on Rat-l fibroblasts (6). Thesrc (MoMuLV) is labile so the
supernatant growth medium should be added to the long-term cul-
tures within 10 min of harvesting the medium from the producer cells.
Add 24 mL DMEM 10% NBCS to control long-term cultures.
 Table 1
Characteristics of Hemopoietic Long-Term Cultures@

Progenitor cell production

Total cell production (nonadherent cells) Differential morphology of nonadherent cells

Early Late
Blast granule- granule-Macm-
Per 105cells Per culture/wk cells cytes cytes @ages
Nonadherent
cells/culture/wk CFU-s 15-30 4!-xxmo (c200)
3-10 x 10’ (406)

Adherent GM-CFC 150300 4500-30,000 (~2000) Activeculture lo-20 l&20 55-7.5 <5
cells/culture
3-6X106 Nonhemopoietic -3 -3 2040 50-70
Cllhl??

‘This table shows some example figures for cell production in active long-term cultures. The figures shown are ap
proximate ranges to illustrate the extent of hemopoietic activity that should be attained in long-term cultures. Figures
in brackets describe the characteristics of nonhemopoietic cultures that may be observed when cultures are very old
or where culture conditions are not adequate to promote hemopoietic activity.
284 Spooncer and Dexter

6. At 4-6 hr later, make the volume of media in the long-term cultures up

to 10 mL with complete long-term culture growth medium.
7. Feed the long-term cultures weekly by removing 5 mL and replacing
5 mL fresh medium. Assess the cultures weekly for signs of viral
transformation (5), which is indicated by the following events:
a. A fall in the nonadherent cell count
b. A rise in the concentration of CFU-S and GM-CFC
c. An increase in the self-renewal capacity of CFU-S (measured
by serial transplantation in vivo) and GM-CFC (measured by
replating in semi-solid media)
d. Changes in the adherent cell morphology, including decrease
in the number of hemopoietic foci and cobblestone areas, de-
crease in fat cells, loss of the cellular organization and appear-
ance of “transformed” stromal cells (Fig. Zb).
The changes in hemopoietic activity may first be observed
5-10 wk after infecting the cultures withsrc (MoMuLV). At this
stage, factor-dependent cell lines may be derived from the cul-
tures.
8. Two methods have been used to establish FDCP-Mix cell lines from
SE (MoMuLV) infected long-term cultures.
a. Method A
(1) Harvest nonadherent cells fromsrc (MoMuLV) infected cul-
tures and plate them in semisolid medium with Fischer’s me-
dium and horse serum (10% v/v). Supplement the medium
with IL3 at a concentration that will promote optimal develop-
ment of GM-CFC colonies from normal bone marrow in semi-
solid media. The number of cells plated should be such that
when colonies have developed on the plate they are well
spaced and individual clones can easily be harvested.
(2) Incubate the plates at 37°C in 5% CO, in air (humidified) for
7-10 d. At this stage inspect the plates using, e.g., an Olympus
zoom stereomicroscope, and pick out individual colonies from
the semisolid medium with a sterile Pasteur pipet.
(3) Disperse single colonies in a 2 mL vol of Fischer’s medium,
horse serum 10% (v/v>, and IL3 (FHS/IL3) in 24-well cluster
plates. Inspect the individual clones daily using an inverted
microscope. Until the cells begin to proliferate, feed the cul-
tures every 34 d by gently removing 1.5 mL FHS/IL3 and re-
placing it with fresh medium. The cells lie at the bottom of the
well and should not be disturbed during a medium change.
Culturing Primitive Hemopoietic Cells

(4) When the cells begin to proliferate, monitor the cell concen-
tration and gradually expand the culture volume, maintaining
the cell count at between 105/mL and 5 x 105/mL. Not every
clone will survive, but in a successful experiment (seeNotes)
>80% of the colonies will establish a cell line.
(5) When th e c1ones are established and growing rapidly,
freeze some in liquid nitrogen (using standard procedures, see
Chapter 1, this vol.) and determine the cell line characteristics.
b. Method B
(1) Harvest nonadherent cells from the long-term cultures.
Wash the cells and suspend them in FHS/IL-3 at approxi-
mately 2 x IO5 cells/mL.
(2) Observe the cultures for growth of cells and when the cells
are growing well, clone cell lines by plating in semi-solid me-
dium with FHS/IL3 exactly as described above in Method A.
9. The cloned cell lines die in the absence of IL3. In liquid culture in FHS/
IL3 they have a doubling time of 15-24 h. The log phase of growth oc-
curs between about 8 x 104-8 x 105cells/mL. The plating efficiency of
the cells in semisolid medium with FHS/IL3 is between 5-250/o. The
plating efficiency tends to increase as the cell lines age. Cells may not
continue to grow if they are diluted below 5 x lo4 cells,/mL in liquid
medium and they begin to die when the cell density is above 106cells/
mL. Maintain the cell lines in active growth by subculturing to ap-
proximately 6 x lo4 cells/mL twice weekly. If the cells are subcultured
only once a week and are therefore in the plateau phase of growth for
long periods, the plating efficiency (CFC content) of the cell popula-
tion declines and the incidence of macrophages increases. Culturing
cells in fetal calf serum causes a decline in the self-renewing cells and
leads to extinction of the cell lines.
10. The FDCFMix cell lines have the following characteristics (Z), many
of which are in common with hemopoietic stem cells.
a. Survive and self-renew in response to IL3 (some cell lines may
also respond to granulocyte-macrophage colony-stimulating
factor).
b. Survive and undergo myeloid differentiation in response to
inductive hemopoietic environment in vitro (e.g., irradiated
long-term culture adherent layer) in the absence of added IL3.
c. The cells can be induced to differentiate into mature macro-
phages, neutrophils, and erythrocytes in the presence of fetal
calf serum and erythropoietin in semisolid medium (the CFC-
286 Spooncer and Dexter

Fig. 3. Typical morphology of FDCP-Mix cells grown in liquid culture with FHS/IL3.

mix assay). Occasional megakaryocytes, eosinophils, and mast

cells are also seen.
d.The cells are not infected with the src (MoMuLV). However,
they do produce the ecotropic MoMuLV (helper virus).
e. The cells are nonleukemic.
f. Early isolates of the cell lines (within3 moof establishment) can
form spleen colonies in potentially lethally irradiated mice.
The ability to grow in vivo is lost as the cell lines “age”. The
early isolates also establish long-term hemopoiesis in the pres-
ence of an inductive hemopoietic environment, i.e.; self-re-
newal of the cells as well as differentiation is maintained for
many months.
11. Figure 3 shows the morphology of FDCP-Mix cell lines grown in liq-
uid culture with FHS/lL3. The majority of the cells have a primitive
appearance and contain many basophilic granules. The occasional
mature macrophage may be observed (~3%).

4. Notes
4.1. Long-Term Bone Marrow Cultures
Establishing long-term cultures is technically straightforward, but
achieving good growth in the cultures can be very difficult for many rea-
sons (some of which are not defined!). Possible problems are listed below.
Culturing Primitive Hemopoietic Cells 287

1. The most likely problemis that the batch of horse serum is suboptimal.
The best solution to this is to obtain a sample of horse serum from a
laboratory that has established the technique and is willing to provide
about 100 mL of their serum. Compare samples against the “good’
batch by establishing cultures and measuring cell production, CFU-S,
and GM-CFC for at least 6 wk.
2. Some strains of mice do not generate hemopoietically active cultures
for a long duration. For example, C3H, CBA mice and C57B1/6 mice.
(7).
3. Growing cultures with loose caps in a humidified gassing incubator
leads to a high level of fungal contamination because of the accumula-
tion of spore-carrying moisture around the neck of the flask. These in-
cubators are best for short-term cultures. If contamination of cultures
is a problem, then the cultures can be established individually by
flushing a single femur into 10 mL of growth medium in the flask, in-
stead of by pooling 10 femurs into 100 mL of medium.
4. If the adherent layer is not well established within the first 2-3 wk of
culture, hemopoietic activity will not occur. These cultures may some-
times be rescued by giving another inoculum of l-2 x 10’ fresh bone
marrow cells at 3-4 wk.
5. Rough handling and frequent disturbance of cultures can inhibit the
development of the adherent layer or even cause detachment of ad-
herent cells. Handle the cultures gently and resist the temptation to in-
spect them on the microscope frequently during the first 3 wk.

4.2. Modifications to the Standard Culture Method

1. The technique can be scaled up if large numbers of cells are required
e.g., flush three femora into 30 mL growth medium for culture in a T75
flask (75 cm2 surface area).
2. Terminalerythroiddifferentiationinlong-termculturescanbeachieved
by supplementing the growth medium with anemic mouse serum
(AMS). The AMS needs to be tested to find an effective batch and is
usually effective at l-2% v/v (8). Higher concentrations of AMS are
toxic to long-term cultures.

4.3. Establishing Cloned, Factor-Dependent

FDCP-Mix CeZZLines
1. In our experience the transformation of long-term cultures by the src
(MoMuLV) does not always follow the pattern described above. We
do not know the reasons for this, but they may be that the long-term
288 Spooncer and Dexter

cultures are suboptimal at the stage of infection or that the producer

cell supernatant does not provide adequate infectious virus. In 12 sep-
arate long-term culture infections performed over the last 4 yr, 50% of
the experiments have led to hematological transformation and the
emergence of FDCP-Mix cell lines.
2. Within a single experiment infecting long-term cultures with SK
(MoMuLV) the responses of individual cultures may vary. It is useful,
therefore, to assess cultures individually and to select single cultures
with particularly high concentrations of CFU-S and CFC from which
to derive FDCP-Mix cell lines.
3. When establishing FDCP-Mix cell lines by plating non-adherent cells
from SYC(MoMuLV)-infected cultures in semisolid medium we have,
in some experiments, serially replated the colonies in semisolid medium
2-3 times before selecting clones for expansion in liquid culture. This
procedure may enhance the concentration of CFC with high self-re-
newal capacity (which can establish FDCP-Mix cell lines) compared
with the selection of clones from primary semisolid media cultures.

References
1. Dexter, T.M.,Spooncer,E.,Simmons, P., and Allen,T. D. (1984) Long-termmarrow
culture: An overview of technique and experience, in Long-Term Bone Marrow CuI-
tureKrocFoundation Series 18 (Wright, D. G. and Greenberger, J. S. eds.), Liss, New
York, pp. 57-96.
2. Spooncer, E., Heyworth, C. M., Dunn, A., and Dexter, T. M. (1986) Self-renewal and
differentiation of Interleukin-3-dependent multipotent stemcells are modulated by
stromal cells and serum factors. Differentiation 31,111-118.
3. Till, J. E. and McCulloch, E. A. (1962) A direct measurement of the radiation sensi-
tivity of normal mouse bone marrow cells. Rud. Res. 14,213-222.
4. Anderson, S. M. and Scolnick,E. (1983) Construction and isolationof a transforming
murine retrovirus containing the SYCgene of Rous sarcoma virus. J. Viral. 46,594-
605.
5. Boettiger, D., Anderson, S., and Dexter, T. M. (1984) Effect of src infection on long-
term marrow cultures: increased self-renewal of hemopoietic progenitor cells with-
out leukemia. Cell 36,763-773.
6. Wyke, J. A. and Quade, K. (1980) Infection of rat cells by avian sarcoma viruses: fac-
tors affecting transformation and subsequent reversion. Virology 106,217-233.
7. Reimann, J., Burger, H. (19791 In vitro proliferation of hemopoietic cells in the pres-
ence of adherent cell layers. Experimental Hematology 7,45-51.
8. Dexter, T. M., Testa, N. G., Allen, T. D., Rutherford, T., and Scolnick, E. (19811 Mo-
lecular and cell biologic aspects of erythropoiesis in long-term bone marrow cul-
tures. Blood S&3,699-707.
 Chapter 23

High Proliferative Potential

Colony Forming Cells

T, Ray Bradley, George S. Hodgson,

and Ivan Bertoncello
1. Introduction
High proliferative potential colony forming cells (HIV-CFC) in mouse
bone marrow (BM) were defined functionally by their ability to form large
colonies, greater than 0.5 mm diameter, and containing an average of 5 x
lo4 cells, in low-cell-density nutrient agar cultures after 10-14 d of incu-
bation (I).
HPP-CFC were initially detected in bone marrow cells taken from
mice after 5-fluorouracil (FU) treatment by virtue of the fact that large
colonies developed in culture in response to a combined growth factor
stimulus of pregnant mouse uterus extract (PMUE) (2) or CSF-1 from L cell
conditioned medium (3) (macrophage colony stimulating factors, M-CSF)
plus a source of synergistic factor(s) (SF) (1,4,5), but not with either factor
alone. Colonies that developed with M-CSF alone were low in numbers,
less than 0.5 mm diameter, and contained from 50 to 5 x lo3 cells with the
majority of colonies at the smaller end of the range (2). These are desig-

289
290 Bradley, Hodgson, and Bertoncello

nated as derived from low proliferative potential colony forming cells

(LPP-CFC). At 2 d after FU treatment, the LPP-CFC are drastically de-
pleted to less than 0.1% of the numbers present in normal bone marrow
whereas the HPP-CFC are 4% of normal. HPFCFC have been shown to
regenerate more rapidly than LPP-CFC, and their regeneration pattern
correlates closely with the marrow repopulating ability (MRA) and the
platelet repopulating ability (IRA) of post-FU BM, but not with CFU-S
developing at 10 or 13 d posttransplantation, and became supranormal in
numbers at 8-10 d post-FU before returning to normal at 14-16 d (4).
The synergistic factor(s) sources used initially were human, rat, or
mouse spleen conditioned media or human placenta (4,5). Conditioned
medium from human bladder carcinoma cells, 5637, was also shown to be
a source of SF in that together with CSF-1 it generated cells from post-FU
BM that could be assayed by determination of the increased binding of Y-
CSF-1 and the factor was designated hemopoietin-l(6). With M-CSF, 5637
CM also generates large colonies inagar cultures (7). Evidence has recently
been produced suggesting that hemopoietin-1 is identical with interleu-
kin-l (8).
Reexamination of the production of large colonies with purified
growth factors has shown that a major HP-CFC population of cells in FQ
BM is optimally stimulated using a combination of the three purified
growth factors, CSF-1 plus IL-l plus IL-3 (9). Cells taken later in post-FU
regeneration, FU,,BM also respond to other combinations of growth fac-
tors, e.g., CSF-1 + ILa, CSF-1 + IL-1 to produce large colonies, but optimal
large colony development still requires the three combined factors (9).
HPP-CFC are detectable in normal bone marrow with the growth fac-
tor combination of CSF-1 plus IL-1 plus IL-3 at a low incidence: approxi-
mately 1 in 400 cells. Other combinations of growth factors can result in
large colony formation, for instance, GM-CSF can substitute for M-CSF,
and although qualitatively no differences are detectable, quantitatively the
colony formation in post-FU bone marrow is less (9).
HPP-CFC are among the most primitive hemopoietic cells in mouse
bone marrow yet shown to proliferate and differentiate in vitro. The exact
relationship of HPP-CFC to CFU-S,, is not rigidly defined at present. Us-
ing multiparameter cell separative strategies, it is possible to obtain frac-
tions containing up to 30% HPP-CFC and at least 50% CFU-S,, (7). Cofrac-
tionation using these techniques indicates a close relationship, but does not
necessarily indicate that they are identical.
Mouse Bone Marrow HPP-CFC 291

2. Materials
1. Agar gels: Stocks of 1 and 0.66% Difco Bacto-agar (seeNote 1) are
made for use over a month. Weigh out the solid agar into a sterile con-
ical screwtopped flask containing a sterile magnetic stirrer bar. Add
the requisite amount of sterile distilled water (DW). Stir and heat until
the agar is melted and just commences to boil. Cool at room tempera-
ture. Repeat melting andboiling. Store at room termperature. It is im-
portant to make sure the caps are loose during melting and boiling to
prevent shattering of the flasks.
2. Culture medium: Double strength medium is made from powdered
media. We use alpha-modification of Eagle’s MEM (seeAppendix),
but any basic medium that can be made to double strength may be
used. Alpha medium has consistently given better results than most
other media. For simplicity of daily use involving large amounts of
medium, we prepare a more concentrated basic stock, with vitamin
supplementation, for freezing and make up to the double-strength
medium each day as follows:
a. Concentrated Stock
Use alpha powder without nucleosides for 10 L of medium (see
Note 2). Stir in approximately 1500 mL sterile DW 4-6 h. Add
100 mL Eagle’s MEM Vitamins (x100) and 10 mL phenol red
(1% aqueous). Determine osmolality and dilute to 1300 mos-
mol (seeNote 2), and gas with CO,. Sterile filter (0.22 pm filters)
and store frozen at -2OOC.
b. Double Strength Medium for Plating
Concentrated stock 32 mL
Sterile glutamine (200 mM) 2mL
Sterile serum (seeNote 3) 40 mL
Sterile sodium bicarbonate (5.6%) 8mL
Gentamycin (seeNote 4) 8000 u
Sterile DW 18 mL
This is 100 mL of double strength medium ready for use (see
Note 5).
3. Balanced Salt Solution: Hepes buffer (1.5M): 35.75 g Hepes (acid salt),
0.2 mL phenol red (l%, w/v>. Dissolve in approximately 60 mL dis-
tilled water and titrate to pH 7.2 with NaOH. Make up to 100 mL and
filter through a 0.22 pm filter and store at 28°C.
292 Bradley, Hodgson, and Bertoncello

4. Concentrated Stock Balanced Salt Solution: This is prepared as a x10

concentrated stock solution that is filtered through a 0.22 pm filter and
stored at 2-BOC. 80 g/L NaC1,4 g/L KCl, 10 g/L glucose, 20 mL/L of
a 1% (w/v) solution of phenol red, 2 x 105U/L gentamycin.
5. Isotonic Balanced Salt Solution (BSS): 10 mL Hepes buffer (1.5M and
5 mLNaHC0, (5.6%, w/v) is added to 100 mL BSS (x 10) and distilled
water is added to bring the volume to 1 L. The solution is filtered
through a 0.22 pm filter and stored in aliquots of 100 mL at 2-8”C.
6. Growth Factors: These are prepared at suitable concentrations (see
text) in single strength medium.
7. Incubation Boxes: Since the dishes should be incubated at 7% oxygen
gas phase, we routinely use 5-L capacity polystyrene plastic boxes (see
Note 6: Stewart Plastics, Purley Way, Croydon CR9 4HS,UK, Cat. No.
212 into which perforated stainless steel trays are placed). The tray is
2 cm above the bottom of the box. Two diagonally opposite holes (18
gage needle size) are drilled in the lids of the boxes.
8. Gas Mixture: Plastic boxes are gassed with a gas mixture of 5% oxy-
gen, 10% carbon dioxide, and 85% nitrogen, which is ordered from
regular suppliers (seeNote 12).

3. Methods
Plating of bone marrow cells in nutrient agar can be made using either
a double or a single layer system. We routinely use a double-layer system
since colony formation is better than in the single layer system.

3.1. Removal of Marrow

1. Groups of at least five mice are used to provide femurs. The femoral
shafts are stripped of all tissue and are flushed from one end and then
the other with chilled BSS containing 2% newborn calf serum using a
1 mL syringe fitted with a 23-gage needle.
2. Marrow cells are collected at a concentration of one femur equivalent
(2.5 x lo7 cells)/mL BSS and kept on ice until used.

3.2. Plating Procedure (Double Layer)

1. Sterile growth factors to give the desired final concentrations are
added directly to the non-tissue-culture-treated 35-mM Petri dishes
(Note 7). Routinely the triple growth factor combination is used at the
Mouse Bone Marrow HPP-CFC

following doses: CSF-1 1 x 103 U, IL-la 1 x lo4 U (conversion assay

units), IL-3 25 U per dish, but all growth factor preparations should be
tested for the concentrations necessary to achieve optimal colony
formation.
2. Melt the 1% agar in a conical flask to just boiling and put in a 37OCbath
to cool.
3. Calculate the amount of double strength medium to be used for the
number of dishes to be plated and allow extra to ensure that the me-
dium-agar mixture will be adequate in volume. Place in a conical
screw topped flask and warm in the 37OC bath.
4. When the agar is cooled to 37OC (test it by agitating it in the flask and
feeling the temperature), add an equal volume of it to the double
strength medium mix thoroughly (either by pipet or with a sterile
magnetic stirrer bar previously placed in the flask) and dispense 1 mL
to each of the dishes, shaking the dishes from side to side to ensure
complete coverage of the dish and mixing with growth factors (Note
8). This basal layer should gel within a few minutes at room temper-
ature (2OOC).
5. For the overlay (containing the bone marrow cells), melt and just boil
the 0.66% agar and place in the bath to cool (Note 9). Add the required
number of bone marrow cells to prewarmed medium (double strength)
and immediately add an equal volume of agar. After mixing thor-
oughly, dispense 0.5 mL aliquots to the dishes; no shaking is required.
After a few min the upper layer should be gelled and the dishes can be
incubated (seeNote 10). For normal bone marrow we routinely use 2.5
x lo3 cells per dish: for FUzd, 2 x 104;for FQ, 1 x 103. For fractionated
enriched bone marrow populations, the numbers are reduced appro-
priately (Note 11).
6. 50 mL of sterile DW is placed under the stainless steel tray of the plas-
tic box to provide humidity during the incubation. Stack the dishes in
the box. Each box can take up to 150 dishes.
7. The box lids are then taped on with three layers of tape which is left
taut and wound around the lid-bottom box junction. No wrinkles
should be evident and the taping should be finished by smoothing it
down over the junction.
8. The boxes are then gassed at the rate of 2.5 L/min for 30 min to give
a final oxygen concentration of 7%, the gassing holes are sealed with
three layers of tape, and the box is placed in the 37OCroom or incuba-
tor (Note 12).
294 Bradley, Hodgson, and Bertoncello

9. At 14 d of incubation, the colony numbers are counted (Notes 13-15)

using a dissecting microscope at 20x magnification with a calibrated
grid in one eyepiece to measure colony diameters. Normally the col-
onies originating from HPP-CFC are clearly visible without magnifi-
cation but are checked to observe that they are tightly packed with
cells and to count any smaller colonies (Note 14).

4. Notes
1. Various agarose preparations may be used instead of agar, provided
that it is one gelling adequately at room temperature, but colony for-
mation with myeloid cells is lower than in agar. Likewise autoclaving
of agar reduces colony formation and should be avoided (10).
2. Various other media have been used. It is important to measure the
osmolality of media and sera to be used in order to ensure uniformity
of culture conditions. On testing a range of media, we have found a
final medium osmolatity of 280-300 mosmol to be optimal for mouse
bone marrow colony formation.
3. Batches of sera must be tested to choose a pool suitable for the next 6
mo to 1 yr work. We find some batches of newborn calf serum to be
better than fetal calf serum.
4. Gentamycin has been used routinely for several years to replace peni-
cillin and streptomycin because of its spectrum and stability.
5. The bicarbonate concentration used gives strong buffering with the
10% carbon dioxide in the gas phase.
6. The plastic boxes are washed and dried before using and between in-
cubations and have several advantages over normal incubators:
a. Each experiment can remain undisturbed over the incubation
period.
b. Contamination during incubation is not a problem.
c. They are cheap, and, if a 37°C room, or nonhumidified incuba-
tor, is available the number of workers who can carry out clono-
genie experiments is greatly increased for little change in cost.
d. They can be used to test numerous gas concentrations for vari-
ous cell and culture types.
7. Growth factor preparations should not be added to the underlay at
more than 0.3 mL/dish and preferably left to 0.15 mL maximum mak-
ing 10% of the total double layer system in order to avoid the gels los-
ing their firmness.
Mouse Bone Marrow HPP-CFC

8. For plating large numbers of dishes, e.g., above 100 dishes, Cornwall
continuous pipets are used for both underlays and overlays with al-
uminum foil covered or cotton plugged flasks replacing the screw
topped flasks.
9, For the single layer system the 0.66% agar is melted and placed in the
37OC bath. The volume of double strength medium is placed in an-
other flask and warmed, the requisite numbers of cells are added, and
the equal volume of agar added immediately mixed and plated.
10. The most common failure of cultures arises from using agar that is too
hot or over-heating pipets during routine flaming of them. On the
other hand, care must be exercised to ensure that the agar-media mix-
ture does not gel before dispensing into the dishes.
Il. It has been observed that colony formation with fractionated en-
riched populations of bone marrow is usually better than with un-
fractionated marrow (7).
12. The gassing of the incubation boxes is calculated to pass 75 L of gas
through the 5-L vol boxes. A brief passage of gas will not suffice to
achieve the final 7% oxygen used nor an adequate CO, concentration.
13. One of the most important aspects of optimal colony formation in vi-
tro is to ensure that low cell densities are used:
a. to prevent possible secondary effects occurring such as stimu-
lation of accessory cell production of growth factors,
b. to prevent the smaller colony formation that can occur when
too many colonies develop in the dish.
When the incidence of cells present at low frequency in the population
is being measured, the number of replicate dishes must be increased
rather than the cell density per dish. In practice, 20-50 colonies/dish
is an optimum to achieve, and especially with an enriched population,
dishes should be set up at different cell densities.
14. Colonies from different bone marrow populations may develop at dif-
ferent rates, e.g., using FU,, BM cell cell’s clones are just starting to de-
velop at 6 d of incubation and achieve their large colony formation
rapidly over the next 4-8 d. FU,, BM develop more rapidly and are
often large in diameter at 8-10 d of incubation.
15. HPP-CFC have been detected in all strains tested, e.g. BALB/C, C57-
B16,AKR, C,H/HeJ, CBA (C57Bl x BALB/C) -Fl, (C57Bl x DBA/2)-Fl,
16. Since colony formation takes place over a lengthy incubation period
(10-14 d), the I-IPP-CFC may develop by sequential action of the
growth factors initially placed in the cultures on cells generated with-
296 Bradley, Hodgson, and Bertoncello

in the colonies during colony development. Also, combinations of

other growth factors than those discussed here may detect cells with
high proliferative potential with their actions being either additive or
synergistic.
17. At the present time the size of the colony, and more particularly the
number of cells generated per single HPP-CFC, are the criteria for de-
tection of these cell types. The development of large colonies in the
primary cultures at 14 d does not exhaust their total proliferative po-
tential since these colonies can be sampled, dispersed into single cells,
and replated to give further colony formation, although the secondary
colonies are smaller and no longer require all of the three growth fac-
tors. For example, large colonies developed from FU,,BM cells with
CSF-1 plus IL-3 plus 5637 conditioned medium can be replated with
CSF-1 alone to yield ultimately 27 x lo6 cells/initial HPP-CFC. This
may demonstrate local exhaustion of the medium during the primary
colony formation and/or local inhibitory effects within the tightly
packed colonies.

References
1. Bradley, T. R. and Hodgson, G. S. (1979) Detection of primitive macrophage progen-
itor cells in bone marrow. BZood 54,1446.
2. Bradley, T. R., Stanley, E. R., and Sumner, M. A. (1971) Factors from mouse tissues
stimulating colony growth of mouse marrow cells in vitro. Amt. J, Exp. Biol. Med. Sci.
49,595.
3. Stanley, E. R. and Heard, P. M. (1977) Factors regulating macrophage production
growth. Purification and some properties of the colony stimulating factor from
medium conditioned by mouse L cells. I. Biol. Chem 252,4305.
4. Bradley, R. R., Hodgson, G. S., and Bertoncello, I. (1980) Characteristics of macro-
phage progenitor cells with high proliferative potential, their relationships to cells
with marrow repopulating ability in 5-FU treated mouse bone marrow. Experimen-
tal Hematiol. Today (1979) (Baum, S. J., Ledney, G. D., and van Bekkum, D. W., eds.),
Karger, NY, p. 285.
5. Kriegler, A. B., Bradley, T. R., Januscewicz, E., Hodgson, G. S., and Elms, E. F. (1982)
Partial purification and characterization of a growth factor for macrophage pro-
genitor cells with high proliferative potential in mouse bone marrow. Blood 60,503.
6. Bartelmez, S. H., Sacca, R., and Stanley, E. R. (1985) Lineage specific receptors used
to identify a growth factor for developmentally early hemopoietic cells: Assay for
hemopoietin-2. I. Cell. Physiol. 122,362.
7. Bertoncello, I., Bartelmez, S. H., Bradley, T. R., and Hodgson, G. S. (1987) Increased
Qa-m7 antigen expression is characteristic of primitive hemopoietic progenitors in
regenerating marrow. I. Immunol. 139,1096.
Mouse Bone Marrow HPP-CFC 297

8. Mochizuki, D. Y., Eisenman, J. R., Conlon, I’. J., Larsen, A. D., and Tushinski, R. J.
(1987) Interleukin-1 regulates hematopoietic activity, a role previously ascribed to
hemopoietin-1. Proc.Nafl. Acad. Sci.84,2567.
9. Bartelmez, S. H., Bradley, T. R., Bertoncello, I., Mochizuki, D. Y., Tushinski, R. J.,
Stanley, E. R., Hapel, A. J,, Young, I. G., Kriegler, A. B., and Hodgson, G. S. (1989)
Interleukin-1 plus interleukin-3 plus colony stimulating factor-l are essential for
clonal proliferation of primitive myeloid bone marrow cells. Exp. Hmafol. in press.
10. Dixon, R. A., Linch, D., Baines, I’., and Rosendaal, M. (1981) Autoclaved agar con-
tains an inhibitor of granulocyte-macrophage colony growth in vitro. Exp. Ceil Res.
131,478.
 Chapter 24

Murine Bone Marrow-Derived

Macrophages

E. Richard Stanley
1. Introduction
The molecular phagocytic lineage comprises, in order of increasing
maturity, the committed macrophage precursor cell, the monoblast, pro-
monocyte, monocyte and the macrophage. Methods for the preparation
and culture of bone marrow-derived macrophages, developed by Stanley
and colleagues (l-3), provide large numbers of mononuclear phagocytes
that are capable of extensive cell proliferation. Since their proliferation can
be stimulated by colony stimulating factor-l (CSF-l), granulocyte macro-
phage colony stimulating factor (GM-CSF), or interleukin-3 (IL-3), they
represent an important primary cell source for studies of the actions and
interactions of these three growth factors. The principles underlying the
method are: (1) to generate and expand primitive mononuclear phago-
cyte precursor cells by culturing bone marrow cells in a combination of
partially purified CSF-I and IL-3 for a period of 3 d, (2) to remove
contaminating red cells, fibroblasts and mature macrophages and disrupt
aggregates of proliferating cells, by proteolytic digestion of the nonad-
herent cells at d 1 and 3 of culture and, (3) to obtain a population of
mononuclear phagocytes that is relatively homogeneous with respect to
their state of differentiation by recovering only those cells (i.e., mono-
blasts, promonocytes) that acquire the capacity to adhere to tissue culture
plastic during d 4-5 of culture. About 95% of the bone marrow-derived
macrophages so obtained possess the CSF-1 receptor, 93-98% proliferatein
299
300 Stanley

response to CSF-1 and 90% of cells die on removal of CSF-1 from the ser-
um-containing medium (I, 4, 5). This latter observation reflects the
absence in these cultures of contaminating, fibroblast-like, CSF-1 pro-
ducing cells and enables this population to be used not only to study the
effects of growth factors on macrophage proliferation, but also on their
survival.
2. Materials
1. Dulbecco’s Modified Eagle’s Medium (DMEM)
2. Pronase solution: 0.02% (w/v) Pronase (B grade, Calbiochem), 1.5
mM EDTA in phosphate buffered saline (PBS)
3. Complete medium: DMEM supplemented with 15% (v/v) fetal calf
serum (FCS), 0.292 mg/mL glutamine, 0.02 mg/mL asparagine 0.5
I.&! 2-mercaptoethanol, 0.2 g/L penicillin, 0.2 g/L streptomycin
4. Stage 1 CSF-1: Stage 1 L cell CSF-1 prepared as in ref. (6).
5. 11-3: Purified recombinant murine IL-3 or medium conditioned by
serum-free cultures of the myelomonocytic leukemia cell line WEHI-
3 (7)
6. Zwittergent stock solution: 1% (w/v) Zwittergent 3-14 (Calbiochem).
Stored at 4°C.
7. 0.005% Zwittergent: Zwittergent stock solution diluted 1 in 200 with
PBS. Stored at 4°C.
3. Method
The method is divided into (1) isolation of bone marrow cells, (2)
generation of mononuclear phagocyte precursor cells, (3) differentiation of
precursor cells to adherent mononuclear phagocytes, (4) culture of bone
marrow-derived macrophages (2,2) and (5) determination of macrophage
concentration (3). The last section is included since adherent cultured
macrophages cannot be easily detached by trypsinization and alternative
methods for quantitating cell number are required.
3.1. Isolation of Bone Marrow Cells (See also
Chapters 22 and 26)
1. Remove the tibias and femurs from C3H/HeJ mice by cutting the
proximal end of the femur and the distal end of the tibia, leaving the
other ends intact.
2. Insert a 23-gageneedle into theintact ends and flush bone marrow out
through the cut ends with ice-cold DMEM.
3. Dispense marrow plug by three passes through a 22-gage needle and
centrifuge (12OOg,5 min, 4°C).
Bone Marrow-Derived Macrophages 301

4. Resuspend the pellet in ice-cold DMEM and determine the nucleated

white cell concentration by counting an aliquot of the resuspended
cells in 2% acetic acid.

3.2, Generation of Mononuclear Phagocyte

Precursor Cells
1. Adjust the cell density to lo6 cells/mL in complete medium contain-
ing 500 U/mL (0.22 nM CSF-1) of stage 1 CSF-1, and 1 nMIL-3 (or 10%
WEHI- conditioned medium).
2. Seed cells into tissue culture flasks at a density of 2.9 x 105 cells/cm*
and incubate (37”C, 10% CO, in air) for 24 h.
3. Collect nonadherent cells and discard adherent cells (fibroblast-like
cells, mature mononuclear phagocytes, and other cells adhering to the
flask).
4. Centrifuge the nonadherent cells (8OOg,10 min, 4OC), resuspend in 1
mL of Pronase solution/107 cells and incubate for 15 min at 37OC.
5. Stop Pronase digestion by the addition of horse serum (0.2 mL/ 10 mL
Pronase solution) and layer the suspension onto 15 mL of ice-cold
horse serum.
6. Incubate on ice for 15 min to allow clumped erythrocytes and debris
to settle through the horse serum.
7. Remove the cell suspension from the horse serum by Pasteur pipet,
overlay onto another 15 mL of ice-cold horse serum, centrifuge (1200
g, 10 min, 4”C), disperse the pellet in DMEM as described above and
seed into the original number of flasks.
8. Incubate the flasks at 37OC for 2 d, collect the nonadherent cells, and
discard the adherent cells. Pronase treat the nonadherent cells as
described above (steps 4-7) and resuspend the pellet in complete
medium containing 0.22 nM CSF-1.
9. At this stage, if desired, adjust the cell suspension to 10% (v/v) with
respect to dimethylsulfoxide and store at -196OC for later use.

3.3. Differentiation of Precursor Cells to Adherent

Mononuclear Phagocytes
1. Adjust concentration of the pronase-treated nonadherent cells to 105
cells/mL.
2. Seed into tissue culture dishes at a density of 1.9 x 104 cells/cm* in
complete medium containing 0.22 nM CSF-1 and incubate for 2 d at
37OC in 10% CO, in air.
3. Remove the nonadherent cells with two washes of sterile PBS.
 Stanley

3.4. Culture of Bone Marrow-Derived

Macrophages
1. Maintain log-phase growth by culturing in 1.3 nM CSF-1 in complete
medium. At high cell densities it may be necessary to periodically
change the medium [see ref. (I) for formula predicting the CSF-I
consumption by bone marrow-derived macrophages].
2. Log-phase cells may be rendered quiescent by washing once withPBS,
replacing the growth medium with complete medium (without CSF-
1) and incubating for 16 h.
3. Remove adherent macrophages for subculture by scraping with a
sterile cell scraper.

3.5. Determination of Macrophage Concentration

1. Wash the cell monolayer with ice-cold PBS.
2. Pipet 0.005% Zwittergent onto the cells.
3. Incubate for 5 min at 20°C.
4. Examine with a microscope to ensure cell detachment and remove by
pipet using several up and down pipetings.
5. Dilute appropriately and count by electronic cell counter or hemo-
cytometer.

References
I. Tushinski, R. J., Oliver, I. T., Guilbert, L. J., Tynan, P. W., Warner, J. R, and Stanley,
E. R. (1982) Survival of mononuclear phagocytes depends on a lineage-specific
growth factor that the differentiated cells selectively destroy. Cell 28,71-U.
2. Guilbert, L. J. and Stanley, E. R. (1986) The Interaction of 1251-Colony-stimulating
Factor-l with Bone Marrow-derived Macrophages. J. BioZ. Chem. 261,4024-4032.
3. Tushinski, R. J. Zwittergent cell counting method, personal communication.
4. Tushinski, R. J. and Stanley, E. R. (1983) The regulation of macrophage protein
turnover by a colony stimulating factor (CSF-1). J. Cell. Physiol. 116,67-75.
5. Tushinski, R. J. and Stanley, E. R. (1985) The regulation of mononuclear phagocyte
entry into S phase by the colony stimulating factor, CSF-1. J. Cell. Physiol. 122,
221-228.
6. Stanley, Il. R. (1985) The macrophage colony stimulating factor, CSF-1, inMethods in
EnzymoZogy-lmmunocheicaI Techniques, vol. 116 (Colowick, S. I?. and Kaplan, N. 0.
eds.), Harcourt Brace Jovanovich, New York, pp 564-587.
7. Guilbert, L. J., Nelson, D. J., Hamilton, J. A., and Williams, N. (1983) The nature of
12-O-Tetradecanoylphorbol-13-Acetate(TPA)-stimulated hemopoiesis,colonystim-
ulating factor (CSF) requirement for colony formation, and the effect of TFA on [12511
CSF-1 binding to macrophages. J. CeZZ.Physiol. 115,276-282.
 Chapter 25

Long-Term B Lymphoid
Cultures from Murine
Bone Marrow
Establishment and Cloning
by Using Stromal Cell Line AC 6.21

Cheryl A. WhitZock
and Christa E. MuZZer-Sieburg
1. Introduction
Nearly all hematopoietic cells in mammals derive from precursors
that undergo much or all of their development in the bone marrow. In vitro
models for many lineages are available and represent modifications of the
original bone marrow culture system designed by Dexter and Lajtha (I)
and described elsewhere in this book. In this chapter, we describe a second
bone marrow culture system, first reported in 1982 (2), that provides an in
vitro environment selectively supporting long-term proliferation and dif-
ferentiation of early B lymphocyte lineage cells. This method can be used
to obtain heterogeneous populations of immature precursors of the B cell
lineage greatly enriched from other hematopoietic cell types. Clonal
populations can also be obtained by extension of this method to limiting
dilution culture.

303
304 Whitlock and Muller-Sieburg

A recent major advance in the technology of B lineage cell culture has

been the isolation of stromal cell lines that support B lymphopoiesis from
the primitive precursors through the pre-B cell stage. These lines can sub-
stitute for the mixed stromal cell layer that provides the supportive micro-
environment in the primary bone marrow cultures. We describe one such
cell line, AC 6.21, and its use for expansion of established heterogeneous
populations of B lineage cells and for limiting dilution culture to obtain
clonal lines or determine the B cell precursor frequency of selected
populations.
The technique of long term bone marrow culture for the study of B
lymphopoiesis is still in its infancy, but much progress has been made since
its inception in 1982. We now know something about the characteristics of
the stem cell precursors that establish the continuous B lineage cell lines
and will hopefully someday learn more about how these precursors are
committed to the B lineage by the culture environment. Many investiga-
tors are also isolating stromal cell lines that are able to support B lympho-
poiesis and, through their efforts, the factors and cellular interactions that
are important in supporting B cell ontogeny will be defined in the future.

2. Materials
1. Culture medium: RPM1 1640 medium,5% (v/v> selected batchof fetal
calf serum (Note 1),5 x 10sM 2-mercaptoethanol, 2 mM glutamine, 1
mM sodium pyruvate, 100 U/mL penicillin, 50 pg/mL streptomycin.
Antibioticsarerecommendedforinitiatingcultures,butmaybeomitted
after 2 wk. Antifungal agents are toxic to the stromal layer and should
not be used. B-mercaptoethanol is essential for viability and growth
of the stromal cell line AC 6.21.
2. Mice: Balb/c mice provide the most consistent results, but continuous
lymphoid cultures have been established using other mouse strains.
Mice must be between 2-l /2 and 3-l /2 wkof age at the time of marrow
harvesting. BALB congenic strains such as BALB.B, BALB.K, and
BAB-14, are useful for preparing feeder layers (Note 3).
3. Enzyme: Dispase-Collagenase is purchased in 100 mg vials from Boeh-
ringer-Mannheim. One vial is dissolved in 20 mL serum-free medium
to prepare a 10x stock, then filtered through a 0.45 pm filter before ali-
quoting and storing at -35 to -70°C. Thawed and unused portions re-
tain significant activity when refrozen or stored for a few days at 4OC.
Long-Term B Lymphoid Cultures 305

Table 1
Antibodies and Their Specificities
Cells in
Antibody Specificity bone marrow % Reference

RA3-6B2 B220 (pre-B and B cells) 25-30

Ml /70 Mac-l (macrophages) 23-28
RA3-8C5 Granulocytes (=Gran-1) 28-33
30H12 Thy-l .2 3-5
53-2.1 Thy-l .2 3-5
19XE5 Thy-l .l 3-5
31-11 Thy-l (nonallelic) 3-5
GK1.5 L3T4, on T, cells l-3
5.3-6.72 Lyt-2, on T, cells 2-3

4. Reagents and equipment for immunofluorescent staining and cell

sorting: Iscove’s medium provides better viability of cells during these
procedures. Propidium iodine is used in the final step of the staining
procedure in order to detect and eliminate dead cells from the sorting.
Stocks are prepared in PBS at 1 mg/mL and stored at -2OOC. Once
stained, the cell suspension is filtered through a nylon screen with
mesh3-60/45 (TetcoInc.,Elmsford,NY) inorder toremovecellclumps.
Sorting requires a fluorescence-activated cell sorter (FACS) equipped
with two lasers, one tuned to 488 nm to excite FITC, and the other to
590 nm to excite Texas red.
5. Antibodies: Antibodies used for enrichment of B lymphocyte progen-
itors are listed in Table 1 with their corresponding references. We have
found that good quality antibody preparations are critical for success-
ful purification of Thy-lloT-BG-M-cells (seeSection 3.2.4 and Note 4).
We use antibodies derived from serum-free hybridoma culture super-
natents that have been -loo-fold concentrated by ammonium sulfate
precipitation (50% w/v). These antibodies can be labeled with biotin,
FITC, or Texas red as described (10) and are always titrated prior to the
FACS sorts to determine the optimal concentration for staining. All
antibodies are diluted in Iscove’s medium with 3 to 5% serum and
sterile filtered through a 0.2 Frn disk filter directly before use. (Pre-
washing the filter with serum-containing medium will reduce loss of
antibody by nonspecific binding to the filter.)
306 Whitlock and Muller-Sieburg

3. Methods
3.1. Lymphoid Long-Term Bone Marrow Cultures
(Whitlock-Witte Cultures)
The following technique is remarkable for its simplicity and repro-
ducibility. The most critical ingredient for success is the batch of fetal calf
serum used. Selection of a suitable serum batch is described below (Note
1). A recommendation for those initiating cultures for the first time or
screening serum batches is to use Balb/c mice because marrow from these
mice has proven to be best for rapidly and reproducibly establishing B line-
age cultures.
3.1.1. Harvesting Marrow and Initiating Cultures
1. Sterilize the surgical area widely with ethanol.
2. Open the skin overlying the femur anteriorly with sterile scissors from
abdomen to well below the knee.
3. Trim the thigh muscles away, then remove the bone by cutting just be-
low the hip and knee joints.
4. Transfer the bones to a Petri dish containing medium or balanced salt
solution with 5% serum and keep on ice until all the bones are re-
moved. The knee joint is preserved until just before marrow harvest-
ing because of the tendency of the marrow plug to be partially ex-
truded when it is cut.
5. Scrape the interior of the marrow cavity with a 25-gage needle while
flushing out the marrow plug with 3-5 mL of medium or balanced salt
solution containing 5% serum. All clumps of cells and bony spicules
should be saved and cultured. (A goal of marrow harvesting is to ob-
tain as many stromal cells as possible.)
6. Pipet the cells and medium up and down to break up as many clumps
as possible, then centrifuge the entire cell suspension at 1200 rpm for
10 min.
7. Suspend the cell pellet in culture medium at lo6 cells/ml.
8. Dispense the cell suspension at approximately 2.5 x 10s cells/cm2 of
the culture vessel (an exception is in vessels less than 10 cm2, such as
in 24- and 96-well plates). The volumes used for most conventional
culture vessels are shown in Table 2.
9. Incubate cultures at 37OC,in 7-8% CO,, and in a well-humidified incu-
bator. Cover multiwell plates in plastic wrap to decrease air circula-
tion around them. Tissue culture plates are best handled in stacks on
trays, and individual trays can be wrapped with foil. Wrapping cul-
Long-Term B Lymphoid Cultures

Table 2
Culture Volumes for Standard Tissue Culture Vessels
Surface area, Volumes for initiation:
Vessel type cm2 mL

24-Well platesb 2 1.0

6-Well plates 10 2.5
60 mm dishes 20 5.0
100 mm dishes 57 14.0
T25 flasks 25 7.0
T75 flasks 75 20.0
“The volume given is cell suspension needed to produce a similar density of cells per
square centimeter in each vessel type. For initiation of high density cultures that will pro-
duce B lineage cells,a cell suspension of l@cells/mL is used. For preparation of low-cell-
density mixed stroma, 3 x 105cells/ml is used.
bFor24-well plates (and smaller vessels,if attempted) a higher density of cells per sur-
face area must be used in order to seed enough cells to prevent well-to-well variability in
the stromal layers.

tures serves to decrease evaporation and concentration of medium, as

well as aid in containing and preventing spread of airborne contamin-
ants, such as mold spores.
3.1.2. Maintenance of Cultures
Once initiated, long-term lymphoid bone marrow cultures need only
to be fed with fresh medium until the time nonadherent lymphocytes are
harvested for analysis. The timetable for feeding long-term cultures is pro-
vided in Table 3. For 60- and loo-mm tissue culture plates, dehydration is
a problem, and these cultures may require feeding twice a week. Flasks
and multiwell plates have less of a dehydration problem and can be fed
only once a week.
1. Initiate cultures as described above.
2. For biweekly feedings, add an amount of fresh medium equivalent to
l/2 of the volume of culture medium used to initiate the culture with-
out removal of any of the spent medium. (For once weekly feedings,
this step is omitted.)
3. At the subsequent feeding, aspirate approximately 80% of the spent
medium (with care not to remove the nonadherent cells), then add a
volume of fresh medium equivalent to the amount used to initiate the
culture. (For once weekly feedings, cultures are aspirated at each feed-
ing and fed with a volume of fresh medium equivalent to the amount
used to initiate the culture.)
308 Whitlock and Muller-Sieburg

Table 3
Schedule of Feeding for Culture Maintenance
Day
0 Initiation of cultures
[3 or 41” [optional] Feed with l/2 volumeb
Removal of 80% spent medium; feed with full volume
LO or 111 Repeat of d 3 or 4
14 Repeat of d 7
‘Tissue culture dishes have a greater tendency for the medium to evaporate; therefore,
they should be fed twice weekly to prevent concentration of the medium and poor culture
growth.
bDetailsof feeding stepsare given in “Maintenanceof Cultures” section.

3.1.3. Collection of Cultured Lymphocytes for Seeding

Secondary Cultures or for Experiments
The whole process of establishment of B lymphopoiesis (Note 2) can
take 3-6 wk, depending on the serum batch. Nonadherent cell numbers
steadily increase, after a brief decline, and reach a “maximum” that is char-
acteristic for each individual culture and ranges roughly between 105and
106/mL. When patches of small- to medium-sized lymphocytes predomi-
nate in the cultures (Fig. l), they can be harvested for assay without sacri-
ficing the primary culture.
1. Gently pipet the surface of the stromal layer to dislodge the lymphoid
cells that are loosely adherent to a subpopulation of stromal cells.
2. Once dislodged, remove 90% of the medium and suspended nonad-
herent cells.
3. Add fresh medium to the primary culture vessel, and the remaining
lymphoid cells will continue to divide until they reach their charac-
teristic “maximum” density in 4-7 d.
4. Transfer the harvested lymphoid cells to a new tissue culture vessel
and incubate at 37OC for 2 h to allow any dislodged stromal cells to
adhere. (It is best to leave the harvested lymphocytes in their con-
di tioned medium during this incubation. The adherence step is not re-
quired for routine expansion of lymphoid cultures.)
5. Harvest the lymphocytes a second time by gently pipeting the medi-
um or swirling the culture vessel.
6. Repeat the adherence steps if a large number of stromal cells appear
to be present.
7. Lymphocytes harvested from the primary cultures can now be used
for experiments or subdivided onto established mixed stromal layers
Long-Term B Lymphoid Cultures 309

Fig. 1. Established long-term lymphoid bone marrow culture initiated at lti cells/mL
(high cell density). Patchesof small- to medium-sized lymphocytes predominate and are
closely associated with underlying stromal cells. Scattered nonadherent cells not associ-
ated with stromal cells have poor viability. Often lymphocytes are seen beneath the
stromal cells (center/bottom), a process termed pseudoimperipolesis.
310 Whitlock and Mulbr-Sieburg

or stromal cell lines (seebelaw) in order to expand the number of cells

that can be obtained. The number of lymphocytes that is optimal to
seed onto low-cell-density, mixed stroma has never been determined,
but we know that primary cultures can be subdivided at least lo- to 20-
fold.
3.1.4. Preparation of Mixed Stromal Cell Feeder Layers
Mixed stromal cell layers devoid of lymphoid cells can be prepared by
initiating bone marrow cultures at a lower cell density than used to initiate
cultures that produce B lineage cells. Low-cell-density, mixed stroma pro-
liferate to cover 50-75% of the culture vessel surface in about 3-4 wk. At
this point, nonadherent lymphocytes harvested from high density cul-
tures, as described above, can be seeded onto them.
1. Harvest bone marrow from young mice of the appropriate mouse
strain or congenic type that will allow differentiation of the seeded
lymphocytes from those that might derive from the feeder layer (see
Note 3).
2. Suspend cells in culture medium at 3 x lo5 cells/ml.
3. Initiate cultures with the same volumes as indicated for high density
cultures (Table 2).
4. Maintain cultures, as indicated, for high density cultures in Table 3
and ‘Maintenance of Cultures” section.
5. After 3-4 wk of culture, check cultures by phase microscopy for degree
of confluence (should be at least 50% confluent), types of stromal cells
present (should be a variety of stromal cell types), and the presence of
foci of nonadherent lymphoid cells (Fig. 1; discard cultures with lym-
phoid colonies since these will contaminate the seeded cells).
6. Continue to regularly feed the feeder layers until they are ready to be
used, and they should be functional for 2-3 mo.
3.1.5. Preparation of Stromal Cell Line (AC 6.21)
Feeder Layers
Stromal cell lines that can substitute for the mixed stroma are cur-
rently becoming available. One of several lines we have prepared, AC 6.21
(II), is especially good at supporting B lymphopoiesis from a very prim-
itive precursor through the cytoplasmic u.-positive stage (seeNote 3).
1. Obtain confluent or nearly confluent stroma cell layers.
2. Aspirate medium.
3. Wash stroma 3x with serum-free RPMI-1640.
4. Add a vol Collagenase-Dispase (0.5 mg/mL in serum-free medium)
Long-Term B Lymphoid Cultures 311

sufficient to cover the bottom of the vessel (approximately l/3-1 /2 of

the vol used to initiate the culture, Table 2).
5. Incubate at 37OC.
6. Wait 15 min.
7. Check by phase microscopy for cell detachment. Typically, AC 6.21
does not completely detach by this treatment until pipeted. If enzy-
matic digestion is complete, the membrane processes will be retracted
to thin branching fibers that give a “spider web” appearance to the cul-
ture.
8. If necessary, mix gently and wait 10-15 min more.
9. Check by phase microscopy again for cell detachment.
10. Once enzymatic digestion is complete, harvest the cells by vigorously
pipeting the vessel surface with aPasteur pipet. (If most of the cells are
still loosely adherent to the vessel before pipeting, then the Collagen-
ase-Dispase solution can be removed by aspiration and the stromal
cells harvested directly into a desired amount of fresh culture medi-
um containing serum. This allows elimination of Step 12 below and
simplifies routine passage of the cells.)
11. Check by phase microscopy for efficient removal.
12. Pellet cells at 800 rpm for 7 min if harvested into the enzyme solution.
13. Once in culture medium, count viable cells.
14. For preparation of feeder layers that are to be used within 1 or 2 d, dis-
pense the harvested stromal cells into the new cultures at a density
that is about l/4 to l/2 the density needed for preparing confluent
stromal cell layers (Table 4).
15. For routine passage of the stromal cell line, it should be carried at sub-
confluency to prevent selection of variants that are not contact-inhibi-
ted. Our routine is to passage the cell line once per week and, at each
passage, set up flasks that are 1:20 and 1:50 splits of a culture that is just
nearing confluency. This is approximately 1 x lo4 and 4 x 1O3cells /mL,
respectively.

3.2. Cloning of B Lineage CeZZs

(Limiting Dilution Culture)
Precursors capableof long-term proliferation inlymphoid bonemarrow
cultures can be cloned by limiting dilution culture on mixed stromal cell
layers or, preferably, on an established stromal cell line, such as AC 6.21.
The source of lymphocytes to be cloned can be either nonadherent cells
from established Whitlock-Witte cultures or fresh bone marrow (Note 4).
312 Whitlock and Muller-Sieburg

Table 4
Approximate Numbers of AC 6.21 Needed for Confluency
in Various Culture Vessels
Vessel No. for confluency Maximum densitv”
96-Well 4x103
24-Well 4x104
6-Well 2x105 4x105
60 mm dish 4x105 8x105
100 mm dish 1.1 x 106 2x106
T25 flask 5x105 1 x106
T75 1.5 x lo6 3x106
“AC 6.21 has an extensive membrane that can cover a large area of the culture dish. Cul-
tures that appear confluent can increase in cell number by reduction of the surface area
occupied by each cell; therefore, the number of cells needed to achieve confluency and the
maximumnumberof cells that canbe harvestedfromadensecultureare slightly different.
This column is provided as an easy reference for determining how many flasks need to be
prepared for large-scale experiments.

3.2.1. Preparation of Mixed Bone Marrow Stroma

for Limiting Dilution
1. Harvest marrow from the appropriate congenic mouse strain (Note 3)
and prepare as for larger bone marrow cultures.
2. Suspend cells in culture medium at 1.5 x lo6 cells/ml.
3. Initiate g&well cultures with 3 x 105cells/well in a vol of 0.2 mL.
4. Wrap stacks of plates in plastic wrap to decrease dehydration.
5. Incubate at 37OC for 2-3 d.
6. Pipet each well to suspend nonadherent cells. This is best accom-
plished by using a 12-space multiwell pipetor set at 0.15 mL.
7. Remove all the medium and add 0.2 mL of fresh medium to each well.
8. At weekly intervals, for a total of 3-4 wk, repeat steps 6 and 7 with the
goal of depleting the stroma of any nonadherent lymphoid precur-
sors prior to cloning.
9. At l-2 d prior to cloning, remove the medium.
10. Add 0.1 mL fresh medium containing 10% fetal calf serum to each
well.
11. During this 2-d period, screen the cloning plates for wells that contain
patches of nonadherent cells that may be mistaken for clones later on
and eliminate these wells from the limiting dilution experiment (that
is, simply ignore these wells when selecting and counting clones).
Long-Term B Lymphoid Cultures 313

3.2.2. Preparation of Stromal Cell Line Feeders

for Limiting Dilution
1. Harvest AC 6.21 stromal cells as described above. One T75 flask will
be adequate for four or five 96-well plates. The approximate numbers
of stromal cells that can be obtained from confluent vessels of differ-
ent sizes are provided in Table 4.
2. Suspend the harvested stromal cells in culture medium with 10% fetal
calf serum at 2 to 4 x lo4 cells/mL.
3. Aliquot 0.1 mL/well(75-100% confluency).
4. Use for limiting dilution experiments within l-2 d of initiation.
3.2.3. Cloning of Established Lymphoid Cell Lines
1. Harvest and adhere lymphoid cells from established bone marrow
cultures as described above.
2. Pellet and resuspend the lymphocytes in fresh medium with 10% serum
at 3 x lo3 cells/mL.
3. Make five threefold dilutions in the same medium to prepare suspen-
sions of 1000,333,111,33, and 11 cells/mL in quantities sufficient to
plate 0.1 mL/well for the desired number of plates. (Often less than
10% of those colonies growing after 3 wk of cloning will continue to
grow after expansion [Note 41; therefore, multiple plates at the lower
cell numbers need to be set up in order to obtain enough colonies that
have a reasonable chance of being clonal and continued growth.)
4. By using a 12-space multiwell pipetor, add 0.1 mL of each cell suspen-
sion to each of the appropriate wells.
5. Stack the plates, wrap in plastic wrap, and culture at 37°C.
6. Weekly, carefully remove 0.1-0.15 mL of spent medium by using the
multiwell pipetor so not to disturb the cells at the bottom of the wells.
7. Add an equivalent amount of fresh medium to each well.
8. After 3 wk, screen wells by phase microscopy to detect wells with lym-
phoid colonies. (The variety of colony types found on limiting dilu-
tion culture are described below [Note 51.)
9. Prepare a plot of the log frequency of negative wells vs the number of
cells added to the wells to check for linearity at the lower cell densi-
ties, particularly in the range from which clones are to be selected.
(Typically, there is negative interaction at high cell densities and devi-
ation from linearity.)
10. To expand the clones, prepare stromal cell layers in 24-well plates and
also in &well plates, 60 mm dishes, or T25 flasks for further expansion.
314 Whitlock and Muller-Sieburg

For mixed stroma, this process takes 3-4 wk, as described above, and
should be started in all sized vessels around the time of initiating the
limiting dilution cultures. Feeders of stromal cell line AC 6.21 can be
prepared l-2 d prior to use.
11. Check each mixed stromal feeder layer carefully for lymphoid cell
growth before use.
12. Change the medium to medium containing 10% serum l-2 d prior to
using the mixed stromal layers for expansion of clones.
13. Vigorously pipet the limiting dilution well with the clone to be har-
vested to disrupt the stromal layer and release any lymphoid cells inti-
mately associated with it.
14. Transfer the entire contents of the pipeted well to one well of a 24-well
plate containing an established feeder layer.
15. Feed the expanded clones weekly by aspirating 80% of the medium
and replacing it with fresh medium. Care should be taken not to cross-
contaminate wells, and the plates should be kept covered in plastic
wrap to prevent dehydration.
16. Any well showing patches of 50 or more nonadherent cells, indicative
of continued proliferation of the transferred cells, can be expanded
further.
3.2.4. Cloning of Selected B Cell Progenitors
from Fresh Bone Marrow
A subpopulation of fresh bone marrow cells, designated Thy-l’”
T-B-G-M- (Note 4) is enriched in progenitors capable of producing long-
term B-lymphoid cultures. Our method for isolating these cells for limit-
ing dilution is outlined below.
1. Harvest fresh bone marrow as described above for the initiation of
long-term lymphoid cultures, except use Iscove’s medium contain-
ing 5% fetal calf serum instead of RPMI-1640. (In Iscove’s medium,
bone marrow cells maintain good viability over a longer period of
time.)
2. Dilute appropriate antibodies (Table 1) in Iscove’s medium and ster-
ile filter through a 0.2 pm disk directly before use.
3. Pellet cells in a conical tube at 1200 rpm for 10 min.
4. Suspend cells in rat anti-Thy-l antibody. (For this and all subsequent
steps, 10 PL of antibody or avidin is used for each lo6 cells.)
5. Incubate on ice for 15 min.
6. Wash (i.e., dilute the antibody/cell suspension with 2 mL Iscove’s
medium containing serum, then underlay with 50-100% fetal calf
serum. Pellet the cells through the serum, then carefully aspirate the
Long-Term B Lymphoid Cultures 315

liquid from the tube being careful not to disturb the cell pellet, but re-
move as much serum and antibody as possible.)
7. Suspend the cell pellet in FITC-Goat anti-Rat Ig and incubate 15 min.
8. Wash as above (Step 6).
9. Suspend the cell pellet in normal rat serum (diluted 1:4 in Iscove’s me-
dium) and incubate 3-5 min.
10. Add a cocktail of biotinylated antibodies: rat anti-B220, rat anti-Gran-
I, rat anti-Mac-l, rat anti-Lyt-2, and rat anti-L2T4. Incubate 15 min on
ice.
11. Wash as above (Step 6).
12. Suspend the cell pellet in Texas red-Avidin and incubate 15 min.
13. Wash as above (Step 6).
14. Suspend the cell pellet in Dulbecco’s modified PBScontaining 20 pg/
mL propidium iodine.
15. Filter the cell suspension through anylon screen to remove cell clumps
before sorting.
16. The cell population of interest is that which stains weakly with anti-
Thy-l and not at all with the antibodies specific for B cells, macro-
phages, mature T cells, and granulocytes. Generally, we obtain 1 to 2
x lo4 Thy-l’” T-B-G-M- cells in a 2 h sort.
17. Collect sorted cells into ice cold Iscove’s medium with 5-10% serum.
18. Count the sorted cells.
19. Suspend the cells in culture medium with 10% fetal calf serum (now
back to RPMI-1640 medium) at 300 cells/ml. Make threefold dilu-
tions in quantities sufficient to plate 0.1 mL in each well of the num-
ber of desired limiting dilution plates. (If bulk cultured at this point,
suspend at 103-lo4 cells/ml.)
20. Aliquot 0.1 mL of each cell suspension on established feeder layers in
96-well plates (seeabove).
21. Feed weekly as described above for limiting dilution cloning of estab-
lished lymphoid cultures (Section 3.2.3).
22. Screen wells after 3 wk of culture for lymphoid colonies, as described
in Section 3.2.3.
23. Select and expand clones as described in Section 3.2.3.

4. Notes
1. Selection of a batch of fetal calf serum that efficiently supports B-lym-
phocyte proliferation in this culture system is the key to the successof
all the above procedures. Screenings we have done suggest that the
highly defined serums do not work well in this system. Of those re-
 Whitlock and Muller-Sieburg

maining, approximately 20% will work. Two basic criteria are useful
for comparing serum lots. One is its ability to permit good establish-
ment (50-75% confluency at 3 wk of culture) of low-cell density, mixed
stromal cell layers using Balb/c bone marrow. The second criterion is
outgrowth of lymphoid colonies in high cell density cultures. In a 60
mm dish, high cell density (IO6 cells/ml in 5 mL) culture at 3 wk, there
should be a minimum of five discreet patches of lymphoid cells con-
taining more than 500 cells each. Serum lots that have proven to work
well for this culture system have been stable for up to 4 yr when stored
at -35°C.
2. Recognition of whether a culture is establishing adequately is im-
portant so that time is not wasted waiting for B lineage cultures to
establish when the potential of successful establishment is low. The
ability to assess how well a culture is doing comes only with exper-
ience, but the following gives a verbal description of how the cultures
should look at different times after initiation, which may aid in assess-
ment. The first phase is outgrowth of the stromal layer and death of
the majority of the nonadherent cells. This progresses rapidly during
the first 7-10 d of culture. Both lymphoid and myeloid precursors
survive this initial phase to proliferate and form discrete foci of non-
adherent cells. Lymphoid colonies tend to be intimately associated
with the stromal cells that underlie them and are composed of cells
that vary in size from small to medium-sized lymphocytes (Fig. 1).
Myeloid colonies are more uniform in cell size and do not conform to
the shape of the stromal cells beneath them. Myeloid cells also are
larger than even the largest lymphoid cells.
The second phase is a crisis phase in which nonadherent cell num-
bers decrease and, with some serum batches, essentially disappear.
This occurs and lasts a variable amount of time after culture initiation,
but a return of nonadherent cells should be seen after 3-4 wk of cul-
ture. Only lymphoid cell proliferation survives this second phase.
With good batches of serum, the crisis phase is short, and often the first
phase overlaps with the third phase, characterized by steady increases
of nonadherent cell numbers to the maximum cell density character-
istic of the individual culture (l-10 x KP/mL).
The last or fourth phase is characteristic of Balb/c bone marrow
cultures, but we have not determined whether it occurs in cultures
initiated with other mouse strains. If the heterogeneity of immuno-
globulin gene arrangements of the nonadherent cells is followed during
phase three, it is found that the culture gradually become pauciclonal
after 4-6 mo of culture. The beginning of the fourth phase is heralded
Long-Term B Lymphoid Cultures 317

by a gradual increase in the maximum number of nonadherent cells in

the culture and coincides with the cultures becoming more and more
pauciclonal. The cultures progress to a stage when the nonadherent
cells must be subdivided in order to maintain viability, and eventually
a clonal, tumorigenic, stromal layer-independent cell line emerges.
Because of the tendency of older cultures to be pauciclonal and less
“normal” in their growth patterns, those interested in studying normal
B cell development should limit their experiments to cultures that are
between 3- and 12-wk-old.
3. Feeder Layers: Mixed stroma vs stromal cell lines and the use of
congenic mouse strains. Proliferation of B lymphoid cells in this cul-
ture system is dependent on adherent stromal cells. In the primary
bone marrow cultures, as described in the previous Note, a mixture of
adherent cells establishes what we call a “mixed stromal cell feeder
layer.” From one established culture, we isolated a number of stromal
cell lines that can substitute for the mixed stroma. One such line, AC
6.21, is especially good at supporting proliferation.
Use of the stromal cell line, AC 6.21, for limiting dilution culture has
many advantages. First, a large number of 96-well plates can be pre-
pared in a relatively short amount of time with a minimum of effort.
Second, the microenvironment within each well is more consistent
than with mixed stroma, and third, there is never any problem with
contamination of the clones with stroma-derived lymphocytes. Ir-
radiation of the stromal cells is not needed since this stromal cell line
is contact inhibited and will remain viable for several weeks without
passage if fresh medium is given weekly. This latter characteristic
makes it highly useful for cloning and limiting dilution analysis of B
lineage precursors.
A disadvantage of this stromal cell line over mixed stroma is that
dead cell debris accumulates on the AC 6.21. The failure of this debris
to accumulate on the mixed stroma probably arises from the presence
of macrophages that phagocytize the debris rather than a lower turn-
over of the lymphoid cells.
We have used the AC 6.21 stromal cell line for most of our limiting
dilution studies and know that it can support proliferation and differ-
entiation of primitive progenitors (Thy-2*OT-B-G-M-) to the pm-B cell
stage of differentiation with synthesis of cytoplasmic immunoglobu-
lin (II). Differentiation to surface Ig-bearing cells does not proceed as
well on the stromal cell line as it appears on mixed stromal cell feeders.
AC 6.21 is grown in the same medium used for long-term lym-
phoid bone marrow cultures. P-mercaptoethanol is essential for via-
 Whitlock and Muller-Sieburg

bility of AC 6.21, and media should always contain 5-10 x WM fresh

2-mercaptoethanol.
Stromal cell lines such as AC 6.21 have a tendency to select for vari-
ants that will overgrow the confluent cell layer if it is carried routinely
in a confluent state. Therefore, cultures for passage should be no more
than around 80% confluent at the time of subculture. It is also advisa-
ble to have a large number of vials of an early passage frozen for future
use in the event the stromal cell line does start to overgrow.
If mixed stromal cell layers are used as feeders for limiting dilution
culture or routine passage or expansion of bulk cultures or clones, it
should be kept in mind that the feeder layer may give rise to lym-
phoid cells that may contaminate the passaged cells. A simple method
for being able to monitor feeder layer contamination is to use congenic
mouse strains to prepare the feeders. Balb/c congenics at the histo-
compatibility locus (e.g., BALB.B and BALB.K) and the immuno-
globulin locus (e.g., BAB-14) work equally well in this culture system.
No problem has arisen in our experiments when we have seeded B
lymphoid cultures onto feeders with a different H-2-type. Histocom-
patibility markers can be easily monitored by surface immunofluo-
rescence. Balb/c and BAB-14 cells can be differentiated by restriction-
length fragment polymorphisms in the heavy chain constant region
genes (12).
4. Limiting Dilution Culture Theories and Practice: Limiting dilution
culture simply means to culture a cell population at progressively
lower densi ties until the cells of interest are at a frequency of less than
l/well. The statistical theory used at the basis for this technique is
beyond the scope of this manuscript, but in practice two points should
be kept in mind. One is that if proliferation of the cell of interest is de-
pendent only on the microenvironment provided by the culture con-
ditions (in this case the medium and stromal cells) and if there is no
negative or positive interaction between the cell of interest and other
cell types in the mixed population cultured at limiting dilution, then
a plot of the log of the frequency of negative wells (those not containing
the cell of interest) versus the number of cells per well cultured will be
linear. The second point to remember is that at a plating frequency of
one cell of interest per well, approximately two-thirds of the wells will
contain that cell. Since one-third of the wells do not contain the cell of
interest, then a significant number of wells will contain two or more.
This latter point must be kept in mind when selecting the plates from
which to choose clones.
Long-Term B Lymphoid Cultures 319

Approximately 1 in 1000 fresh bone marrow cells, when seeded at

limiting dilution on an established bone marrow stromal layer, have
the capacity to proliferate up to4 wk. As detailed above, this frequency
has been derived from limiting dilution analysis using AC 6.21.
Although the vast majority of nonadherent cells found in a well-es-
tablished long-term lymphoid bone marrow culture are pre-B and B
cells, bone marrow derived pre-B cells (purified by flow cytometry as
cells that express B220) do not have the capacity to initiate long-term
lymphoid cultures. We have found this capacity exclusively in a small
population of bone marrow, characterized by expression of low levels
of the cell surface antigen Thy-l. These cells lack B220, Mac-l, Gran-
1, L3T4, and Lyt-2 antigens, and are thus termed Thy-PTB-G-M-cells.
One in 15 Thy-PTBG-M- cells, which comprise 0.1-0.2% of total bone
marrow, give rise to a lymphoid culture (Table 5). Thy-llDT-B-G-M-
cells are also highly enriched for pluripotent hematopoietic stem cells
(13). This indicates that long-term lymphoid bone marrow cultures
provide an in vitro system that allows the study of B lineage differen-
tiation from stem cells through mature B cells. When established
cultures are used for limiting dilution culture (3-8 wk post initiation),
approximately 1 in 10 will proliferate to form a colony of 50 to several
thousand cells after 3 wk. Only 10% or fewer of these will continue to
proliferate after expansion. Therefore, when planning a limiting dilu-
tion experiment for obtaining clones for future experiments, a large
number of plates at the dilutions that should have approximately 0.3
clonogenic cells/well should be initiated to obtain a large number of
clones for expansion. Which clones will have long-term proliferative
potential should be obvious after 2 wk of culture in the 24-well plate.
5. Selecting Lymphoid Clones from the Variety of Colonies Seen in
Limiting Dilution Cultures: The most striking finding on limiting di-
lution culture of bone marrow cells on AC 6.21 is the complexity of the
colony types observed. The colonies not only differ greatly in cell
number, but also in cell morphology. The most prominent colony that
will be observed is one that consists of large cells, approximately three
times the size of a lymphocyte and are round and nonadherent in the
1st wk in culture (Fig. 2A and 2B). With time, the cells in these colo-
nies become granular and vacuolated, and they adhere to the stromal
cells beneath them (Fig. 2C). About 1 in 35-50 fresh bone marrow cells
will form such a colony; therefore, these colonies present a problem in
limiting dilution culture of lymphoid cells in whole bone marrow.
Large numbers of them in a well will compete with growth of lym-
 Whitlock and Muller-Sieburg

Table 5
Frequency of B Cell Precursors in Sorted Bone Marrow Populations
Population Bone marrow, % 1/frequency”

Total bone marrow 100 1200

B220+ 30 >4000
B-M-G” 16 150
Thy-l+b 5 70
Thy-l ‘“T-B-G-M- 0.1-0.2 15f5
“B-M-G: cells depleted of B220, Gran-1, and Mac-l expressing cells.
Thy-l’: contains both T lymphocytes and Thy-l’“TB-G-M- cells.
‘Frequencies of wells containing lymphoid colonies were determined at 2 wk.

phoid colonies. Since lymphoid cell precursor frequency in whole

bone marrow is 1 in 300-1000, then every well containing a lympho-
cyte precursor also potentially contains six or more colonies of these
large, granular cells. Fortunately, the large, granular cells do not have
long-term growth capacity and, therefore, will be easily depleted with
passage of the lymphoid clones.
Lymphoid colonies can be recognized as those that consist of very
small cells that vary in size and tend to adhere to the stromal cells (Fig.
2A). All colonies with this gross appearance can be shown to contain
lymphocytes by Wright’s-Geimsa staining and immunofluorescent
staining of B220, a B lineage specific surface marker.
A third group of colonies consists of cells that are more uniform in
appearance and intermediate in size when compared to lymphocytes
and large, granular cells (not pictured). Wright’s-Geimsa staining of
these cells most often shows them to contain predominantly primi-
tive cell types with loose chromatin and basophilic cytoplasm. In
some cases, these primitive cells are mixed with some lymphocytes or
polymorphonuclear leukocytes. Attempts to expand such colonies
have failed thus far.
If nonadherent cells from primary lymphoid long-term bone mar-
row cultures are used as the source of cells for cloning, then the distri-
bution of colonies obtained is more uniform. If young cultures are
used (3-wk-old or less), then there is still a significant frequency of
large, granular cell colonies that are obtained, but the frequency is
about lo-fold less than for fresh bone marrow. In addition, the fre-
quency of clonable B lineage cells is IO-fold higher; therefore, it is
simpler to obtain a large number of pure B lineage clones by using es-
tablished B lineage cultures. However, our experience has shown that
 Fig. 2A Fig. 2B Fig. 2C

Fig. 2. Colony types in limiting dilution culture of fresh bone marrow cells on AC 6.21. (A) A colony of small lym-
phocytes is shown adjacent to a young colony of “large-granular cells.” The latter cells are at a stage in which most
are still nonadherent, but many are beginning to increase their granularity. (B) A higher power view of the large-
granular cell colony from (A). (C) Pictured is a high power view of a ‘large-granular cell” colony that is more ma-
ture than shown in (B). All of the cells are now highly granular and adherent. Many are swollen by a large intra-
cytoplasmic vacuole that displaces the nucleus to one side (signet ring formation). Magnifications: (A) 25x, (B) 50x,
(Cl 50x.
322 Whitlock and Muller-Sieburg

clones obtained from a single experiment can have the same immuno-
globulin heavy chain gene rearrangements and, thus, are siblings.
There are other cell types that undergo limited proliferation on AC
6.21 to give colonies of less than 50 cells. Once such cell type is oblong
in shape and forms a colony where the cells characteristically migrate
away from each other (not pictured). We have not been able to obtain
enough of these cells for WrighYs-Geimsa stain or other methods of
characterization.

References
2. Dexter,T.M.andLajtha,T.G. (1974) Proliferationofhaemopoieticstemcellsintitro.
Br. I. Haematol. 28,525-530.
2. Whitlock,C.A.andWitte,O.N. (1982)LongtermcultureofBlymphocytesand their
precursors from murine bone marrow. Proc. Natl. Acad. Sci. USA 79,3608-3612.
3. Coffman, R. L. and Weissman, I. L. (1983) Immunoglobulin gene rearrangement
during pre-B cell differentiation. I. MoI. Cd. Zmmunol. 1,31-38.
4. Springer, T., Galfre, G., Secher,D. S.,and Milstein, C. (1979) Mac-l: a macrophage
differentiationantigenidentifiedbymonoclonalantibody. Eur.J. Zmmunol. 9,301-306.
5. Holmes, K. L., Langdon, W. Y., Fredrickson, T. N., Coffman, R. L., Hoffman, P. M.,
Hartley, J.W., and Morse, H. C. (1986)Analysis of neoplasmen induced by CAS-BR-
M MuLV tumor extracts. 1. ZmmunoE. 137,679-688.
6. Ledbetter, J. A. and Herzenberg, L. A. (1979) Xenogeneic monoclonal antibodies to
mouse lymphoid differentiation antigens. Zmmunol. Rev. 47,63-90.
7. Lostrom, M. W., Stone, M. R., Tam, M., Burnette, W. N., Pinter, A., and Nowinski,
R. C. (1979). Monoclonal antibodies against leukemia viruses: identification of six
antigenicdeterminantsonthep15(E)andgp7Oenvelopeproteins. Virology98,336-350.
8. McGrath, M. S., Pillmer, E., and Weissman, I. L. (1980) Murine leukemogenesis:
Monoclonal antibodies to T cell determinants arrest T lymphoma cell proliferation.
Nafure 285,259-261.
9. Dialynos, D. P., Wilde, D. B., Marrack, P., Pierrces, A., Wall, K. A., Havran, W.,
Otten, G., Loken, M. R., Pierres, M. R., Kappler, J.,and Fitch, F. W. (1983) Character-
izationof the murine antigenic determinant, designated L3T4a, recognized by mono-
clonal antibody GK1.5: Expression of L3T4 by functional Tcell clones appears to cor-
relate primarily with classII MHC antigen restriction. Zmmunol. Rev. 74,29-55.
10. Hardy, R. (1984)Purification and coupling of fluorescent proteins for use inflow cy-
tometry, in HandbookofExperimentaZZmmunology, fourth edition (Weir, D. M., Black-
well, C., and Herzenberg, L. A., eds.), Blackwell Scientific, Oxford, pp. 146-155.
11. Whitlock, C. A., Tidmarsh, G. F., Muller-Sieburg, C. E., and Weissman, I. L. (1987)
Bone marrow stromal cell lines with lymphopoietic activity express high levels of a
pre-B neoplasia-associated molecule. CeII 48,1009-1021.
12. Nottenberg, C. and Weissman, I. L. (1981) Ccl gene rearrangements of mouse im-
munoglobulin genes in normal B cells occurs on both expressed and nonexpressed
chromosomes. Proc. Natl. Acad. Sci USA 78,484-488.
13. Muller-Sieburg, C. E., Whitlock, C. A., and Weissman, I. L. (1986) Isolation of two
early B lymphocyte progenitors from mouse marrow: A committed pre-pre-B cell
and a clonogenic Thy-l” hematopoietic stem cell. Cell 44,653-662.
 Chapter 26

CFU-C in Agar

Vincent Praloran
and Anna Bartocci
1. Introduction
Agar culture systems for the clonal growth and differentiation of
hemopoietic cells were first described 20 yr ago (1). The progenitor cells
that developed into colonies in agar after several days of culture in the
presence of a source of hemopoietic growth factor (2,3) were initially called
“Colony Forming Units in Culture” (CFU-C). They are found in bone
marrow, spleen, blood, fetal liver, and yolk sac. It was subsequently
demonstrated that the CFU-C population was heterogeneous and con-
tained progenitors giving rise to granulocyte/macrophage (CFU-GM),
granulocyte (CFU-G), andmacrophage (CFU-M) colonies. Progenitor cells
of other lineages (erythroid and megakaryocytic) have also been similarly
demonstrated in hemopoietic organs.
The progressive identification of different types of progenitors, the
demonstration of their hierarchical distribution, the purification of the
growth factors regulating their proliferation and differentiation, and the
results of other in vitro and in vivo experiments led to a schematic three-
compartment model for hemopoiesis involving:

323
324 Praloran and Bartocci

1. The stem cell compartment, with extensive self-renewal and commit-

ment potential.
2. A committed progenitor cell compartment that is heterogeneous and
hierarchically distributed from multipotential to unipotential pro-
genitors. The differentiation of cells in this compartment is associated
with a progressive loss of proliferative potential and is dependent
upon the presence of specific growth factors.
3. A compartment of maturing and mature cells, composed of morpho-
logically identifiable cells restricted to one lineage, endowed with
very limited proliferative potential. These cells quickly acquire the
phenotypic and functional properties of circulating mature cells
(Fig. 1).
The in vitro culture technique (1) described in this chapter has pro-
vided much information on the hemopoietic system over the last 20 yr.
It will remain useful for:
a. Determining the frequencies of different types of progenitors and
their modification in vitro or in vivo by different agents.
b. The screening of various substances for their effects on pro-
genitor cell proliferation and/or differentiation.
c. The identification and isolation of hemopoietic stem cells and
the analysis of their growth factor requirements.

2. Materials
1. Media:
a. Alpha (a) medium stock solution: dissolve the powder for 1 L
of medium in 177 mL of deionized distilled water (DDW), filter
through a 0.22 pm sterile filter and store at -2OOC (seeNote 1).
b. x2 alpha medium:
Alpha stock solution 32 mL
Glutamine solution (200 mM) 2mL
Fetal calf serum (seeNote 2) 40 mL
Sodium bicarbonate
(5.6% w/v solution) 8mL
Gentamycin solution (10 mg/mL) 0.5 mL
DDW 17.5 mL
Mix thoroughly, filter through a 0.22 pm filter, and store at 4OC
for less than 1 wk.
c. xl Medium for cell collection and dilution: x2 alpha medium
diluted with an equal volume of water and adjusted to pH7.35.
CFU-C in Agar 325

-. -,
PI Er Eo Ba G Bly Tly

Fig. 1. Schematic representation of the hemopoietic system emphasizing granulocyte

and macrophage production. IL1 actson very primitive hemopoietic cells, both directly
onprimitiveprogenitorsandindirectly through the stimulationof theproductionof other
growth factors by accessorycells. IL-1 alone does not stimulate hemopoietic cell prolif-
eration. Furthermore, the proliferation of very primitive cells is not stimulated by either
IL-l, IL-3, GM-CSF or CSF-1 alone. Combinations of IL-3, GM-CSF and CSF-1 with IL-
1 are required (seeChapter 23). IL-3 is a multipotential growth factor, active on all
hemopoietic cell lineages. GM-CSF, CSF-1 (M-CSF), and G-CSF act on granulocyte or
macrophage progenitor cells and maturing or mature granulocytes and macrophages.
They not only stimulate proliferation and differentiation, but also survival and activation
of functional properties of the mature cells of these lineages. (Pl): platelets; (Er) erythro-
cytes;(Eo) eosinophils: (Ba) basophils; (MB) monocytes-macrophages; (G) granulocytes;
(Bly) B lymphocytes; (Tly) T lymphocytes.

2. 1% Agar Solution: In a 500 mL conical flask, mix 2 g of agar with 200

mL of DDW. Heat with stirring to boiling; allow to cool and reboil a
second time. Allow to cool to 50°C (seeNote 3).
3. 0.66% Agar Solution: Add 33 mL of prewarmed DDW to 66 mL of 1%
agar at 50°C and reboil once.
4. Ficoll Paque.
5. Gas Mix: 7% 02, 10% CO, 83% N,.
326 Praloran and Bartocci

3. Methods
3.1. Bone Marrow CelZ Collection
3.1.1. Mouse Bone Marrow
Sterilize surgical instruments by boiling or in 70% ethanol.
1. Kill the mice by cervical dislocation (seeNote 4). Sterilize the skin with
ethanol and, using sterile scissors, remove the muscles from the femur
and around the knee. Disarticulate at ball and knee joints. Cut the
lower extremity of the femur at the epiphyse-cartilage junction.
2. Using a 3 mL sterile plastic syringe with a 23-gage needle, flush the
femoral cavity several times with medium, collecting cells and me-
dium in a 5 mL sterile plastic tube.
3. After flushing all femurs, disperse the cells to form a single cell sus-
pension (seeNote 5).
4. Count the white cells and adjust the cell concentration, as desired,
with medium.
3.1.2. Human Bone Marrow
1. Samples are aspirated from the posterior iliac crest or sternum in
syringes previously rinsed with preservative-free heparin diluted
1:lO in heparinized (50-100 U/mL) xl medium.
2. Carefully layer this aspirate on a Ficoll-Paque cushion and centrifuge
at 400g for 30 min at 15OC (4). The mononuclear cells are aspirated
from the interface, washed two times by centrifiguration in xl me-
dium and diluted to the desired concentration prior to plating as de-
scribed below.
3.2. PZating Procedure
1. Number the Petri dishes (35 mm Falcon) and add growth factors
(GF:0.15 mL/dish) (seeNote 6) according to the appropriate experi-
mental protocol. Each experimental condition must be set up at least
in triplicate, depending on the expected progenitor cell frequency.
2. Calculate the total volume of agar-medium needed for underlayers
and overlayers according to the total number of dishes to be used. In
order to have enough agar medium for all dishes, prepare the under-
layer and overlayer agar-medium mix for 10 more dishes than re-
quired in the protocol, e.g., for 50 dishes:
2 x a medium, Agar,
GF 37-40°c 37-39°C Cells
Underlayer ++ 30 mL 30 mL (1%) -
Overlaver - 15 mL 15 mL (0.6%) ++
CFU-C in Agar 327

3. Underlayer: Keeping the pool of agar medium at 37OC, distribute 1

mL/dish in all dishes using a pipet or an automatic pipeting device.
Allow the underlayer to gel at room temperature for 5-15 min.
4. Overlayer: Prepare the agar medium (0.33% final) and quickly add the
calculated volume of cell suspension, mix thoroughly (seeNote 7) and
dispense 0.5 mL/dish keeping the temperature of the agar medium
pool at 37OC (seeNote 8) and swirling the dishes as you proceed to
ensure even distribution.
5. Allow the overlayer to gel at room temperature and incubate imme-
diately.
3.3. Incubation Procedure
1. Transfer the dishes to grids over a thin DDW layer in plasticboxes (Fig.
2) (for a detailed description, seeChapter 23).
2. Seal the boxes with plastic tape and using holes in the cover (one as an
inlet, the other as an outlet) gas with a 7% 02, 10% CO, 83% N, (5) gas
mixture for 20 min, at a flow rate of 3 L/min. At the end of this time,
the medium in the dishes should look slightly orange.
3. Immediately seal the two holes in the cover with plastic tape.
4. Incubate at 37OCin the dark for 7-10 d depending on the type of pro-
genitor cell colony.
5. Use a dissecting microscope at x20 magnification to count colonies
(>50 cell/aggregate) and clusters (4-50 cells/aggregate).

3.4. Morphological Analysis of Colony CeZZs

Procedures
._ _ for whole plate staining __are time
._ consuming
_. and give poor
results with the agar system. For morphological studies, we suggest a col-
lagen gel culture technique that allows good quality in situ staining of
colonies (6):
1. Prepare a collagen solution from rat tail tendons according to the tech-
nique of Lanotte and Schor (6,7, seeChapter 19, this vol.). Briefly, the
tail tendons are cut out, dropped in absolute ethanol for sterilization
and transferred to 0.5M acetic acid for depolymerization. When all the
collagen is solubilized, it is dialyzed against distilled water, filtered
through a 0.22 pm sterile filter and sterilely stored at 4°C.
2. The culture medium mix contains:
x2 Medium 25%
Collagen solution 25%
FCS 20%
xl Medium 20 or 25%
Cell suspension 5%
Growth factors 5% (or added separately to the dishes)
328 Praloran and Bartocci

SUMMARY
1) PLATE UNDERLAYER.
+015ml +lml Cool Down
GF 0 5% Agar
at 40°C

2) PLATE OVERLAYER:

+ 0.5ml Cool Down

0.33% Agar + Cells
at 40°C

3) TRANSFER DISHESTO BOXES AND SEAL WITH PLASTICTAPE

-PLASTIC
TAPE

4) GAS BOXESFOR 20’ WITH GAS MIXTURE

5) COMPLETELYSEAL BOXESAND INCUBATE

6) RECORD:

Clusters

Fig. 2. Schematic summary of plating and incubation procedure.

CFU-C in Agar

All reagents, including the mixed medium, must be kept at 4°C until
distribution in dishes. Polymerization of collagen (gel formation)
occurs when the temperature rises above 15OC.
3. Incubate as for agar cultures.
4. To prepare the gel for staining, it must be transferred from the plate to
a glass slide, then dried and fixed. This part of the method is delicate
and is described in detail in the figure in ref. 8.
5. The dried gels yield a very thin film of collagen containing the colo-
nies and clusters. Standard staining andmost cytochemical reactions
are as easily performed on the dried gel as on a smear and allow easy
cellular identification (8).

4. Notes
1. The Alpha medium stock solution must be prepared without sodium
bicarbonate in order to avoid amino acid precipitation upon thawing.
2. Several batches of fetal calf serum must be tested in this culture system
in order to eliminate batches with inhibitory activity.
3. Agar must not be sterilized by autoclaving. Care must be taken to
avoid burning during preparation. Discard the preparation if burning
occurs.
4. For mouse bone marrow cells, even if a low number of cells is needed,
it is necessary to use cells pooled from three different mice.
5. The cell suspension has to be carefully dispersed to single cells by
repeated and patient pipeting. Before diluting for culture, tiny cell
aggregates and debris may be eliminated by allowing them to settle
under unit gravity.
6. If several growth factors are to be added in the same dish, be careful
not to contaminate pipet tips with other growth factors. When dis-
pensing the underlayer, avoid touching the dishes or the dispensed
growth factor.
7. A homogeneous cell distribution in the agar mix culture medium is
necessary for reliable results; because of its high viscosity, it requires
careful mixing.
8. To prevent premature gelling, the temperature of the agar medium
mix must be kept between 37 and 40°C. A previously adjusted water
bath is very useful for this purpose. Gelling of the agar underlayer
must be complete before addition of the overlayer and the overlayer
must gel before incubation at 37OC. Approximately 5 min at 20°C is
required for gel formation.
330 Praloran and Bartocci

References
1. Bradley, T. R. and Metcalf, D. (1966)The growth of mouse bone marrow cells in vitro.
Austalian Journal of Experimental Biology and Medical Science 44,287-299.
2. Metcalf, D. (1986) The molecular biology and functions of the granulocyte macro-
phage-stimulating factors. Blood 67,257-2&I.
3. Clark, S.C. and Kamen, R. (1987) The human hematopoietic colony stimulating fac-
tors. Science 236,1229-1237.
4. Boyum, A. (1968) Isolation of leukocytes from human blood. Scandinavian ]ournal of
Clinical and Laboratory Investigation 21: Supp. 97,31-50.
5. Bradley, R. R., Hodgson, G. S.,and Rosendaal, M. (1987) The effects of oxygen ten-
sion on haemopoietic and fibroblast cell proliferation in vitro. I. Cell Physiol. 94,
517-522.
6. Lanotte, M., Schor, S.,and Dexter, T. M. (1981) Collagen gels as a matrix for haemo-
poiesis. Journal of Cellular Physiology 106,269-277.
7. Schor, S. L. and Court, J. (1979) Different mechanisms in the attachment of cells to
native and denatured collagen. Journal of Cell Science 38,267-281.
8. Lanotte, M. (1984)Terminal differentiation of hemopoietic cell clones cultured in tri-
dimensional collagen matrix in situ cell morphology and enzyme histochemistry
analysis. Biology of the Cell 50,107-120.
 Chapter 27

Human Long-Term
Bone Marrow Culture

Armand Keating and Paul Toor

1. Introduction
Major advances in our understanding of human hematopoiesis have
come from the development of semisolid in vitro culture techniques for the
detection of progenitor cells capable of colony formation (seeChapter 28,
this volume). However, colony assays select for hematopoietic precursors
committed to a specific lineage(s) and, therefore, are of limited value in the
assessment of very early progenitor cells. Also, they do not provide a
means of assessing the influence of microenvironmental cells deemed
essential for maintaining hematopoiesis in vivo (I). Furthermore, for the
assays to be valid, cells must be plated at a sufficiently low concentration
to ensure clonality (a condition in which each colony has arisen from a
single progenitor cell). Consequently, cell-cell interactions that may be
important in exerting regulatory effects on hematopoiesis are likely to be
minimal.

331
332 Keating and Toor

These limitations, in part, have been overcome by the adaptation of

Dexter’s murine long-term marrow culture system (2) to human marrow
(3). An essential feature of this liquid culture system is the development
of an adherent layer of heterogeneous composition within 2-3 wk after cul-
ture initiation (3). The predominant cell type in the adherent layer shares
many of the characteristics associated with smooth muscle cells (1). Loosely
adherent and nonadherent hematopoietic cells residing above the adher-
ent layer can be removed and assayed for colony forming progenitor cells
(CFC). Under optimal conditions for the initiation and maintenance of
human long-term marrow cultures (LTMC), granulocyte-macrophage
progenitors (CFU-GM) can be regularly detected for 14 wk and erythroid
(BFU-E) and multipotent progenitors (CFU-GEMM) for up to 8 and 4 wk,
respectively (4). More primitive progenitors present within the adherent
layer can only be assayed by disrupting the cultures (5).
Thus, the human long-term marrow culture system enables the in-
vestigator to examine and manipulate interacting cell populations in a
more physiologic manner and permits an assessment of the adherent cell
layer, regarded as a suitable in vitro model of the hematopoietic micro-
environment (1). Moreover, long-term marrow cultures may be initiated
by precursor cells earlier in ontogeny than the progenitors detectable in
semisolid clonal assays, since successful cultures can be generated from
marrow depleted of CFU-GM (6). Although the system has limitations,
such as a tendancy to favor granulopoiesis (4), it remains the best in vitro
approximation currently available to in vivo hematopoiesis.

2. Materials
1. Fetal bovine serum.
2. Horse serum.
3. Percoll (Density 1.077 g/mL).
4. McCoy’s 5A tissue culture medium.
5. Sodium bicarbonate (7.5% w/v>.
6. 100 mM sodium pyruvate.
7. Media additions: Vitamins (100x concentrate), essential amino acids
(50x concentrate) glutamine (200 mM) (as provided by the manufac-
turer, Gibco).
8. Antibiotic-antimycotic solution (each mL contains 10,000 U penicil-
lin, 10,000 U streptomycin, 25 yg amphotericin B).
9. Hydrocortisone, 2.75 mM in Dimethylsulfoxide.
Long-Term Human Bone Marrow Culture 333

3. Methods
The procedure for LTMC generation is outlined schematically in Fig.
1 (Note 1).
3.1. Long-Term Culture Medium
1. Place the following ingredients in a 500 mL Erlenmeyer flask:
mL
Hydrocortisone (1 mg/mL
Dimethylsulfoxide) 0.18
Sodium bicarbonate 5.0
Sodium pyruvate 5.0
Vi tamins 5.0
Essential amino acids 4.0
Nonessential amino acids 2.0
Glutamine 5.0
Antibiotic-antimycotic solution 5.0
Horse serum 62.5
Fetal bovine serum 62.5
McCoy’s 5A medium 343.82
500.00
2. Stir the contents of the flask gently for a few minutes to ensure
thorough and uniform mixing.
3. Filter-sterilize the medium through a 0.2 pm filter, and store at 4OCfor
use within 3 wk.

3.2. Isolation of Cell Populations (Note 2)

1. Aspirate 3-5 mL of bone marrow into a syringe containing 500 U pre-
servative-free heparin.
2. Dilute the marrow 1:l with McCoy’s 5A medium.
3. Isolate the mononuclear cells by density centrifugation over Percoll
(Note3). Collect the interface mononuclear cells after centrifigation at
4008 for 30 min, wash three times in medium, count, and resuspend in
McCoy’s 5A medium.

3.3. Generation of Long-Term Culture

1. Place the cells into long-term culture medium in any of the following
systems (Note 4) (Fig. 2):
334 Keating and Toor

Bone marrow cells (Collected In preservative-free hepann, 10 UnltSlmL,

and diluted 1 1 with McCoys 5A medium)

Density centrifugation
(PercoIl* 1077 g/mL, 30 mmutes, 400g)
I
Mononuclear cells (Suspended in long-term culture medrum)

25 cm2 tissue culture 20 x lo* cells in 10 mL

flask

Flat-sided tissue
culture tube 3-5 x 10’ cells in 2.5 mL
(“Ambitube”)
or

6 well multi-well 10 x lo6 cells In 5 mL per well

plate

or

Multi-chamber/glass slide 2 x 10e cells in

micro-culture system 1.0 mL per chamber

Incubate cultures at 37°C In a

5% COB humidified atmosphere for 7 days, then
transfer to Incubator at 33°C

Feed cultures weekly by replacing half of

supernatant with fresh medium

Weekly plate cells collected from removed supernatant

for presence of CFC.

Fig. 1. Scheme for generating human long-term marrow cultures.

a. 20 x lo6 cells in 10 mL medium in a 25 cm2 tissue culture flask.

b. 3-5 x lo6 cells in 2.5 mL medium in a flat-sided tissue culture
test-tube (“Ambitube”).
c. 10 x lo6 cells in 5 mL medium in each well of a six well multi-
well plate.
Long-Term Human Bone Marrow Culture 335

Fig. 2 Light micrograph of an adherent layer from a hematopoietically active human

long-term marrow culture. (x100).

d. 1 x 106cells in 0.5 mL medium in each chamber of a multicham-

her/glass slide microculture system (Notes 5,6).
2. Incubate the cultures at 37oC in a humidified incubator containing 5%
co,
3. Feed the cultures after 7 d by replacing half of the supernatant with
fresh medium. Thereafter, transfer the cultures to an incubator set at
33OC.
4. Continue feeding the cultures weekly in the same manner (Note 7).
5. Assay nonadherent cells present in the supernatant removed during
weekly culture feeding for CFC (CFU-GM, BFU-E, CFU-GEMM) (see
Chapter 28 for method).

4. Notes
1. ThemethodforgeneratingL~Cdescribedinthischapterisbasedon
our modification (6) of the Gartner and Kaplan technique (3). Assess-
ment of hematopoiesis in LTMC with colony assays is generally con-
fined to the investigation of the nonadherent cell population. How-
336 Keating and Toor

ever, hematopoietic progenitors can also be detected and enumerated

from the adherent layer, but this involves sacrificing the culture (5).
The cells to be assayed can be released by disrupting the adherent
layer mechanically after incubation with 0.25% trypsin for 5 min.
After washing in media containing 10% fetal bovine serum, the cells
are assayed for CFC as previously described (seeChapter 28, this vol-
ume).
2. Several factors are particularly important for the generation of suc-
cessful LTMC. Our practice has been to screen for the most suitable
lots of both FBS and horse serum, although in our experience the qual-
ity of the FBS tends to be the more critical factor. Serum batches are
screened by establishing LTMC with normal donor marrow using the
protocol described. A culture generated with medium containing a
suitable batch of serum gives rise to a confluent adherent cell layer as
shown in Fig. 2, with numerous hematopoietic islands and round cells
within 2-3 wk of culture initiation.
3. The method of isolating nucleated marrow cells also influences the
quality of the culture. We have found that the use of the discontinuous
Percoll gradient is more consistent in producing healthy cultures than
other cell isolation methods, such as buffy coat generation or the use
of Ficoll-hypaque gradients.
4. Cultures also vary according to the brand of the plastic culture flasks
employed. We have found that Corning TMplastic flasks permit the
best growth most consistently.
5. The water source for the preparation of culturemedium has long been
known to be an important factor in successful tissue culture. Conse-
quently, when the technique is first established in the laboratory,
water from different sources should be screened in the same manner
as described for serum.
6. A wide selection of tissue culture vessels offers considerable flexi-
bility in the design of studies using the long-term marrow culture sys-
tem. For example, small flat-sided tissue culture flasks (ambitubes)
are particularly useful for generating LTMC from limited cell num-
bers. The culture of a large quantity of marrow as in the case of clinical
in vitro purging (7) can be accomplished in 150 cm2 CorningTM flasks.
Multichamber/glass slide microcultures permit the study of the ad-
herent cell population with immunofluorescent or autoradiographic
techniques.
7. A variation of the long-term marrow culture system (the two-stage
system) involves the use of allogeneic or autologous adherent layers
on which can be superimposed a marrow or peripheral blood-derived
hematopoietic cell population of interest (8). The adherent layers are
Long-Term Human Bone Marrow Culture 337

derived from healthy 34wk-old LTMC generated from normal

marrow and are irradiated (1100 cGy) to prevent proliferation of the
hematopoietic cells within the layer. The overlying hematopoietic
cells can then be assayed weekly for CFC in the usual manner for up
to 6 wk. This two-stage LTMC system can be used as an indirect means
of assessing multipotent stem cell toxicity. For example, the ability of
marrow, depleted of CFC by a particular modality (cytotoxic mono-
clonal antibody, drug, antineoplastic agent), to regenerate CFC (CFU-
GM, BFU-E, CFU-GEMM) after incubation over an irradiated LTMC
adherent layer may serve as a measure of the viability of early precur-
sor cells not measurable in colony forming assays. Treated cells at a
concentration of 2 x lO”cells/mL (assuming no loss as a result of treat-
ment with the particular modality) are placed over the irradiated
layer, incubated at 33”C, and assayed for CFC at weekly intervals for
up to 4-6 wk. Alternatively, assays can be performed every 2 wk de-
pending on the activity of the culture. The two-stage LTMC system
may be particularly suitable for the assessment of hematopoiesis in
marrow destined for autologous transplantation and treated ex vivo
to eliminate neoplastic cells.

References
I. Singer, J. W., Keating, A., and Wright, T. H. (1985) The human hematopoietic micro-
environment, in Advances in Haematofogy, vol. 4 (Hoffbrand, V., ed.) Churchill, Lon-
don, pp. l-24.
2. Dexter, T. M., Allen, T. D., and Lajtha, L. G. (1977) Conditions controlling the pro-
liferation of hematopoietic cells in vitro. J. CeZZ.Physiol. 91,335-344.
3. Gartner, S. and Kaplan, H. S. (1980) Long-term culture of human bone marrow cells.
Proc. Natl. Acad. Sci. USA 77,4756-4759.
4. Powell, J., Keating, A., Singer, J. W., and Adamson, J. W. (1983) Analysis of he-
matopoiesis in human long-term marrow cultures. International Society of Experi-
mental Hematology, London, Exp. Hemutol. II, 6.
5. Coulombel, L.,Eaves, A. C.,and Eaves, C. J. (1983) Enzymatic treatment of long-term
marrow cultures reveals the preferential location of primitive hemopoietic progeni-
tors in the adherent layer. Blood 62,291-297.
6 Keating, A., Powell, J., Takahashi, M., and Singer, J. W. (1984) The generation of hu-
man long-term marrow cultures from marrow depleted of Ia (HLA-DR) positive
cells. Blood 64,1159-l 162.
7. Chang, J., Morgenstern, G., Deakin, D., Testa, N. G., Coutinho, L., Scarffe, J. H.,
Harrison, C., and Dexter, T. M. (1986) Reconstitution of haemopoietic system with
autologousmarrow takenduringrelapseofacutemyeloblasticleukemiaandgrown
in long-term culture. hncet 1, 294,295.
8. Takahashi, M., Keating, A., and Singer, J. W. (1985) A functional defect in irradiated
adherent layer from chronic myelogenous leukemia long-term marrow cultures.
Exp. Hematol. 13,926931.
 Chapter 28

In Vitro Clonal Culture

of Human Hematopoietic
Progenitor Cells

Armand Keating and Paul Toor

1. Introduction
Human hematopoietic tissue represents a complex developmental
system that is closely regulated to ensure the maintenance of appropriate
levels of circulating blood cells. At least nine distinct lineages of cells at
various stages of maturation have been described (1). Substantial progress
in the study of hematopoietic stem cell ontogeny has resulted from the
development of semisolid culture methods for the detection of committed
progenitor cells. The assays define specific individual precursor cells in-
directly by the ability of these cells to form colonies consisting of clonally
expanded progeny at varying stages of differentiation. In turn, these
methods have led to the identification, characterization, and synthesis by
recombinant DNA techniques of some of the growth factors capable of
stimulating colony formation and infer to an in vivo regulatory role for
such factors in hematopoiesis.
Assays employing different semisolid supports have been established
for the clonal culture of progenitor cells committed to granulocyte- macro-
phage (CF’U-GM) (Z), erythroid (BF’U-E, CFU-E) (3), megakaryocytic (4),
and multipotential (CF’LJ-GEMM) (5) lineages.

339
340 Keating and Toor

In this chapter, we describe a single method that conveniently allows

the detection and enumeration of progenitor cells (CFC or colony forming
cells) capable of colony formation. Thus, in a single dish, CFC can be de-
tected with multlineage potential (CEU-GEMM), as well as progenitors
committed to megakaryocytic (CFU-Meg), erythroid (BFU-E), and granu-
locyte-macrophage (CFU-GM) pathways. With this technique, the fre-
quency of all colony types is similar to that obtained in theindividual clonal
assays.

2. Materials
1. Erythropoietin (Step 1, lOOOU/mg: Terry FoxLaboratory, Vancouver,
British Columbia).
2. Fetal bovine serum.
3. Ficoll-Paque (Density 1.077 g/mL).
4. Iscove’s Modified Dulbecco’s Medium (IMDM).
5. “Lux” Petri dishes.
4. Methylcellulose powder (density, 4000 cl?).
7. Percoll (density, 1.077 g/mL).
8. Phytohemagglutinin.

3. Methods
The procedure is outlined diagrammatically in Fig. 1 (Note 1).

3.1. Isolation of the Cell Population

Best results are obtained if the marrow or peripheral blood specimen
is used within 2-3 h after collection. If this is not possible, the specimen,
anticoagulated with preservative-free heparin, may be left overnight at
room temperature for use in the morning.
1. Draw 3-5 mL bone marrow or 5-10 mL peripheral blood into a syringe
containing 500 U preservative-free heparin.
2. Dilute 1:l with Iscove’s Modified Dulbecco’s Medium (IMDM).
3. Isolate the mononuclear cells by density centrifugation over Percoll.
Collect the interface mononuclear cells after centrifugation at 4008 for
30 min, wash three times in media, count, and resuspend in IMDM.
Human Hematopoietic Progenitor Cells

Bone marrow cells (Collected m preservative-free hepann, IO units/ml,

and diluted 1:l with IMDM)

Density oentrlfugation (Percoil: 1077 g/mL,

I 30 minutes, 4OOg)

Mononuclear ceils (Final concentration 2 x 105/mL mixture)

Mixture
0.9% methyl cellulose
30% human plasma
5 x 1g5M 2-mercaptoethanoi
95 mm x 12 mm test-tube 5% leukocyte-conditioned medium
2 umts erythropoietin/mL
cells in IMDM to volume

Vortex 15 seconds and plate

35 mm petri-dash
,-*I----..
(
-‘-------:
.-.- -..*
l.

#I

Q
, --.- . .* ’
.-

14 day mcubation at 37°C In a

i 5% CO2 humidified atmosphere
. ..---..
00-a.--..‘;,
/.-
c ,.- *:-. 7
,’ .--def - -;*
-we .s- _.- .
- ---._-
Q
Observe erythroid, granulocyte-macrophage,
megakaryocyte and mixed ceil colonies under
inverted microscppe x 37

Fig. 1. Scheme for CFC assay.

342 Keating and Toor

3.2. Preparation of Leukocyte-Conditioned Medium

Although recombinant growth factors are becoming commercially
available, leukocyte-conditioned medium (LCM) containing hematopoi-
etic stimulating factors is readily prepared (6,7) and may be used in the
CFC assay.
1. Collect 50 mL blood samples in 500 U preservative-free heparin from
each of three normal donors.
2. Dilute the blood 1:l with IMDM.
3. Isolate the lymphocytes by density centrifugation over Ficoll-Paque.
Collect these cells from the interface after centrifugation at 4008 for 30
min, wash three times with media, count, and resuspend in IMDM.
The lymphocytes from the three blood samples can then be pooled.
4. Resuspend the cells at a concentration of 1 x 106/mL in IMDM with
10% fetal bovine serum. Add reconstituted phytohemagglutinin
stock solution to a concentration of 1% (v/v>.
5. Incubate 100 mL vol in tissue culture flasks for 7 d at 37OCin a humid-
ified atmosphere containing 5% CO,.
6. Harvest the superna tant, filter-sterilize, and store at 4OCfor use with-
in 4 wk, or at -7OOC for use within a year.
3.3. Preparation of Methylcellulose (2.1%)
1. Place 42 g of methylcellulose powder and a stir bar in a 2 L Erlenmeyer
flask and autoclave, using the dry cycle.
2. Add 979 mL sterile distilled water to the flask and bring to boil. Let the
water boil gently for 5-10 min, or until all the clumps of methylcellu-
lose are broken up.
3. Place the flask on a magnetic stirrer for 3 h until cooled to room tem-
perature.
4. Add 979 mL sterile double-strength IMDM, and stir for 1 h at room
temperature.
5. Continue stirring overnight at 4°C.
6. Aliquot the methylcellulose in 100 mL vol, and store at 4°C for use
within 4 wk, or at -70°C, for use within 1 yr.

3.4. Collection of Human PZasma (Note 2)

1. Draw 50 mLor blood from a consenting adult donor into a syringe con-
taining 500 U preservative-free heparin.
2. Centrifuge the plasma at 2008 for 10 min to remove the plasma.
3. Centrifuge the plasma at 4008 for 5 min to remove the platelets.
Human Hematopoietic Progenitor Cells 343

4. Filter-sterilize the plasma, aliquot it, and freeze at -7OOCfor up to 6 mo.

5. Prior to use, thaw out the plasma, and filter-sterilize it through a 0.2
l..trnfilter to remove cryoprecipitates.

3.5. Assay for CFC

The culture conditions employed for the detection of CFU-GM, BFU-
E, CFU-Meg, and CFU-GEMM are based on a modification of the proce-
dure originally described by Fauser and Messner (5).
1. Suspend the mononuclear cells (~10~ cells/mL final concentration:
Note 3) in IMDM in a tube containing methylcellulose as viscous sup-
port (0.9% final concn,), 30% human plasma, 5% LCM, 5 x 10-W 2-mer-
captoethanol, and 2U/mL humanurinary erythropoetin. Themethyl-
cellulose can be handled easily with a syringe and 1Bgage needle.
2. Vortex the tube gently for 10 s to ensure thorough mixing.
3. Plate 1 mL amounts of the mixture into 35 mm Petri dishes. A mini-
mum of three, but preferably six, dishes should be plated.
4. Incubate the dishes up to 16 d at 37OC in a humidified incubator con-
taining 5% CO, (Note 4).
5. Examine and enumerate the colonies at d 14 using an inverted micro-
scope (Figs. 2,3; Note 5).

4. Notes

1. The CFC assay described above offers several advantages over indi-
vidual colony assays. The most obvious is convenience. A major ad-
vantage resides in the use of methylcellulose as the semisolid support.
Methylcellulose is easy to handle and, unlike agar, will not harden if
a delay occurs during plating. Moreover, colonies can be readily
plucked from methylcellulose plates and the material removed with-
out difficulty, thus permitting cytologic and cytogenetic assessment
of the cells within colonies. Also, cell transfer and replating experi-
ments can be more easily performed. Finally, multilineage colonies
appear to grow better in methylcellulose.
2. Fetal bovine serum can be used instead of human plasma in the assay.
However, inclusion of human plasma results in superior erythroid
and, in particular, megakaryocyte colony growth. In our experience,
few batches of fetal bovine serum facilitate the assay of CFU-Meg. The
presence of human plasma in the assay tends to cause the assay mix-
ture to “gel.” Consequently, colony morphology differs; the colonies
344 Keating and Tow

Fig. 2. Light micrograph (x125) of mixed cell colony (CF’U-GEMM). The colony con-
tains erythroid cells, granulocytes, megakaryocytes, and macrophages.

Fig. 3. Light micrograph (x80) of erythroid colony (BFU-E, lower left) and a macro-
phage colony (CFU-M, note dispersed monocytes-macrophages, upper right).
Human Hematopoietic Progenitor Cells 345

tend to be “tighter” and the cells within them less dispersed, particu-
larly in the case of BF’U-E. The major drawback in the use of human
plasma is the need for extensive screening to determine specimens
giving optimal colony growth. Indeed, some specimens inhibit col-
ony formation, but this appears to be unrelated to the blood group of
the plasma donor. The most suitable plasma is obtained from fasting
donors and often a pooled batch from two or more donors gives the
best results.
3. Marrow or peripheral blood cells can be depleted of macrophages in
order to enrich the target population for CFC. Eliminating macro-
phages in the cells to be plated also reduces the background single cell
population and facilitates enumeration of cluster (5-39 cells) as well as
colonies (more than 39 cells). Adherent cells are removed by incubat-
ing the mononuclear cell fraction in IMDM with 10% fetal bovine
serum for 2 h at 37OCin polystyrene culture dishes. The macrophage-
depleted cells then can be obtained by gentle pipeting and rinsing of
the dishes with medium.
4. A common technical pitfall in the CFC assay is to provide inadequate
humidity during incubation, leading to drying of the methylcellulose
and poor colony formation. This can be readily prevented by arrang-
ing, within a large covered transparent Petri dish, the assay dishes (35
mm Petri dishes) around a central open assay dish filled with sterile
distilled water.
5. Colonies can be screened with an inverted microscope through the
large transparent dish containing themethylcellulose dishes from d 10
onward without unduly affecting colony development. When colony
formation is judged optimal, the dishes are removed and the colonies
assessed and enumerated at magnification x 37. Figure 2 shows a
mixed cell colony (CFU-GEMM), whereas Fig. 3 contains a hemoglo-
binized erythroid colony (BFU-E) and a loose macrophage colony
(CFU-M).
References
2. Metcalf, D. (1985) The granulocyte-macrophage colony-stimulating factors. Science
229,x5-22.
2. Pike, B. L. and Robinson, W. A. (1984) Human bone marrow colony growth in agar
gel. J. Cell. Physiol. 76,77-84.
3. Tepperman, A. D., Curtis, J. E., and McCulloch, E. A. (1974) Erythropoietic colonies
in cultures of human marrow. Blood 44,659-669.
4. Vainchenker, V., Bouguet, J., Guichard, J., and Breton-Gorius, J. (1979) Megakaryo-
cyte colony formation from human bone marrow precursors. Blood 54,940-945.
 Keating and Toor

5. Fauser, A. A. and Messner, I-I. A. (1978) Granuloerythropoietic colonies in human

marrow, peripheral blood and cord blood. Blood 52,1243-1248.
6. Aye, M. T., Niho, Y., Till, J. E., and McCulloch, E. A. (1974) Studies of leukemic cell
populations in culture. Blood 44,205-219.
7. Parker, J. W. and Metcalf, D. (1974) Production of colony-stimulating factor in mito-
gen-stimulted lymphocyte cultures. J. ImmunoI. 112,502-510.
 Chapter 29

Flow Sorting
for Isolating CFU-E

Suzanne M. Wat-t
and John M. Davis
1. Introduction
Erythroid progenitor cells have been classified into three groups of in-
creasing maturity: the primitive burst forming unit (p-BF’U-E), the mature
burst forming unit (m-BFU-E), and the erythropoietin responsive colony
forming unit (CFU-E). This classification is based on their time of matur-
ation in vitro, their proliferative capacity, and their responsiveness to
growth factors (I). The CFU-E can be distinguished from the more primi-
tive erythroid progenitors by their ability to proliferate and mature in re-
sponse to a single growth factor, erythropoietin. In clonal assays in vitro,
the CFU-E form single or double clusters characteristically containing 8-64
mature or maturing erythroid cells 2 d after cultures have been initiated
with mouse bone marrow or fetal liver (2,3).
The availability of enriched populations of CFLJ-E is important to our
understanding of the function and mode of action of growth factor recep-
tors in normal cells, the influence and regulation of molecules, genes and
viruses that are specific to the erythroid lineage, and as a baseline for
understanding errors in gene regulation or function that govern the devel-
opment of leukemic or preleukemic states. One approach to purifying
CFU-E has relied on the use of fluorescently tagged monoclonal antibodies
as probes to cell surface molecules together with flow cytometry (4,5). Al-

347
348 Watt and Davis

though probes that specifically identify CFU-E are not available, pheno-
typic analysis has revealed that CFU-E can be segregated from moreprimi-
tive erythroid precursors, from morphologically recognizable erythroid
cells, and from mature myelomonocytic cells and their progenitors with a
series of rat monoclonal antibodies (4-6). Hemopoietic tissues vary in their
content of different types of hemopoietic progenitor cells and of maturing
cells of particular lineages (6). Since mouse fetal liver is a major site of ery-
thropoiesis and contains high numbers of CFU-E, the strategy for isolating
CFU-E described here relies on the fractionation of low density fetal liver
cells on the basis of their forward light scatter characteristics and differen-
tial reactivity with two rat monoclonal antibodies, YBM 42.2.2 and YBM
10.14.9 using flow cytometry (4-6). The anti-T200 antibody YBM 42.2.2
does not react with the CFU-E or more mature erythroid cells (4) but reacts
with all leukocytes, thus allowing segregation of CFU-E from both myelo-
monocytic and lymphoid cells and from all other hemopoietic precursors.
The YBM 10.14.9 antibody is then used to separate CFU-E from more ma-
ture erythroid cells (5). This approach yields cell populations containing
at least 60% CPU-E, whereas 80% of the cells have the morphology of early
erythroid blast cells and do not stain with benzidine, which identifies
hemoglobin containing cells (5-7; seeNote 1).

2. Materials
2.1. Reagents for Media Preparation
1. Powdered Iscove’s Modified Dulbecco’s Medium (IMDM) contain-
ing (per L) 3.024 g NaHCO, 60 mg penicillin, 100 mg streptomycin,
5 x 105M 2-mercaptoethanol. After preparation, do not adjust pH.
Store at 4OC. Light sensitive. This should be prepared at lx and 2x
strength.
2. IMDM without bicarbonate but with all other additions. Adjust to pH
7.3 with 5M NaOH or pH 5.1 with HCl.
3. Powdered modified Eagle’s minimal essential medium with Earle’s
salts, but without phenol red containing 20 g/L BSA and sodium azide
(0.02%).
4. Sodium bicarbonate (NaHCO,).
5. Sodium benzylpenicillin.
6. Streptomycin sulphate (745 U/mg).
7. 2-Mercaptoethanol.
8. IM HEPES buffer pH 7.3 (commercially available).
CFU-E Isolation 349

9. 10x PBS: 0.2M sodium phosphate buffer with 1.48M sodium chloride
pH 7.3.
10. 4% (w/v> sodium azide.

2.2. In Vitro Culture Reagents

1. Methylcellulose (65 HG, 4000 ml?a, Fluka, AG, Buchs, Switzerland)
prescreened to support CFU-E growth as described below and in
Note 2.
2. Bovine serum albumin (BSA). Store at 4OC.
3. Lipids
a. Cholesterol (5-cholesterol-3-B-01).
b. Oleic acid (cis-9-octadecanoic acid).
c. L-a-phosphatidyl choline dipalmitoyl.
4. Human transferrin.
5. Erythropoietin (Epo). Epo step 1 (lOOOU/mg) or pure Epo (8O,OOOU/
mg) obtained from Terry Fox Laboratory Media Preparation Service,
Vancouver, Canada or commercial recombinant Epo. Store in 200~PL
aliquots at -70°C in 0.1% BSA or prescreened FCS. Do not refreeze
after thawing.
6. Commercial Ficoll-Hypaque (density 1.077 g/cc). Light sensitive.
Store at room temperature.
7. Fetal Calf Serum (FCS). Prescreened for supporting CFU-E growth.
Store in lo-mL aliquots at -2OOC.

2.3. AnimaZs
Day 12-13 pregnant CBAf/CaH mice 8-12 wk of age. Day 0 of preg-
nancy is taken as the day of appearance of the vaginal plug.

2.4. Hybridomas
1. YBM 42.2.2 (Rat IgG2a antibody). This does not bind protein A at
neutral PH.
2. YBM 10.14.9 (Rat IgG2c antibody). This antibody binds protein A at
neutral PH.

2.5. Antibody Purification

1. Commercial rabbit anti-rat Ig.
2. Commerical Protein A-Sepharose CL-4B.
350 Watt and Davis

3. O.lM phosphate buffer pH 8.

4. O.lM glycine-HCl buffer pH 3.
5. 2M Trizma base in water.
6. PBS: 0.02M phosphate buffer with 0.148M sodium chloride pH 7.3.
7. Saturated ammonium sulfate solution pH 6.8.
8. O.lM borate buffer pH 8.2.
9. 0.2M triethanolamine pH 8.2.
10. 20 mM dimethylpimelimidate dihydrochloride in 0.2M triethanol-
amine pH 8.2.
2.6. Antibody Labeling
1. Commercial fluorescein isothiocyanate (FITC) coupled protein A.
Store at 4OC.
2. Fluorescein isothiocyanate (Isomer 1). Store dessicated at 4OC.
2.7. Stains
1. Commercial May-Grunwald Stain prefiltered through a Whatman
1MM filter paper.
2. Commercial Giemsa R66 improved stain.
3. Benzidine stock solution: 0.2% (w/v) benzidine hydrochloride in
0.5M acetic acid. This can be stored in the dark at 4OCfor 34 wk. Cau-
tion, benzidine is a carcinogen.
4. 30% (w/w) 1$02 solution.
2.8. In Vitro Culture Reagents
2.8.1. 2% Methylcellulose
1. Add 20 g methylcellulose to 500 g sterile glass freshly double-distilled
deionized boiling water.
2. Boil for 2 min with great care to avoid excess foaming. Control the
level of heating.
3. Add sterile double-distilled deionized water to 520 g to correct for
water loss resulting from evaporation.
4. Cool to approximately 37°C.
5. Add 500 mL of double-strength IMDM. Keep covered with foil.
6. Cool on ice with mixing for 2-3 h.
7. Stir on a magnetic stirrer at 4°C overnight to allow the methylcel-
lulose solution to clear.
8. Store in 50-mL aliquots protected from light at -2OOC for up to 4 wk.
CFU-E Isolation 351

2.8.2. Deionized and Delipidated BSA

1. Dissolve400 mg of Dextran T40 in 400 mL of glass double-distilled de-
ionized water.
2. Add 4 g of activated charcoal (Norit A or SX-1) and leave at room tem-
perature for 30 min with occasional mixing.
3. Add 20 g BSA to the surface of the dextran-coated charcoal solution
and leave for 2-3 h at 4OCto dissolve without mixing.
4. Titrate to pH 3.0 with concentrated hydrochloric acid to inhibit heat-
induced polymerization of the albumin.
5. Incubate for 30 min at 56OCin a shaking water bath.
6. Centrifuge at 10,000 rpm for 20 min and Millipore filter the super-
natant.
7. Adjust the pH to 5.5 with 2M NaOH.
8. Deionize the BSA solution overnight at 4OCover 40 cm3 of Amberlite
MB-l mixed in exchange resin.
9. Concentrate the solution to 150 mLon an Amicon UMlO membrane at
4OC.
10. Adjust the pH to 7 with 2M hydrochloric acid.
11. Millipore filter and store in lo-mL aliquots indefinitely at -2OOCor
4OC.
2.8.3. Lipids
1. Dissolve 4 mg L-a-phosphatidyl choline dipalmitoyl, 3.9 mg cho-
lesterol, and 2.8 mg oleic acid in a few drops of chloroform at room
temperature or ethanol at 50°C in a 25-mL glass beaker. Completely
evaporate the solvent under a stream of nitrogen, leaving the film of
lipid on the bottom of the beaker.
2. Add 10 mL of bicarbonate-free IMDM pH 5.1 containing 1% of the
deionized and delipidated BSA.
3. Immerse the beaker containing the lipids in ice and sonicate under air
for 10 min at maximumenergy just below the foaming point so that the
lipids form small micelles.
4. Millipore filter through 1.2 pm and then 0.45 pm filters. Store in-
definitely at 4OC.
2.8.4. Other Reagents
1. 5 x 10-2M 2-mercaptoethanol in double-glass distilled deionized water.
Millipore filter. Prepare fresh stocks before use.
2. Ferric chloride stock (FeCS). Since ferric chloride is hygroscopic,
352 Watt and Davis

weigh out a large piece of FeC1,*6%0 and immediately dissolve in

103M hydrochloric acid (HCl). Dilute to a 7.9 x lO3M stock solution
in 1O3M HCl. Store at -2OOC.
3. Transferrin: Dissolve 360 mg of human transferrin in 4 mL of bi-
carbonate-free IMDM pH 7.4 and 1.15 mL of 7.9 x lO”M FeCJ in 103M
HCl. Millipore filter and store indefinitely at 4OC.

3. Methods
3.1. Antibody Preparation
1. Collect the supernatant from the hybridoma cell lines grown in IMDM
with 1% FCS, 24 h after cells reach confluency.
2. Precipitate the immunoglobulin (Ig) by adding an equal volume of
saturated ammonium sulfate solution (pH 6.8). Mix for 1 h at 4°C.
3. Centrifuge at 10,OOOgfor 10 min at 4OC.
4. Resuspend the pellet to one-tenth the original volume in PBS and
dialyze against O.lM phosphate buffer pH 8.0 at 4OC.
5. Estimate the protein concentration by measuring the absorbance at
280 nm.

3.2. Antibody Purification

1. Mix purified rabbit antirat Ig with Protein A-Sepharose to a final con-
centration of 11 mg Ig/mL of Sepharose beads in O.lM borate buffer
pH 8.2 for 30 min at room temperature.
2. Wash the beads in excess borate buffer and then in 0.2M triethanol-
amine pH 8.2.
3. Resuspend the Sepharose in 20 vol of 20 mM dimethyl pimelimidate
dihydrochloride freshly made in 0.2M triethanolamine pH 8.2. Mix
for 45 min at room temperature. This will covalently crosslink the Ig
to the protein-A and prevent it from leaching from the column.
4. Spin the beads at 5008 for 1 min and resuspend in an equal volume of
20 mM ethanolamine pH 8.2 for 5 min at room temperature.
5. Wash the beads three times in O.lM borate buffer pH 8.2.
6. At the same time, equilibrate Protein A-Sepharose with O.lM borate
buffer pH 8.2. YBM 10.14.9 will bind to Protein A-Sepharose at neu-
tral pH, whereas YBM 42.2.2 will not.
7. Apply the concentrated YBM 10.14.9 sample to a Protein A-Sepha-
rose column and the YBM 42.2.2 to the rabbit anti-rat Ig-Protein A
Sepharose column; lo-20 mg of Ig can be applied/mL of beads.
CFU-E Isolation 353

8. Elute the bound Ig with O.lM glycine-HCI buffer pH 3, and neutralize

the eluted material immediately with Trizma base.
9. Dialyze the antibodies against PBS and store in small aliquots with
0.1% BSA.
10, If the purified antibodies are to be coupled to fluorescein isothiocya-
nate (FITC), dialyze against O.lM bicarbonate buffer pH 9.3 instead of
the PBS, and couple with FITC immediately. Do not store antibodies
for an extended period of time in the bicarbonate buffer.

3.3. Fluorescein Labeling of Antibodies

1. Dialyze the antibodies against O.lM bicarbonate buffer pH 9.3 for 2-5
h at 4OC.
2. Dissolve thefluoresceinisothiocyanateisomer 1 (FlTC) at 1 mg/mLin
DMSO. Add 25 pg of FITC/mg of purified YBM 42.2.2 Ig for 2 h at
room temperature with constant rotation to give a fluorescein to pro-
tein ratio of approximately 3:l.
3. Separate the FITC conjugated Ig from the free FITC by passing
through a 5-mL Sephadex G-25 column equilibrated with PBS. Collect
the Ig fraction and store in small aliquots containing 1% BSA and
0.02% NaN3 at -2OOC.

3.4. Cell Preparation and Labeling

1. Dissect the livers from d 12-13 mouse fetuses using cataract knives
and place in chilled bicarbonate-free IMDM containing 10% FCS.
2. Prepare a single-cell suspension by gently syringing the fetal livers
through 19-, 21-, and 25-gage needles sequentially attached to a 2-mL
syringe.
3. Place the cells in a IO-mL centrifuge tube and allow cell clumps to set-
tle at 4OCfor 5 min. Pass the supernatant through sterile nylon gauze
to remove smaller clumps.
4. Centrifuge 5 mL of cells (10’ cells/ml) over 4 mL of Ficoll-Hypaque
(density 1.077 g/mL) at 1600 rpm for 30 min at room temperature.
5. Collect the low density cells from the Ficoll-Hypaque interface and
wash three times in Eagle’s-HEPES medium with BSA and sodium
azide. Approximately lo5 cells are recovered/fetal liver processed.
6. Add normal mouse serum, heat inactivated at 56°C for 30 min to the
cell suspension at a final concentration of 10% to block Fc receptors.
Incubate the cells for 20 min at room temperature, and wash the cells
in Eagle’s-HEPES medium containing BSA and sodium azide.
354 Watt and Davis

7. Centrifuge the antibodies in a Beckman airfuge at 26 lb/i$for 10 min

at room temperature to remove aggregrates and further minimize Fc
binding.
8. Label cells (lO’/mL) with saturating levels of FITC-tagged YBM 42.2.2
(approximately 400 pg/mL) at 4°C for 30 min. Include propidium
iodide (50 pg/mL) during the incubation.
9. Wash and resuspend the cells to 2 x lo6 cells/mL in Eagle’s-HEPES
medium containing BSA and sodium azide at 4OC.

3.5. Cell Sorting (See Note 1)

1. Separate the labeled cells on a fluorescence-activated cell sorter by the
two parameters of forward light scatter and fluorescein fluorescence.
2. For a FACS-II cell sorter, cellular excitation is achieved with an argon
ion laser at an output power of 0.3 W and an emission wavelength of
488 nm. Set the light scatter gain at approximately 4. For fluorescein
fluorescence, set the photomultiplier voltage at 650 V with an ampli-
fier gain of 8-16. These voltages and gains will vary with the instru-
ment used. The fluorescein fluorescence is detected by placing a 530-
nm long pass interference filter and a 530-nm long pass filter in front
of the appropriate photomultiplier tube.
3. Sterilize the tubing and nozzle by passing ethanol through the cell
sorter for 30 min. Wash out the ethanol with sterile distilled water, and
pass sterile distilled water through the tubing and nozzle for at least
1 h, following this with a 0.9% saline wash for 30 min prior to sorting.
4. Run the cells at 4°C through the cell sorter, collecting cells with inter-
mediate to high forward light scatter characteristics that are negative
for YBM 42.2.2 labeling when compared to control cells labeled with
an irrelevant FITC-tagged rat monoclonal antibody of the same sub-
class. SeeFig. 1A and 1B.
5, Collect the sorted cells in bicarbonate-free IMDM pH 7.3 containing
20-50% FCS at 4OC in earthed siliconized glass tubes. Note: CFU-E
will die if the collection medium becomes alkaline.
6. Spin the sorted cells at 800 rpm for 15 min at 4OC.
7. Resuspend the cells to 2 x lo6 cells/ml in Eagle’s-HEPES medium
with BSA and sodium azide.
8. Label with FITC-tagged YBM 10.14.9 or with unlabeled YBM 10.14.9
(using 400 pg/107 cells) for 30 min at 4OC.
9. Wash the cells twice with Eagle’s-HEPES medium with BSA and so-
dium azide. Resuspend to 2 x lo6 cells/ml in the same medium.
CFU-E Isolation 355

1sT SELECTION

b
FORWARD LIGHT SCATTER

SORTED B
0. e

2ND SELECTION
Lymphold, myelold FITC-YBM42 2 2 NEGATIVE
& early progen,tor cells

- SORTED -
C

Erythrold
cells CFU-E 3RD SELECTION
FITC-YBMlO 14 9 POSITIVE

0 100 256
b
FLUORESCENCE INTENSITY

Fig. 1. Typical histograms showing the regions selected for isolating CFU-E. The fetal
liver cells are first selected on the basis of their intermediate to high forward light scatter
characteristics (A) and the YBM 42.2.2 negative cells (B) are sorted. The isolated cells are
labeled with FITC-tagged YBM 10.14.9, and the positive cells (0 are collected. Care is
taken to exclude the highly positive cells that are stained with propidium iodide and rep-
resent the nonviable cell fraction.

10. When using the unlabeled YBM 10.14.9, a two-stage indirect labeling
procedure is necessary. For this, incubate the cells with FITC-Protein
A (20 pg/mL final concentration) for 30 min at 4OCprior to washing,
and resuspension in Eagle’s-HEPES medium with BSA and sodium
356 Watt and Davis

azide. Include propidium iodide (50 pg/mL) in the final incubation

step. Dead cells labeled with propidium iodide give a very high fluo-
rescence signal in the fluorescein channel and can be gated out when
only two parameters are available for sorting.
11. Sort the labeled cells selecting the YBM 10.14.9 positive cells (avoid-
ing the very highly propidium iodide labeled nonviable cells) using
conditions described for the first sort and shown in Fig. 1C.
12. Centrifuge the sorted cells at 800 rpm for 10 min at 4OC. Resuspend in
bicarbonate-free IMDM containing 10% FCS at 105cells/mL. Keep at
4OCuntil cultured.

3.6. CFU-E Culture (See Note 2)

1. Thaw the 2% methylcellulose stock at room temperature.
2. Set up triplicate cultures in 35-mm plastic Petri dishes in a final vol of
1 mL containing: 0.8% (w/v) methylcellulose, 10% (w/v) FCS, 1%
(w/v) deionized and delipidated BSA, 0.3 mg transferrin saturated
with FeCS, 5 xlOsM 2-mercaptoethanol, 0.05 U erythropoietin, 8 PgL-
a-phosphatidylcholine dipalmitoyl, 7.8 pg cholesterol, 5.6 pg oleic
acid, and lo*-lo3 sorted fetal liver cells or 104unsorted fetal liver cells
all in single-strength IMDM pH 7.3.
3. Incubate the cultures at 37OC in a humidified incubator gassed with
5% CO, in air for 2 d.
4. Score single- or double-cell clusters containing 8-64 mature or matur-
ing erythroid cells using an inverted microscope with a loo-fold mag-
nification.

3.7. Morphology
1. Cytocentrifuge sorted cells and air dry.
2. Fix the cells with methanol for 10 min at room temperature.
3. Incubate in May-Grunwald stain for 20 min at room temperature and
then in 3% Giemsa in tap water for 20 min at room temperature.
4. Wash the slides in tap water. Air dry and mount the coverslips with
DPX.

3.8. Ben&dine Staining

1. Suspend the cells in tissue culture medium (without NaN3) to a con-
centration of 1.5 x lo6 cells/mL.
CFU-E Isolation 357

2. Pipet 150 PL of the cell suspension into the well of a 96-well-flat bot-
tom microtiter plate.
3. Prepare the staining solution; add 10 PL of 30% I$OZ to 1 mL of benzi-
dine stock solution. Mix and use within 30 min.
4. Add 50 PL of staining solution to cells, and mix quickly by pipeting up
and down several times.
5. Wait 5 min. At this time, cells containing at least 10% hemoglobin will
have stained a dark blue. Estimation of the number of hemoglobin
containing cells is best achieved by photographing the cells at this
stage, since both the color of the stain and the number of stained cells
will change with time.
6. The early erythroid blast cells represent the most immature and
largest erythropoietic precursors recognizable, containing a promin-
ent nucleolus, basophilic cytoplasm, and loose chromatin pattern, and
are negative for benzidine staining. More mature erythroid cells con-
taining hemoglobin will stain with benzidine.

4. Notes
1. The method describes the isolation of CFU-E using sequential sorting
with two monoclonal antibodies, YBM 42.2.2 and YBM 10.14.9. Rel-
atively high recoveries of CFU-E (40%) can be achieved in this way.
These studies could be done equally well using a single multiparam-
eter sort with antiisotypereagents labeled with fluorescein and phyco-
erythrin or directly conjugated reagents. Substantial enrichment for
CFU-E from both normal fetal liver and bone marrow can also be
achieved with a set of monoclonal antibodies listed in reference 7. In-
deed, more efficient purification may be obtained for CF’U-E by com-
bining three probes, such as YBM 10.14.9, YBM42.2.2, and YW 13.1.1,
since all these antibodies exhibit different patterns of reactivity with
normal hemopoietic cells. Studies using simultaneous two- and three-
color sorting with a variety of antibodies to human erythroid precur-
sors show the potential benefit of such approaches to cell fractionation
(8). Other procedures that allow substantial enrichment for CFU-E
include multiparameter sorting using fluorescein-conjugated poke-
weed mitogen and rhodamine-conjugated antineutrophil/monocyte
antibodies (9). In addition, Nijhof and Wierenga (10) have obtained
sufficient numbers of highly purified CFU-E for biochemical analysis
358 Watt and Davis

in a relatively short time by using density separation and elutriation

of spleen cells from thiamphenol-treated mice.
2. The isolated cells are analyzed for CFU-E by their growth in meth-
ylcellulose (3). Details of methylcellulose preparation are also given
in reference Il. Fetal calf serum can be omitted from the cultures, since
the BSA, lipid, transferrin, and erythropoietin additives have been de-
signed to allow CFU-E growth in serum-free conditions (3). The ap-
proximate concentrations of each additive are described, but it is
essential to test and titrateeach additive in order to obtain the best con-
ditions for CPU-E growth.
3. Details of antibody purification are given in reference 12. The ben-
zidine staining technique described here was adapted from reference
13. Single cells from the purified CFU-E can also be sorted directly into
150 PL of the methylcellulose-supplemented culture medium in a
microtiter tray. The colonies are allowed to develop for 2 d at 37OC,
and the maturing erythroid cells can be stained after cellulase diges-
tion of the methylcellulose (14). For this, 75 PL of FCS containing 1.08
mg/mL cellulase (1943 cellulase U/g> is added to each well, and the
cultures are incubated overnight at 37OC. The following day, the con-
tents of each well are transferred as drops to a glass microscope slide
and allowed to air dry. The cells are fixed and stained with May-
Grunwald/Giemsa stain. Alternatively, the cells may be stained in
situ with benzidine with or without digestion of the methylcellulose
with cellulase.

References
I. Gregory, C. J. (1976) Erythropoietin sensitivity as a differentiation marker in the
hemopoietic system: studiesof three erythropoietic colony responsesin culture. 1.
Cell Physiol. 89,289-301.
2. Eaves, A. C. and Eaves, C. J. (1984) Erythropoiesisin culture, in Clinics in Hematology
(McCulloch, E. A., ed.), WB Saunders, London, pp. 371-392.
3. Iscove, N. N., Guilbert L. J., and Weyman, C. (1980) Complete replacement of serum
in primary cultures of erythropoietin-dependent red cell precursors (CFU-E) by
albumin, transferrin, iron, unsaturated fatty acids, lecithin and cholesterol. Exp. Cell
Res. 126,121-126.
4. Watt, S. M., Gilmore, D. J., Metcalf, D., Cobbold, S. J., Hoang, T. K., and Waldmann,
H. (1983a) Segregation of mouse hemopoietic progenitor cells using the mono-
clonal antibody YBM/42. J. Cell Physiol. 115,37-45.
5. Watt, S. M., Metcalf, D., Gilmore, D. J., Stenning, G. M., Clark, M. R., and Wald-
mann, H. (198313) Selective isolation of murine erythropoietin-responsive progen-
itor cells (CFU-E) with monoclonal antibodies. Mo2. Biol. Med. 1,95-115.
CFU-E Isolation 359

6. Watt, S.M., Gilmore, D. J.,Clark, M. R., Davis, J. M., Swirsky, D. M., and Waldmann,
H. (1984) Hemopoietic progenitor cell heterogeneity revealed by a single mono-
clonal antibody YW13.1.1. Mol. Biol. Med. 5351-368.
7. Watt, S. M., Gilmore, D. J., Davis, J. M., Clark, M. R., and Waldmann, H. (1987) Cell
surface markers on hemopoietic precursors. Reagents for the isolation and analysis
of progenitor cell subpopulations. Mol. Cell Probes 1,297-326.
8. Loken, M. R., Shah, V. O., Dattilio, K. L., and Civin, C. I. (1987) Flow cytometry an-
alysisof humanbone marrow: I.Normalerythroiddevelopment. Blood69,255-263.
9. Metcalf, D. and Nicola, N. A. (1984) The regulatory factors controlling murine ery-
thropoiesis in vitro. Proceedings, NIH Conference on Aplastic Anemia, Airlie House
(Young,N.S.,Levine,A.S.,andHumphries,R.K.,eds.),Liss,NewYork,pp.93-105.
20. Nijhof, W. and Wierenga, P. K. (1983) The isolation and characterization of the ery-
throid progenitor cell: CFU-E. J. Cell Biol. 96,386-392.
II. Davis, J. M. (1986) A single step technique for selecting and cloning hybridomas for
monoclonal antibody production. Mefhods in Enzymology 121,307-322.
12. Schneider, C.,Newman,R. A.,Sutherland, D.R., Asser,U., and Greaves,M. F. (1982)
A one step purification of membrane proteins using a high efficiency immunoma-
trix. 1. BioZ. Chem. 257,10766-10769.
23. Orkin, S. H., Haroshi, S.I., and Leder, I’. (1975) Differentiation in erythroleukemic
cells and their somatic hybrids. PYOC.NatZ. Acad. Sci. USA 72,98-102.
24. Shillingstad, R. B. and Ragan, H. A. (1987) Cellulase slide preparation of methylcel-
lulose cultures of hemopoietic cells. BZoodCeZZs12,657-660.
 Chapter 30

Production of Human
and Murine Eosinophils
In Vitro and Assay
for Eosinophil
Differentiation Factors

Malcolm Strath,
Elaine J. Clutterbuck,
and Colin J. Sanderson

1. Introduction
Bone marrow cells from a number of animal species have been used
extensively in liquid and semisolid cultures to study hemopoiesis and to
produce functional mature cells and factor-dependent cell lines (for review
seeref. I). Neutrophils and macrophages are produced without added
growth factors from murine long-term bone marrow cultures (Z), while
lymphoid cells (3,4) and megakaryocytes (reviewed in ref. 5) can be in-
duced under certain conditions.
When bone marrrow cells from mice or humans are established in
tissue culture in the presence of eosinophil differentiation factor (EDF),
mature functional eosinophils are produced and liberated into the non-

361
362 Strath, Clutterbuck, and Sanderson

adherent cell population (6,7). This factor was originally proposed as inter-
leukind (a), but is now accepted as interleukin-5. Other names for IL5 are
eosinophil colony stimulating factor (CSF-Eo), B-cell growth factor type II
(BCGFII), and T-cell replacing factor (TRF).
Eosinophils are produced for only a relatively short time when mar-
row is cultured in the presence of IL5. This suggests that IL5 stimulates the
differentiation of eosinophils from existing progenitor cells present in the
marrow with no recruitment of eosinophil progenitor cells from stem cells.
When the animal has a pronounced eosinophilia, subsequent culture of the
marrow results in enhanced eosinophil production compared to cultures
established from normal marrow. This is interpreted as an increase in
eosinophil progenitors in the animal. We use the Helminth parasite Meso-
c&&es corti Hoeppli 1925 to effect such an increase in eosinophil progeni-
tor numbers in the mouse. This parasite has been used extensively to study
host/parasite interactions and causes a well-documented eosinophilia (9),
with peak numbers of eosinophil progenitors appearing in the bone mar-
row 10 d after infection (10).
The assessment of eosinophil numbers from cultures can be done in
one of two ways: (1) total cell counts (obtained from an electronic particle
counter, e.g. Coulter counter, or by using a hemocytometer) and the per-
centage of eosinophils (obtained from a differential count on Giemsa-
stained smear or cytocentrifuge preparation) are used to calculate the
number of eosinophils, or (2) by assaying for eosinophil peroxidase, which
can be related to eosinophil numbers (12).
A microplate modification of the culture system (12) provides a
method for assaying sources of IL5 since there is as yet no IL5-dependent
cell line as there are for IL2 and IL3. This IL5 assay is based on assessment
of eosinophil numbers by quantifying eosinophil peroxidase in microplate
bone marrow cultures that have been incubated with IL5 samples. This
assay also detects the eosinophil differentiation activity of GM-CSF and
IL3. The B cell activity of mouse IL5 can be measured using the BCL, cell
line (8) though this assay also detects interleukin-4 (B-cell-stimulatory
factor [BSFl]).
2. Materials
1. Parasite: The second stage larvae (tetrathyridia) of the Cyclophylli-
dean Cestode M. corfi are maintained by intraperitoneal passage in
mice, where it reproduces itself vegetatively. The larvae can be stored
in Dulbecco’s Phosphate Buffered Saline “A” (PBS) at 4OCfor several
weeks, and for longer periods if fetal calf serum is added (13). To our
knowledge there have been no recorded cases of human infections
Human and Murine Eosinophils 363

with M. corti, although there have been several cases of infection by

other species of Mesocestoides (24,25). Parasites for reinfecting mice are
harvested from mice that have been infected for several weeks. The
only noticeable visible effect of the infection is an increasing abdom-
inal distension. It may be necessary to occasionally passage the para-
site into another strain of mouse (e.g., CBA) as it seems to lose the
ability to stimulate high eosinophilia after several months passage
through the same strain. For availability of the parasite, refer to “Inter-
national Register of Living Helminth Species and Strains,” published
by WHO; or “Register of Parasitic Protozoa, Helminths, and Arthro-
pods of Medical and Veterinary Importance,” produced by the British
Society for Parasitology.
2. Mice: We routinely use Balb/c.nimr mice, 6-8 wk-old, which are
maintained under SPF conditions and are allowed free accessto food
and water. Any infections the mice contract before the marrow is har-
vested for culture may result in changes in the cell numbers and types
seen in the marrow cultures.
3. Media: The basic medium used is RPM1 1640 purchased in powder
form and reconstituted as recommended by the manufacturer. We
use two basic forms of RPM1 as indicated below:
(1) HEPES Medium (2) Culture Medium
HEPES buffer 20 mM 1OmM
Sodium bicarbonate None 24mM
Glutamine 2mM 2mM
Sodium pyruvate None 1m.M
Penicillin 100 U/mL 100 U/mL
Streptomycin 100 pg/mL 100 pg/mL
Monothioglycerol (Sigma) None 75 w
The above media are stable at 4°C for several months. Glutamine is
unstable, so fresh glutamine is added from a frozen (-20°C) 200 mM
stock solution and the medium used within 2 wk. Glutamine is a gen-
eral requirement in tissue culture medium and has been found to be
necessary for hemopoietic cell differentiation (26).
The media are further supplemented and used as follows:
a. Bench medium: HEPES medium is supplemented with new-
born calf serum to 5% v/v. This is used for cell and tissue
collections and preparation, and is stored at 4°C. It has the
advantage over bicarbonate buffered medium that it does not
change its pH while outside a CO, environment.
364 Strath, Clutterbuck, and Sanderson

b. Bone marrow culture medium; Culture medium is supple-

mented with lO4M hydrocortisone and 15% fetal calf serum.
The pH of this medium is maintained by culturing in an at-
mosphere of 5% CO, in air.
(1) Media additives:
(a) Hydrocortisone: A UYzM (48.45 mg/mL) stock solution of hy-
drocortisone sodium succinate in PBS is filter sterilized (0.22
pm pore) and stored in 50-100 PL aliquots at -20°C. 10 PL
stock solution is added to 100 mL bone marrow culture me-
dium. Any remaining stock hydrocortisone solution is not
refrozen, but discarded.
(b) Fetal calf serum: This has to be selected for optimal eosinophil
growth, since some batches result in negligible eosinophil
production in cultures. Whether this is caused by inhibitors
or lack of growth factors is not known. It is preferable to test
the fetal calf serum in both the EDF assay and in long-term cul-
tures.
(c) Gentamicin: A stock solution of 5 mg/mL gentamicin sulfate
in PBS is filter sterilized, stored at -20°C, and diluted 1:lOO
into medium for use. Gentamicin is occasionally required
when assaying column fractions for IL5.
4. Sources of Interleukin-5: Native murine IL5 can be obtained from
mitogen-or antigen-stimulated spleen cells (Zl), T cell clones (17), or
the EL4 lymphoma cell line (18). However, these sources also contain
other lymphokines. The T cell hybrid NIMP-THl produces IL5 in the
apparent absence of other known lymphokines (29) and has provided
the most useful source of IL5 until recombinant material from trans-
fected monkey COS cells became available (20). There is no known
source of native human IL5 and work on human eosinophils has been
based on the cross reactivity of murine IL5 (21). However, recombin-
ant human IL5 is now available from transfected COS cells (22). IL5
has also been detected in the serum of parasitized animals (B), and re-
cent work has demonstrated the presence of a factor stimulating hu-
man eosinophil differentiation in serum from patients undergoing
eosinophilia (23).
Interleukin-3 and GM-CSF have some eosinophil differentiation ac-
tivity (6) so limited number of eosinophils can be produced in cultures us-
ing WEHI- conditioned medium or commercially available GM-CSF.
When sources of IL5 such as crude spleen conditioned medium that con-
tain other lymphokines as contaminants are used, large numbers of neutro-
phils and/or macrophages are produced so that eosinophils represent
only a small percentage of the nonadherent cells.
Human and Murine Eosinophils

Methods detailing production of the above sources of IL5 are given in

ref. 24.
5. Eosinophil cultures: A Class II Microbiological safety cabinet is re-
quired for human cultures, and all waste materials should be auto-
claved. Murine cultures are established under standard tissue culture
conditions. A gassed 37°C incubator is required, which has to be hu-
midified for agar, cluster, and microplate cultures.
a. Agar cultures: We use Difco Bacto-Agar that is preselected for
colony growth. A 5% w/v stock solution is prepared by sus-
pending the agar in distilled water and placing in a boiling
waterbath for 5 min. The agar is stored at room temperature.
We do not autoclave the agar since this seems to introduce
some toxicity into the system. Leukocyte migration plates
(Sterilin, Teddington) are used in place of Petri dishes as this
allows the whole agar culture to be easily recovered and
stained for assessing colony number and type. The small cul-
ture volume (400 PL) results in savings in materials and rea-
gents. A square 1OOmMPetri dish is used as the container for
the leukocyte migration plate.
b. Long-term cultures: These cultures are established in either
flasks or Cluster plates. We use 25-80 cm2 flasks from which
large numbers of eosinophils can be produced for functional
studies. Most experimental work is done using 24-well Clus-
ter plates. A cytocentrifuge is useful to prepare slides for
staining and differential counts, the morphology being clearer
than on smears.
c. Microplate cultures: The cultures are established in 96-well
round (U) bottomed microtiter plates. The flat bottomed 96-
well microtiter plates are not satisfactory for this assay sys-
tem.
6. Eosinophil Peroxidase Assay
a. Peroxidase buffer: 0.05M Tris-HCl pH 8.0. Filtered (0.45 pm
pore) and stored at room temperature.
b. o-Phenylenediamine (OPD): a stock solution of 10 mg/mL in
distilled water is stored in 1 mL aliquots at -70°C, where it is
stable for several months.
c. 30% w/v hydrogen peroxide stored at 4°C.
d. Triton X-100: a 10% stock solution in water is stored at 4°C.
e. 4M sulfuric acid.
f. Complete substrate solution: to 48.5 ml peroxidase buffer add
1 mL OPD stock solution, 0.5 mL Triton X-100 and 6 PL hydro-
gen peroxide. OPD is light sensitive and so this solution
366 Strath, Clutterbuck, and Sanderson

should not be prepared until it is required. An automatic mi-

croplate reader with a 490~nm filter is recommended for read-
ing the plates.
7. Eosinophil Stains
a. Giemsa: Buffer; Sorensens buffer concentrate pH 6.8, diluted
to 3.3 mM in distilled water. Giemsa (10%) stock solution is
diluted 1:5 with buffer for use.
b. Congo Red: Dissolve 5 g Congo Red in 50 mL distilled water
then add 50 mL ethanol. The solution is stable and can be re-
used. Different batches of Congo Red seem to vary in their
ability to stain eosinophils.
c. Toluidine Blue: Add 1 g Toluidine Blue to 100 mL methanol.
Acidify by adding 5 mL2MHCl. The solution is stable and can
be reused.
d. Luxol-Fast-Blue: Add urea to 70% ethanoluntil saturated (ap-
proximately 250 g/L), then filter through a Whatman No. 1 fil-
ter paper. Dissolve 1 g Luxol-Fast-Blue in 100 mL of urea sat-
urated 70% ethanol. The solution is stable and can be reused.
e. Harris’ Hematoxylin: Dissolve 1 g hematoxylin in 50 mL eth-
anol, and 100 g aluminum ammonium sulfate (or aluminium
potassium sulfate) in 1 L distilled water (with gentle heating).
Add the hematoxylin solution to the salt solution and bring to
the boil rapidly. CAREFULLY add 2.5 g mercuric oxide (a
violent reaction may occur if added too rapidly) and allow to
cool. Filter. Add 4 mL glacial acetic acid to each 100 mL of
stain and store at room temperature. The stain can be reused
extensively.
3. Methods
3.1. Parasite Passage
With the parasite in a plastic Universal tube in PBS:
1. Aspirate 100-200 PL of PBS into a 1 mL syringe and expel the air
bubble.
2. Allow the parasites to settle at unit gravity.
3. Insert the end of the syringe into the parasite pellet and fill the syringe.
4. Invert the syringe and allow the parasites to settle. Expel excess PBS
and refill with more parasites if necessary.
5. Fit a 0.8 x 40 mm needle and expel air and excess PBS.
6. Inject 100 JJLparasite ip into each mouse.
Human and Murine Eosinophils 367

3.2. Parasite Harvest

1. Kill the mouse by cervical dislocation.
2. Cut the abdominal skin and expose peritoneal wall.
3. Using a 5 or 10 mL syringe with a 0.8 x 40-mm needle inject 5-10 mL
of PBS into the peritoneum.
4. Withdraw the parasite/PBS into the syringe.
5. Expel parasite/PBS into a sterile plastic Universal bottle.
6. Wash the parasites several times by filling the bottle with PBS, allow
the parasites to settle, and pour off supernatant.
7. Use the parasites to infect more mice (seeSection 1) and store the re-
mainder at 4OC.
3.3. Marrow Collection and Cell Preparation
3.3.1. Mouse Marrow (See Note 1)
1. Fill a suitable container with 70% ethanol and immerse a pair of scis-
sors and forceps. Use a tissue to dry the instruments before use and
replace them into the alcohol between each procedure.
2. Kill mice by cervical dislocation.
3. Fill a 5-mL syringe with bench medium and fit a 0.4 x 12-mm needle.
We use the syringe type, which is packed in an outer polypropylene
case, the case being used as a sterile “home” during the procedures.
4. Wet theanimal’sfur with70% ethanol and place theanimalonits back.
5. Pull up the abdominal skin with forceps and make a cut across the ab-
domen with the scissors held vertically.
6. Pinch the skin on each side of the cut between thumb and forefinger,
and enlarge the incision by pulling the skin toward the head and tail.
The incision should enlarge around the animal and the skin finally
break at the animal’s back. Continue to pull the skin back toward and
over the tail and legs until the muscles of the upper and lower legs are
exposed.
7. Complete the skinning by grasping the tibia/fibula with forceps and
the skin with the other hand and pulling the leg from the skin until the
foot is clear and the whole leg completely skinned.
8. Remove the legs from the animal by a combined movement that both
dislocates the leg from the pelvis and cuts the muscles, ideally leaving
the leg complete with femoral head.
9. Holding the femur with forceps, cut across the femur with the scissors
to remove the head.
10. Flush out the marrow by holding the femur with forceps (foot held
368 Strath, Clutterbuck, and Sanderson

upwards) and inserting the needle as far as possible up into the femur,
being careful not to put the needle through the knee joint. With the
needle in the femur, bend the needle to an angle of about 45Oto prevent
the cells running down the outside of the needle and onto the syringe.
Position the leg above the collection tube and when the medium is in-
jected into the femur thecells will be flushed out and drip into the tube.
The cells often congregate as a lump at the end of the bone, so with-
draw the needle and touch the cut end of the femur onto the tube and
dislodge the cells using the syringe needle and flush them into the
tube. Pool the marrow from several mice into one tube.
11. Centrifuge the cells (3008 for 8 min) and remove the supernatant.
12. Resuspend the cells in 5 mL of medium and break up any large cell
clumps by repeated aspiration into a 10 mL pipet. If the cells are not
to be established in culture immediately, then resuspend them in
bench medium; otherwise use bone marrow culture medium.
13. Aspirate the cells into a 10 mL syringe via a 9.8 x40-mm needle and ex-
pel them back into a tube through a 0.4 x 12-mm needle.
14. Count the cells and adjust to the required density.
15. A differential count can be done to assess the degree of marrow eosin-
ophilia (seeNote 2).
3.3.2. Human Marrow
Marrow can be obtained from aspiration of the iliac crest or the
sternum (during cardiothoracic surgery) or from ribs removed during
thoracotomy. Use of marrow samples for research purposes requires the
informed consent of the patient and the approval of the local Hospital
Ethics Committee.
1. Resuspend marrow in bench medium to 1-2 x 10’ cells/mL.
2. Layer 10 mL cells onto 10 mL Ficoll-Paque in Universal tubes.
3. Centrifuge at 6008 for 35 min at room temperature.
4. Collect theinterface cells (mononuclear cells) and add bench medium.
5. Centrifuge (6008 for 10 min) and resuspend pellet in bench medium.
6. Centrifuge (5008 for 10 min) and resuspend pellet in bone marrow cul-
ture medium at lo6 cells/mL.
Approximately 0.5-l .5 x 10’ mononuclear cells are obtained from 1mL
of normal marrow aspirate.
3.4. Eosinophil Cultures
3.4.1. Murine Microplate Cultures
1. Adjust the marrow cell concentration to 1 x lo6 cells/mL in bone mar-
row medium.
Human and Murine Eosinophils 369

2. Aliquot 10 j.tL of sample (or sample dilutions if titrating) into duplicate

microplate wells.
3. Add 100 PL cells/well (including several wells without IL5 to use as
control cultures) and incubatein a humidified atmosphere at 37OCand
5% CO, in air.
4. After 5 d assay the cultures for eosinophils by either total cell counts
and differential counts or by peroxidase (seeSection 3.7).
3.4.2.Human Microplate Cultures
1. Set up cultures as for mouse microplate cultures (steps l-3 above).
4. Every 7-10 d carefully aspirate 50 PL of medium and replace with
fresh medium containing sample.
5. After 21-28 din culture, determine total cell numbers (Coulter counter
or hemocytometer) and perform differential cell counts on a cytocen-
trifuge or smear preparation, or perform a peroxidase assay.
3.4.3. Agar Cultures
1. Melt the agar in a boiling waterbath and hold at 45OC.
2. Adjust the bone marrow cells to 2 x lo5 cells/mL (or dilutions of cells
if required) in bone marrow medium and hold in a 37OCwaterbath.
3. Aliquot 40 PL of IL5 or other growth factors or their dilutions into each
well of a leukocyte migration plate.
4. Pipet 1 vol of agar into 15 vol of the cell suspension, mix well, and im-
mediately aliquot 400 pL/well.
5. Place the plate into the upturned lid of a lOO-mm-square Petri dish
containing a moist piece of tissue or filter paper and cover with the
square Petri dish base.
6. Allow the agar to solidify, putting the plates for a short time at 4°C if
the laboratory temperature is high.
7. Place in a humidified incubator at 37OCin 5% CO, in air.
8. Fix and stain mouse cultures after 5-7 d, and human cultures after
14-21 d (Sections 3.53.6).
3.4.4. Long-Term Cultures (See Note 6)
1. Adjust the bone marrow cells to 1.5 x lo6 cells/mL. If large numbers
of eosinophils are required within 1-wk use marrow from parasitized
animals, 9-20 d postinfection.
2. Aliquot the cells into flasks or Cluster plates, 10 mL of cells for a 25 cm2
flask, 30 mL for an80 cm* flask or 1 mL/well for 24-well cluster plates.
3. Add IL5 at a predetermined optimal dilution and place in a 37°C incu-
bator in 5% CO, in air. The Cluster plates will require a humidified in-
cubator.
 Strath, Clutterbuck, and Sanderson

4. Every 6-7 d, gently tap the cultures to dislodge the nonadherent cells
and remove all the supernatant and nonadherent cells.
5. Add fresh medium containing IL5 and place back into the incubator.
6. Determine the total cell number of nonadherent cells and prepare a
smear or cytocentrifuge preparation.
3.5. Fixing the Agar Cultures
We find it necessary to stain the cultures to accurately count eosinophil
colonies since assessing the cultures by morphology alone under the in-
verted microscope is misleading-we have found both tight and very loose
colonies of eosinophils. Mouse cultures are stained using Congo Red and
counterstained with either Harris Hematoxylin or Toluidine Blue. We
have had little success using Luxol-Fast-Blue for staining mouse agar cul-
tures. Human cultures can be stained with any of the above stains includ-
ing Luxol-Fast-Blue. Colonies are defined as clusters containing more than
40 cells.
1. Set the leukocyte migration plate at an angle of about 30” in a retort
stand. With a slide held horizontally and resting against the rim of a
well, direct the agar disk onto the slide by flushing with a stream of
PBS from a wash bottle. It is possible to mount three agar disks on one
standard 26 x 76 mm microscope slide.
2. Cover the agar disk(s) with a piece of dry Whatman No. 1 filter paper
and put onto a warm plate to dry. Do not have the warm plate too hot.
3. Remove the filter paper just before it is completely dry and allow the
agar disk to dry completely.
4. Fix in fresh methanol for 15 min before staining (Section 3.6.).

3.6. Eosinophil Stains

3.6.1. Harris Hematoxylin and Congo Red
1. Place the fixed slide in acidified Harris Hematoxylin for 5 min.
2. Wash under gently running tapwater until blued.
3. Place in Congo Red and stain for 15 min.
4. Wash briefly under gently running tapwater and dry.
5. Nuclei stain blue, eosinophil granules red.
3.6.2. Congo Red and Toluidine Blue
1. Stain the fixed slide for 10 min in Congo Red.
2. Wash in 50% ethanol for 5 min.
3. Stain in Toluidine Blue for 5 min.
Human and Murine Eosinophils 371

4. Rinse in water and dry.

5. Eosinophil granules stain reddish-brown, mast cell granules dark
blue. Neutrophils and macrophages should not have any cytoplasmic
staining.
3.6.3. Luxol-Fast-Blue
I. Stain the fixed slide for 1.5 h in Luxol-Fast-Blue.
2. Wash under gently running tapwater for 3 min.
3. Dry.
4. Stain with Harris Hematoxylin for 2 min (30 s for cytocentrifuge
preps).
5. Wash under gently running tapwater for 3 min.
6. Nuclei stain blue, eosinophils have green granules.

3.7. Eosinophil Peroxidase (EPO) Assay (See Note 4)

1. Aspirate most of the medium from the bone marrow cultures, taking
care not to suck out the nonadherent cells of the pellet. The medium
does not interfere with the assay, but the wells become overfilled if it
is not removed.
2. Add 100 u.L of substrate solution to each well and leave at room temp-
erature for 30 min. It is not necessary to incubate in the dark for this
short period.
3. Stop the reaction by the addition of 50 ~.LLof 4M sulfuric acid.
4. Determine the absorbance at 490 nm. Once the acid has been added
the color is stable for several hours.

3.8. Fixing and Staining Cytocentrifuge

Preparations
1. Ensure the cytocentrifuge preparation (or smear) is dry then fix in
methanol for 5 min.
2. Stain the fixed slide in Giemsa for 2 min.
3. Wash under tapwater.
4. Blot dry.
5. Examine with oil immersion, a green Kodak Wratten filter No. 11 may
assist in identifying eosinophils.
6. Nuclei are reddish-purple, eosinophils have red to orange granules.
Basophilic granules are blue. The eosinophil stains in Section 3.6. can
be used instead of Giemsa for staining cytocentrifuge preparations
and smears.
 Strath, Clutterbuck, and Sanderson

40

30

20

10

0~
1 I I I I I I I I I
0 2 7 10 14 17 21 24 28 31

Days Post-infection of Marrow Harvest

Fig. 1. Number of eosinophils in supernatant from murine bone marrow cultures after
7 d incubation. Marrow was harvested from mice on different days post-infection with
M. corti. All cultures were established in the presence of IL5 from a T-cell clone condi-
tioned medium. Eosinophil numbers represent the mean f 1 standard deviation of three
replicate cultures.

4. Notes
1. When taking bone marrow we work in the open laboratory, using al-
cohol dried instruments. There are very few problems with contam-
ination at this stage. The animal’s fur is wetted with alcohol to prevent
possible contamination by the fur being flicked.
2. The importance of using parasitized mouse marrow for eosinophil
production is illustrated in Fig. 1. It shows the production of eosino-
phils after 7 d incubationusing marrow harvested from mice on differ-
ent days after infection. Uninfected mice usually produce relatively
few eosinophils. We generally use marrow from mice infected for
9-20 d.
3. Use of donor horse serum (DHS) in long term cultures: If a good batch
of fetal calf serum has been obtained, we do not find it necessary to
Human and Murine Eosinophils 373

Table 1
The Production of Eosinophils in the Presence
and Absence of Hydrocortisone and Donor Horse Serum
Medium additives’ Days in culture

+ + :z
DHS Hc 7
35.90
58i9
41.94
43f6
14 21
42.3
Sk4
28
42.33
0.5 f 1
35
42.34
0.2 f 0.4

+ (a> 27.10 37.50 38.61 38.63 38.64

09 41 f 10 33*4 7f4 0.3 f 0.5 0.2 f 0.1

+ (a) 46.50 54.00 54.84 54.96 55.02

W 63*5 38f12 llf5 2913 If1

52.50 53.19 54.26 54.30 54.30

64f7 32f 18 Sk3 If1 o*o
“(+) indicates medium supplemented with 5% donor horse serum (DHS) or hydrocor-
tisone (Hc), f-1 indicates absence from medium.
b(a) = Cumulative total of eosinophils x lod/mL. (b) = eosinophils as percent of total
nonadherent cells (mean + 1 standard deviation of six duplicate cultures). Bone marrow
from parasitized mice was incubated in the presence of T-cell clone conditioned medium.

include DHS in the bone marrow culture medium. Table 1 shows the
cumulative numbers of eosinophils produced over 35 d in culture to-
gether with the percent of nonadherent cells that are eosinophils. It
can be seen from Table 1 that the presence of DHS suppresses eosino-
phi1 numbers, though the percentage of nonadherent cells that are
eosinophils is increased because of reductions in the other cell types.
Hydrocortisone has a short-term suppressive effect in cultures with-
out DHS, but is beneficial for the longer term cultures.
4. The short-term microplate culture system is used to assayIL5 samples,
to find the optimal amount of IL5 to produce eosinophils in the longer-
term cultures, and to test fetal calf serum batches. In assaying for IL5
in column fractions, it is advisable to include gentamicin in the bone
marrow culture medium. Use marrow frommice infected for between
9-20 d for establishing the cultures. Cells of the neutrophil lineage
constitute 30-60% of total normal human marrow cells and, in the ab-
sence of a suitable growth factor, such cells do not survive in themicro-
plate cultures for longer than 21 d. Since the eosinophil perioxdase
(EPO) assay detects human myeloperoxidase (even in the presence of
the enzyme’s inhibitors such as KCN), the assay is only suitable for use
in the human microplate culture system after 21 d culture with sam-
 Strath, Clutterbuck, and Sanderson

80-

60-

40-

20-

0’
I I I I 1 I I I I
0 7 14 21 28 35 42 49 56

Days in Culture

Fig. 2. Production of eosinophils from bone marrow cultures established fromM. c&i
infected mice. Cultures were incubated continuously in the presence of NIMP-THl con-
ditioned medium from day 0 of culture (01; day 7 of culture (A); and day 14of culture (W.
Eosinophil numbers represent the mean + 1 standard deviation of three replicate cultures.

ples that are known not to contain neutrophil growth factors. Figure
2 shows the production of human eosinophils (and total cells) from
microplate bone marrow cultures stimulated with NIMP-THl condi-
tioned medium. Samples that are cytotoxic and kill the marrow ino-
culum may appear positive in the peroxidase assay, since eosinophil
peroxidase has not been degraded during the culture period. For this
reason it is prudent to examine the cultures before assaying for EPO,
and wells that have no visible cell pellet noted.
5. Agar Cultures: Some workers use slides pre-coated with 0.3% agar to
mount their agar disks on. This is to ensure that the agar sticks to the
slide and is not pulled off when the filter paper is removed.
6. Figure 3 shows the time course of eosinophil production of cultures
established in the presence of NIMP-THl conditioned medium. The
Human and Murine Eosinophils 375

0 7 14 21 28 35 42 49

Days in Culture

Fig. 3. Production of total cells G-1 and eosinophils (- - -1 from microwell cultures of
normal human bone marrow established in the presence (Cl) and absence (0) of murine
EDF (NIMP-THl conditioned medium).

marrow was harvested from mice infected for 14 d with M. corti with
amarrow eosinophilia of 45%. When IL5 was added from day 0 of cul-
ture, most of the eosinophils were produced in the first 2-3 weeks
when over 80% of the nonadherent cells were eosinophils (85.7f 5.1%
at 7 d and 83.7 f 1.2% at 14 d. Without IL5 the eosinophils present in
the marrow samples disappear at the end of 7 d, and addition of IL5
for the first time at this point ensures that subsequent eosinophils are
produced in vitro from the precursors and are not surviving eosino-
phils from the inoculum. In cultures that have the addition of IL5 de-
376 Strath, Clutterbuck, and Sanderson

Table 2
Number of Eosinophils Produced from Mouse Bone Marrow
After 7 d Incubation in the Presence of IL5
Source of IL5
EDF NIMP-THl rmIL5
dilution Number” % NumbelP %
1:lOO 63.36 f 6.99 52.2 f 5.1 nd
1:200 nd 20.52 f 1.89 25.9 f 3.0
1:300 22.00 f 0.40 28.0 f 2.6 nd
1:400 nd 18.49f 0.54 21.4 f 4.8
1:lOOO 3.95 f 0.30 7.0 f 1.1 22.60 f 1.56 34.3 f 1.8
1:3000 0.76 f 0.17 1.9fO.l 10.14 f 0.65 14.3 f 0.1
none 0.11 f 0.15 0.2 f 0.3
“Number given as eosinophils x lV/mL. Both number and percent given as mean rt 1
standard deviation of 4 replicate cultures.

layed for 7 d, it takes another 14 d before peak eosinophil numbers are

produced, when the eosinophils may represent up to 90% of nonad-
herent cells. Fewer eosinophils are produced if the addition of IL5 to
the cultures is delayed for 14 d, when at peak production time the eo-
sinophils only represent 27% of supernatant cells.
Recombinant mouse IL5 (rmIL5) is active in stimulating the pro-
duction of eosinophils in murine marrow cultures. Table 2 shows the
effect of different concentrations of rmIL5 and NIMP-THl condi-
tioned medium on eosinophil production after 7 d culture.
It must be noted that eosinophil production is very variable be-
tween experiments, as well as having large variations between “repli-
cate” cultures. This variation between replicates tends to increase as
the cultures get older, particularly in Cluster plate cultures. This may
represent cloning of the very young committed eosinophil progeni-
tors that are present in the marrow at a low frequency.
A total lack of eosinophil production in the presence of a known IL5
sample may be due to an unsatisfactory batch of fetal calf serum, or a
fault in the medium or its preparation.
7. Other sources of eosinophil progenitors: Spleen and fetal liver are
focal points of hemopoiesis. We have no experience with fetal liver,
but spleen cells have been used in agar cultures and in the IL5 assay.
Uninfected mouse spleen cells produced no eosinophil colonies when
incubated in agar in the presence of EL4 conditioned medium, but
spleen cells from mice infected for 12 d do produce eosinophil col-
onies. Cells from the same spleen gave a positive assay result when
Human and Murine Eosinophils 377

used in theIL5 assay, but had to be partially purified and the EPO level
above background (i.e., unstimulated cultures), was not as great as
marrow cells.

References
1. Greenberger, J. S.(1984) Long term hematopoietic cultures, inHuematupoiesis(Meth-
odsin Huemafology, vol. 11) (Golde, D. W., ed.), Livingstone, New York, pp. 203-242.
2. Motomura, S.,Katsuno, M., Kaneko, S.,and Ibayashi, H. (1983) The effect of hydro-
cortisone on the production and differentiation of granulocyte/machrophage pro-
genitor cells in long-term marrow cultures. Exp. Hematol. 11,56-62.
3. Dorshkind, K. and Phillips,R. A. (1982) Maturational stateof lymphoid cellsin long-
term bone marrow cultures. 1. Immunol. 129,2444-2450.
4. Aspinall, R. and Owen, J. T. T. (1983) An investigation into the B lymphopoietic ca-
pacity of long-term bone marrow cultures. Immunology 48,9-15.
5. Levin, J. (1983) Murine megakaryocytopoiesis in vitro: An analysis of culture sys-
tems used for the study of megakaryocyte colony forming cells and of the character-
istics of megakaryocyte colonies. Blood 61,617-623.
6. Sanderson, C. J., Warren, D. J., and Strath, M. (1985) Identification of a lymphokine
that stimulates eosinophil differentiation in v&o. Its relationship to interleukin3,
and functional properties of eosinophils produced in cultures. J, Exp.Med. 162,60-
74.
7. Clutterbuck, E.J. and Sanderson, C. J. (1988)Human eosinophil haemopoiesis stud-
ied in vitro by means of murine eosinophil differentiation factor (IL5): Production
of functionally active eosinophils from normal human bone marrow. Blood71, M6-
651.
8. Sanderson, C. J.,O’Garra, A., Warren, D. J.,and Klaus, G. G. B. (1986) Eosinophil dif-
ferentiation factor also has B-cell growth factor activity: Proposed nameinterleukin-
4. Proc.Natl. Acad. Sci.83,437-440.
9. Johnson, G. R., Nicholas, W. L., Metcalf, D., McKenzie, I. F. C., and Mitchell, G. F.
(1979) Peritoneal cell population of mice infected withMesocesfoidescorti as a source
of eosinophils. hf. Arch. Allergy Appt. Immunol. 59,315-322.
20. Strath, M. and Sanderson, C. J. (1986) Detection of eosinophil differentiation factor
and its relationship to eosinophilia in Mesocestoides corti infected mice. Exp. Hema-
to1.14, X-20.
22. Strath, M. and Sanderson, C. J.(1985) Production and functional properties of eosin-
ophils produced from bone marrow cultures. J. Cell Sci.74,207-217.
12. Strath, M., Warren, D. J.,and Sanderson, C. J. (1985) Detection of eosinophils using
an eosinophil peroxidase assay. Its use asan assayfor eosinophil differentiation fac-
tors. J. Immunol. Methods 83,209-215.
23. Mueller, J. F. (1972) Survival and longevity of Mesocesfoidestetrathyridia under ad-
verse conditions. J. Parasitol.58,228.
14. Turner, J. A. (1975) Other Cestode Infections, in DiseasesTransmitted from Animals to
Man (Hubbert, W. T., McCulloch, W. F., and Schnurrenberger, P. R., eds.), Charles
C. Thomas, Illinois, pp. 708-744.
25. Hutchinson, W. F. and Martin, J. B. (19801Mesocestoides (Cestoda) in a child in Mis-
sissippi treated with Paromomycin sulphate (Humatin), Am. 1. Trap. Med. Hyg. 29,
478,479.
378 Strath, Clutterbuck, and Sanderson

16. Dass, P. D., Murdoch, F. E., and Wu, H. -C. (1984) Glutamine promotes colony for-
mation in bone marrow and HL-60 cells; Accelerates myeloid differentiation in in-
duced HL-60 cells. In Vitro 20,869-875.
17. Sanderson, C. J., Strath, M., Warren, D. J., O’Garra, A., and Kirkwood, T. B. L. (1985)
The production of lymphokines by primary alloreactive T cell clones; a coordinate
analysis of 223 clones in seven lymphokine assays. Immunology 56,575- 584.
18. Dutton, R. W., Wetzel, G. D., and Swain, S. L. (1984) Partial purification and charac-
terisation of a BCGFII from EL4 culture supematant. I. Immunol. 132,2451-2456.
19. Warren, D. J. and Sanderson, C. J. (1985) Production of a T-cell hybrid producing a
lymphokine stimulating eosinophil differentiation. Immunology 54,615-623.
20. Campbell, H. D., Sanderson, C. J., Wang, Y., Hort, Y., Martinson, M. E., Tucker, W.
Q. J., Stellwagon, A., Strath, M., and Young, I. G. (1988) Isolation structure and ex-
pression of cDNA and genomic clones of murine eosinophil differentiation factor.
Comparison with other eosinophilopoietic lymphokines and identity with Inter-
leukin-5. Eur. J. Biochem. 174,345-352.
21, Lopez, A. F., Begley, C. G., Warren, D. J., Vadas, M. A., and Sanderson, C. J. (1986)
Murineeosinophil differentiation factor: An eosinophil-specific colony-stimulating
factor with activity for human cells. J. Exp. Med. 163,1085-1099.
22. Campbell, H. D.,Tucker, W. Q. J., Hort,Y., Martinson, M. E., Mayo,G., Clutterbuck,
E. J., Sanderson, C. J., and Young, I. G. (1987) Molecular cloning, nucleotide se-
quence, and expression of the gene encoding human eosinophil differentiation fac-
tor (interleukin-5). hoc. Nutl. Acud. Sci. 84,6629-6633.
23. Kern, P., Horstmann, R. D.,and Dietrich, M. (1987) Eosinophil production in human
bone marrow cultures induced by 80-85 kDa serum component(s) of patients with
eosinophilia. Br. 1. Huemafol. 66,165-172.
24. O’Garra, A. and Sanderson, C. J. (1987) Eosinophil Differentiation Factor and its as-
sociated B cell growth factor activities, in Lymphokines and Interferons: A p~uctical up-
proud (Clemens, M. J., Morris, A. G., and Gearing, A. J. H., eds.), IRL, Oxford, pp.
323-343.
 Chapter 31

Clonogenic Assays
for Hematopoietic
and Tumor Cells Using
Agar-Containing Capillaries

H. Rainer Maurer
and Christine Echarti
1. Introduction
The so-called hematopoietic stem cells from mouse and human bone
marrow can be induced in vitro to form colonies, provided that the appro-
priate hormones (colony stimulating factors, CSFs) and culture conditions
have been selected. Such hematopoietic colony cultures are of interest for
in vitro cytotoxicity testing of a great number of drugs. In fact, it has been
increasingly recognized that serious side effects of many drugs involve
injury to hematopoietic organs. Moreover, several hormones and in-
hibitors (including chalones) that regulate hematopoiesis have attracted
clinical interest and their identity can be determined by such in vitro
cultures. Therefore hematopoietic stem cells seem to provide identified
targets to assay for either stimulatory, inhibitory, or cytotoxic drug effects.
Based on the pioneering studies by the groups of Sachs, Bradley, and
Metcalf, (seeChapters 23 and 26 in this volume) methods have been devel-
oped during the last decades to clone, in agar or methylcellulose, the stem

379
380 Maurer and Echarti

cells of virtually all hematopoietic cells (erythrocytes, lymphocytes, gran-

ulocytes, macrophages, megakaryocytes) (2). Moreover, similar clono-
genie cell assays in agar or methylcellulose using human tumor cells have
been developed (human tumor clonogenic cell assay; HTCA) and are
widely used to study tumor biology, to predict the sensitivities of patient’s
tumors to anticancer drugs and to determine the potential activity of new
cytostatic drugs (2,3).
The in vitro colony assays of most research groups utilize Petri dishes
into which the stem cells are seeded in an agar-medium on top of another
agar-layer serving as a feeder layer (seeChapter 26 this volume). Since this
double-layer technique is tedious for large-scale routine work, our group
since 1976, has adopted and greatly refined a microprocedure (4). Instead
of Petri dishes, glass capillaries are utilized, and these offer a number of
advantages:
1. Less materials are consumed (-lo%), which includes cells, media, and
test samples, with the result of a lower cost per assay and an increased
number of samples that can be tested.
2. No agar double layer: Thus an easier and quicker test setup.
3. Lower agar concentration (0.18 vs 0.3% in Petri dishes) with the result
of better cell staining after culture.
4. Linear alignment of colonies: Thus easier and quicker colony count-
ing (with a dissecting microscope) and the potential for optical eval-
uation by simple light-scattering densitometry (5).
5. Reduced risk of contamination by bacteria and fungi.
6. Higher cloning (plating) efficiencies of up to fivefold (6; seebelozu).
The steps of the procedure for cloning cells in agar-containing glass
capillaries are shown in Fig. 1 for granulocyte-macrophage (7) and T-
lymphocyte (a), in Fig. 2 for human tumor clonogenic assays (9).
The dimensions of the capillaries, open at both ends, are selected such
that homogeneous colony growth is maintained along the entire gel. The
coefficient of variation of colony numbers does not exceed +_5% from
capillary to capillary (20). Filling of the capillaries is easy: The capillaries
are put onto the tip of an automatic pipet, by which the cell suspension in
liquified agar is sucked in.
Optical evaluation by light scattering densitometry is suited for
special tasks (10,11). For this the capillaries are drawn through the light
path of a spectrophotometer by means of a densitometer attachment. The
light scattered by the colonies inside the capillaries is received via two
mirrors (5). The agreement of visual colony counting by microscope and
Clonogenic Assays 381

T- Lymphocyfe - Colony -Assay

Humon bloof

Isopoque - Ficoll-centrifugofion
4
PHA-slimulalion in suspension
culture for 24 h /37 OC
4
Add/lion of lest sample
and liquefied agor
1
‘;“‘., .:‘..*. ; 1 SeedIng inlo capWvres
1
I,....,,._.,.1 1 Colony -counting ofler 7 days

Granulocyfe - Co/any -Assay

i-: Mouse bone ,marrow ceils

:>;
li’
** Addilion o/ CSA, fesf sump/e
; and liquified ogar
.-.
a )I)) 1
*a’ .*...
. *, ::. Seeding inlo capillaries
1
0 - ..*, .etc. 1 Colony (and clusfer ) counfing
aller 7 duys

Fig. 1. Standardcolony assaysin glasscapillaries.

light-scattering densitometry is excellent. The latter method allows the

follow-up of daily growth of individual colonies of one capillary (IO).
The cloning efficiencies obtained with the agar capillaries were in all
cases tested higher than those found with the standard Petri dish (35 mm)
or microtiter plate technique (96 wells) (6).
Potential drawbacks of the agar capillary technique should not be
ignored, however. These include:
1. It is impossible to substitute agar for methylcellulose, since highly
viscous solutions of the latter cannot be drawn into the capillaries.
2. The closed incubation form of the capillary does not permit admini-
stration of samples after start of incubation.
382 Maurer and Echarti

Solid tumor e---+ Pathology

1 mechanically
enzymat/cally
or

Single cell suspension

A;;.:-..:\*

..‘:; :
;,y :.:
8 .,.

Con- +Cytostatic (div. doses)

trot
I hr rncubafron
I
2 x Washing and resuspending,
addition of agar (0.2 % )
1
0 :..*a . . ...y.* 7
(I . . . . ., *., .-.. I Seeding in to
0 . .*.*‘. . ..*. J glass capillaries
a *,.a;. ..a.. 1

7 -21 Days incubofion

I
0 1.4 rlr~ak1 J Counting of colonies
I (>40 cells )
1 Evaluafron:
Sensitivity Index

1, 2-L
AlOO

Fig. 2. Tumor stem cell assay as oncobiogram (cytostatic drug sensitivity testing in
vitro).

3. Since fewer colonies are usually counted per capillary (20-60) than in
Petri dishes, it is essential to determine the required number of cells to
be seeded in order to obtain statistically reliable data.
Though one or the other obstacle may be relevant under particular
conditions, it is certainly worthwhile to balance the pros and cons in each
case. Nevertheless major advantages of the agar capillary assays have been
acknowledged for large-scale clinical studies (12).
Clonogenic Assays 383

In this chapter we describe a microclonogenic assay developed for

mouse hematopoietic cells. A similar technique for human bone marrow
cells has been reported in detail (13). Further, we describea procedure that
allows the cultivation of T-lymphocyte colonies in agar-capillaries (8). The
method has proven suitable to screen for calf thymus inhibitors (chalones)
(14) and for drug side effects (15). Similar techniques have been developed
for T-lymphocytes from mouse thymus (16) and for B-lymphocytes from
mouse spleen (17). The lymphocytes are isolated from human peripheral
blood and prestimulated with phytohemagglutinin for 24 h. After adding
a test sample, the lymphocytes are seeded into agar-containing glass
capillaries, and the effect of the sample on colony formation is evaluated
after 7 d. Finally the microtechnique for cloning tumor cells is described.
It works with established tumor cells as well aswith cells from fresh tumor
explants or biopsy material. Methods to establish in vivo-in vitro correla-
tion rates and to determine in vitro sensitivity indices have been described
and discussed (18). The reader is particularly referred to the pharmacoki-
netic problems involved with physiologically relevant drug exposure (19).

2. Materials
2.1. Materials Needed in All Procedures
1. Glass capillaries: inner diameter, 1.38 mm; length, 126 mm. Before
use, the capillaries are put in a beaker containing double distilled
water and cleaned in an ultrasonic bath at 90°C for 30 min, rinsed twice
with double distilled water, dried, and sterilized at 180°C for 2 h.
2. Capillary holders (seeref. 5)
3. Dissecting microscope with special stage holder for capillaries and
indirect lighting.
4. Bacto agar: 3 g of agar are dissolved in 100 mL of double distilled water
using a boiling water bath. The solution is then distributed into 5 mL
aliquots and autoclaved at 121OCfor 15 min.

2.2. Materials for the Granulocyte-Macrophage

Assay
1. Male C57B1/6 mice, ideally 6-8 wk old, but not older than 6 mo.
2. Dulbecco’s Modified Eagle’s Medium (DMEM) with glucose (1 g/L),
bicarbonate (3.7 g/L), glutamine (4 x lo3 M), penicillin (500 U/mL),
and streptomycin (500 yg/mL); pH 7.2.
3. Eagle’s MinimumEssentialMediumIscovemodification (MEMIscove)
384 Maurer and Echarti

with bicarbonate (3.0 g/L), glutamine (4 x lOaM), penicillin, and

streptomycin; pH 7.2.
4. 1:l mixture of DMEM and Ham’s nutrient mixture F 12 (DMEM/F12)
with bicarbonate (1.1 g/L) glutamine (2.5 x lOaM>, penicillin (500 U/
mL), and streptomycin (500 pg/mL).
5. Saline: 9.1 g of NaCl/L of double distilled water.
6. Lipopolysaccharide W (LPS) from Salmonella typhosu.
7. Horse serum: inactivated at 56OCfor 30 min and pretested for optimal
colony growth; stored at -2OOC.
8. Glutamine solution: 200 mM in saline, stored at -2OOC.
9. Bovine serum albumin (BSA) solution: 100 mg/mL from Boehringer
Mannheim.
10. Transferrin solution: 30 mg/mL, l/3 iron saturated from Boehringer
Mannheim (seeChapter 21 for preparation).
11. Soybean lipids: 2 mg/mL liposome suspension from Boehringer
Mannheim.
12. Bovine insulin: 500 pg/mL in 12 mM HCl, stored at -2OOC.
13. Turk’s staining solution: 10 mg of gentian violet are dissolved in a few
mL of double distilled water. After adding 1 mL of glacial acetic acid,
the vol is made up to 100 mL with double distilled water. Stored at
room temperature.
All media and solutions are stored at 4OCunless otherwise stated.

2.3. Materials for the X-Lymphocyte Assay

1. Dulbecco’s Modified Eagle’s Medium (DMEM; see Materials for
granulocyte-macrophage assay, item 2).
2. Hanks’ balanced salt solution (HBSS; see Appendix).
3. Saline: 9.1 g NaCl/L %O.
4. Ficoll-Paque (density 1.077 g/mL).
5. Phytohemagglutinin M @HA-M): dissolve contents of one vial in 5
mL of sterile double distilled water; 300 PL aliquots are stored at
-2OOC.
6. Heparin solution: 5000 U/mL.
All media and solutions are stored at 4OC unless otherwise stated.

2.4. Materials for the Tumor Cell Assay

1. Female Balb / c mice, 6 wk old.
2. Rl?MI 1640 medium with bicarbonate (0.98 g/L), glutamine (2.1 x
Clonogenic Assays 385

104M), penicillin, and streptomycin; pH 7.2.

3. CMRL 1066 medium with bicarbonate (2.17 g/L) and glutamine (7 x
10 -4M); pH 7.2.
4. Enriched CMRL 1066 medium: Supplements for 100 mL of CMRL
1066 liquid medium:
7% NaHCO, solution, 3.1 mL.
2.2% CaC1, solution, 4.0 mL.
0.6% asparagine solution, 1.5 mL.
4.4% sodium pyruvate solution, 0.5 mL.
2.1% serine solution, 0.2 mL.
3% Tryptic Soy Broth (TSB), 10.0 mL.
2.9% glutamine solution, 10.0 mL. (All % = w/v.>
Insulin (e.g. Depot-Insulin Hoechst; 10 mL ampule/400 IU solution
for injection), 5.0 mL.
Horse serum, 15.0 mL.
Penicillin-streptomycin solution 50,000 IU/mL and 0.05 g/mL, re-
spectively, 1.0 mL.
pH 7.2.
5. Mineral oil (e.g. Esso Extra).
6. Metal filters 100 and 50 mesh.
7. Collagenase-DNase solution (in CMRL 1066 medium; ref. 21): col-
lagenase type I, 0.15%, and DNase type III, Sigma, 0.015% (final con-
centrations); prepared freshly.
8. Dispase: Grade I, from Boehringer Mannheim, 2.4 U/mL; prepared
freshly.
9. Horse (HS) and fetal calf (FCS) serum: inactivated at 56OCfor 30 min
and stored at -2OOCin 1 mL aliquots. Sera should be pretested for
optimal colony growth.
10. Saline: 9.1 g NaCl/L H20.
11. Trypan blue solution: 0.2% in saline.
All media and solutions are stored at 4°C unless otherwise stated.

3. Methods
3.1. Granulocyte-Macrophage Colony Assay
3.1.1. Preparation of Colony Stimulating Factor
The type and source of colony stimulating factor (CSF) will determine
which species of colony cells will develop in the cultures (see2 for more
386 Maurer and Echarti

details). Highly purified recombinant CSFs are now available from vari-
ous commercial sources and may be used. In this section we describe a
relatively simple procedure to prepare CSF from mouse-lung-conditioned
medium (MLCM) according to a modified method described by Sheridan
and Metcalf (20). Injection of endotoxin in vivo induces formation of CSF
in mouse lung cells incubated in vitro. The CSF produced may then be
purified from the culture supernatant (conditioned medium). Low mo-
lecular weight inhibitors, which are also formed, are removed by dialysis.
1. 60 to 70 strain C57B1/6 mice are injected intraperitoneally with 100 PL
each of an endotoxin solution containing 50 pg/mL of LPS in sterile
saline.
2. The mice are killed 4 h later by cervical dislocation. The lungs are
removed and washed in saline.
3. Each lung is placed in a Petri dish containing 7 mL of DMEM and
minced to small pieces with sterile scissors and forceps. The cultures
are incubated at 37°C and 7.5% of CO, in a fully humidified atmos-
phere for 48 h.
4. After incubation the conditioned medium (MLCM) is separated from
the lung tissue pieces by filtration. The filtrate is inactivated at 56OC
for 30 min. Precipitates formed are removed by centrifugation at
30,OOOgat 4°C for 30 min. The supernatant is then dialyzed against
saline for 3 d. The saline should be changed frequently.
5. It is recommended to concentrate the MLCM by l/3 of its volume by
ultrafiltration through a YMlO membrane in an Amicon chamber at
4OC. The filtrate is discarded. The concentrated MLCM is sterile
filtered, distributed into 1 mL aliquots and stored at -2OOC.
6. The optimal dose of CSFis determined by performing a dose-response
curve. About 5-30 PL of MLCM/300 VL assay mixture usually are
sufficient for maximal stimulation of granulocyte-macrophage col-
ony growth.
3.1.2. Assay
Granulocyte-macrophage colonies from mouse bone marrow may be
grown in serum-containing and serum-free culture conditions. Both sys-
tems are described here.
1. Before use, add 100 PL of glutamine solution/l0 mL of MEM Iscove.
2. A mouse is killed by cervical dislocation, the femur removed, and
adhering muscle tissue is scraped off with a scalpel. The bone marrow
Clonogenic Assays 387

cells are flushed out aseptically from the marrow cavity with 2 mL of
culture medium (MEM Iscove for serum-containing cultures, DMEM/
F12 for serum-free ones) using a disposable syringe.
3. The cells are suspended by forcing them through a al-gage needle
several times.
4. 90 PL of Turk’s staining solution are mixed with a 10 PL aliquot of cell
suspension. The cells are counted in ahemocytometer (e.g., Neubauer
chamber) and the remainder adjusted to 1.2 x lo6 cells/ml.
5. For a serum-containing assay the following components are pipeted
into sterile Eppendorf reaction tubes:
a. Saline, 100 PL.
b. Horse serum, 60 FL.
c. CSF, 30 pL.
d. MEM Iscove, 10 pL.
e. Cell suspension, 25 pL.
f. Agar-medium mixture, 75 pL.
Final vol, 300 pL. The saline may be replaced by a solution of test agent
(seeNote 1).
6. For a serum-free assay (seeNote 2) the composition of assay mixtures
is as follows:
a. DMEM/F12,96 pL.
b. BSA solution, 30 pL.
c. Transferrin solution (diluted 1:lOO with D/F12), 35 pL.
d. Soybean lipids, 3 pL.
e. Insulin solution, 3 PL.
f. L-Glutamine, 3 pL.
g. CSF, 30 pL.
h. Cell suspension, 25 PL.
i. Agar-medium mixture, 75 pL.
Total vol, 300 pL. The medium may be replaced by test agent solution.
7. For preparation of the agar-medium mixture 3% agar is melted in a
boiling water bath. 1.2 mL are mixed with 3.8 mL of the medium to be
used prewarmed at 37OC. The mixture is then added to the compo-
nents listed above, which have also been prewarmed to 37OC.
8. After mixing carefully, three capillaries are filled with 75 PL of assay
mixture from each Eppendorf tube. For this each capillary is put on a
plasticpipet tip fitted onto an automatic microliter pipet (200 pL), and
the assay mixture is sucked up. The capillary is then removed hori-
zontally and one end lowered, so that the gel moves to the middle.
388 Maurer and Echarti

Then the capillary is left on a cooling plate for a few minutes,

9. For incubation the ends of the capillaries are wiped with 70% ethanol
and the capillaries mounted in special capillary holders disinfected
with 70% ethanol.
10. Incubation is for 7 d at 37OC, 7.5% of CO,, and relative humidity of <
100%.
11. Using a dissecting microscope granulocyte-macrophage colonies are
counted at 15 x magnification (objective 1.5 x, eyepiece 10x). A colony
is defined as an aggregate of more than 40 cells; smaller aggregates
(840 cells) are termed clusters. There are both diffuse and compact
colonies. In control capillaries without test agent 18-25 colonies are
usually found.
12. From the colony counts of three capillaries the median and mean
absolute deviation are calculated. The median of the untreated control
(without test agent) is set equal to 100%. For dose-response curves the
percentages of colony numbers compared to the control are calculated
from the medians and mean absolute deviations of the assay mixtures.
13. Gels containing colonies may be transferred to microscope slides and,
after drying, may be fixed with methanol and stained using the May-
Griinwald-Giemsa method. Peroxidase staining with benzidine and
H20z is also possible.

3.2. T-Lymphocyte Assay

3.2.1. Prestimula tion
1. 20 mL of peripheral blood are obtained by venipuncture from a
healthy donor and mixed with 0.1 mL of heparin solution.
2. A 1 mL sample is removed to obtain erythrocytes (seebelow, step 5).
The remaining blood (about 19 mL) is diluted with 13 mL of HBSS.
3. In each of 4 centrifuge tubes, 8 mL of the blood-HBSS mixture is
carefully layered on top of 6 mL of Ficoll-Paque. The tubes are closed
with parafilm and centrifuged at 400g at room temperature for 40 min.
After this, four layers appear in the centrifuge tubes. The autologous
plasma is removed with a sterile Pasteur pipet and kept at room
temperature.
4. The lymphocyte layer is removed, washed twice with DMEM, and
resuspended in 2 mL of DMEM, and the total number of lymphocytes
Clonogenic Assays 389

determined. From this the final volume of the prestimulation mixture

is calculated. The final cell density is adjusted to 1.25 x lo6 cells/mL.
5. For prestimulating the lymphocytes, erythrocytes are added at a ratio
of 1:l. For this purpose the erythrocyte concentration of the 1 mL
blood sample has to bemeasured. A small aliquot is diluted 1:lOO with
saline and counted. After determining the concentration the needed
volume of blood (= erythrocyte suspension) is calculated.
6. The following components are also needed for prestimulation:
autologous plasma: 375 pL/mL of final volume.
PHA-M solution: 12.5pL/mL of final volume.
To prepare the prestimulation mixture the calculated amounts of
plasma, erythrocytes, PHA-M, and the 2 mL of lymphocyte suspen-
sion are mixed and made up to final volume calculated with DMEM.
7. The mixture is incubated in plastic tubes at 37OC,5% CO,, and I 100%
humidity for 24 h.
3.2.2. Assay (See Note 3)
1. The cells are washed twice with DMEM, resuspended in 1 mL of
DMEM and readjusted to 1.2 x lo6 cells/mL.
2. For the assay the following components are pipeted into sterile Eppen-
dorf reaction tubes:
a. Saline, 100 PL.
b. Autologous plasma, 63 pL.
c. PHA-M solution, 2 pL.
d. DMEM, 23 /JL.
e. Cell suspension, 37 PL.
f. Agar-medium mixture, 75 pL.
Final vol, 300 pL. The saline may be replaced by a solution of test
agent.
3. Preparation of the agar-medium mixture and filling the capillaries are
performed as in steps 7-9 of the granulocyte-macrophage assay
except that 30 PL of assay mixture are sucked into the capillaries.
4. Incubation and evaluation are done as described in steps 10-13 of the
granulocyte-macrophage assay. Colonies are counted at a magnifica-
tion of 37x (objective3.7x, eyepiece 10 x). Normally40-60 colonies are
counted per capillary, usually located at the bottom of the agar gel. A
colony is defined as an aggregate of more than 50 cells. T-lymphocyte
colonies are mostly compact (seeNote 4).
390 Maurer and Echarti

3.3. CZonogoenic Assay Using Tumor Cells

for Cytostatic Drug Screening
3.3.1. Preparation of lkmor Growth Factor (TGF)
TGF is used for activating the tumor stem cells in cloning assays with
fresh human tumor stem cells.
1. About 20 Balb/c mice are injected intraperitoneally with 0.1 mL/
mouse of mineral oil. After 5 wk the mice are killed by cervical
dislocation. The spleens are removed aseptically, minced with scis-
sors and forceps, and sieved through a metal filter (100 pm mesh). The
resulting cell suspension is adjusted to 2 x lo6 cells/mL in RPM1 1640
medium.
2. 5 mL aliquots of this suspension are pipeted into 60 mL Petri dishes
and incubated at 37OC and 5% of CO, for 2 h. The supernatant is then
discarded.
3. Adherent cells are carefully washed twice with fresh RPM1 1640
medium by gently moving the Petri dishes and sucking off the
medium with a Pasteur pipet. Fresh RPM1 1640 medium supple-
mented with 15% of FCS is added to the dishes.
4. The cultures are now incubated for 3 d under the same conditions
mentioned in step 2. After this the supernatant is collected and filter-
sterilized.
5. The activity of this conditioned medium is measured by performing
a dose-response curve on a standard tumor cell line (e.g., mouse lym-
phocytic leukemia L 1210). From the curve, its optimal concentration
for cloning fresh tumor cells is determined (e.g., 50 u.L/30Oj~L of
cloning mixture).
3.3.2. Established Tumor Lines:
Propagation of Suspension Cultures
Stock suspension cultures are kept in 25 cm2 tissue culture flasks.
1. A 1-mL sample is removed from the culture to be passaged, the cells
suspended evenly, and counted with Trypan blue for assessment of
viability.
2. The cell density is then adjusted in such a way that the final concen-
tration in the new culture will be a value dependent on the cell line
concerned, e.g., 1 x lo* viable cells/ml for L 1210. 1 mL of HS or FCS
is added, and the volume made up to 10 mL with medium (e.g., RPM1
1640).
Clonogenic Assays 391

3. After incubation at 37OC and 5% CO, for 3-4 d the procedure is

repeated.
3.3.3. Established Tumor Lines: Cloning Assay
1. The following components are mixed in sterile Eppendorf reaction
tubes:
a. RPM1 1640 medium, 141 pL.
b. HS or FCS, 45 pL.
c. Cell suspension, 60 PL.
d. Agar-medium mixture, 54 PL.
Final vol, 300 pL. The medium may be replaced by a solution of test
agent.
2. The cell density is adjusted depending on the cell line used. For
instance, L 1210 cells are seeded at a final concentration of 2.4 x 103
cells/ml; thus the suspension used for the assay is adjusted to 1.2x 104
viable cells/mL to allow for dilution. For this purpose, 1 mL is re-
moved from a liquid stock culture, a small sample of this is mixed with
Trypan blue solution, and viable cells counted. The viability of the
established cell line and the primary human tumors should not be
below 20%.
3. The agar-medium mixture consists of 1 mL of 3% agar and 2 mL of
RPM1 1640 medium and is prepared as described in step 7 of the
granulocyte-macrophage assay.
4. Capillaries are filled with 30 PL of assay mixture as described in step
8 of the granulocyte-macrophage assay.
5. Incubation and evaluation are done as described in steps 10-13 of the
granulocyte-macrophage assay. Normally colony yields should be
30-60 colonies/capillary. Colonies are aggregates of more than 50
cells.
3.3.4. Cloning of Primary Human Tumor Cells
1. Solid tumors should be placed into sterile screwcap tubes containing
RPM1 1640 medium at room temperature, and if possible, sent to the
laboratory immediately after removal. Before processing the tumor is
thoroughly washed with RPM1 1640 medium. Macroscopically vis-
ible necrotic tissue, remainders of tumor capsules, and connecting
tissue are removed with sterile scissors and forceps.
2. The solid tumor is then placed in 5-10 ml of RPM1 1640 medium in a
60 mm Petri dish and cut with sterile scalpels and scissors, until a
seemingly homogenous cell suspension arises. Do not stop cutting the
392 Maurer and Echarti

pieces too early! The cells are sucked up several times through first 21-
gage and then 23-gage needles, and finally filtered through a metal
filter (50 pm pore size) or through several layers of gauze.
3. The mechanically pretreated tumor sample is now transferred to a
plastic tissue culture flask (25 cm2) with sterile forceps and Pasteur
pipets.
4. An 8-mL portion of sterile collagenase-DNase solution or 5-10 mL of
sterile dispase solution is added, and the mixture is stirred at room
temperature for 90 min. Then the suspension is homogenized with 21-
gage and 23-gage needles and filtered as described above.
5. The cells are washed twice with RPM1 1640 medium and resuspended
in 1 mL of enriched CMRL 1066 medium. Cell concentration and
viability are determined as described for established tumor cell lines.
6. For the cloning mixture the following components are mixed in sterile
Eppendorf reaction tubes:
a. Enriched CMRL 1066 medium, 136 PL.
b. TGF, 50 pL.
c. Cell suspension (106-107viable cells/mL), 60 PL.
d. Agar-medium mixture, 54 pL.
Final vol, 300 pL. The medium may be replaced by a solution of test
agent.
7. Further steps are as described for established tumor cell lines (steps
3-5), except that the capillaries are filled with 50 PL of cloning mixture
and the incubation is for up to 3 wk.

4. Notes
1. In all instances, it is recommended to compare unknown test agents
with a substance of known effect on the same test cells (internal
standard, positive control).
2. In serum-free granulocyte-macrophage assays the capillaries should
be left on the cooling plate for lo-15 min before mounting them in the
holders. They should then be incubated in an exactly horizontal
position. Otherwise it is possible that the agar gel will slip out of one
end of the capillary during the incubation.
3. In the T-lymphocyte assay, the prestimulated cells are very delicate to
handle. Although they are sometimes difficult to resuspend they
must not be vortexed. Pasteur pipets should be used for this purpose.
Vibration of the incubator, especiallv during the first few days, will
Clonogenic Assays 393

lead to irregular colony growth. Therefore the capillaries must not be

moved during this time. Concerning the reproducibility of the assay
and the use of cryoconserved lymphocytes, seeref. 22.
4. The median and mean absolute deviation are determined as follows,
First, the three colony counts from each assay point are put into
sequential order. The value in the middle is the median. If there are
an even number of colony counts (e.g., six values from a control run
in duplicate), the median is defined as the mean from the two middle
values. For the mean absolute deviation the absolute differences be-
tween the colony counts and the median are determined and the mean
calculated.

References
1. Metcalf, D. (1984) TheHemopoietic Colony Stimulating Factors (Elsevier, Amsterdam).
2. Salmon, S. E. and Trent, J. M. (eds.) (1984) Human Tumor Cloning (Grime &Stratton,
Orlando, Florida).
3. Courtenay,V. D. and Mills, J. (1978) An in vitro colony assay for human tumors
grown in immune-suppressed mice and treated in viva with cytotoxic agents. Br. 1,
Cancer 37,261-268.
4. Maurer, H. R. (1979) In vitro colony growth of granulocytes, macrophages, T- and
B-lymphocytes in agar capillaries. Acfu Haemafol. 62,322-325.
5. Maurer, H. R. and Henry, R. (1976) Automated scanning of bone marrow cell
colonies growing in agar-containing glass capillaries. Exp. Cell Res. 103,271-277.
6. Echarti, C., and Maurer, H. R. (1989) Manuscript in preparation.
7. Maurer, H. R. and Henry, R. (1976) Colony growth of mouse bone marrow cells in
agar contained in glass capillaries. Bluf 33,11-22.
8. Maurer, H. R., Maschler, R., Dietrich, R., and Goebel,B. (1977)In vitro culture of
lymphocyte colonies in agar capillary tubes after PHA-stimulation. J. Immunol. 18,
353-364.
9. Maurer, H. R. and Ali-Osman, F. (1981)Tumor stem cell cloning in agar-containing
capillaries. Naturwissenschaften 68,381-383.
10. Maurer, H. R. and Henry, R. (1978) Drug evaluation on haemopoietic cells in vitro.
Arzneimittel-Forsch. (Drug Red 28,601-605.
11. Maurer H. R., Henry, R., and Maschler, R. (1978) Chalone inhibition of granulocyte
colony growth in agar: Kinetic quantitation by capillary tube scanning. CeZ2Tiss.
Kinet. 11,129-138.
12. Von Hoff, D. D., Forseth, B. J., Huong, M., Buchok, J. B., and Lathan, 8. (1986)
Improved plating efficiencies for human tumors cloned in capillary tubes versus
Petri dishes. CancerRes.46,40124017.
13. Neumeier, R., Maurer, H. R.,Maschler,R., and Bombik, B. M. (1981)A micromethod
to culture granulocyte colonies from human bone marrow in agar-containing glass
capillaries. Exp. Hemufol. 9,44-51.
14. Maschler, R. and Maurer, H. R. (1978) Screening for specific calf thymus inhibitors
(chalones) by different assays.Hope-Seyler’s Z. Physiol. Chem. 359,825-834.
394 Maurer and Echarti

15. Tortarolo, M., Braun, R., Hiibner, G. E., and Maurer, H. R. (1982) In vitro effects of
epoxide bearing valepotriates on mouse early hematopoietic progenitor cells and
human T-lymphocytes. Arch. Toxicol. 51,3742.
26. Matthiessen, I?.,Echarti, C., Gerber, J., and Maurer, H. R. (1978) T-lymphocyte col-
ony formation of murine thymocytes in agar contained in glass capillaries. Nut 55,
517-522.
17. Ulmer, A. and Maurer, H. R. (1978) The formation of B-lymphocyte colonies in agar-
containing glass capillaries. Immuno&y 34,919-925.
18. Ali-Osman, F., Bier, J., and Maurer, H. R. (1983) In tifvo cytostatic drug sensitivity
1 testing in the human tumor stemcell assay:amodified method for the determination
of the-sensitivity index. Tumor Diagn. Ther. 4,1-6.
19. Ali-Osman, F., Giblin, J., Dougherty, D., and Rosenblum, M. L. (1987) Application
of in vivo and in vitro pharmacokinetics for physiologically relevant drug exposure
in a human tumor clonogenic cell assay.CancerRes.47,3718-3724.
20. Sheridan, J.W. and Metcalf, D. (1973) A low molecular weight factor in lung-condi-
tioned medium stimulating granulocyte and monocyte formation in vitro. J. CelI.
Physiol.Bl, ll-24.
21. Leibovits, A., Liu, R., Hayes, C., and Salmon, S.E. (1983) A hype-osmotic medium
to disaggregate tumor cell clumps into viable and clonogenic single cells for the hu-
man tumor stem cell clonogenic assay.Infer&. J. Cell Cloning 1,478-485.
22. Maschler, R. and Maurer, H. R. (1981) Lymphocyte chalone from calf thymus: prob-
lems of large scale extraction and assay.Z. Nafurforsch. 36 c, 568-571.
 Chapter 32

Cytogenetic Techniques
for Human Leukemias

G. John Swansbury
1. Introduction
It is appropriate that a chapter on cytogenetic techniques for studying
leukemic tissue appears in a series on molecular methods since these
subjects are closely related, both in history and in application. It is through
molecular methods that the significance of specific recurring chromosome
abnormalities begins to be understood, but it is these specific recurring
chromosome abnormalities that suggest the areas of interest for molecular
studies.
For example: when the breakpoint cluster region (bcr) for the Phila-
delphia chromosome translocation, t(9; 22) (q34; qll), was identified and
DNA probes were developed (I), it was supposed that the labor-intensive
cytogenetic screening for the presence of the Philadelphia translocation
(and similarly for other translocations) might become redundant. Events
have shown, however, that the correspondence between abnormal bcr and
detected Philadelphia translocation is not 100% (2), so for the time being,
at least, cytogenetic and molecular studies must continue to support,
confirm and stimulate each other.
It is said that if there are many methods towards one end, then not
only is none much better than the others, but none is very good at all.

395
396 Swansbury

This currently happens to apply to cytogenetic studies of leukemic tissues:

there are probably as many methods as cytogeneticists and none gives the
consistently good quality that can be obtained with, for example, PHA*-
stimulated lymphocytes. The observation that consistently applied tech-
nique can give vastly different results in consecutive cases and that
populations of cells with different morphology can occur on the same slide
indicates that there is an intrinsic, or disease-, or treatment-derived effect
on marrow kinetics and chromosome morphology. Most effects obtained
by varying techniques are slight and are swamped by these variations. A
survey of laboratory methods undertaken by the Association of Clinical
Cytogeneticists (3) revealed that almost every aspect of the processing of
marrows for cytogenetic studies was subject to great variation-and it can
be assumed that all the variations worked for at least some cases. One of
the most significant factors in improving results is to set up multiple
cultures using different media, culture conditions, and harvesting pro-
cedures-a dozen cultures per sample is not unusual in some laboratories
(Hagemeijer, A., personal communication). This, however, requires that
an adequate sample is provided and that there is time available to cope
with this large amount of work-beyond the scope of most cytogeneticists
in routine or service laboratories.
It used to be considered that direct preparation of marrow samples
gave the most reliable picture of the in vivo state of the marrow. However,
it has been shown that using various culture times is highly desirable since
different cell lineages predominate at different times (4,5). Erythroid cells
tend to predominate in the direct marrow preparation and granulocyte
cells after culture. Except in erythroleukemia, where the converse is true,
increased proportions of chromosomally abnormal cells have been di-
rectly associated with loss of erythroblasts. Berger et al. (4) also observed
that normal, erythroblast mitoses had good chromosome definition whereas
myeloblast/monoblast mitoses had fuzzy chromosomes. Insome ANLLs,

* Abbreviations frequently used are:

ALL Acute lymphoblastic leukemia;
ANLL Acute nonlymphocytic leukemia;
CLL Chronic lymphocytic leukemia;
CML Chronic myeloid leukemia;
FdUr FluorodeoxyUridine;
PHA Phytohemagglutinin;
TPA 12-O-tetradecanoylphorbol-13-acetate.
Cytogenetic Techniques for Human Leukemias 397

and especially those with a prior myelodysplastic phase, all lineages are
affected; erythroblasts have been shown to be involved by using radio-
active Fe uptake and heterozygosity for the enzyme GGPD (6).
The importance of considering the cytogenetic effect of culturing was
clearly demonstrated in the study of the t(15; 17) translocation in acute pro-
myelocytic leukemia (M3). It was thought to have a geographic variation
in incidence (7) but this is now thought to arise from varying culture
techniques. It is possible that it occurs in 100% of M3 if enough material is
studied (4,8). The translocation is found much more often in overnight
cultures than in direct preparations. In addition it can easily be missed
because mitoses with it often have poorer chromosome morphology than
accompanying normal cells and the translocation produces only small
changes that can easily go undetected in the morphology of chromosomes
15 and 17.
Divisions from ALL often have particularly poor morphology, es-
pecially when from a clone with high hyperdiploidy. A very detailed
methodology has been published by Williams et al. (9) who have a
relatively high successrate.
The techniques for cytogenetic studies in leukemia are simple and
robust; the analysis of the chromosomes obtained comes with a little
practice; the interpretation of the results, however, requires experience
and so this aspect is given a substantial proportion of the notes section.

2. Materials
Most of the solutions should be kept in the dark at 4OC. Unless stated
otherwise, add 0.1 mL to a 10 mL culture.
1. Containers: Sterile, capped, plastic, centrifuge tubes, “universals,” or
sterilized glass bottles that have been siliconized and have silicone cap
liners.
2. Pipets: Plastic or siliconized [with dimethyldichlorosilanet, e.g.,
“Repelcote”] to prevent adhesion of fixed cells.
3. Medium: RPM1 1640, McCoy’s 5A, Ham’s FlO or F12, TC199, Iscove’s,
and so forth, are all usable. This laboratory now routinely uses Hepes-
buffered RPM1 1640, which was developed specifically for leukemic
cells, except where described below. To each 100 mL bottle, add 1 mL

+Indicates known or potential carcinogens/poisons: handle with care!

 Swansbury

of preservative-free heparin and antibiotics (e.g., 1 mL penicillin +

streptomycin). Medium used for longer cultures should have L-
glutamineadded (final concentration 0.15 mg/mL); this is an essential
amino acid that is unstable and has a short life at room temperature.
Iscove’s low folate medium can be used without serum.
4. Serum: Fetal calf serum is preferred, but human serum of blood group
AB can be used. The proportion routinely added is 1 part serum to 4
parts medium but 5 or 10% may be adequate. For longer-term cul-
tures, each batch of serum should be tested to ensure that it supports
cell growth.
5. Mitogens: Phytohemagglutinin (PHA) stimulates T-lymphocytes to
divide, acting via monocytes that produce interleukin II.
For B-cell disorders, other mitogens are used, usually pokeweed
mitogen, which indirectly stimulates both T and B-lymphocytes, and
TPA, which is specific for B cells. PI-IA and pokeweed mitogen for
cytogenetic use are obtained freeze-dried or lyophilized, ready for
reconstitution to the appropriate concentration.
12-0-tetradecanoylphorbol-13-acetate (TPA)? (also Phorbol 12-
myristate 13-acetate): 10 yg/mL in 0.5% ethanol, stored in aliquots at
-2OOC. Prepare by dissolving in 10% ethanol and then further dilut-
ing 1:19 with water.
6. FluorodeoxyUridine (FdUr)+: Stock solution: 1 part FdUr (25 pg/mL)
to 3 partsuridine (1 mg/mL), giving final concentrations of 0.1 and4.0
mM.
7. Thymidine: 0.05 g in 100 mL distilled water; filter sterilize (0.22 I.IM
millipore filter); store in aliquots at -20°C (do not refreeze); thawed
solution keeps at 4OCfor at least 1 mo.
8. Arresting agent: Colcemid or colchicine: Colcemid+: stock solution 1
pg/mL. Also called demecolchicine, from deacetylmethylcolchicine.
Colcemid at greater than 0.1 pg/mL can be toxic (20) but colcemid is
said to be less toxic than colchicine, so it may be more suitable if a long
exposure (e.g., overnight) is used.
Colchicine+: Stock solution: 20 pg/mL in saline (0.9% NaCl).
9. Ethidium bromide+: 10 pg/mL in saline (0.9% NaCl).
10. Hypotonic solution: 0.075M KC1 (5.59 g/L). For delicate membranes,
and if Iscove’s medium has been used, a mixture of three parts KC1 to
1 part 1% aqueous citric acid may be preferable. Note that the effec-
tiveness doesn’t derive from just the osmolarity: the K+ ions have a
physiological action, so no advantage is obtained by diluting further.
Cytogenetic Techniques for Human Leukemias 399

11. Fixative: Absolute methanol and glacial acetic acid (3:l). This should
be freshly prepared just before use, although it may be kept for a little
while if chilled.
12. 2 x SSC: NaCl(17.53g) + sodium citrate (8.82g) made up to 1 L of
aqueous solution.
13. Trypsin: (2.5%) stored frozen in 1 mL aliquots. Diluted 1:50 in PBS
(Ca*+and MgZ+ free) when required.
14. Phosphate buffered saline (PBS; seeAppendix): pH 6.8 for stains, pH
7.3 for trypsin, and pH 5.2 and 6.5 for fluorescence.
If necessary, use N%HI?O, (0.07M) and K2HI?04 (0.07M) in the ratio
1:49 for pH 5.2 and 8:17 for pH 6.5.
15. Slides: Frosted-end variety are preferable for convenience of labeling.
The slides must be free of dust and grease. Wash in detergent, rinse
well in water, then in dilute HCl and alcohol. Store dry, protected
from dust, or in acidified alcohol.
16. Coverslips: 1.5 grade thickness or less for G-banding, grade 0 for
fluorescence work; 22 x 50 mm size.
17. Stains: Giemsa, Wright’s, and Leishman’s stains (all commercially
available) are all suitable. Leishman’s stain is said to be better for the
fuzzy chromosomes that occur in leukemias.
Fluorescence stains: Atebrin: (quinacrine dihydrochloride)+ 50 pg/
mL in 2% KCl. A&dine Orange? 0.01% in phosphate buffer at pH 6.5.
18. McIlvane’s buffer (pH 5.6): Atebrin mounting solution: 0.466 g citric
acid, 0.825 g N%HPO,, 50 mL glycerol; make up to 100 mL of solution
with distilled water.
19. Mounting medium: Gurr’s neutral mounting medium is routinely
used in this laboratory; other suitable mountants are XAM, DPX,
Histamount, and so on.

3. Methods
As already stated, every cytogeneticist has his/her own variations on
basically similar processes. This is a summary of the general technique;
variations may need to be used to suit particular conditions.
In summary, chromosomes are prepared from dividing cells (mitoses)
since at metaphase, just before division, they shorten and become recog-
nizable, discrete units. The cells may be already dividing in the tissue
supplied or, in certain circumstances, may be stimulated into division.
They are arrested and accumulated in metaphase or prometaphase by
 Swansbury

destroying the spindle, e.g., with colcemid. The cells are treated with a
hypotonic solution to encourage spreading of the chromosomes. They are
then fixed, after which they can be stored indefinitely. Fixed cells are
spread on slides and air-dried. They can be stained immediately, but are
usually first treated to induce banding patterns on the chromosomes to
assist in their identification.

3.1. Collection of Samples

Bone marrow is the preferred tissue for investigation of most hemo-
poietic disorders. Valuable results can sometimes be obtained from blood,
spleen, and so forth, and these are better than marrow for CLL and myelofi-
brosis. Details relevant to culturing blood are given in Note 1; for the
methods described here it will be assumed that a sample of bone marrow
is being studied. It is very important that the marrow aspirate for
cytogenetic study is taken before any cytotoxic therapy is given: Except for
CML, clones in leukemias almost always disappear during remission,
often reappearing at relapse.
Cytogenetic studies of bone marrow are expensive because they are so
labor-intensive, and a lot of time can be wasted on inadequate samples.
Ideally, a generous portion of the first spongy part of the biopsy should be
sent: later samples tend to be heavily contaminated with blood. Resiting
the needle, through the same puncture if necessary, gives better results
than trying to get more material from the same site. If there is plenty of
material, consider storing some in liquid nitrogen (details in Note 5).
Heparinized bone marrow samples can be transported with or without
medium. However, use of medium may reduce the likelihood of the
sample clotting or drying out. The samples should be sent to the labora-
tory as quickly as possible without exposure to extremes of temperature.
However, a result can usually be obtained even from samples a few days
old (except for ALL, which seems to need more immediate attention),
blood being more robust than marrow. The samples should not be refrig-
erated: the cells may take several hours to start dividing again after ex-
posure to cold conditions.

3.2. Handling Samples

The risk from aerosol formation from marrow/blood is low but the
risk following accidental introduction of infected serum (e.g., through a cut
on the skin or by penetration with a needle or glass pipet) is high. All
samples should be handled carefully, as if they might have hepatitis B or
Cytogenetic Techniques for Human Leukemias 401

HTLV/HIV (AIDS). Gloves should be worn and “quills” should be used

(rather than pipets or needles) while processing unfixed tissue. A laminar
flow cabinet should be used if available, otherwise a clear, draft-free
bench not used for other purposes. Careful sterile technique should be
developed. Since most cultures are short-term, contamination is not
usually a problem.
3.3. Setting Up the Cultures
I. If the marrow aspirate arrives in medium known to be fresh and the
patient is not on chemotherapy, then serum and more medium can be
added directly. Otherwise, spin down the sample and resuspend it in
warmed medium and serum.
2. At least two different cultures should be set up, the two most easy
being a direct and a 24-h culture, incorporating ethidium bromide as
described below. (For ALLs, direct and 2-4-d cultures may be better.)
The cultures most likely to succeed, however, are exposure to col-
cemid overnight and FdUr synchronization. Where possible, it is well
worthwhile using all four types of culture. In addition, using various
culture times increases the likelihood of detecting abnormal clones, as
well as of getting any result at all, since divisions from different
lineages tend to occur at different times. The longer the cells are
exposed to colcemid, the more divisions will be collected. However,
the chromosomes continue to contract and the metaphases may
eventually become unusable.
3. Use 5 or 10 mL cultures with 5-20% serum and a cell density of up to
106/mL. Stand cultures at an angle, rather than upright, since this
increases the surface area of the deposit and reduces local exhaustion
of the medium.
4. For the direct preparation, incubate at 37OC for an hour or so, if
possible, before adding colcemid.
5. Cultures can be gassed with 4% carbon dioxide in air to help maintain
the pH of the culture medium if it is bicarbonate buffered. An in-
creased partial pressure (pp) of CO, is not as important for cell growth
as a decreased pp of oxygen -2.5% may be optimal for longer cultures.
3.4. Harvesting
Use capped 10 mL plastic centrifuge tubes and do all centrifugation
steps at 1000 rpm for 10 min. Avoid fierce treatment of living cells with
hypodermic needles or fine-bore pipets as far as possible: use a syringe
fitted with a plastic filling tube (e.g., a “Kwill”).
402 Swansbury

1. Add 0.1 mL of colcemid (the syringe need not be sterile for this) and
incubate at 37OC for 10 min to 2 h (or overnight). A very useful
variation is to add 0.1 mL ethidium bromide (11) with the colcemid
and leave the culture for 2-4 h. This retards chromosome conden-
sation while allowing divisions to accumulate.
2. Spin down the cells and add hypotonic solution (at not less than room
temperature or above 37OC),mixing thoroughly by tapping the baseof
the tube after the first few drops have been added. Make up to 10 mL.
3. After lo-15 minutes, spin again, remove the supernatant, and tap the
base of the tube to loosen the pellet. Add a drop or two of hypotonic
solution if necessary to ensure dispersal of the cells, since they are
useless if they are fixed in a lump. Add a few drops of fresh fixative
while tapping the tube. Make up to 10 mL and invert the tube to en-
sure thorough mixing. If the volume of red blood cells was large and
inadequately dispersed in the hypotonic solution, they may fix into a
gelatinous mass, trapping the white cells. If this happens, decant off
(and use) the supernatant and then try adding 1:l hypotonic solution:
fix to the clump, which may then be dispersed.
4. Leave in this first fix for a little while before spinning again and add-
ing fresh fixative. After this, the sample can be stored at 4OC with
further changes of fix when convenient. The fix should be changed at
least three times (until the solution is completely colorless) before
spreading. Dehydration of the cells improves if left in fix for at least
overnight and up to a week if time permits. For longer periods, store
at -2OOC. If there was a lot of fat in the specimen, it should be removed
by washing once with Carnoy’s fixative (ethanol 60%, chloroform
30%, acetic acid 10%).

3.5. Spreading
1. A supply of clean slides should be kept either dry in a dustproof
container or in a jar of industrial alcohol acidified with a little HCl. Dry
slides are usually left in a freezer for a while prior to use, as the film of
frost helps spreading. If glass pipets are used, they must be silicon-
ized since the cells are particularly sticky at this stage. Change the
fixative shortly before spreading, then spin again and add any of the
following to obtain a slightly cloudy suspension:
a. 6:l methanol:glacial acetic acid. The slide is then held at an
angle and briefly passed through a gentle Bunsen or spirit
Cytogenetic Techniques for Human Leukemias 403

lamp flame. It is usually possible to see the fixative vapor

burn. This method usually gives good spreading, but possi-
bly at the expense of some banding quality and definition of
outline.
b. Fix: The suspension is dropped onto dry or cold-frosted
slides. Spreading may be improved by “huffing” on each
slide immediately or waving in the hot moist air above a flame
or by adding a drop of fresh fix on top of the spread.
c. 60% aqueous acetic acid. This method usually gives fair
spreading of chromosomes, but also has the disadvantage of
sometimes affecting the quality of banding. Prolongation of
the time in 60% acetic acid before spreading may improve
spreading, but possibly at the expense of further deteriora-
tion in banding quality. The slides can be dried in the hot air
above a flame.
2. A couple of drops onto each slide is adequate. Any loss of band defini-
tion caused by heating is of more consequence when fluorescence
stains are to be used. Spreading may be helped if the suspension is
dropped from a foot or more above the slide, but this is risky and
should not be done if the sample is small and precious. Place the slides
on a hotplate or in an oven (not greater than 60°C) for a few hours to
dry*
3. If possible, check the slides immediately under phase-contrast micro-
scopy for cell density and chromosome spreading. If these are not
optimal, try varying the spreading procedure. An added factor that
affects spreading is the weather-particularly the humidity.
4. Some of the slides may be stained immediately, but banding is not
very effective until the slides have “aged” for a few days. Some aging
effect can be obtained more quickly by incubating the slides at 60°C
overnight in an oven.
3.6. Banding
There are many methods of producing banding on chromosomes.
Band patterns are broadly grouped into G (Giemsa) bands (seeFig. 1) and
R (reverse) bands which are largely complementary (12). Some methods
specific for identifying certain chromosome regions (seeFig. 2) are not de-
scribed here (see,for example, 13). Sequential banding is possible in the
order fluorescence -+ GTG + C-bands, but not in the reverse.
404 Swansbury

Fig. 1. GTG-banded karyotype, showing duplication (arrow) of part of the long arms
of a No. 1 chromosome.

Cell “synchronization” techniques are described in Notes 2 and 3.

This laboratory routinely uses GTG-banding produced by incubation in 2
x SSC followed by treatment with trypsin (modified from ref. 24). By this
method the trypsin serves to enhance the bands produced by the 2 x SSC,
and may be omitted. Trypsin may be used alone, but without the incuba-
tion the banding tends to be unpredictable and it is difficult to control the
amount of digestion needed. Chromosomes banded with SSC may have
a slightly hairy appearance; trypsin alone tends to produce rounded-off
bands and thus an apparently clearer outline.
3.6.1. SSC Banding
1. Heat up a Coplin jar containing 2 x SSC to 60°C in a water bath. Incu-
bate the slides for 60-90 min for bone marrow, an hour for blood.
2. Cool the slides in SSC and rinse in water. (They may be dehydrated
through alcohols and left to dry-overnight if wished-before con-
tinuing with trypsin treatment.)
Cytogenetic Techniques for Human Leukemias 405

Fig. 2. C-banded metaphase of the same case.

3. Add 1 mL of trypsin to 50 mL of PBS (Ca*+- and MgWree): Hanks’

buffered salt solution may also be used. The temperature should not
be over 10°C.
4. Transfer the slides to a jar containing buffer. Each slide is then
immersed in the trypsin solution for a few seconds, washed in alcohol
or in a copious volume of buffer to arrest the enzyme action, and
placed in a rack to dry.
The time needed in trypsin is quite variable and needs to be deter-
mined for each laboratory. It is affected by variations in spreading
technique, age of the slides, degree of contraction of the chromosomes,
general chromosome morphology, and so forth. If the trypsin appears to
act very quickly (within a few seconds), some of the conditions can be
altered to slow it down-e.g., use a lower temperature or a suboptimal
buffer (pH 6.8).
3.6.2. Dypsin Banding
Slides should be incubated for at least an hour in an oven at 60°C
immediately before banding. The duration of exposure to trypsin needs
406 Swansbury

careful attention, being about 5-10 s for BM. A variation is to dip the slide
for 30 s into 25 vol hydrogen peroxide (1:3 in tap water) then rinse in tap
water before proceeding to use trypsin for 8 s for both blood and marrow.
It may be helpful to watch a metaphase under x40 phase contrast to
see when the chromatids start to swell.

3.7. Staining and Mounting

of GTG Banded Preparations
1. Dip the slides in PBS first, to reduce the risk of picking up oxidized
stain from the surface. Use freshly made stain, diluted in pH 6.8
buffer:
Giemsa: 1 part stain to 25 parts buffer for 10-15 min.
or Leishman: 1 part stain to 4 parts buffer for 3-4 min.
or Wright: 1 part stain to 4 parts buffer for 3-5 min.
2. Rinse the slides in buffer, blot gently, and stand to dry.
3. Mount in a little Gurr’s neutral mounting medium (in our experience,
this does not lead to leaching of stain, but XAM, DPX, Histamount,
and so forth, are probably also suitable) using grade 1.5 (or less)
thickness coverslips. If possible, leave to dry overnight before exam-
ining the slide with an oil immersion objective, since oil and soft
mountant tend to mix. Do not leave the slides exposed to strong sun-
light. Do not thin the mounting medium with xylene unless necessary,
since it may result in loss of stain.

3.8. Fluorescence Banding

Fluorescence is not recommended for routine cytogenetic work with
leukemic marrow. The interpretation requires considerably more experi-
ence and, because the abnormalities found are often complex, the rapid
fading during exposure to ultraviolet light is a major drawback to the pro-
longed analysis of complex abnormalities down the microscope, and pho-
tography gives a poor record because of the low contrast. However, there
are instances when fluorescence is helpful inidentifying particular parts of
chromosomes involved in translocations. For example, Atebrin causes het-
erochromatin, satellites, and Y-bodies to fluoresce particularly brightly,
whereas DA-DAPI (20) (seeFig. 3) shows only the heterochromatic regions
of chromosomes 1,9, and 16 and the short arms of chromosome 15. Fluor-
escent R-banding is particularly useful for checking on abnormalities near
the ends of the chromosome arms.
Cytogenetic Techniques for Human I;eukemias 407

Fig. 3. DAPI-banded metaphaseof a PI-IA-stimulated blood cell.

3.8.1. Fluorescent G-Banding

The simplest method of fluorescence G-banding, adapted from ref.
(22), is:
1. Stain the slide in Atebrin for 5-20 min. (The stain may be kept and re-
used.)
2. Rinse in 2% KCl.
3. Add a drop of mounting solution (M&we’s buffer), place a grade 0
coverslip, press, blot, and seal the edges with rubber solution or nail
varnish.
4. The slides can be examined immediately, using ultraviolet light, or
first refrigerated overnight at -2OOC to improve the result.
5. Restaining is simple: peel off the sealant, float off the coverslip, and
restain as before.
408 Swansbury

3.8.2. Fluorescent R-Banding

For R bands with good contrast, use the following procedure, adapted
from ref. (22):
1. Incubate in phosphate buffer pH 5.2 at 87°C for 20 min (longer for
fresh slides, less for old),
2. Stain in 0.01% A&dine Orange in phosphate buffer, pH 6.5, for 5 min.
3. Rinse in pH 6.5 buffer for 2 min.
4. Mount in buffer and seal the coverslip with rubber solution or nail
varnish.
5. An alternative method of producing R bands is to substitute bromo-
deoxyuridine for thymidine when rescuing after treatment with
Methotrexate or FdUr.
3.8.3. Chromosome Analysis
Use a 10x objective for screening and an oil-immersion 100x objective
for studying the metaphases. Analysis directly with the microscope is
possible with experience, but at least one cell from each case should be
photographed as a record. If the workload permits, at least one cell should
be karyotyped, the photographed chromosomes being cut out and ar-
ranged in order (12), especially in apparently normal cases. When screen-
ing, remember to bias toward cells with poor morphology where there is
a mixed population. Selection of only good mitoses can lead to failure to
detect the presence of an abnormal clone. If the quality of metaphases is
poor, then full analysis may not always be possible. However, even if the
chromosomes can only be counted and/or grouped, useful information
can sometimes be obtained.
The analysis of polyploid mitoses may appear daunting, but in some
cases their morphology is better and so abnormalities may be more appar-
ent. In some cases, especially after exposure to radiation, a high frequency
of chromosome abnormalities is found. Note that a clone is assumed to be
present if at least two cells have the same abnormal or increased chromo-
some or at least three cells with the same gained chromosome, assuming
that random loss is not at a high frequency (seeNote 4).

4. Notes
1. Blood cultures are used for the following purposes:
a. T-Lymphocytes which are cultured with PHA (0.1 mL/lO mL
culture) in the presence of monocytes for at least 48 h are
Cytogenetic Techniques for Human Leukemias 409

transformed and start dividing. These cells can usually be used

to establish the patient’s constitutional karyotype, against
which any acquired abnormalities found can be compared.
Other tissues, e.g., skin, may be used instead, but blood lym-
phocytes are the most convenient. If the patient is on chemo-
therapy or the T-cells are affected by the disease, then PHA-
stimulated cultures may not be successful. A Philadelphia-
chromosome-positive result may sometimes be obtained from
PHA-stimulated T-lymphocytes, useful if other cultures fail.
b. In the leukemias, cultures of leukocytes unstimulated by any
mitogens may be useful if marrow is not available or has failed
to produce a result. Mitoses may be found in unstimulated
cultures in approximately 10% of cases that have primitive cells
in the blood from the bone marrow or extramedullary sites of
hemopoiesis. If a chromosomally abnormal clone is found,
these cells can be scored as bone marrow cells. It must be borne
in mind, however, that some of these cells may have left the
marrow weeks or months previously, so that the population of
mitoses found may not represent an up-to-date picture of the
state of the marrow. Mitoses also occur as a result of in vivo
transformation of normal lymphocytes, e.g., as a result of in-
fection, so the origin of cells with normal chromosomes cannot
generally be determined. These must be interpreted with care.
Other malignancies, e.g., multiple myeloma, may also release
dividing cells into the blood.
c. Cytogenetic studies of CLL and B cell disorders can be made on
4-5-d blood cultures, in preference to marrow, stimulated with
pokeweed mitogen, which transforms both B and T lympho-
cytes, or TPA, which is specific for B cells. Use of TPA also
reduces the toxic effect of colchicine on CLL cells (15). Difficult
cases may respond to “cocktails” of two or three mitogens.
2. The methotrexate (MTX) method of obtaining better-quality chromo-
somes was introduced for blood lymphocytes and adapted for other
tissues (16). Although promising results were initially claimed for
hematologic disorders, the method is falling into disuse since it does
not appear to be helpful in most cases. MTX was thought to arrest cell
division at a stage when thymidine is required, thus bringing divid-
ing cells into synchrony, the block being later released by adding
thymidine, The results are now thought to derive from a slowing-
410 Swansbury

down of early S-phase, retarding the contraction of the chromosomes,

or through an effect on the chromosome morphology and the cell
membrane. The time after rescuing with thymidine is different for
different tissues. A release of 5 h was defined for normal lympho-
cytes. Gallo et al. (27) have shown that the time should be 9.5-11.5 h
for myeloid and leukemic cells; Morris and Fitzgerald (18) have
shown that the time varies between patients and that the cell cycle
time is generally shorter in CML than ANLL.
3. The fluorodeoxyuridine (FdUr) (29) method is gaining popularity
over the MTX procedure, particularly for ALL. FdUr is less toxic than
MTX, so it is not necessary to wash the cells after exposure and cell loss
is reduced.
a. Culture the sample for 6-24 h in a low-thymidine medium (i.e.,
not Ham’s FlO or F12; seeAppendix, this vol).
b. Add FdUr + Uridine stock solution: 0.1 mL/lO mL culture.
c. Incubate overnight.
d. Add thymidine or bromodeoxyuridine (BrdU: stock solution
200 mg/mL) to release the block or spin down and resuspend
in Ham’s FlO or F12.
e. Harvest 7-8 h later, with colcemid for the last half hour.
f. Further processing is as already described.
4. Interpretation of results: Most centers find that about 50% of ANLLs
have a detectable chromosome abnormality, rising to about 90% if
enough cells can be analyzed. It has been suggested that all cases may
eventually be shown to have an abnormality, but in practice there will
always be some cases in which an abnormality cannot be demonstra-
ted. Detection of a karyotypically abnormal clone is evidence for the
presence of a malignancy. Note that it doesn’t automatically mean
that the patient has a poor prognosis: some abnormalities are actually
associated with a better prognosis than a normal karyotype. The
finding of only karyotypically normal cells does not, however, mean
that there is no malignant clone present. If the patient has acute
leukemia and has had any cytotoxic therapy, for example, then it is
highly probable that any clone will become undetectable until the
disease relapses and so a normal result will be obtained. In contrast,
in CML, the Philadelphia chromosome persists after treatment but
subsequently acquired abnormalities disappear until the acute phase
supervenes.
Cytogenetic Techniques for Human Leukemias 411

Bear in mind the possibility that the clone was present, but not
detected: it has been estimated that in an adult about 40 thousand
million (billion) new marrow cells are produced every hour, so only
a very small proportion is being examined. The more cells examined,
the more confident you can be that there isn’t a clone present. At least
25 metaphases should be analyzed, whenever possible, unless the
presence of a clone can be established with fewer metaphases. It has
been calculated that if 29 cells are normal then you can be 95%
confident that a clone involving 10% or more of the population is not
present (23). Even if a cloneis detected after analyzing just a few cells,
more should be analyzed in caseof clonal variation and even multiple
clones.
In some cases karyotypically normal cells of good morphology are
found on the same slide as karyotypically abnormal cells of poor
morphology. When two morphologically distinct populations occur
without any detectable chromosome abnormality, one may assume
that the poorer cells derive from the leukemic clone. However, the
converse, that the good-morphology, karyotypically normal cells are
not leukemic, does not follow. In fact, in some instances, e.g., ANLL
subsequent to a myelodysplasia, it is quite likely that all the divisions,
abnormal or apparently normal, derive from the neoplasia.
It can be discouraging to work in leukemia cytogenetics when even
experienced workers can find that up to one third of samples may fail
to give a useful result because of (a) inadequate material, (b) absence
of mitoses, or (c) mitoses being completely unanalyzable. The work
remains interesting, however, because of the high frequency and
variety of chromosome abnormalities and because there is still so
much to be learned about their biological and clinical significance.
5. Use of frozen material: The custom, in some hospitals, of freezing
samples of blood and marrow in liquid nitrogen has been of great help
to those working with DNA. It is also possible to perform cytogenetic
studies on frozen material if sufficient care is taken in the freezing and
thawing processes. A minimum procedure is given here: refinements
are according to the facilities available.
a. Freezing: Suspend the cells in: 10% Dimethylsulfoxide (DMSO),
20% serum and 70% RPM1 or other medium. Start freezing
straight away: 1 degree/min for 30 min then 3 degrees/min to
the end.
412 Swansbury

b. Thawing: Warm up quickly (shake the vial in warm water) and

wash with medium + serum to remove the toxic DMSO.
Cultures may need an extra day before harvesting.
Although some samples fail to recover, this Department has occa-
sionally obtained chromosomes from frozen material that were of better
morphology than those from a fresh portion of the same biopsy.

Acknowledgments
Methodology is a constant topic of conversation among cytogen-
eticists and I am indebted to the many, particularly fellow-workers in this
Department, who have contributed to this chapter by sharing experiences,
theories and advice.
I am also grateful to Professor Sylvia Lawler, former head of the
Department of Cytogenetics, and the Royal Marsden Hospital, London,
UK, under whose auspices this chapter was written.

References
2. Groffen, J., Stephensen, J. R., Heisterkamp, N., de Klein, A., Bartram, C. R., and
Grosveld, G. (1984) Philadelphia chromosome breakpoints are clustered within a
limited region-bcr--on chromosome 22. Cell 36,93-99.
2. Bartram, C. R. and Carbonell, F. (1986) bcr rearrangement in Ph-negative CML.
Cancer Genet. Cyfogenet. 21,183,184.
3. Harrison, C. J., Fitchett, M., Potter, A. M., and Swansbury, G. J. (1987) A guide to
cytogenetic studies in haematological disorders. Eugenics Society OccusiomZ Papers,
New Series No. 1.
4. Berger, R., Bernheim, A., Daniel, M. T., Valensi, F., and Flandrin, G. (1983) Cyto-
logical types of mitoses and chromosome abnormalities in acute leukemia. I..e~kx~&
Res. 7,221-235.
5. Keinanen, M., Knuutila, S., Bloomfield, C. D., Elonen, E., and de la Chapelle, A.
(1986) The proportion of mitoses in different cell lineages changes during short-term
culture of normal human bone marrow. Blood 67,1240-1243.
6. Fialkov, P. J., Singer, J. W., Adamson, J. W., Vaidya, K., Dow, L. W., Ochs, J., and
Moohr, J. W. (1981) Acute non-lymphocytic leukemia: heterogeneity of stem-cell
origin. Blood 57,1068-1073.
7. Mitelman, F. and Levin, G. (1978) Clustering of aberrations to specific chromo-
somes in human neoplasms. III. Incidence and geographic distribution of chromo-
some aberrations in 856 cases. Hereditus 89,207-232.
8. Swansbury, G. J., Feary, S. W., Clink, H. M., and Lawler, S. D. (1985) Cytogenetics
of acute promyelocytic leukemia: incidence of t(15; 17) at the Royal Marsden Hos-
pital, London. Leukemia Res. 9,271-278.
9. Williams, D. L., Harris, A., Williams, K. J., Brosius, M. J., and Lemonds, W. (1984) A
Cytogenetic Techniques for Human Leukemias 413

direct bone marrow chromosome technique for acute lymphoblastic leukemia.

Cancer Genet. Cytogenet.13,239-257.
10. Wiley, J. E., Sargent, L. M., Inhirn, S. L., and Meisner, L. F. (1984) Comparison of
prometaphase chromosome techniques with emphasis on the role of colcemid. In
Vitro 20,937-941.
11. Misawa, S., Horiike, S., Taniwaki, M., Abe, T., and Takino, T. (1986) Prefixation
treatment with ethidium bromide for high resolution banding analysis of chromo-
somes from cultured human bone marrow cells. Cancer Genet. Cytogenet. 22,
319-329.
22. ISCN (1978) An international system for human cytogenetic nomenclature. Birth
Defects: original article series, XIV, No. 8, New York, The National Foundation.
23. Rooney, D. E. and Czepulkowski, B. H. (1986) Human Cytogenetics: a practical
approach. IRL Press,Oxford, UK and Washington, DC.
14. Sumner, A. T., Evans, H. J., and Buckland, K. A. (1971) A new technique for dis-
tinguishing between human chromosomes. Nature (Nezu Biology) 232,31,32.
15. Connor, T. W. E. (1985)Phorbol ester-induced lossof colchicine sensitivity in chronic
lymphocytic leukemia lymphocytes. Leuk. Res. 9,885-895.
16. Yunis, J.J.(1981)Newchromosome techniquesinthestudyof humanneoplasia. Hu-
man Pathol. 12,540-549.
17. Gallo, J.H., Ordonez, J. V., Grown, G. E., and Testa, J. R. (1984) Synchronization of
human leukemic cells: relevance for high-resolution banding. Hum. Genet. 66,
220-224.
18 Morris, C. M. and Fitzgerald, l?.H. (1985) An evaluation of high resolution chromo-
some banding of hematologic cells by methotrexate synchronization and thymi-
dine release. Cancer Genet. Cytogenet. 14,275-284.
19. Webber, L. M. and Garson, 0. M. (19831Fluorodeoxyuridine synchronization of
bone marrow cultures. Cancer Genet. Cytogenet. f&123-132.
20. Schweizer, D. (1980) Simultaneous fluorescent staining of R bands and specific
heterochromatic regions (DA-DAPI bands) in human chromosomes. Cytogenet. Cell
Genet. 27,190-193.
21. Caspersson, T., Gahrton, G., Lindsten, J., and Zech, L. (1970) Identification of the
Philadelphia chromosome as a number 22 by quinacrine mustard fluorescence
analysis. Exp. Cell Res. 63,238-240.
22. Verma, R. S.and Lubs, H. A. (1975) A simple R banding technic. Am. J. Hum. Genet.
27,11&117.
23. Hook, E. B.(1977) Exclusion of chromosomal mosaicism: tables of 90%,95% and 99%
confidence limits and comments on use. Am. 1. Hum. Genet. 29,94-97.
 Chapter 33

Time-Lapse
Cinemicroscopy

Peter N. RiddZe
1. Introduction
Cinematography commenced as a scientific technique used as a
system for “slowing down” observed movement. Marey in 1888 (1) con-
structed, following a number of other ideas, a “Chambre Chronophoto-
graphique,” which had practically all the elements of the modern tine
camera, With this he made serial photographs (not transparencies) of
various biological phenomena (2).
The mounting of the tine camera to the microscope was achieved
shortly after the turn of the century, and apparatus producing film that
would be difficult to surpass today was in use in the early 1920s. No
opportunity to see the films made by such workers as Canti at that time
should be missed. Such equipment in fact differs from present-day cine-
microscopes only in general sophistication and the (possibly excessive) use
of “chips” where motor-driven switches are quite adequate. (A motor-
driven control unit lives at the back of my laboratory shelf and is still
brought out when the electronic units fail. It is more clumsy to use, but it
does work on and on and on with no attention.) Also, more sophisticated
optical systems such as phase contrast and DIC have become available
415
416 Riddle

since earlier cinemicroscopes came into use. It is less obvious that some
older but equally valuable methods have been largely forgotten in the
flurry of modern developments. An example is Rheinberg illumination
(3), which is k nown to relatively few users of microscopes (as opposed to
microscopists). Similarly, it is rarely remembered that a thin biological
subject can often be seen in considerable detail using simple but properly
adjusted bright field optics. Most workers turn to the phase contrast
(despite its inherent artifacts) without considering possible alternatives.
Thus, much of the work done now would have been possible with equip-
ment available half a century or more ago. It is hoped that this chapter may
show that the system is far from demanding, and that some dusty but per-
fectly adequate equipment will be brought back into use from laboratory
cupboards.
The maxim that seeing is believing is frequently difficult to apply to
scientific observation. So often, results are inferred from effects that are
secondary to the data actually sought. One example of this is the estimation
of cell growth by measuring the incorporation of Tritiated Thymidine in
the (not unreasonable) hope that the number of cells entering the S phase
is an accurate predictor of ensuing divisions. Such a procedure not only
loses the culture, but also adds extra variables to the system that may cause
changes to the very parameter being measured (4). In contrast, cinemicro-
graphic recording of cellular activity yields a record that can be seen di-
rectly while not destroying the specimen. Data can be reassessed later for
different parameters, and last but not least, the visual record often allows
unsuspected activity to be seen and investigated if relevant.

1 .I. Basic Considerations

The theory of conventional cinematography is that a series of individ-
ual photographs (transparencies) are taken with a brief time spacing (usu-
ally l/25 s) apart. When projected at the same time spacing and in register,
the image appears to move in the same way and at the same rate as the
original. The fact that the record consists of disconnected images is hidden
by the persistence of vision of the observer. If the speed of projection is
greater than the speed used when filming, the apparent rate of movement
of the image may be changed either to appear slightly more rapid as in the
“Keystone Cops” or increased much more, as for cinemicroscopy of cells
in tissue culture. Alternatively, if the filming interval is shorter than the
projection interval, the movement is slowed and high-speed films can even
record the passage of a bullet. When observing cells in tissue culture, it is
Time-Lapse Cinemicroscopy 417

not unusual to have a time-lapse interval of between 1 and 300 s which

causes an apparent increase in rate of between 25x and 7500x. Films taken
at these intervals show movement and serial divisions. The development
of cell clones can be recorded with the longer intervals so that a 30-m length
of film can record over 2 wk of activity. Thus, the time-lapse cinemicro-
scope offers a method to “compress” time and subsequently extract a
vastly increased amount of data.
The majority of movements within cells and cultures are so slow that
it is difficult to perceive them at all by eye, although the associated bio-
chemical phenomenamay occur much more rapidly. With gross transloca-
tional movement, even the fastest moving of mammalian cells, leukocytes,
do so at speeds measurable in pm/min (5), whereas fibroblasts seldom ex-
ceed 60 l..tm/h (6); many remain practically stationary. These rates are con-
siderably affected by environmental factors such as medium, substrates,
temperature, toxins, and so on (7). The division of the majority of cells from
start to finish of mitosis takes about 1 h, whereas in contrast, the whole
period between cell divisions lasts from 8 h upward, for example (8). Un-
fortunately, it is rarely possible to obtain worthwhile data about one slow
and one fast parameter on the same film. It is essential to recall, when
looking for rapid changes in a film that has been taken at a long time inter-
val, that the individual frames are brief records of conditions at widely
separated times and that, in consequence, there are many intermediate
situations that will not have been recorded.
Optical considerations also affect changing appearance and cell be-
havior as seen at the projection screen. The most important of these is the
relation of total magnification of the image to the time interval. For a par-
ticular interval, the speed of a point on the screen travels proportionately
to the linear magnification. The image of a cell organelle that may travel
10 cm on the screen when photographed with a 10x objective will move
1 m in the same time if a 100x lens is used. Although these considerations
appear obvious, they can be overlooked when time intervals and magni-
fications are being chosen at the start of an experiment. What data are re-
quired from the film dictates the parameters when filming, but the basic
principles are similar for all time-lapse photography.

2. Methods
Cultures for cinemicroscopy are assumed here to be already growing
in a suitable container (see below) and ready for examination.
418 Riddle

2.1. Filming
2.1.1. Parameters
Before any further bench work is attempted, the experimental param-
eters to be studied should be listed. Without these details, it is not possible
to choose suitable microscope and camera settings. Many biologists want
to “have a film” of some phenomenon without previous thought of what
they want from the result. Many even seem surprised that films can be of
greater analytical than pictorial value. Such unplanned material is seldom
of little worth other than of idle enjoyment to those who have not seen such
films before. Sound technique demands an accurate appreciation of what
must be filmed, whether it is the length of time between cell divisions, the
rate of cell movement, or another feature.
2.1.2. Equipment
The microscope and photographic equipment for cinemicroscopy can
be purchased, in most cases, from the manufacturer as a unit. Alterna-
tively, it can be assembled piecemeal from components of one’s own pref-
erence. Experience with both approaches suggests that the former is likely
to give more consistent results and the latter more flexibility (and satisfac-
tion). It is essential that the system should be dedicated to this particular
use since films can take a long time to produce and cannot be casually inter-
rupted for irrelevant purposes.
The entire system, however, requires the following basic elements
listed “outward” from the specimen:
1. specimen chamber
2. environmental control for specimen
3. microscope with appropriate lens systems and illumination
4. comfortable operator seating
5. tine camera
6. film
2.1.2.1. Specimen Chambers. More than ten years ago in the early
days of science, articles on time-lapse cinemicroscopy usually contained a
description of at least one typeof cell growth chamber designed for specific
observations. Generally, these consisted of two coverslips or glass plates
separated by a washer or gasket to contain a minute amount of medium
with the cells adherent to one surface. Microscopes generally used were of
conventional rather than theinverted form and required the specimen cells
Time-Lapse Cinemicroscopy 419

(and the medium) to be on the upper surface and not where gravity nor-
mally dictated. This involved growing the cells on the surface to be ob-
served and then inverting it. Medium was contained by surface tension or
total filling of the space. Some special chambers were, however, essentially
different in that they could provide conditions such as raised pressure,
which was not possible in other designs (e.g., 9,10).
The more general use of the inverted microscopes that are now offered
by the majority of microscope manufacturers has removed the need for
growing the cells on the ceiling. Conventional Petri dishes or plastic cul-
tureflasks canbeputonto themicroscopestage. Lossof detail by distortion
from less than perfect plastic is seldom a problem, and optical distortion
resulting from increased thickness can be compensated for by the use of
lenses corrected for thicker substrates. If plastic surfaces are found to be
inadequate, there are optically flat glass Petri dishes in which the base is
designed to suit the long focus of such special lenses (seeNote I).
Use of the same type of container as that in which other cells are grown
in parallel experiments removes the risk of differences, resulting from sub-
strates, both in cell behavior and in more subtle effects. In the author’s
opinion, there is seldom any need now to venture into the realms of special
chambers as long as inverted microscopes are used (for those who wish to
use special chambers (see10). It should be noted also that the use of conven-
tional plastic dishes is not possible for high-magnification photography
with oil immersion objectives. Even so, high magnification is now possible
if dishes with bases made of thin stretched plastic membranes are used.
These are gas permeable but oil resistant (Petriperm, made by Heraeus),
and allow the immersion objectives to come as close to the cells as do con-
ventional coverslips. The image quality is occasionally somewhat marred
by faults in the plastic, which is also easily damaged, but such faults, even
if an unspoiled area cannot be found, are usually outside the depth of field
of the system when the specimen is in focus.
2.1.2.2. Environmental Control
2.1.2.2.1. TEMPERATURE.Temperature requirements for growth are
well known for most cell lines, although in the author’s experience, most
laboratory incubators seem to be set at the “normal” human temperature
of 37OC, rather than that appropriate to any other species being grown.
Usually, provided the temperature is stable the cells do not appear to
suffer. In addition, the ability to change the temperature of the culture en-
vironment rapidly may be required experimentally with, for example,
420 Riddle

temperature-sensitive mutants. Several methods are available for heating

cultures. The once popular “heated stage” has drawbacks for long-term
work, and the crucial part of the culture, that under observation, gets no
heating at all. The most useful systems are those that enclose the micro-
scope stage or the stage and as much of the body as is physically necessary.
A box of Perspex with doors to allow access to the controls and insertion
of the culture is convenient and easy to construct. Some manufacturers
supply these “incubators” with the microscopes, but it is as well to ensure
that the design was not created around small hands, making the access to
the stage very difficult. With large access ports opened, it has been found
that temperature fall is seldom significant, provided that adequate heat-
ing backup exists. It is as well to be sure that the available stage equipment
will not need to be replaced with something larger, requiring complete
rebuilding.
Temperature within the box can be maintained by any available heat
source, provided that it does not cause local overheating or vibration.
Heaters inside the box can be the source of considerable convection and
poor distribution unless a fan is included too. Although fans that are nearly
vibration free are available for this role, the risk of slight vibration evident
at high magnification remains, especially if the subject is suspended in a gel
medium and not attached.
External blown heat has been found to be an excellent way to warm the
incubator box. The domestic hair dryers mounted in a box with a long flex-
ible tube connected to a plastic bag for the head have proven to be both
incredibly long lasting and easy and cheap to convert to this use. The tube
is led to the incubator box through a hole and an air baffle placed within to
disperse the hot-air stream. The fan motor is run continuously, often for
years on end (its life being mainly limited by environmental dirt). Motor
maintenance, of occasional lubrication, should prevent the motor from
seizing up, overheating, and in serious cases, melting through the bottom
of the box. The heating element is controlled by a solid-state thermostat (9)
and a simple thermistor device has given temperatures stable at the stage
to within O.lOC as indicated by a conventional thermometer. This is ade-
quate for the most demanding tissue culture. A hair dryer heating unit
when used in this way causes no vibration to the microscope unless it
shares the same bench. In practice, mounting the heaters on a separate,
wall-mounted shelf beneath the microscope bench, with the pipes coming
up through holes, is both convenient and adequate (seeNote 2).
2.1.2.2.2. GAS. The medium in use dictates the gas requirement.
Some media containing buffers such as HEPES or TRIS, which are used in
Time-Lapse Cinemicroscopy 421

the absence of any carbon dioxide, might appear at first sight to be the ideal
solution to the problem of pH control during cinematography. However,
there is little point in introducing what can only be described as a further
variable into conditions that are of themselves different to those in which
the cells are normally grown. With sealed culture chambers, there may s till
be a need for a minute flow of gas past the drop of medium to compensate
for any loss of carbon dioxide. The Roberts/Trevan chamber (II) is one
that is designed to have the medium in the center of the space between the
cover slips with the gas circulating around it. Such gas flow must be very
small (of the order of 1 cc/5 min) and has proven very difficult to regulate
and measure with any accuracy.
Gas supplies cannot be fed directly to the incubator box surrounding
the microscope, since problems of leakage, corrosion, and condensation
demand that the humid gas be restricted to a smaller, sealed space. It has
been found that Petri dishes can be gassed adequately in boxes small
enough to fit on the stage. These have optically flat glass or plastic win-
dows in the lid and open viewing ports in the base. The holes are covered
by the dish. Two such designs are shown (Fig. la and lb), and variations
could be made according to particular circumstances. The square design
was thought to be unlikely to be sufficiently gas tight to maintain pH, but
in practice it has been found to be simple, easy to clean, and effective. The
alternative stainless-steel circular boxes (which are far more elegant) are
little if at all more effective. Neither chamber can be purchased commer-
cially but can be made from readily available materials (seeNote 3).
2.1.2.2.3. HUMIDITY. Whether small or large, the gassed volume must
be maintained as near 100% relative humidity as possible to avoid concen-
tration of the medium with all the untoward chemical and biological
changes. This humidity is almost impossible to achieve, but something
very close can be reached by: (1) placing a reservoir of water close to or
around the culture, and (2) bubbling the incoming gas supply through a
Wolff bottle containing water. It is desirable to heat the water slightly
above the temperature of the culture to slightly super-saturate the gas.
Extra complications can be avoided by mounting the humidifier bottle
within the incubator, but even then a small amount of additional heat is
needed to compensate for evaporative cooling. A torch bulb in a “finger”
worked into the side or base of the bottle (1 or 2 W) is usually enough. A
wire around the bottle would serve equally well or a painted on heating
element as used in ref. 12. Any connecting pipes between the humidifier
bottle and the environmental chamber must be kept short, of small internal
diameter, and warmed to avoid moisture resulting from condensation in
422 Riddle

4a

‘4ia

Fig. la. Circular metal environmental chamber. Above, lid seen vertically. Middle,
chamber in section. Below, base from above. (1) Optically flat glass insert (cemented in),
(2) stainless-steel top, (3) inlet hole for humidified gas, (4a) hole for locating pin, (4b) pin,
(5) metal base, (6) water reservoir in moat form, (7) viewing port.

the tube. The tube itself should be of PVC or polythene, rather than silicone
rubber since the latter is somewhat permeable to carbon dioxide.
A warmed chamber of water, such as the humidifier reservoir, does
tend to acquire unwanted flora and fauna in a remarkably short time. Cop-
per sulfate in the reservoir (1% would be adequate) is nonvolatile and
Time-Lapse Cinemicroscopy 423
lb

06
2
,--- I
I 4 I I
,-- -1

/ -\
8 / \\I
5

\0-A//
‘;::j
0
-t&l-
4

6
7

!
Fig. lb. Dismountable square environmental chamber. Above, dismantled. Middle,
assembled, in section. Bottom, vertical appearance. (1) Glasslid, (2) perspex body, (3) alu-
minum alloy base, (4) water reservoir (bottle lid), (5) recessed corner for stage clips, (6)
inlet hole for gas supply, (7) viewing port, (8) spring clip (ex. microscope stage) on ex-
tended spigot.
424 Riddle

suppresses such growth, but it reacts with the incoming carbon dioxide to
form a very unsightly blue/green precipitate. Potassium dichromate has
been found to be equally effective, and the same reservoir solute is simply
topped up to replace evaporation losses without any need to clean out the
bottle. More sophisticated antibacterial agents, e.g., Dichlorophen have
proved effective but quite unnecessary.
2.1.2.2.4. LIGHT. Some types of cells are sensitive to light, and labor-
atory lights should be dim except when setting up the apparatus. An al-
ternative light source from a low-power lamp not only keeps sensitive
cultures in the best condition, but also helps viewing when only low levels
of microscope light are available for viewing (seeNote4). “Environmental”
light from the laboratory is of low intensity at the culture surface compared
to the amount that assaults the cells each time the microscope lamp is
turned on. The effects of this concentrated beam of light can sometimes be
seen as physical damage (inactivity or even death), but such extremes are
more usually the result of local heating in the light path close to the micro-
scope condenser. It is salutory to place one’s hand in the same position as
a culture to understand this.
There are two ways to counteract the heating resulting from the con-
ventional tungsten light beam: (I) insert a heat filter (heat absorbing or di-
chroic heat reflecting glass) in the light path, and/or (2) have the light
source playing on the culture for minimum periods of film exposure and
observation only (13).
Most modern camera control units have means to switch the illumi-
nating beam on or operate a shutter in the light path. If the latter method
is used, the lamp should be allowed to heat up to full brightness before the
exposure is made. If this is not done, exposures made when the lamp is on
for observation may be significantly more heavily exposed than those
when the switching is in operation. Film made then has sharp variations
in intensity between frames, which is very distracting during projection.
2.1.2.2.5. VIBRATION. Vibration is an ever-present artifact that is very
difficult to eliminate. It affects both cultures and microscopes.
Some (most) cultures when settled on the culture dish are little af-
fected by background vibration, but those that settle slowly take much
longer to settle and may even be made to collect in the center of the dish if
the movement is severe. The effect of vibration more usually considered
is that upon the equipment itself. Microscope images are spoiled by vibra-
tion especially at high magnification. Some types of culture are relatively
resistant to such effects, but the image of cells grown in agar gel can become
Time-Lapse Cinemicroscopy 425

a blur if the natural period of resonance of the system approaches the

vibration frequency. Elimination of vibration depends upon its source and
the way in which the equipment is mounted. The first precaution when
setting up such apparatus is to ensure that the bench is adequate. Obser-
vation of a sensitive specimen (e.g., at high magnification) soon settles the
question of how much of the image is being spoiled. It should be realized
that some forms of vibration are relatively short-lived and may last only a
few seconds. Others are persistent, or only occur when the microscope is
in use. The persistent vibration that affects a building (traffic shake, ele-
vators, central heating fans, and so on) is often only present in certain
places, and a search reveals that more suitable siting of sensitive equip-
ment is the best way to overcome the problem (seeNote 5).
When a suitable position is found (a basement often being ideal), the
use of a firm bench is advisable. It should be free-standing if possible and
heavy, although exceptions have been seen where cinemicroscopes have
been mounted on relatively flimsy tables and yet showed no evidence of
vibration. Extremely heavy and relatively inexpensive benches can be
made from concrete garden furniture, and a “pub” table weighing some
100 kg, with the seat removed has proved very stable indeed. Vibration
caused by the electrical equipment in use (intervalometer cooling fans,
heaters, and so on) is best eliminated by placing the offending source on a
separate shelf.
Finally, there is a vibration caused by the camera shutter mechanism.
This is brief, and occurs at the beginning and end of an exposure. Where
the camera and microscope are integal, this cannot be eliminated by chang-
ing the bench, although where the camera is mounted on a separate pillar
(not necessary in the author’s opinion), the two units can be separated on
two distinct surfaces. (The relative movement of camera and microscope
form a separate problem here.)
Cures for camera-induced shake are: (1) using a relatively long ex-
posure. Short periods of vibration occur while the camera and microscope
are still moving because of the operation of the camera shutter. A longer
exposure (greater than l/2 s) allows the majority of the exposure to be
taken after movement has died away, (2) using flash exposures that freeze
the relative movement. Although this measure ensures that the image on
each frame is not blurred, it does not compensate for differences in register
caused by vibration between each frame. None of these measures are
needed for the majority of low-power cinemicrographs where surprising
stability is achieved on the simple bench.
426 Riddle

Fig. 2. Inverted microscope with tine camera mounted on triple eyepiece. Camera
drive is situated on the camera body and the projection from the microscope is used to in-
sert titles. Camera control is governed by the “box” on the right.

2.1.2.3. Microscopes. It has already been mentioned that the conven-

tional microscopes, with objectives above the stage, are of less use than the
inverted types to the biological cinemicroscopist. Earlier examples of in-
verted cinemicroscopes were conversions of inexpensive conventional
types (14), and the modern instruments have only been common labora-
tory equipment for a relatively few years. However, unless special cell
chambers are to be used, they are essential. Many of the well-known man-
ufacturers now offer inverted microscopes, some with tine attachments
and control units (Fig. 2). Choice is as much a matter of preference as avail-
able budget. Units from the larger makers do not differ very greatly in
terms of their potential optical quality-which is always high.
Time-Lapse Cinemicroscopy 427

It should be remembered that, if a culture is grown in a plastic dish of

less than perfect optical quality and where the thickness is not especially
matched to the objective lens, then no amount of microscope “superiority”
will improve on the image obtained. Also, the image is usually quite ade-
quate for analytical purposes, even if less than perfect photographically.
Choice should depend upon such features as accessibility of controls,
robustness, ease of cleaning, and available accessories such as a separate
still (PolaroidTM) camera. Optical components must be adequate. Of the
accessories to check, a very long working distance condenser for phase
contrast (clearance at least 2.5 cm) and very low magnification objectives
with phase contrast are probably the most important. This microscope
must be dedicated to its time-lapse role, preferably used by only one or two
people. Any wider use nearly always leads to deterioration of the micro-
scope and consequently to inferior or even ruined film. It cannot be over-
emphasized that, if second, third, or more cinemicroscopy units are avail-
able, the experimental time can be cut disproportionately. The single unit
can never produce a true control film to compare with one of a test situation
because it is made at a different time. For further reading upon microscopy
in general and biological film production, the reader is referred to James (3)
and Michaelis (2) (seeNote 6).
2.1.2.4. Seating. Films show progress of the subject afterward. How-
ever, prolonged periods of observation by eye are inevitable, especially
when the subject changes frequently. Suchobservationis an invaluable aid
for personal “education” about cells. There are few subjects more exciting
to observe than the mitosis of a cell seen at high magnification, yet only a
very few biologists can claim to have watched this elemental phenomenon.
To do so, however, requires some hours on the stool by the microscope.
Similarly, for observation and during the setting up and adjustment for
filming, it is essential to have comfortable seating. The most important fac-
tor is to have seat and eyepieces at a relative height that allows the user to
sit comfortably without the back being stretched or arched. Only experi-
ence with mismatched combinations of laboratory stool and microscope
can impress this fact on the mind and elsewhere.
2.1.2.5. Cameras. A tine camera can be mounted permanentlyon the
microscope using the microscope body to support or at least hold the
camera body. There must be integral viewing to see and focus the actual
image for the film. The image seen through the normal microscope lenses
covers a larger field and is seldom in the same plane of focus. The majority
of bioscientific filming is currently done on 16-mm film since the format
428 Riddle

and size are both convenient for handling and storage and are about the
minimum that will give results of acceptable standards. The camera func-
tions, namely time between exposures and length of exposure, together
with the film advance are controlled by an Intervalometer, recently pro-
duced units being electronic rather than electromechanical. These may be
small and mounted upon the camera itself or almost ludicrously large and
smothered with flashing lights. It is usual to incorporate here a means to
“run the film on” without delay, and in the more modern instruments,
automatic exposure control and even an indication of color temperature
for the users of color film.
The cameras do not need to be capable of functioning as normal cine-
cameras, because the only part of the mechanism (usually) used in time-
lapse work is the film advance and shutter. The remaining components
(for focusing and motor speed control) are not used. Where the camera
control unit operates a motor drive that is fitted to a conventional camera
back, the motor and control components are not used. As a result, it is
sometimes possible to obtain the “camera back,” i.e., all but the motor drive
and without any lens systems, for a price that is far less than that of the
entire camera. Choice of camera is often dictated by the microscope manu-
facturer, but several types of varying degrees of sophistication are avail-
able. For laboratory work, it is unlikely that anything more than the most
basic and simple camera back is required (seeNote 7).
2.1.2.6. Film. Film is available in both color and black and white
types. For the majority of routine cell culture uses, black and white is ade-
quate. Most culture specimens have no natural color and the image is
monochromatic if obtained by phase contrast. Admittedly, the color ef-
fects produced by DIC (Differential Interference Contrast) are sometimes
very elegant, but that often indicates that the system is not being used
properly!
As the majority of tine cameras are 16-mm format, the film does need
to be of fine grain to avoid loss of detail. Also, the contrast of the optical sys-
tem is such that a low-contrast film is not adequate for negative viewing
(see Note 8) (although the additional contrast achieved during printing
makes more conventional types suitable, the grain is more pronounced
than that of microfilm). For this reason, it has been found that the film
emulsion used in microfilms is adequately fast, practically free of grain,
and nowadays produced on a very strong base that avoids damage in the
projector. One such film (if not the only one now available in perforated
form) is KodakTM Infocapture AHU microfilm 1454. Kodak has ceased to
Time-Lapse Cinemicroscopy 429

market this microfilm (in UK if nowhere else) in perforated form. Should

a large enough order be required, the company has been willing to per-
forate a special batch. This has given most satisfactory results.
Closed circuit videorecording of microscope images is also amenable
to the time-lapse mode. The method currently used involves stopping and
starting the movement of the normally continuously moving tape, while
the rotating tape head over which the tape passes during recording re-
mains in motion. This causes relatively rapid wear of that most expensive
component of the equipment. Also, the tape suffers. In addition, reverse
viewing of the resulting record is not as convenient as with film, except
where very expensive videorecorders are used. There is the compensating
advantage in that the record is available immediately, but given in-lab pro-
cessing facilities for film, the differences are not great. These factors may
not be of great significance where very short time intervals are required,
but the longer intervals more often associated with the filming of tissuecul-
tures can produce problems. Nevertheless, the method is used. The great-
est disadvantage in videorecording lies in the rate at which changes are
made to standard techniques. X-mm tine film is much the same now as
it was 50 years ago and will still fit the projectors of that era, Video tech-
niques are changing, and it is quite possible for two adjacent laboratories
to be using incompatible apparatus. At present, the VHS system seems to
be the most popular, but what the industry is going to offer us 10 years
hence is open to debate.
2.1.2.6.1. DEVELOPING FACILITIES. Where only a small number of
films are produced from time to time, development of the film should be
entrusted to a specialist film laboratory. It should be noted that some of
these now deal exclusively with color film, and obtaining first-class prints
or copies of black and white films is much more difficult than some few
years ago. Within the laboratory, short lengths (up to 10 m) can be devel-
oped in some special spiral developing tanks. This length is too short for
most subjects, but there is one type that is capable of accepting a full 30-m
length. Even so, the film has to be cut halfway along its length, which does
not improve its analytical potential. More usefully, bench top developing
units are available, and if space (not necessarily in a dark room) can be
found, then the time from camera to projector can be reduced to 20 min. It
is also cheaper provided enough developing is required.
2.1.2.6.2. SPECIMEN MARKER. Filming of cultures can be extended over
many days, and the culture may need to be removed from the microscope
for change of medium or treatment. Proper analysis demands that the
 Riddle

,.-s-4
, 6 1

;z'
\ I
,7
0
:

I 2
3

c
3

4 5
a

3a 3b
Fig. 3. Objective markers. (3a) Eccentric diamond type. Mounted on the nosepiece, the
diamond canbe rotated manually to mark the dish. (1) Diamond scribe, (2) adjustment for
needle position, (3) ring for turning needle, (4) nosepiece thread. (b) ink marker type, (1)
Lid(mustbereplaced whennotinuse),(2) triple fiber-tippenintriangularform, (3)plastic
replaceable pen (Zeiss, Jena), (4) body to hold pen, (5) nosepiece thread.

same area be filmed before and after treatment. It is essential to mark the
area under study to relocate it after such procedures. Simple location by
stage coordinates is rarely adequate. Several microscope manufacturers
offer “objective markers” at a great variety of prices. These units are of two
types: (I) a rotatable diamond (or tungsten carbide) pointer mounted ec-
centrically on a block (Fig. 3a). The block is screwed into the microscope
nosepiece like an objective and is concentric with the optical path. To mark
a selected area, the marker is swung into position and a circle is scribed
around the area under study, (2) a similarly mounted nosepiece block that
bears three felt-tip pens in a triangle (Fig. 3b). These mark the area by sim-
ple pressure, but are not resistant to water or even wiping. It is possible to
create a simple version of such a device by mounting the “0” of a hand
printing set (e.g., John Bull printer) on an old objective and then inking it
before bringing into contact with the surface.
With both methods, when using an inverted microscope, the dish is
lifted from the stage before a mark is made. One way to avoid this is to rack
the microscope condenser down firmly onto the top of the dish (before in-
serting in an environmental chamber) before making the mark. Alterna-
Time-Lapse Cinemicroscopy 431

tively, a sheet of plate glass (warmed to stop condensation) on the top of an

exposed dish will hold it in place.
2.1.2.6.3. TIME SWITCH. Where the intervalometer lacks any means
to stop the film at a given time, a simple household time-switch of the sort
used to turn lamps or heating units on and off is suitable. If a process under
study is going to extend for some inconvenient length of time into the
evening, let alone a weekend, a lot of film can be saved by switching off the
intervalometer at an appropriate time.
2.1.2.6.4. PERFUSION SYSTEMS. Perfusion of the cell culture with fresh
medium would appear at first sight to be the ideal method to obtain either
constant conditions (16) or rapid changes in conditions.
When fitted to a cinemicroscope and perfusing a Petri dish, the
amount of “plumbing’ required can be forbidding. It is necessary to ar-
range a pump to take medium from a reservoir to the dish, warming the
medium en route, and a separate (or integral) pump to suck off medium
from the dish at a predetermined height to maintain level. Failure of either
can lead to a very messy disaster. Gassing of the reservoir and the piping
to the dish is required where flow is slow enough for carbon dioxide escape
through the tube walls. This involves a gassed jacket around the pipe,
inside the pump, and around the culture. The medium must remain sterile
and it is advisable to have a means to stir the medium while in the dish.
2.1.2.6.5. GRATICULE. In many films, the actual screen dimensions are
of vital importance, as in the measurement of cell movement (6). It is
possible to calculate the dimensions of the projected image using knowl-
edge of the various lenses used, the microscope conformation, projection
lens details and its “throw, “ but these parameters are complex to calculate
and relatively inaccurate in the answer obtained. In addition, factors such
as film shrinkage (slight) at development all introduce further errors. It is
always useful to have a list of various field sizes available with the
microscope lenses available, but these are of use only when selecting lenses
during setting up. It is better and absolutely accurate to insert at the
beginning of any film (preferably just after the title and at the same
magnification to be used) a short section of 3-5 frames of the image of a
“stage micrometer.” These are scales photographed or etched on glass and
mounted in a holder like a conventional microscope slide. For most cell
biological purposes, a scale of 1 mm in length, subdivided into 1Oths and
IOOths, gives a suitable image.
2.1.2.6.6. DEMISTING SYSTEM. Where culture chambers have an air
space above the medium, as in those using Petri dishes, any maneuver that
cools the lid or its container will cause misting. This can be so severe that
 Riddle

the specimen is impossible to see because scattered light degrades the per-
formance of the condenser. Before the dish is placed on the microscope, the
mist can be dispelled rapidly, if somewhat hazardously, by gentle applica-
tion of a Bunsen burner flame. For the less adventurous microscopist, a
warm (not hot) plate of glass or metal, kept nearby can be placed on the lid
to dispel1 the mist in a few moments. Such plates can be kept in an
incubator or heated by an electric element (22). Once the dish is on the
microscope, the mist is usually kept at bay by the close proximity of the
warm condenser. This is usually enough to clear already misted dishes if
not noticed before a film was started, but demisting by this method may
take a long time.

2.2. Cinematography
2.2.1. Preliminary Set Up
Itis anadvantage to put a titleonto thefilmbeforesettingup themicro-
scope, especially if several cameras are in use. If there is no optical port
available in the microscope system to this, the camera must be removed
from the microscope and a film made of a title written on a board using a
conventional lens. An alternative system is to make a short length of film
with an “England Finder” on the stage. These finders are slides marked
with circles containing rows of miniature letters and figures. Films can
then be numbered serially from Al, A2, and so on. If the required magni-
fication is known in advance, a short section showing a graticule is invalu-
able. Decisions made after preliminary examination either require the
culture to be removed, the graticule inserted at the selected magnification,
and then the culture to be replaced. An alternative is to put the graticule
onto the film at the end of the film with the risk that there may not be
enough left (e.g., after overnight or weekend filming).
Knowing the basic requirements, the culture is then placed on the
microscope stage and environmental requirements are adjusted. Follow-
ing approximate microscope adjustment, a preliminary observation shows
whether the culture is uniform with evenly distributed cells or whatever
particular feature is to be demonstrated.
The choice of a field almost inevitably introduces considerations of
random fields. In some situations, it is difficult not to “select” a “random”
field, which is a contradiction in its own terms. There is always the “green-
er grass in the next field,” and the only way to avoid such hazards is to en-
sure that there are simply enough cells present, that none are atypical, and
Time-Lapse Cinemicroscopy 433

that the photogenic aspect of the film is, at this stage, considered to be more
important than the scientific.
2.2.2. Location
The field to be filmed must be marked, either by hand using a felt-tip
pen inserted under the stage or with one of the devices described under
equipment. This allows the same field to be located at a later date. A
“l?olaroidTM” picture of the field is useful for accurate location before any
operation involving the removal of the culture. Preferably this should be
made through the viewing eyepiece, so that the eyepiece graticule is super-
imposed upon the picture of the cells. After a procedure involving the re-
moval of the dish, the area can be relocated and orientated so accurately
that it is hard to detect any movement on the subsequent film. A brief blank
of unexposed frames to mark the time of change is valuable. This proce-
dure usually involves a specially adapted PolaroidTM camera. The extra
time involved saves a lot later.
2.2.3. Magnification
The parameters studied dictate the magnification to be used for film-
ing. If estimates of statistical value are needed, the sheer number of cells
in the field is vitally important. Conversely, detailed study of single areas
may need maximum magnification. Knowing the type of image to be seen
from the film, the relevant selection of lenses is made and changes aremade
at either the objective or projection eyepiece (projective) or using either of
the others with a magnification control built into the microscope, which
usually allows linear magnification changes from 1-1.5x.
2.2.4. Eme Interval
Since various cell processes take a great variety of times, the interval
must be chosen accordingly. No rules can be given, but remember that 30m
of film carries 4000 frames plus a “lead” and “tail” each of 200 frames. A
frame interval of 90 s uses 40 frames/h, and one of 60 s uses 60 frames/h.
Whatever procedure needs to be filmed should be placed on a manageable
length of film for analysis. It is time-consuming, boring, and wasteful to
have to run repeatedly over more film than is needed. It has been found
that a length of 10 ft, i.e., 400 frames, gives a manageable rate for analysis
of a single parameter (for example, a complete cycle from one division to
the next). The procedure then takes 16 s to project. Where the same process
is repeated (e.g., repeated cell divisions), the time is added onto and not
compressed into the original time. Full analysis of a film 4000 frames in
434 Riddle

length is possible but incredibly wearisome, largely because of the time

taken in rewinding the film. (There are systems to give continuous circular,
end to end projection of a film, but these are not appropriate for long films.)
It is not possible to study two parameters of incompatible length, e.g.,
the total cell cycle and thelength of mitosis, on single film. Any attempt to
do so results in one proceeding too fast for analysis and the other too slow
for comfort.

2.2.5. Equilibration

Aculture inevitably suffers a relatively violent upheaval on its way to

the microscope and it is desirable to allow it a period to equilibrate with its
new surroundings before starting the film. If nothing else, aPetri dish that
has been in a moist incubator will tend to dry on its lower surface if used
in the system described above (seeEquipment); this differential moisture
situation does lead to slight distortion and loss of focus. Any equilibration
minimizes the effect.

2.2.6. Microscope Adjustment

It is only possible to suggest a system for microscope adjustment in the

most general terms as different equipment varies. However, a strict rou-
tine for the setting up of such systems is invaluable. It is helpful to record
the various settings on a note sheet, which is a checklist and future ref-
erence of both culture and microscopy (Fig. 4).
A useful routine for phase contrast microscopy of a tine specimen
would be:
1. Locate the field.
2. Focus on specimen.
3. Mark the field (seeabove).
4. Close field stop to within field of view.
5. Move condenser to bring edge of field stop into sharp focus.
6. Open field stop to be outside area used either in filming or in photom-
etry (if automatic exposure is used).
7. Insert phase telescope and superimpose images of annulus and ring.
8. Observe image for defects not visible before adjustment.
Strict adherence to such a routine avoids errors of setting up with loss
of image quality. It must be repeated whenever the specimen is moved or
the objective changed.
Time-Lapse Cinemicroscopy 435

Number DM.

Date Collaborators

Disposal

Purpose

Film I I I I
Unit
,.....*.......
,.....*... .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..I.........
....................................... .................. . . . . . . . . ..a......
.................

5Pe

Cells
..a....... ..
..,...........,......................................,,.......,.........I.................
.......................................... ....................................

Msdlum
.*....,.,.,...
.*....,.,. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..a....
. . ....................................... .................. .................

Treatment

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..I.................
.......................................... ........ ......... .................

Gas
. . . ,. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .*...........**..
.......................................... ........ ......... .................

Temp.

Illumination
. . . . . . . . . . . . ..I...................................... ...................................

Filter
. . . . . . . . . . . . ..I...................................... ........ ......... .................

E$psure
. . . . . . . . . ..)..f................................... ........ ......... . . ..............

or ASA
. . . . . . . .. .. . . . .. . . .. . .. . . . . . .. .. .. . . . . . . . .. .. . . . . . . . . ........ ......... ...... ..........

Intensity
,....,.. *...*,.,,*......,......,..***..............# .................. .................

Condanrar
.,....*......* I...,................................,, ..................................

Objective
,. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ......... ......... .................

Prqective
. . . . . . . . . . . ..I...................................... ........ ......... .................

Intsrv.31
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..I ......... ......... .................

Prism
. . . . . . . . . . . . ..I.....................................~ ......... ..........................

Fieldwidth
. . . . . . . . . . ...* I...........,......................,.., ...................................

Pilm
stat
. . . . . . frame
. . . -7. . . . . ..,......*........ I I ..*....*a c *.*....*. c . . . . . . . . I . . . . . . . . . I . . . . . . . . . I...**...
.............. .................. ......... ......... ........ ......... .................
g;;;%s: .... / ..................
1 1......... / ......... 1........ 1......... / .................
/

Fig. 4. Sample data sheet used to record experimental details that are otherwise so
easily forgotten. It also actsas a checklist during setting up.length is possible but incred-
ibly wearisome, largely because of the time taken in rewinding the film. (There are
systems to give continuous circular, end-to-end projection of a film, but these are not
appropriate for long films.)

2.2.7. Photometry
Some systems available for control of exposure have been mentioned
under equipment. With an automatic system, it is only necessary to ensure
that the light level is adequate for the exposure chosen and that the correct
436 Riddle

film speed rating is set on the intervalometer. Manual adjustment must

follow the instructions for the particular photometer in use.
The speed rating marked on the film packet has been found to be an
unreliable guide in cinemicrographic use (of one manufacture). In micro-
scope use, it has been found better to make a “test strip” using a similar
specimen to that for future study. Exposures are then based on the most
satisfactory of the images obtained.
To do this, select a test field, and either expose at the same light level
and vary the film speed settings (on an automatic unit) while exposing a
few frames at each or, on a manually controlled intervalometer, vary the
length of exposure with the light constant. Put a blank frame between each
change for location purposes. Changes of one full stop (i.e., a doubling of
exposure at each change) is best on the first test, and if the result leaves any
uncertainty between two levels, a more finely graded test at l/3 stop inter-
vals should resolve all doubts.
2.2.8. Start of Film
It has been found useful to record on the checklist (Fig. 4) both the time
of start of an experiment and also the frame number indicated on the inter-
valometer(s) at the start of an experiment. These indicate whether film
transport is occurring at the expected rate and also whether enough film is
left for any period when the unit is to be left running unattended.
2.2.9. Surveillance
The amount of attention required by the cinemicroscope when run-
ning depends (apart from individual idosyncrasies of the equipment) up-
on the specimen and to a large extent on the magnification. Once a culture
has equilibrated with its surroundings, very little change may be visible in
focus at low magnifications (less than 10x objective). At 4-10x that magni-
fication (possibly using oil immersion), the relatively short depth of field
may cause an unacceptable drift of focus within as short a period as a few
minutes.
Unless automatic focusing is introduced (a possibility that is scarcely
worth considering because of the expense, even if reliable), the only answer
for high-power work has been the adjustment at frequent intervals. Over-
frequent readjustment, unless done very gently, can ruin the appearance
of a film almost as much as a slight focus error.
Fortunately, high-magnification filming is usually relatively brief. In
contrast, the studies at lower magnifications can last for days to weeks and
fortunately can often be left over a weekend without any checking.
Time-Lapse Cinemicroscopy 437

2.2.10. Removal of Film

16-mm tine cameras take daylight loading film. The spools have solid
sides, and the outer five or so turns of film are sacrificed as a leader that
keeps light from the exposed portion. The camera can be loaded in the light
and-after the end of a film-a further 200 frames should be run on as a tail
so that the film can be taken from the camera in the light. If no such lead
is left and the film is developed at a processing laboratory, the last section
will be lost. Be sure therefore to replace film before the final 400 frames
have been used, or be prepared to remove film and start development in
total darkness.
2.2.11. Development
Assuming that development is being done in one’s own dark room, it
is essential to have used several test lengths of film to establish such vari-
ables as temperature of the baths of developer. Any variation in the qual-
ity of routine output calls for a repeat test. X-ray type developer and rapid
fixer have been found ideal for the high-contrast film used in phase-
contrast filming. The developer and fixer both have a limited life in terms
of the amount of film they can process before exhaustion and also in
absolute age once diluted. As a rough guide, assume for 1.5-2 L of working
solution that the passage of 2 d is equivalent to the development of 30 m of
film. The capacity of such developers varies. Although a developer that
is approaching the end of its lifetime is capable of producing apparently
satisfactory negatives, a follow-on film developed in fresh materials may
show an unacceptable change in density when the two sections are spliced
together. Fixers (seeNote 9) used as recommended usually last the equiv-
alent time to the developer.
2.2.12. Storage
Film for scientific purposes is seldom required in lengths of greater
than 30 m, The cans in which the film is supplied are ideal for storage and
can be kept in the original boxes, which have space for a note of the content
on the outside. Although there are optimum conditions of humidity for
film storage, it has been found that film kept in normal laboratory condi-
tions does not suffer unduly. For demonstration, longer lengths of film are
sometimes needed, and it is wise to have a small stock of 60 and 120 m
spools available.
2.2.13. Splicing
To join lengths of film, for whatever purpose, a splicer is needed. The
most usual types, which work with conventionally backed film, use film
438 Riddle

cement to make the join. The ends to be joined are cut to overlap with the
perforations in correct register for both sections. The overlapping portions
are then scraped or ground to remove the emulsion from the one side and
a little surface material from the other side. Film cement is applied to one
or both abraded surfaces before bringing the two into firm contact. Some
types use a heated anvil under the mating surfaces to speed up the setting
time. Pressure must otherwise be maintained for a minute or so until the
cement hardens.
Film made upon a base that is not soluble in conventional film adhe-
sives requires a different technique (which can also be used for conven-
tional film). The joins, which are every bit as satisfactory as the cemented
type, are made with an adhesive tape. The splicer for this method is used
to trim the ends of the film, so that they butt accurately (usually across a
pair of perforations). The film is then clamped with the ends in contact and
a clear adhesive tape is placed over the joint. Downward pressure with a
trimmer block presses the tape on, cuts the tape from the side of the film and
perforates the tape in the same place as the film. Although it is not con-
sidered necessary to tape both sides, it is the author’s practice to do so, since
film for laboratory use has much more severe treatment than that for sim-
ple viewing.
When purchasing a tape splicer, a double- or single-perforation type
should be specified. Use of a single-perforation model involves an extra
stage to perforate the second piece of tape, but a twin-perforated type is not
suitable for some situations. However, these are unlikely to be encoun-
tered in time-lapse work.
2.3. Observation and Analysis
2.3.1. Projectors
It is advisable to have a projector that is capable of both normal and an-
alytical viewing. Unfortunately, all the projectors available use the con-
ventional intermittent feed systems and rotating prism viewers, since
those used in the editing of professional films are both very expensive and
not as capable of producing a large, detailed image. These machines ad-
vance the film continuously (with consequently less wear and often tear),
and it is surprising that a large screen unit of this type has not emerged for
analytical cinematography.
Several types of analytical projectors are available. Some of these have
been developed from conventional projectors and others developed purely
for analysis. Since the behavior of the various makes can be very different,
Time-Lapse Cinemicroscopy 439

it is advisable to consult other users before purchase. Some are superb;

others tear, burn, or scratch film, and/or fail to operate for long periods.
The conventional projector is, after all, only expected to run for half an hour
at a stretch, whereas film analysis can continue all day and every day.
An ideal analytical projector should be practically silent (few are, but
some are worse than others, making work very tiring). They should be able
to project the film in forward or reverse mode both at the normal projection
speeds of 18 and 25 frames/s (fps), and a range of longer intervals, e.g., 12
fps down to 1 fps. They should be able to project a single frame (“still
frame”) without any physical change such as warping to the film, let alone
damage. More rapid film movement in either direction would be an ad-
vantage, but is seldom found. Control of the projector functions should be
on a remote hand-held unit to allow the viewer to sit close by the screen.
A frame counter mounted on the hand-control unit (as opposed to the pro-
jector itself) is invaluable to avoid ‘Wimbledon neck.” This is said in all
seriousness and with experience. Automatic feed projectors are of dubious
advantage in the laboratory situation, especially if it is required to remove
lengths of film without rewinding onto either spool. Some automatic pro-
jectors are also very difficult to unload if a film has become damaged or dis-
placed during projection. The projectors and projector analyzers in the
laboratory are best used with a back projection system in such a manner
that the screen is on the bench in front of the observer. Units to hold pro-
jector, mirror, and screen are available commercially. Alternatively, the
projector can be mounted at a convenient height above the bench and a mir-
ror fitted to project the image down onto a paper screen on the bench. A
photographic tripod head with ball joint is a convenient way to mount the
mirror to adjust it as it is often displaced by the head of the observer! With
this arrangement, the image falls with the long axis across the bench and
not along it as is customary. This does not matter when film of cells in
culture is the subject, but can be confusing when trying to decipher a
written title (seeNote 10).

2.3.2. Analytical Procedures

It is difficult to make any recommendations about the way to ap-
proach the observation of a film. The method depends on the parameters
being investigated. However, it may help to indicate some general prin-
ciples that have been found valuable. If the film is for demonstration pur-
poses, a copy negative or at the very least a print should be obtained before
putting the negative through the projector. The copy should be used for
440 Riddle

analysis if the negative is of great value.

When ready to project, a preliminary scan of the film at normal pro-
jection speeds (25 fps) should bemade several times in forward and reverse
modes to obtain a mental picture of what is happening. It is also possible
at this stage to notice additional points of interest for the future. For exam-
ple, a film made to study intermitotic times under changing conditions
may show concommitant changes of movement and shape that were not
suspected when the experiment was designed. Later investigationof these
points may be fruitful. After the original survey has revealed whether the
cells have shown the required behavior, a systematic analysis, cell by cell,
is nearly always needed. Just what parameter is followed should have
been decided before the culture ever reached the microscope. The amount
of potential data on a film of crowded cells can appear daunting to the
beginner and the experienced observer alike, but a systematic approach
nearly always proves matters to be far more easily handled than feared.
Common to most analyses, it is necessary to locate the cells at the first
frame position, and the projector is used in the still frame mode while this
is done. A projection screen upon which the observer can write becomes
essential. For back-projection sys terns, remove the usual ground glass, and
replace that with clear glass holding tracing paper held by clips or sticky
tape. Number the sheet, and then before starting analysis, choose a frame.
With the projector in still-frame mode, mark two recognizable and well-
separated fixed points. This enables the screen to be aligned for the same
film with the minimum of effort should the systembe disturbed or another
film have to be put temporarily in its place.
If a simple growth curve is required (a useful starting point in studies
of proliferation), a grid drawn on a clear plastic sheet can now be superim-
posed. The cells are ticked off with a wipe-off pen making counts at times
chosen to be suitably separated for the rate of growth. This method has
been found far easier than attempts to count the cells, even against a grid,
without marking off, and a crowded field containing over 1000 cells be-
comes manageable. For other analyses, e.g., of intermitotic times, the pa-
per screen becomes a “cell map” and the cells are numbered at their starting
positions. Each cell is followed individually, and the place at which, e.g.,
mitosis occurs is noted and given its cell number. The beginner will find
difficulty in following more than one cell at a time, but with experience and
given a not too active culture, up to four cells (daughters and granddaugh-
ters of a division) can be followed at one time. Data can be written onto a
“family tree” in this instance (26).
Time-Lapse Cinemicroscopy 441

Another more sophisticated analysis is that of speed and direction of

movement of whole cells or cell components. A simple but time-consum-
ing method is to draw the tracks of the cells with pencil upon the cell map
and then to return to the beginning. Then advancing the film to frames that
correspond to appropriate times, the point reached is marked on the origi-
nal trace. Subsequently, each trace is measured with amap measure. Final
conversion to pm/h is made with reference to the image of the graticule
filmed at the start of the film.
A far quicker method can be applied when a digitizer/computer com-
bination is available by the screen. Given suitable software, such appara-
tus can measure the tracks of the hand-held cursor and also will convert the
data into statistically analyzed summaries. The time taken to gathering the
data onto disk, using one such system in the author’s laboratory is ap-
proximately l/2 to 3/4 h/50 cells. Print out takes a short time.
One form of analysis that is difficult if not impossible in any true sense
is that of cellular morphology. The various individual factors that combine
to make up this rather vague concept are all amenable to analysis, e.g., area,
number of pseudopodia, number and shape of nuclei, and a host of others
but it is seldom possible to reduce this to a numerical description. One
exception to this rule occurs where a distinctive feature can be used to de-
note a particular morphology, and the change from motile, to non motile
cells with a web of proliferations as described by Small et al. (17) is an
example.

3. Notes
1. If long focus lenses are not available, it is possible to introduce a thin-
ner area into the base of a plastic Petri dish. A hole is made with a
heated cork borer or better with a hole saw. The hot tube method can
leave a burr around the hole and often causes cracking as the plastic
cools (remove the burr, carefully, with a scalpel). The hole saw
method often leads to shattering of the dish. The hole is covered with
a conventional coverslip that is sealed on using either a silicone rubber
or two-part epoxy resin adhesive. Subsequent sterilization after the
adhesive has cured is achieved by thorough washing, concluding with
double-distilled water and finally with 70% ethanol for some 30 min.
The vessel is then allowed to dry in clean air. A microwave oven can
also be used to sterilize this dish.
2. Hair dryers of this type are now less popular, since high wattage mini-
442 Riddle

ature designs are the current fashion. If a suitable unit cannot be pur-
chased, a box-mounted blower with a heater in the air path can be
made into a similar device.
3. Problems of changing focus while cultures are being filmed (and
which usually arise on Friday evening) have, in some cases, been
found to be the result of the container. Considerable effort revealed
that the focus varied because of distortion of the base on which the
culture chambers were placed. This was caused less by temperature
change than by humidity variation and differential humidity effects
either side of the perspex bases that were first used. Substitution of
firm, stainless-steel, or aluminium alloy bases has cured this problem
almost completely. If the surface upon which the dish is seated cannot
bend and the microscope controls do not slip, unexpected alteration in
focus must arise in the culture dish base. It has been observed that
dishes that are freshly seeded and previously dry do distort over a
period of several hours (presumably) as a result of absorption of water
from the medium into the wet side. Further slight changes can some-
times be seen with changes in atmospheric humidity and are avoided
by full laboratory air-conditioning.
High-magnification filming shows up these distortions far more
than low magnifications, and there comes a time when constant cor-
rection at intervals of a few minutes (or automatic focus) must be used.
Otherwise, variation must be accepted. The dishes with the stretched
membrane bases (seeabove)are less prone to distortion causing change
of focus.
4. A thought: Total darkness, even if possible in the technique being de-
scribed, may be almost as “unnatural” as bright light. Skin cells can
be exposed to very high light levels at times, and the light undoubtedly
penetrates several cm below the surface. It is doubtful if the very cen-
ter of a laboratory mouse is ever in total darkness during daytime!
5. Tests for vibration do not require the microscope to be moved from
place to place. Simple feeling (as lightly as possible) with the back of
the hand will detect slight movement. It is said that the tip of the nose
or the forehead is more sensitive to vibration than other parts, but
somewhat embarrassing to use if seeking a suitable site around a
building! An easier test is to place a shallow dish of water on the sur-
face being examined and to observe the water surface by reflected
light. When there is vibration present, concentric rings form on the
surface; that test is unfortunately sensitive to the periodicity of the
Time-Lapse Cinemicroscopy 443

movement of the water in the dish, and several different size dishes
should be tried.
6. Filming of two microscope fields with one camera using a comparator
head has been advocated but is complex. The only other possible
means to obtain images of several areas on the same film is to use a
computer-controlled stage that tracks repeatedly from area to area. If
the wish is to make a final film comparing two different situations side
by side, it is easier to prepare separate negatives and then have split
screen prints made, showing half of one film and half of the other on
the same print (15).
7. The differences in final film obtained with cameras of the various
types can only be summarized under the headings of imperfections.
Faults that are usually, but not always, brought about by the camera
itself are:
a. Scratching-check the various places where the film comes
into camera contact to make sure that they are clean, not dam-
aged, and free to rotate if in the form of rollers. Do not confuse
this with scratches caused by processing systems (unlikely if
made at a professional film laboratory), but easily caused in a
bench unit if not perfectly clean.
b. Inadequate or irregular film advance-this may appear as a
wedge- shaped clear or dark area on some or all frames. Check
that the shutter movement is correct, as seen when the camera
is removed from the microscope and any lenses. Do this with
film in place since the drag of the film can slow shutter move-
ment. If faulty, first check the loading of the film and that the
camera is clean. If still faulty, the relevant servicing should be
obtained. Some camera-control systems can also cause similar
effects and also require servicing.
c. Unequal exposure produces single or multiple frames appear-
ing darker or lighter than the rest (seeunder Environment). The
only way to detect the cause is to observe the camera action, and
again servicing may be indicated. This fault is also seen occa-
sionally in film made with units that have automatic exposure.
The cause there can be the result of extra light reaching the
photometry system through secondary light paths (titling sys-
tems, other eyepieces, and so on). It can also stem from systems
that have photometric devices that switch between sensitivity
levels. Where the exposure being used is on the borderline at
444 Riddle

the changeover point, the exposure can alternate. The answer

here is to set the light level so that the critical threshold is not
approached. A somewhat different variation of film density
can be the result of unequal rates of passage of the film through
a laboratory developing system. The indication here is that
darker sections have not been drawn throughout the developer
at a steady rate. These sections (on the negative) tend to start
at any part of the frame and are not related to fogging of the
border. Overall fogging caused by light leakage onto the film
appears somewhat similar, but in this case, the darkened areas
cover both frames and the film between and around the perfo-
rations.
d. Unequal spacing between frames. This can be because of dirt
on the film guides, excessive film tension from any cause, or
even badly perforated film. The latter does occur, but is very
rare indeed.
In all cases where faults are found, the value of more than
one identical cinemicroscopy unit is incalculable. To be able to
exchange components is the best way to locate the one that is
faulty.
8. For much of biological microscopy, where phase contrast is used, the
light densities of the images are not “true” black to white density dif-
ferences, but contrasts induced by optical means. In this case, there is
no theoretical objection to looking at the images in negative or positive
form. Whereas such crimes are totally unacceptable in the realms of
entertainment or illustrative photography, the scientist who wishes to
extract data can do so even if the film becomes somewhat scratched,
and can save a lot of time, space, and expense by viewing and analyz-
ing the negatives. This has the added advantage that the image in
many cases is more acceptable visually, because the grain of the film
is less accentuated (prints are not made upon the fine grain stock of the
microfilm) and the phase contrast “halo” of brightness around a
darker area becomes an apparent shadow that is more acceptable to
our perception of an image.
9. Some fixers are normally used with a hardener that is stored as a sep-
arate concentrate. Making up solutions without preliminary dilution
of the fixer before adding the hardener causes a precipitation of sulfur
(from the thiosulphate), which damages the film. Since fixer tanks in
Time-Lapse Cinemicroscopy 445

the small bench-top developing units do collect an encrustation of

sulfur in the most awkward of places, a routine rinse out between
changes of fixer may prove insufficient to remove anything but the
more floccular debris. It is sometimes necessary to chip rock-solid
sulfur away. A note of how much film can be developed before any
undesirable deterioration occurs is invaluable. A record of the amount
of film that has been processed after each change of solutions allows
changing the solutions in adequate time.
10. Suitable mirrors can be made from bathroom mirror tiles, available in
a variety of sizes at hardware shops. The inevitable ghost image
caused by reflection from the unsilvered front surface is not signifi-
cant. However, superior mirrors can be made of plate glass sheets
surface coated with aluminium in a standard EM evaporation unit.
The reflectance is superb and undistorted, but the surface is easily
damaged.

References
1. Marey, E. J. (1894) Le Mouvement (Masson, G., ed.), Paris.
2. Michaelis, A. R. (1955) Research Films in Biology, Anthropology, Psychology, and Med-
icine (Academic Press, New York).
3. See J. James (1976) Light Microscupy Techniques in BioZogy and Medicine, Martinus
Nijhoff Medical Division, Netherlands, pp. 151,152.
4. Richmond, K. M. V., Riddle, P. N., and Brooks, R. F. (1984) Apparent desensitization
of Swiss 3T3 cells to the mitogens FGF and vasopressin. J. CeZI Physiol. 121,547-557.
5. Wilkinson, C. and Lackie, J. M. (1983) The influence of contact guidance on chemo-
taxis of human neutrophil leukocytes. Exp. Cell Res. 145,255-264.
6. O’Neill, C. H., Riddle, P. N., and Rozengurt, E. (1985) Stimulating the proliferation
of quiescent 3T3 fibroblasts by peptide growth factors or by agents that elevate
cellular cyclic AMP level has opposite effects on motility. Exp. Cell Res. 156,65-78.
7. T.ackie, J. M. (1986) Cell Movement and Cell Behavior (Allen and Unwin, London,
Boston, and Sydney).
8. Brooks, R. F., Riddle, P. N., Richmond, F. N., and Marsden, J. (1983) The Gl distri-
bution of “Gl-less” V79 Chinese hamster cells. Exp. Cell Res. 148,127-142.
9. Salmon, E. D. and Ellis, G. W. (1975) A new miniature hydrostatic pressure cham-
ber for microscopy. J CeZZBioZ. 65,587-602.
10. Riddle, P. N. (1979) Time-Lapse Cinemicroscopy (Traherne, J. E. and Rubery, P. H.,
eds.), Academic Press, London and New York.
11. Roberts, D. C. and Trevan, D. J. (1961) A versatile microscope chamber for the study
of environmental changes on living cells. 1. R. Microsc. Sot. 79,361366.
12. Riddle, P. N. (1983) A device for demisting Petri dishes. Lab. Practice 32,80.
13. Riddle, P. N. (1977) Beam deflection as an alternative to light shuttering in cinemi-
croscopy. Lab. Practice 26,865.
446 Riddle

24. Trevan, D. J. (1961) A simple inverted microscope for use with a tine camera. 1. R.
Microsc. Sot. 79,367,368.
25. Riddle, P. N. (1974) Comparison cinemicrographs prepared by optical split screen
printing, Report of2nd International Colloquium of Znterkamers‘73 (Serb, V. and Siba-
lova, J., eds.), In Vitro vs CSSRHradci Kralove.
26. Brooks, R. F. and Riddle, P. N. (1988) Differences in growth factor sensitivity be-
tween individual 3T3 cells arise at high frequency: possible relevance to cell senes-
cence. EXQ. Cell Res. 17~4378-388.
17. Small, R. K., Riddle, P.N., andNoble, M. (1987) Evidence for the migration of oligo-
dendrocyte-type-2 astrocyte progenitor cells into the developing rat optic nerve.
Nature 328,155-157.
 Chapter 34

Transmission and Scanning

Electron Microscope
Preparation of Whole
Cultured Cells

Josef Neumiiller
1. Introduction
Despite the enthusiasm of the first investigations of cell ultrastruc-
ture, morphological studies have since lost some of their importance for
biomedical research. The development of quantitative biochemical meth-
ods has been the cause of this reduced interest in morphology. Biochemical
reactions, however, take place in compartments of the cell and the extra-
cellular matrix. This compartmentation, in an ultrastructural dimension,
is the prerequisite for a systemic discharge of metabolic processes in
temporal continuity. This compartmentation is provided by phospholi-
pid biomembranes serving as support for sets of enzymes or receptors and
as vessels for internalized or synthesized substances. Investigations of the
biochemical processes in compartments are only possible if these com-
partments can be separated by ultracentrifugation or by ultrahistochemi-
cal methods. Routine preparation for transmission electron microscopy
(TEM) preserves only the morphology of cells and limits the possibilities

447
448 Neumilller

for staining to electron-dense reaction products. Furthermore, the elec-

tron beam causes considerable contamination and damage to the object. In
the past decade several attempts have been made to overcome these
problems in order to obtain realistic morphological representations of cells
and to perform ultrahistochemistry. First of all the development of cryo-
methods has made it possible to retain and preserve substances at their ori-
ginal site. However, the use of only “soft aldehyde fixatives”and the re-
nunciation of crosslinking reagents such as 0~0, causes a loss of contrast
of the ultrastructural components. The newer techniques of: (a) cryo-
transfer from the ultramicrotome to the EM in a vitrified state, (b) cryosub-
stitution that involves a gradual substitution of ice with organic solvents
and finally with resins that polymerize at low temperatures, (c) controlled
freeze drying over a molecular sieve, and (d) the replica techniques taken
from freeze-fractures can partially resolve the problems of conventional
preparation methods. The problem of contrast can also be solved in some
cases by electronic contrast enhancement and image processing.
The lack of information about absent ultrastructural details in sec-
tioned cells can to some extent be compensated for by investigations using
SEM (Scanning Electron Microscopy) or HVEM (High Voltage Electron
Microscopy).
The following chapter deals with preparation schedules for several
anchorage dependent and suspension cell cultures as well as for cell cul-
tures grown in gels.
In the early days of EM, whole cultured cells on formvar coated grids
were investigated but gave poor resolution under the normal 50 kV elec-
tron beam. After the development of ultramicrotomy these methods
receded into the background. Metallurgists, however, were faced with the
need to develop a HVEM allowing penetration of thin metal foils. Appro-
priate electron microscopes using a voltage of I MeV were constructed.
Biologists, too, became interested in the features of these microscopes.
Above all, the microtrabecular lattice was clarified by Porter and his col-
leagues. Only few biological investigations have been carried out because
of the small number of HVEMs available in the world. New generations
of high resolution electron microscopes provided a compromise by work-
ing at a high voltage (HV) up to 400 kV. With this HV setting it is also
possible to penetrate through cells that have been spread on grids (1-5).
This method is very powerful, particularly for investigation of cell motil-
ity, the cytoskeleton, and its interactions with the extra-cellular matrix
(ECM) using immunogold particles and contrasting with silver stain.
Transmission and Scanning Electron Microscope 449

2. Materials
1. Fixative: 2% glutaraldehyde (GA) in O.lMNa-cacodylate-HCl-buffer
+ O.lM sucrose (pH 7.2), total osmolarity: 510 mOsM, vehicle osmolar-
ity: 300 mOsM. GA is commercially available as a 25% aqueous solu-
tion (prepare fresh before each use); 1% 0~0, in Na-cacodylate-HCl
buffer (can be stored at 4OCfor some months).
2. Na-cacodylate-HCl buffer as in ref. (I) without GA and sucrose as a
washing solution (can be stored at 4OCfor some months).
3. A gradual series of ethanol: 30,50,70,80,90, and 95% ethanol in double
distilled water; absolute ethanol that has been previously purified of
any water and particles, using a molecular sieve that absorbs molec-
ules with an effective diameter ~0.3 nm (store at 4°C).
4. Intermediate fluids: Freon TF; gas flasks with Freon 13 or 116, or pure
co,.
5. Reagents for embedding: Propylene oxide (1,2-epoxypropane); N-
butylglycidyl ether; ultralow viscosity resin (ULVR): 100 g ERL 4206
resin (vinylcyclohexene dioxide), 200 g HXSA (hexenylsuccinic anhy-
dride), 25 g Araldite RD2 (DY026) resin, and 25 g DMAE (dimethyl-
aminoethanol) (store at 4OC).
6. Formvar resin, 0.2%, in water-free chloroform in order to prepare sup-
porting film on EM grids (prepare fresh).
7. Materials spheroids, microcarriers, or coverslips for sticking cells,
onto the specimen mount for SEM: 10 mg poly+lysine in 100 mL dou-
ble distilled water; a double-sided adhesive tape; a silver-loaded
epoxy adhesive.
8. Reagents for removal of resin from embedded cells in gels: 0.5 g cry-
stalline NaOH in 50 mL absolute ethanol; a gradual series of 30,60,70,
and 90% amyl acetate in absolute ethanol and pure amyl acetate (pre-
pare fresh).
9, Chemicals for the freeze-fracture: A 5 mM aqueous solution of poly-
L-lysine (MW 2000-4000); 0.5% boiled and filtrated starch solution in
double distilled water (both to be prepared fresh); Freon 22; 10% hy-
drofluoric acid (store at 4OC).
10. Materials for the preparation of microchambers: Beem capsules (size
00); nylon gaze (mesh width 100 pm); dialysis tubes (10 mm diemeter,
separation at MW 50,000).
11. Small Beem capsules (size 3) for direct embedding of selected areas of
monolayers.
450 Neumiiller

12. Other materials: Round coverslips (10 mm diameter); glass fiber grids
(mesh width 0.5 mm) for handling of floating monolayers; gold grids
G 200; copper grids with a hexagonal pattern (repeat distance 460
lines/in.); appropriate flat embedding forms made of silicone rubber.

3. Methods
3.1. Method 1: Simultaneous Preparation
of Cell Cultures for SEM and TEM (6)
3.1.1. Fixation
1. Fix in 2% glutaraldehyde (GA) in O.lM Na-cacodylate-HCl buffer plus
O.lM sucrose (pH 7.2), total osmolarity: 510 mOsM, vehicle osmol-
arity: 300 mOsM (seeNote 1). Add prewarmed (37OC)fixative in equal
volume to the culture medium for 5 min without moving the culture
vessel (seeNote 2).
2. Replace the fixative/mediummixture with pure fixative at room tem-
perature for 1 h and gently agitate the culture vessel.
3. Change the fixative and fix for a further 24 h at 4OC.
4. Rinse 3 x with Na-cacodylate-HCl buffer at room temperature.
5. Carry out postfixation in 1% 0~0, in O.lM Na-cacodylate-HCl buffer
(pH 7.2) at room temperature.

3.1.2. Dehydration
Use gradual steps of ethanol as follows: 2 x 5 min in 30% and in 50%;
10 min in 70%, 80%, 90%, 95%; 2 x 10 min in 100% (seeNote 3).

3.1.3. Drying at the Critical Point-

Preparation for SEM
1. Incubate with Freon TF as intermediate fluid (can be omitted if poly-
styrene vessels are used).
2. Transfer the sample to the specimen boat of the critical point drying
apparatus (CPDA). In order to avoid artifacts the specimen has to be
kept immersed in Freon TF or absolute ethanol (Fig. 1).
3. Put the specimen boat into the CPDA cooled to IO-15°C. Close CPDA
and fill with CO, or Freon 13 or 116.
4. Allow the fluid in the specimen boat to escape and change theinterme-
diate fluid 3x at 5 min intervals.
Transmission and Scanning Electron Microscope 451

Fig. 1. Human skin fibroblasts grown on glass coverslips. Artifact by drying

(bar=10 m).
5. Increase the temperature until the critical point is reached (it differs in
respect to the intermediate medium used):
Carbon dioxide (CO,>: Tc = 31.3OC
PC = 75.5 kg/cm2
Freon 13 (CClFJ: Tc = 28.9OC
PC = 39.5 kg/cm2
Freon 116 (CF,-CF,): Tc = 19.7OC
PC = 30.4 kg/cm2
(Tc = critical temperature; PC = critical point.)
6. When the critical point is reached, the liquid interface becomes opales-
cent and disappears. Allow the gas to escape slowly in order to avoid
recondensation.
3.1.4. Sputter Coating
1. Place the specimen onto a specimen mount using a conductive adhe-
sive.
2. Put the specimen mount into a sputter coater and cover it with a gold-
palladium @O/20) layer of 1.5-2 nm (seeNote 4).
452 Neumiiller

3.1.5. Alternative Embedding-Preparation for TEM

(continue after dehydration step described above)
1. Put the specimen into propylene oxide for 10 min (this step may be
omitted if polystyrene vessels are used).
2. Change the ULVR (ultralow viscosity resin) 3x at 4 h intervals or trans-
fer the specimen directly from 100% ethanol to the resin in a flat em-
bedding form. The resin is miscible with absolute ethanol.
3. Polymerize in an appropriate oven at 70’ C for 12 h.

3.2. Method 2: Preparation of Cell Pellets,

Spheroids, Cell Suspensions and Cells Grown
on Microcarriers for SEM
3.2.1. Cell Harvesting
1. Filter the cell pellets, spheroids, and microcarriers through a nylon
gaze (mesh width: 100 l.rrn) placed in a small glass funnel.
2. Rinse with 20 mL culture medium (without FCS), prewarmed to 37OC.
3. Incubate a dialysis tube (separation at MW 50,000,10 mm diameter)
in PBS for 30 min before use. Knot one end of the tube, cut out the tip
of the filter with the cell aggregates or microcarriers, and place it into
the tube. Fill the tube with medium without FCS and knot the second
end of the tube.
4. Alternatively, centrifuge cell suspensions at 800x g for 10 min, resus-
pend in a small amount of culture medium without FCS, and fill di-
rectly into the dialysis tube.
3.2.2. Fixation, Dehydration, and Critical Point Drying
Perform according to Method 1. The cells, aggregates, or microcar-
riers remain in the dialysis tubes, which are placed into 20 mL tumbler
vials, completely filled with the respective liquid and moved by means of
a rotation tumbler. Avoid flushing too rapidly and changing the pressure
in the CPDA too fast during outstreaming of gas in order to prevent burst-
ing of the tubes.
3.2.3. Coating with Gold-Palladium
1. Cover the SEM specimen mounts with a double-coated adhesive tape
(the tape should be somewhat smaller than the top surface area of the
specimen mount). Open the dialysis tube and disperse the contents of
the filter over the sticky surface of the tape.
Transmission and Scanning Electron Microscope 453

2. Invert the specimen mounts and remove the excess of particles by

generating of a soft air stream from an air puffer.
3. Connect the edge of the tape to the specimen mount with a droplet of
conductive epoxy adhesive.
4. Put the samples in a desiccator which contains silica beads, evacuate
and allow the glue of the tape to dry for at least 24 h.
5. Perform sputter-coating according to Method 1.

3.3. Method 3: Preparation of Cell Pellets,

Spheroids, Cell Suspensions (Fig. 2a-c) and Cells
Grown on Microcarriers (Fig. 3a-f) for TEM
1. Cell pellets from monolayers or centrifuged cell suspensions, sphe-
roids and sectionable microcarriers can alternatively be processed for
TEM. Instead of the dialysis tubes, Beem capsules covered with a lOO-
pm-mesh-width nylon gauze are used. Cut the gauze to an appropri-
ate size and clamp to the capsule with the bored capsule cap (Fig. 4).
2. Fix and dehydrate according to Method 1 (the chambers made from
the Beem capsules must be submersed completely in the respective
fluids of the tumbler vessels). If polydextran beads are used, the incu-
bation time should be extended 2-3x to permit sufficient infiltration.
3. After dehydration in absolute ethanol submerse the capsule chambers
in changes of ULVR for 4 x 4 h. Each change of resin should take place
in a separate tumbler vial moved by means of a rotation tumbler.
4. Perform the infiltration with resin at 4OC. After the last incubation
bring the vials into a vertical position and clean the outsides with a
filter paper wetted with propylene oxide.
5. Allow the particles in the capsule to settle and perform polymeriza-
tion at 70°C for 12 h.
6. Trim the blocks. This is quite easy because the particles are concen-
trated in the tip of the capsule.
7. For sectioning, the use of a diamond knife is recommended, above all
if microcarriers are embedded.

3.4. Method 4: Preparation of Monolayers Grown

on Coverslips or Plastic Petri Dishes for SEM
1. Use plastic Petri dishes with four ring-divisions fitted with 1 cm dia-
meter coverslips previously marked with an asymmetric letter and
sterilized (Fig. 5) or 6-cm diameter Petri dishes to grow cells.
454 Neumiiller

Fig. 2a. Attachment of human erythrocytes to an adhesive tape. Preparation by using

dialysis tubes @ar = 5 pm).

Fig. 2b. The same preparation as Fig. 2a, but of human granulocytes (bar = 5 ~1.

Fig. 2c. The same preparation as Fig. 2a, but of human buffy coat cells (bar = 4 ~1.
Transmission and Scanning Electron Microscope 455

Fig. 3a. HeLa cells grown to a high density on glass microcam ‘ers (bar = 10 pm).

Fig. 3b. Cell contactsand surface projections of HeLacells (detail of Fig. 3a;bar = 5~1.
456 Neumiiller

Fig. 3c. Human skin fibroblasts grown on micmcarriers (Cytodex 3) (bar = 100 pm).

2. Fix and dehydrate cells after spreading and multiplication according

to Method 1.
3. (a) Displace the coverslips carefully from the dish with a preparation
needle and a tweezer with angled parallel edges and put into a special
coverslip tray (Fig. 6) that fits the specimen boat of the CPDA. Drying
can be easily performed (Method 1) in this tray that can be made by
anyone with do-it-yourself experience.
(b) Ifcellswereplatedonplastics,stripsofthebottomofthedishescan
be cut out with a hot scalpel blade. This has to be carried out very
quickly to avoid desiccation of cells. The plastic strips can be marked
as previously mentioned and processed in tumbler vials. They can be
dried without the use of Freon TF, e.g., directly from absolute ethanol
to co*.
 Fig. 3d. HeLa cells grown on microcarriers (Biosilon) (bar = 30 pm).

Fig. 3e. TEM at low magnification of human tendon fibroblasts grown to a high density
on Biosilon microcarriers which have not been stirred (* = microcarrier; bar = 1.5 pm).

I
I

Fig. 3f. As Fig. 3e, but at higher magnification (bar = 5 prr$

457
458 Neumiiller

Fig. 4. Beem capsule with a filter (F) clamped to the perforated cap (C) containing
microcarriers (ML

4. Mount the coverslips or plastic pieces on to the specimen mounts with

a conductive epoxy adhesive, place into a desiccator under vacuum
for 24 h, and sputter coat as described under Methods 1 and 2.

3.5. Method 5: Simultaneous TEM Preparation

of Cultured Cells
Process as in Method 4 until dehydration in 100% ethanol.
3.5.1. Cells Grown on Round Coverslips
1, Fill the rings containing the coverslips with a large drop of ULVR,
which is changed 3x at 10 min intervals at 4OC.
2. Fill a Beem capsule completely with ULVR, inverse rapidly and place
over the coverslip in a vertical position. The diameter of the Beem cap-
sule should be cl cm.
3. Polymerize in this position at 70°C for 12 h.
Transmission and Scanning Electron Microscope 459

Fig. 5. Petri dish with ring-divisions containing marked coverslips. (F)

4. Break away the Beem capsule from the Petri dish. Split the coverslip
from the resin block with a scalpel blade or by rapid cooling with CO,
snow.
5. Carry out appropriate trimming after orientation under an inverted
light microscope.

3.5.2. Cells Grown on Plastic Dishes (7)

1. Mark the edge of a selected area of cells inside the Petri dish with a thin
preparation needle.
2. Replace the 100% ethanol in the dish by 2 mL ULVR.
3. Change it at 3x 10 min intervals at 4OC.
4. Cover the marked area with an inverted and resin filled Beem capsule
as mentioned in step 2 above (3.5.1) and then process further as de-
scribed there.
5. When the Beem capsule containing the resin block is split from the
dish, the marks are visible as replicas on the flat surface at the aperture
of the Beem capsule and serve as a guideline for trimming.
6. Orientate the block in the ultramicrotome at an exact 90’ angle to the
cutting axis. After one semi-thin section, perform ultrathin sections
parallel to the monolayer plane.
460 Neumftller

Fig. 6. Specimen boat of a CPDA with metal grid trays CT).

3.6. Method 6: ‘!Floating Sheet* Preparation

of Monolayers for TEM (Fig. 7a, b) (8) (See Note 5)
1. Grow cells in 6-cm diameter plastic dishes.
2. Fix and dehydrate according to Method 1.
3. Add 5 mL propylene oxide or N-butylglycidyl ether to the dishes. Put
a white porcelain plate under the dishes.
4. Move the dishes gently when the plastic appears rippled (after about
5 min). The monolayer floats up to the surface of the liquid.
5. Harvest the floating monolayer with a glass fiber grid (mesh width 0.5
mm), invert the grid and place it carefully in a flat embedding form
filled with ULVR. Press the grid slightly against the resin surface us-
ing a glass rod in order to separate the monolayer from the grid.
6. Polymerize at 70°C for 12 h.

3.7. Method 7: Whole Mount Preparation

of Cultured Cells on Grids (9)
3.7.1. Preparation of Grids
1. Wash slides in distilled water and dry with Kleenexm. Remove rem-
nant dust particles by using a gas jet duster. Do not clean with soap,
Transmission and Scanning Electron Microscope 461

Fig. 7a. TEM preparation of “floating sheets” of human ligamentous fibroblasts. Note
the electron-dense glycoprotein layer (AA,) and the overlapping of these cells of a conflu-
ent monolayer (bar = 2 PI.

Fig. 7b. TEM preparation of “floatingsheets” of human ligamentous fibroblasts. Note

the production of collagen fibrils (?I in deep recesses of the cell (bar = 0.5 ~1.
462 Neumiiller

solvent, or acids because in this case the separation of films is very

difficult.
2. Prepare a 0.2% solution of formvar resin in water-free chloroform. The
resin must be dissolved completely in the solvent. Use a magnetic
stirrer.
3. Run a thin preparation needle along the edge of the slide. Seize the
coated surface and dip the slide slowly at an angle of 45Ointo a glass
trough filled with water. Under appropriate illumination the floating
of the formvar film separating from the slide can be distinctly seen.
4. Transfer the films to gold grids using a grid filming device. The grids
are held by a TeflonTM plate at the bottom of the trough. The surface
of the liquid is slowly depressed and the films come in contact with the
grids without any folds forming.
5. Remove excess water and dry the grid plate.
6. Stabilize the filmed grids with a carbon layer using a vacuum coater.
7. Make the carbon surface hydrophilic by irradiation with a UV lamp
for 30 min.

3.7.2. Attachment of Cells to Grids

1. a. Cells are grown on the slides by incubation in small plastic
dishes. Avoid sliding together of grids. Petri dishes with divi-
sions (Fig. 4) may be used.
b. Nonadherent cells can be attached to the carbon coated sur-
face by dipping the grids into 10 mg/mL 30 poly+lysine and
washing 3x in distilled water. Place the grids into the plastic
dishes and allow the cells to settle and adhere to the coated
surface.
2. Transfer the grids for further processing into small Beem capsules that
have been perforated with some holes at the conical part and on the
sides in order to permit a good fluid exchange.
As an example, HeLa cells grown on filmed grids are shown in SEM
(Fig. 8a) and in TEM mode as whole cell mounts (Fig. 8b).

3.7.3. Fixation, Dehydration, Critical Point Drying,

and Carbon Coating
All techniques are performed according to Method 3. The grids with
the attached cells are coated a second time with carbon.
Transmission and Scanning Electron Microscope 463

Fig. 8k HeLa cells in TEM mode prepared by the whole cell mounting technique (bar
=lprn).

Fig. 8b. HeLa cells prepared in SEM mode by the whole cell mounting technique (bar
=5op.m).
464 Neumiiller

3.8. Method 8: Preparation of Cells Cultured on, or

in, GeZs for TEM and SEM (See Note 6)
1. Perform fixation in situ according to Method 1.
2. Extend the rinsing with Na-cacodylate-HCl buffer to twice the usual
time.
3. a. Cut the gel into small cubes that then are transferred to tumbler
vessels and dehydrate according to Method I.
b. If the cells are grown on the surface of the gels, dry the cubes
in small gaskets in the CPDA.
4. a. Embed cells grown inside the gel in ULVR after dehydration
in tumbler vials:
100% Ethanol 3x 1 h
Propylene oxide 1 h
1 part ULVR + 3 parts propylene oxide 1 h
1 part ULVR + 1 part propylene oxide 1 h
3 parts ULVR + 1 part propylene oxide 1 h
ULVR alone 2x 3 h
b. Prepare cells grown on the surface of the gels according to Method
1.
5. Place the cubes in silicone embedding molds containing ULVR and
polymerize at 70” C for 12 h.
6. Trim the blocks and make sections for TEM.
7. After the TEM observation, shorten the resin block to a flat disk on
which the trimmed pyramid remains (compare with [lo]).
8. Incubate the disk in 0.5 g of crystalline NaOH in 50 mL of absolute
ethanol for 30-60 min while the solution is agitated with a magnetic
stirrer. Avoid evaporation of the solvent.
9. Perform critical point drying when the removal of the resin is com-
pleted, using intermediate fluids as indicated in Method 1.
10. Incubate 3x in absolute ethanol for 10 min, and change stepwise to
amyl acetate (7:3,4:6,3:7,1:9, and 100% amyl acetate) for 10 min at
every step.
11. Dry in the CPDA with CO, for 30 min.
12. Sputter with gold-palladium (seeMethod 1).

3.9. Method 9:
Freeze Fracture of Monolayer (ll) (See Note 7’)
1. Attach the cells to a glass coverslip using an aqueous solution of 5 mM
poly-L-lysine (MW 2,000-4,000).
Transmission and Scanning Electron Microscope 465

2. Coat the slide with a thin film of a 0.5% boiled and filtered starch
solution.
3. Bring the coverslip side with the attached cells into contact with the
filmed side of the slide without applying any pressure.
4. Place the slide with the attached coverslip in a freezing container filled
with Freon 22 that has been precooled with liquid nitrogen to -160°C.
5. Separate the coverslip from the slide using a razor blade mounted on
a polyethylene sheet.
6. Place the broken parts in a freezing stage that is put in a vacuum unit
on a liquid N2 cooled shield. N2 is released and the system evacuated
at 10” torr.
7. Increase the temperature to -8OOCusing an internal heater inside the
vacuum sys tern.
8. Remove the cooling shield and shadow the specimen with platinum
at a 20° angle and with carbon at a 90° angle.
9. Float off the replica in 10% hydrofluoric acid, rinse with distilled water
and mount on grids according to Method 7 (3.7.1[4]).

4. Notes
1. There is one basic principle that needs to be followed in order to get
good preparation results. As mentioned already above, cultured cells
are very sensitive to changes in pH, osmolarity, and variations in tem-
perature. These parameters have to be maintained during the alde-
hyde fixation steps at a level that is similar to that found in living cells.
Aldehyde fixatives do not completely break down the semipermeabil-
ity of the plasma membrane. Therefore the osmolarity should be at 510
mOsM (vehicle osmolarity 300 mOsM) (6,12) and it is recommended
to check the osmolarity with an osmometer. Even in cryomethods, in
which cells are frozen with the addition of a cryoprotectant (13), the os-
molarity is of great importance.
2. During fixation, changes in temperature or vigorous shaking can com-
pletely alter the shape of the cells. Therefore one should add GA at
37°C to the complete medium for only a few minutes. This precaution
stabilizes the plasma membrane and its projections. If the mixture of
GA and medium remains in the culture vessels for more than a few
min, protein aggregates from fetal calf serum cover the cells and can-
not be removed. In a second step GA is added to an adequate fixation
buffer (always at 37°C) that is allowed to cool down to room tempera-
ture after 1 hr, the fixative is replaced again and the fixed cells are kept
at 4°C for 12 h.
466 Neumiiller

If one is interested in a good preservation of microvilli and other cell

membrane projections, postfixation with 0~0, is indispensable in or-
der to avoid bubble-like swelling. 0~0, on the other hand, increases
the rigidity and fragility of the cell membrane. This leads to clefts and
fragmentation. Therefore, the appropriate fixation and the right con-
centration of 0~0, have to be found (Fig. 9a-d).
3. For dehydration in ethanol or acetone the cells do not need longer than
a few min, but care has to be taken to make sure that the substratum
is also well infiltrated and dehydrated, above all if microcarriers made
of DEAE cellulose are used.
4. After critical point drying the specimen is placed onto a specimen
mount, using a double-sided adhesive tape or a drop of silver-loaded
epoxy adhesive. In order to avoid charge problems it is recommended
that conductivity is provided between the specimen and the specimen
mount by a droplet of a conductive adhesive. Such adhesives are
available and they dry in a few min. The specimen mount with the spe-
cimen should be stored in a vacuum desiccator overnight before
coating. For coating a sputter coater is usually used. In this apparatus
the specimen is put into a vacuum jar in a holder that is at the same time
the anode of alarge diode. The cathode is situated above the specimen
and holds a gold target.
The system is evacuated until co.05 torr are reached. Pure argon is
let in several times in order to replace the remaining air, When the
vacuum of 0.05-0.07 torr is reached again, a high voltage of about 1 kV
is set. Because of the electron-push the argon atoms are ionized before
they reach the gold target and cause an emission of gold from the
cathode to the specimen surface at the anode. This process causes a
violet light in the jar. The particles collide and are thereby slowed
down. Thus the vacuum decreases and is kept at 0.1-0.2 torr by a mini-
mal influx of argon through a needle valve. The argon influx triggers
the electric current that should be maintained at about 10 mA during
sputtering. By these three parameters (voltage, electric current, and
sputter time) the thickness of the gold layer can be gaged. It should be
about 0.15 nm in order to avoid “burying” fine surface detail under the
metal coating. The size of the gold grains can be diminished by using
gold-palladium targets (22).
Mostly the observation in SEM is performed using secondary elec-
trons. In immunohistochemical methods with gold particles bound to
receptors or antigens, the detector for back-scattered electrons is used,
Transmission and Scanning Electron Microscope 467

Fig. 9a. Spread human skin fibroblast. Cell fixed with GA and OsO,. Note the well
preserved microvilli (bar = 4 WI.

Fig. 9b. Human fibroblast grown on a microcarrier coated with gelatin (Cytodex 3)
fixed only with GA. Note the bubble-like contracture of microvilli (bar = 4 ~1.
468 Neumuller

Fig. 9c. Human synovial fluid cells attached to glass coverslips. Fixation with cooled
GA in cacodylate buffer (4°C). Note the retraction of the cytoplasm at thelobopodia. Only
radial remnants of the cytoskeleton are visible (bar = 10 cun).

Fig. 9d. Good preservation of the cell shape of human skin fibroblasts (bar = 20 ~1.
Transmission and Scanning Electron Microscope 469

Fig. 10. Macrophage attached to a glass coverslip labeled with a monoclonal antibody
against HLA DR (BMA 020) and in a second step with a goat-anti-mouse immunogold
probe (VI (bar = 2 ~1.

which gives good contrast between the gold particles and the rest of
the cell surface (14-22) (Fig. 10).
5. For the simultaneous processing of cell cultures for TEM, there are
numerous possibilities. Cells can be grown in a second Petri dish of
the same size but without divisions, and prepared as a floating sheet
(8). By this method cells are placed in several layers in which they
maintain their contacts and are sitting on a thin glycoprotein film,
which in cross-sections is visible as a distinct, electron-dense line (Fig.
7a). With this method it is possible to show the formation of collagen
fibrils in deep recesses of the cell surface of fibroblasts in confluent cell
cultures (Fig. 7b).
6. If cells are grown in semisolid gels no cubes can be cut. Therefore, em-
bedding is performed in situ in the culture vessel. Appropriate cubes
are cut after polymerization. Further processing is the same as the
preparation of cells in gels.
7. The main problem in cryomethods is the formation of ice crystals. To
overcome this, sucrose, DMSO, or glycerol is added to the culture
medium before freezing (13,23). These substances, however, particu-
470 Neumiiller

larly alter the lipid and protein constitution of the cells. Therefore
freezing methods without cryoprotectants have been developed. Im-
mediate immersion into liquid nitrogen is not useful because the con-
tact of the warm surface bearing the monolayer causes the formation
of bubbles. For this reason precooled polished copper plates are com-
monly used to which the monolayers are rapidly attached (24,25). The
procedure described in Method 9 has the advantage that the two com-
plementary parts of the broken cells can be investigated separately.
This method is very powerful in combination with immuno-histo-
chemical techniques using antibody or protein A labeled gold par-
ticles.
References
1. Wolosewick, J. J. and Porter, K. R. (1979) Microtrabecular lattice of the cytoplasmic
ground substance. Artifact or reality. J. Cell Bid. 82,114-139.
2. Pawley, J. and Ris, H. (1987) Structure of the cytoplasmic filament system in freeze-
dried whole mounts viewed by HVEM. J. Microsc. 145 (Pt 3), 319-332.
3. Schliwa, M. (1986) Whole-mount preparations for the study of the cytoskeleton in
Electron Microscopy 2986 Vol. 3 (Imura, T., Marusche, S., and Susuki, J., eds.), J. Elec-
tron Microsc. 35, Suppl. pp. 1905-1908. Japanese Sot. Electr. Microsc. Tokyo, Japan.
4. Porter, K. (1986) Section VII: High Voltage Electron Microscopy. J. Electron Microsc.
Techn. 4,142-145.
5. Buckley, I. K. (1975) Three dimensional fine structure of cultured cells: possible im-
plications for subcellular motility. Tissue 6 Cell 7,51-72.
6. Collins, V. I’., Fredriksson, B. A., and Brunk, U. T. (1981) Changes associated with
the growth stimulation of in vitro cultivated spheroids of human glioma cells. Scan.
Electr. Microsc. 1981/Q 187-196.
7, Miller, G. J. and Jones, A. S. (1987) A simple method for the preparation of selected
tissue culture cells for transmission electron microscopy. 1. Electron Microsc. Techn.
5,385,386.
8. Arnold, J. R. and Boor, I’. J. (1986) Improved transmission electron microscopy
(TEM) of cultured cells through a “floating sheet” method. J. Uftrusfruct. Molec.
Sfruct. Res. 94,30-36.
9. Hyatt, A. D., Eaton, B. T., and Lunt, R. (1987) The grid-cell-culture technique: the
direct examination of virus-infected cells and progeny viruses. J Microsc. 145 (I? l),
97-106.
10. Cajander, S. B. (1986) A rapid and simple technique for correlating light microscopy,
transmission and scanning electron microscopy of fixed tissues in Epon blocks. J.
Microsc. 143 (Pt 3),265-274.
22. Edwards, H. H., Mueller, T. J., and Morrison, M. (1986) Monolayer freeze-fracture-
a modified procedure. J. Electron Microsc. Techn. 3,439-451.
12. Arro, E., Collins, V. I’., and Brunk, U. T. (1981) High resolution SEM of cultured cells:
preparation procedures. Scan. Electr. Microsc. 198l/II, 159-168.
Transmission and Scanning Electron Microscope 471

23. Gelderblom, H. R., Kocks, C., L’Age-Stehr, J.,and Reupke, H. (1985) Comparative
immunoelectron microscopy with monoclonal antibodies on yellow fever virus-
infected cells: pre-embedding labeling versus immunocryoultramicrotomy. J. Viral.
Meth. 10,225-239.
24. Hodges, G. M., Southgate, J., and Toulson, E. C. (1987) Colloidal gold-a powerful
tool in scanning electron microscope immunocytochemistry: an overview of bioap-
plications. Scanning Microscopy 1,301318.
15. Goode, D. andMauge1, T. K. (1987) Backscatteredelectron imaging of immunogold-
labeled and silver-enhanced microtubules in cultured mammalian cells. 1.Elecfron
Microsc. Techn. 5,263-273.
16. ,Handley, D. A. (19851Ultrastructural studies of endothelial and platelet receptor
binding of thrombin-colloidal gold probes. Europ. J. Cell Bid. 39,391398.
17. Handley, D. A. (1987)Receptor-mediated binding, endocytosisand cellular process-
ing of macromolecules conjugated with colloidal gold. Scanning Microscopy 1,359-
367.
18. Bohn, W., Mannweiler, K., Hohenberg, H., and Rutter, G. (1987) Replica-immuno-
gold technique applied to studies on measles virus morphogenesis. ScanningMicro-
scopy1,319-330.
29. Paatero, G. I. L., Miettinen, H., Klingstedt, G., and Isomaa, B. (1987) Scanning elec-
tron microscopic detection of colloidal gold labelled surface immunoglobulin on
mouse splenic lymphocytes following treatment with the amphiphilic agent CTAB.
Cell. Molec. Biol. 33,13-20.
20. Handley, A. D., Arbeeny, C. M., and Witte,L. D. (1985)Intralysosomal accumulation
of colloidal gold-low density lipoprotein conjugates in chloroquine-treated fibro-
blasts, in Proceedings of fhe 43rd Annual Meeting of the Electron Microscopy Society of
Amerzku (Bailey, G. W., ed.), San Francisco Press, San Francisco, pp. 546,547.
21. de Harven, E.,Soligo, D., and Christensen, H. (1987) Should we be counting immu-
nogold marker particles on cell surfaces with the SEM?J.Microsc. 146 (Pt 2),183-189.
22. Silver, M. M. and Hearn, S.A. (1987) Postembedding immunoelectron microscopy
using protein A-gold. Ulfrastrucf. P&d. 11,693-703.
23. Linner, J.G., Livesey,S. A., Harrison, D. S.,andsteiner, A. L. (1986) A new technique
for removal of amorphous phase tissue water without ice crystal damage: a prepara-
tive method for ultrastructural analysis and immunoelectron microscopy. J. Hisfo-
them. Cytochem 34,1123-1135.
24. Bearer, E. L. and Orci, L. (1986) A simple method for quick-freezing. 1. Electron
Microsc. Techn. 3,233-241.
25. Lawson, D. (1986) Myosin distribution and actin organization in different areas of
antibody-labeled quick-frozen fibroblasts. J. Cell Sci. Suppl. 5,45-54.
 Chapter 35

Double Indirect-
Immunofluorescent
Labeling of Cultured Cells

Christine A. Boocock
1. Introduction
Immunofluorescence is a powerful technique for identifying and
localizing intra- and extra-cellular components both in histological sec-
tions and in cultured cells of plant or animal origin. Briefly, an antibody,
raised against a specific component, is used as a label to map the distribu-
tion of the component in the specimen, and then visualized under the light
microscope using a fluorescent dye (a “fluorochrome”) such as rhodamine
or fluorescein. These dyes are excited to fluoresce by microscope illumina-
tion of the appropriate wavelength. By using fluorochromes that differ
both in the wavelength required for excitation and in the color of light
emitted, several components can be mapped within the same specimen.
In this way it is possible, for example, to distinguish cell types within
tissues, to identify components involved in cell motility, adhesion and
cell-cell recognition or, using monoclonal antibodies, to detect small vari-
ations in antigen structure. Specimens must first be fixed to preserve
structure and to immobilize components that would otherwise be cross-
linked and aggregated by the antibodies used to label them. Cell mem-

473
474 Boocock

branes are then permeabilized by detergent extraction. This allows anti-

bodies to reach and label components within the cell. These procedures
can be carried out with minimal damage to cellular structure, so that the
distribution of an antigen labeled by immunofluorescence can be related to
information given by other optical techniques, such as phase contrast and
interference reflection microscopy (2,2). Furthermore, by fixing a series of
replicate cultures at intervals during an experiment, active cellular proc-
esses can be observed, the components involved being recognized by their
molecular structure (1,3). Some of the different methods available for im-
munofluorescent labeling are briefly compared below.
In direct immunofluorescence, a fluorochrome is linked to the pri-
mary antibody. This fluorescently conjugated antibody binds directly to
antigenic sites in the fixed and permeabilized specimen. Direct immuno-
fluorescence is the quickest, simplest method of immuno-labeling, but
gives rather weak specific labeling and a relatively high background from
nonspecific binding.
In indirect immunofluorescence, the fluorochrome is linked instead to
a secondary antibody directed against the primary antibody. Several sec-
ondary antibody molecules bind to each primary antibody molecule, thus
enhancing the brightness of specific labeling and reducing the relative con-
tribution of background labeling. To distinguish several components in
the same specimen, primary antibodies must be raised in separate species,
and secondary antibodies must be strictly species-specific and conjugated
to different fluorochromes whose excitation and emission spectra overlap
to a minimal extent (seeTable 1).
Protein A (4,5) is a cell wall protein (mol wt 42,000) of Stu@zyZococc~s
uureus and binds with high affinity the Fc regions (seeNote 8) of immuno-
globulin (Ig) molecules, especially of the class IgG. Fluorescently conju-
gated protein A can be used instead of a secondary antibody to cut down
background labeling, and is especially useful where background labeling
results from the binding of secondary antibodies to cell surface Fc receptors
(found on granulocytes, B-lymphocytes, and macrophages) in the speci-
men. However, only about twofold amplification is obtained in this way,
compared with the 7- to 8-fold amplification obtained using a secondary
antibody. Moreover, the usefulness of protein A is limited by its lower af-
finity for immunoglobulins of some species, particularly sheep, goat, and
rat, and for some subclasses of IgG in other species (4,6).
Biotin-avidin and biotin-streptavidin systems (5,7&biotin, a small
vitamin (mol wt 224) that can be linked to antibodies with minimal effect
Immunofluorescent Labeling 475

Table 1
Excitation and Emission Maxima of Common Fluorochromes
Abs. Emiss. Color of
max. max. observed
Fluorochrome (nm) (nm) fluorescence
Texas red 596 615 Deep red
Lissamine-rhodamine-B 570 590 Red
Rhodamine isothiocyanate (RITC) 554 573 Red
Fluorescein isothiocyanate (FITC) 492 515 Green
Aminomethyl coumarin acetic acid 350 450 Blue

on their biological activity. Egg-white avidin and streptavidin (from

cultures of Sfrepfomyces avidinii) are tetrameric proteins, both able to bind
four molecules of biotin. Streptavidin (mol wt 60,000) binds biotin with an
affinity 6-10 orders of magnitude greater than that of an antigen-antibody
interaction, and with less nonspecific binding than avidin. Using a
biotinylated primary (or secondary) antibody and a fluorescently labeled
streptavidin, three (or four) amplification steps are achieved: each antigen
molecule binds several antibodies; each antibody has several biotins
attached, each of which can bind one streptavidin molecule; and each
streptavidin can have several fluorochromes attached. Alternatively, a
bridging streptavidin (7) can be used to achieve a further threefold ampli-
fication: when bound by a biotinylated primary (or secondary) antibody,
the bridging steptavidin retains three free sites able to bind fluorochrome-
conjugated biotin molecules.
Whichever of these methods is used, the quality of immunofluo-
rescent labeling depends ultimately on that of the antibodies. Polyclonal
antisera generally give a high density of labeling because they contain
different antibodies able to bind several sites (determinants) on each anti-
gen molecule. Affinity purification removes most of those that bind to
molecules other than the intended target. Monoclonal antibodies have the
advantage of specificity for a single antigenic determinant and therefore
generally cross-react less with other molecules. Another critical factor is
the affinity of the antibody for its antigen. This varies greatly according to
the species and method of preparationof specimen material, so that quanti-
tative comparisons can be misleading. Optimal conditions and antibody
dilutions for immunofluorescence must be found, with a certain amount of
trial and error, for each new specimen and antibody, and those given here
are intended as a guide, not as a set of hard-and-fast rules.
476 Boocock

2. Materials
1. Cells cultured on coverslips. Quantities used in this procedure are
suitable for 15-mm diameter coverslips (circular coverslips are easiest
to handle, and most microscope objectives are optically corrected for
coverslip thickness no. 1.5).
2. A coverslip rack is useful to ensure equivalent processing of numer-
ous replicate cultures.
3. PBS (phosphate-buffered saline): 0.14M NaCl, 2.7 mM KCl, 1.5 mM
KH,PO,, 8.1 mM Na,HPO,.
4. PBSB: PBS containing 0.05% bovine serum albumin (BSA). Protein is
added to all antibody solutions to compete with nonspecifically bind-
ing antibodies. Fetal calf serum (10% w/v) may be used instead of
BSA.
5. Fixation buffer: 0.25% (v/v> glutaraldehyde in PBS.
6. Extraction buffer: 0.1% (v/v) Triton X-100 in PBS.
7. Sodium azide: 0.01% in PBS. (EXTREMELY TOXIC. Handle with
great care.)
8. Sodium borohydride: 0.05% in distilled water. (TOXIC and liberates
hydrogen on contact with water.) Store sodium borohydride des-
sicated, open as briefly and infrequently as possible, and only dilute
immediately before use. Add a small spatula-full to 15 mL of ice-cold
water and shake in a sealed container.
9. Clean microscope slides.
10. A water-miscible mountant, such as “Uvinert-” (Gurr) or ‘Citifluor”
(Goodwin & Davidson, Department of Chemistry, The City Univer-
sity, London). These contain “anti-fade”ingredients that retard the
fading of fluorochromes in light of their excitation wavelength.
11. Primary antibodies (either antisera or monoclonal antibodies purified
from ascites fluid) raised against each antigen to be labeled. Try to ob-
tain antibodies raised against material of the same species as your spe-
cimen. Antibodies raised against an avian protein may, for example,
have very weak affinity for the mammalian equivalent. Use of affin-
ity-purified antisera can eliminate labeling of cross-reacting antigenic
determinants. However, even monoclonal antibodies can cross-react
with similar determinants on completely different molecules. Cross-
reactivity of antibodies can be checked on immunoblots of specimen
material.
12. Preimmune sera obtained from the same host species (preferably from
the same animal) as each orimarv antibodv. On control specimens, to
Immunofluorescent Labeling 477

test for nonspecifically binding antibodies, these are substituted for

the corresponding primary antibodies.
13. Fluorescently conjugated secondary antibodies raised against im-
munoglobulins of the same species as each primary antibody. If more
than one component is to be labeled, secondary antibodies must be
sufficiently species-specific to distinguish totally between primary
antibodies. To test for cross-reactivity between secondary antibodies,
each should be used, on separate control specimens, in combination
with each primary antibody. The use of secondary antibodies pre-
adsorbed against sera of other species can obviate the need for this
control. Secondary antibodies must be conjugated with fluoro-
chromes separated as far as possible inexcitation and emission spectra
(seeTable 1). In combination with fluorescein, Texas red is better than
rhodamine in this respect. The blue dye coumarin, recently applied in
double-label immunofluorescence, additionally eliminates the exci-
tation of one fluorochrome by fluorescence emitted by the other (8).

3. Methods
3.1. Double Indirect-ImmunofZuorescent Labeling
of Cultured Cells (See Note 1)
3.1.1. Dilution of Antibodies
1. Antisera vary greatly in antibody titer and the ideal dilution to use
varies also with antibody affinity, specimen thickness, and so on. In
a preliminary trial, test on replicate specimens a few dilutions be-
tween, 10 and 100 pg/mL. Store all antibodies frozen or lyophilized,
and dilute only just before use. Solutions for double-labeling of tu-
bulin and vimentin are described in Table 2. These are given merely
as examples and are not intended for use as general recipes.
2. Centrifuge diluted antibodies 30 min at 10,000 x g, to pellet any pre-
cipitates. Keep antibody solutions on ice, and take care not to disturb
the pellet.
3.1.2. Fixation and Permeabilization of Cells
Volumes of solutions are not critical. Use enough to immerse cover-
slips and avoid their drying out. About 1 mL/coverslip should suffice (see
Note 2).
1. Rinse coverslips briefly in PBS, then pre-fix 5 min at 37*C in fixation
buffer.
478 Boocock

Table 2
Suggested Dilutions of Antisera for Double Indirect Immunofluorescent
Labeling of Tubulin and Vimentin in Cultured Cells
Primarv antibodies Volume Final dilution
Rabbit antitubulin antiserum 10 pl l/25
Goat antivimentin antiserum 5 PI l/50
PBSB 235 ~1
Preimmune sera
Rabbit preimmune serum 10 kl l/25
Goat preimmune serum 5 PI l/50
PBSB 235 J.L~
Secondary antibodies
FITC-mouse IgG antirabbit IgG 6 11 l/40
RITC-rat IgG antigoat IgG 4 PI l/60
PBSB 230 ~1

2. Rinse briefly in PBS, and then permeabilize for 15 min in extraction

buffer. Steps 2-4 may be performed at room temperature.
3. Rinse briefly in PBS, and then fix 5 min in fixation buffer.
4. Dip briefly in PBS,and then rinse in three changes of PBSover 30 min.
5. Reduce any remaining aldehyde as follows: Immerse coverslips in
FRESHLY MADE UP, ice-cold sodium borohydride solution. Incu-
bate 5 min on ice, shaking dishes continuously. Bubbles should form
if the solution is fresh. Drain dishes, add fresh borohydride solution,
and repeat.
6. Rinse as in Step 4. Continue rinsing until autofluorescence (resulting
from residual aldehyde) is completely quenched (check with a fluo-
rescence microscope).
3.1.3. Antibody Labeling
1. Drain coverslips thoroughly and place, close together but not touch-
ing, in lidded Petri dishes lined with moist filter paper (filter paper
may be omitted if a humidified incubator is to be used for incubations).
2. Add to each coverslip 60 PL of primary antibody solution (preimmune
sera to controls) and incubate 30 min at 37OC(or 1 h at room temper-
ature).
3. Drain excess antibody solution from coverslips (touch edge to filter
paper), dip briefly in PBS, and then rinse 45 min to 1 h, using at least
three changes of PBSand agitating occasionally. This rinsing must be
Immunofluorescent Labeling 479

very thorough. Because IgG antibodies are divalent, secondary anti-

bodies can cross-link any primary antibodies remaining in solution,
forming immunoprecipitates, which will appear as spurious fluores-
cent structures.
4. Repeat Steps 1-3, using secondary antibodies.
5. Continue rinsing in distilled water, checking occasionally under the
microscope for removal of background fluorescence. Use 0.01% so-
dium azide for the final rinses, to prevent bacterial growth after
mounting.
6. Invert each coverslip onto a small drop of water-miscible mountant on
a clean glass slide.
7. Observe, using a microscope equipped with epifluorescence illumi-
nation and “fluor” type objectives (seeNotes 3 and4). Onlyonefluoro-
chrome may be visualized at a time, although photomicrographs of
structures labeled with different fluorochromes may be superim-
posed on color film (seeNote 5). The corresponding beam-splitter/
interference-filter combination is inserted (and must be correctly
oriented) into the microscope body, just behind the objective. This
transmits light of the excitation wavelength to the specimen and light
of the emitted wavelength to the microscope eyepiece or camera.

4. Notes
1. The procedure given is for cultured cells, but very similar methods are
applicable to sections (l-10 pm thick) of frozen or wax-embedded tis-
sue (8,9). The use of frozen-sectioned material avoids some of the loss
of antigenicity incurred by fixation.
2. Rigorous cleanliness is important to avoid contamination with cross-
reacting antigens. Use acid-washed glass culture substrata (plastics
are easily scratched, trapping reagents, and may themselves be fluo-
rescent), separate, disposable dishes for all incubations and rinses,
and separate, clean instruments to avoid cross-contamination. Wear
gloves at all times. The amplification inherent inmost methods makes
early contamination more serious, but dust or air-bubbles included at
any stage, and in the optical system, can cause reflections with ruin-
ous effect on image contrast.
3. If labeling is very faint, first check that the fault is not in the optical
system. Second, try repeating the method using a higher concentra-
tion of the primary antibody. Third, raise the concentration of the sec-
ondary antibody. If all else fails, try a third incubation with a tertiary
480 Boocock

antibody directed against the secondary antibody and conjugated

with the same fluorochrome. The resulting chain of three antibodies
should be stabilized, before mounting, by post-fixation: Swirl for 15 s
in 5% acetic acid/80% ethanol, and then briefly in (large volumes of)
70% ethanol, 50% ethanol and water.
4. High background labeling may be the result of insufficient rinsing,
specimen contamination, cross-reactivity or nonspecific binding of
antibodies, or antibody binding to cell surface Fc receptors. Figure 1
shows the effect of these factors on the fluorescence image. Unwanted
labeling of Fc receptors can be diminished by using a biotin-avidin
system, or by replacing the secondary antibody with fluorescently
labeled protein A. Alternatively, all antibodies can be replaced with
corresponding Fab or F(ab’), fragments (seeNote 8). However, some
of the antigenicity of the primary antibody resides in the Fc region, and
fluorescently conjugated secondary antibodies (or their Fab frag-
ments) raised against Fab or F(ab’), fragments are rarely commmer-
cially available.
5. Fluorochromes, especially fluorescein, gradually fade when exposed
to light of the excitation wavelength, or even to daylight or bright ar-
tifical light. Keep specimens in the dark, e.g., foil-wrapped, and avoid
illuminating for longer than is necessary in order to focus before tak-
ing micrographs. Successive exposures of the same area will need in-
creasing exposure times. If using several fluorochromes, don’t panic:
illumination of the excitation wavelength for one fluorochrome
should not cause fading of another. Fading can be retarded by “anti-
fade” mountants although these tend to have high refractive indices,
marring the image obtained with phase-contrast optics. Some water-
miscible mountants also allow bacterial growth, which may in time
destroy the specimen. Avoid this tragedy by using 0.01% sodium
azide for the final rinses before mounting.
6. Poor fixation can impair both the antigenicity of cellular components
and the structural integrity of the specimen, and several buffering sys-
tems have been developed specifically to preserve cytoskeletal struc-
ture during fixation (10). Glutaraldehyde is a good fixative for pre-
serving cellular structure. Alternatives are paraformaldehyde (2,3),
ice-cold absolute methanol and/or acetone (3), and ethylene glycolyl
bis (succinamidyl succinate) (EGS). EGS preserves antigenicity better
than either glutaraldehyde or paraformaldehyde, because it forms
more widely spaced cross-links in the specimen material. For this
Immunofluorescent Labeling 481

reason, fixed specimens remain rather fragile. Dissolve EGS initially

in dimethyl sulfoxide and dilute to 5-10 mM in warm PBS immedi-
ately before use: it tends to precipitate once diluted if allowed to cool
below 37OC. Fixation with methanol or acetone obviate the need for
a separate permeabilizing step. Another alternative is to include de-
tergent in the fixation buffer, although this accelerates fixation, mak-
ing timing and temperature more critical. Try 0.05% Triton X-100,
0.5% glutaraldehyde in the buffer described by Small (10). Fix for 4
min at 37OC.
7. Permeabilization extracts a significant amount of cell-surface protein,
indirectly disrupting cellular structure. Cell-surface determinants,
such as growth factor, hormone, or neurotransmitter receptors and
cell adhesion molecules are best labeled on cells that have not been
permeabilized. Using permeabilized and unpermeabilized replicates,
they can be distinguished from receptors that the cell has internalized.
8. Digestion of the immunoglobulin molecule with papain gives rise to
fragments of two types: Fc (fragment crystallizable), which contains
domains able to bind staphylococcal protein A and cell-surface Fc
receptors, but has no antigen-binding specificity; and Fab (fragment
antigen-binding), which is a monovalent antigen-binding fragment,
consisting of just one of the two antigen-binding arms of the immuno-
globulin molecule. F(ab’), is a divalent antigen-binding fragment pre-
pared by digestion of immunoglobulin with pepsin.
4.1. Variations of the Technique.
Immunofluorescence has many applications beyond the mapping of
a few antigenic determinants within fixed cells and tissues. Modified
methods can be applied to living cells (12,12) or permeabilized but unfixed
“cell models” (23), giving a more dynamic picture of cellular processes.
Antibodies can also be used as tools to manipulate cellular structures, e.g.,
to cross-link cell-surface receptors (14) or to immunoprecipitate intracellu-
lar elements (25), and thus investigate the connections and functional
relationships between cellular components. Unless (monovalent) Fab
fragments are used, the (divalent) primary antibody cross-links cell sur-
face components to form a “patch.” The movement of patched determi-
nants in the cell surface membrane can be followed by subsequent fixation
and labeling with a secondary antibody (14) or by image-intensified video
microscopy (H), and is of interest in understanding the role of the cell
membrane in cell motility.
482 Boocock

Fig. 1. (a) Chick heart fibroblast labeled with antibodies to tubulin. Cells were fixed 4
min in 0.05% Triton X-100,0.5% glutaraldehyde in the buffer described by Small (1981)
(10). (b) Chick heart fibroblast badly fixed with cold methanol and labeled with antibod-
ies to tubulin. Poor fixation is indicated by disintegration of microtubules, and distortion
of cellular structure was apparent by phase-contrast microscopy. (cl Chick heart fibro-
blast labeled with antibodies to tubulin. Incomplete removal of unbound primary anti-
body has allowed the formation of fluorescent immunoprecipitates, obscuring detail of
the cell. Note: Incomplete removal of unbound secondary antibody would instead have
caused overall brightness of cells and background. (d) Chick heart fibroblast labeled with
antibodies to vinculin, a component of cell-substratum adhesions. Not only the marginal
adhesions of this cell are labeled but also its nucleus, indicating cross-reactivity with com-
ponents other than vinculin. Indeed, this antiserum cross-reacted with several nuclear
Immunofluorescent Labeling 483

The links between cytoskeletal components and the ventral mem-

brane of the cell are particularly important in the study of cell adhesion and
motility. In immunofluorescently labeled preparations, detail is often
obscured by out-of-focus images from other planes within the cell. This
problem has been overcome first by studies of isolated ventral membranes
of adhering cells (16), and second by an ingenious optical technique giving
a three-dimensional view of intact, fluorescently labeled cells (17).
Further variations of the technique have been used to throw light on
the mechanism of cell adhesion in motile and phagocytic processes. Cells
are plated onto surfaces derivatized with a “carpet” of a primary antibody
(E?), or of a protein that can later be immunofluorescently labeled (19). The
resulting pattern of dark patches reveals the regions where tight adhesion
of the cell restricts access of the secondary antibody, or where the cell has
removed the protein carpet.
The spatial relationships between cytoskeletal and extracellular com-
ponents have been investigated in ever-increasing detail by the more re-
cently developed technique of immuno-electron microscopy (20,21 and
Chapter 38, this vol). This differs from immunofluorescence in that elec-
tron-dense markers such as colloidal gold or ferritin replace fluoro-
chromes, allowing cellular components to be mapped at the much higher
resolution of the electron microscope (21). Because of the different artifacts
produced by preparation for different types of microscopy, immunofluo-
rescence and immuno-electron microscopy are especially valuable in com-
bination (20).
Other techniques closely related to immunofluorescence include the
labeling of cellular components with fluorescent cytochemical markers,
such as the phallotoxins (3,22) and heavy meromyosin (14), or their substi-

proteins on immunoblots of chick heart material. (e) Mouse macrophagelabeled with

antibodies to the membraneprotein, talin, associatedwith cell-substratum adhesions.
The antiserum was raised againstchick gizzard talin and showedlow affinity for mam-
malian talin. Cell-substratumadhesionsarenot notice-ablylabeled. Instead, the whole
cell is labeled by antibody binding to cell surface Fc receptors. Note especially bright la-
beling of membranous folds-“ruffles’‘-at both ends of the cell. (0 Mouse macrophage
labeled with antibodies to the colony-stimulating factor, CSF-1. Although Fab fragments
of the primary antibody were used to avoid labeling cell-surface Fc receptors, specific la-
beling of cell-associated CSF-1is obscured by a general cytoplasmic fluorescence because
of incomplete quenching of residual glutaraldehyde used in fixation. Exclusion of this
fluorescence from intra-cellular vesicles indicates that, unlike that in (e), it is not confined
to the cell surface.

All scale bars 10 pm.

484 Boocock

tution with fluorescent analogues, which may be incorporated into living

cells directly (12) or by microinjection (23,24). These techniques again are
most useful in combination with immunofluorescence. The high contrast
of the fluorescence image makes it suitable for video image analysis, and
the clear distinction between labeled and unlabeled cells is useful in sys-
tems for cell-counting (flow cytometry) and fluorescence-activated cell-
sorting (FACS) (seeChapters 40 and 41, this vol.).

Acknowledgment
This work was supported by the Cancer Research Campaign.

References
1. Schliwa, M., Nakamura, T., Porter, K. R., and Euteneuer, U. (1984) A tumor-promo-
ter induces rapid and coordinated reorganization of actin and vinculin in cultured
cells. 1. Cell Biol. 99,1045-1059.
2. Wehland, J., Osborn, M., and Weber, K. (1979) Cell to substratum contacts in living
cells: A direct correlation between interference reflexion and indirect immunofluo-
rescence microscopy using antibodies against actin and alpha actinin. I. Cell Sci. 37,
257-273.
3. Schlessinger,J.andGeiger,B. (1981)Epidermal growthfactor inducesredistribution
of actin and alpha-actinin human epidermal carcinoma cells. Exp. Cell Res. 134,273-
279.
4. Goding, J. W. (1978) Use of staphylococcal protein A as an immunological reagent.
J. lmmunol. Mefh. 20,241-253.
5. Hsu, S. M. and Raine, L. (1981) Protein A, avidin and biotin in immunohistochemis-
try. J Hisfochem. Cyfochem.29,1349-1353.
6. Lindmark, R., Thoren-Tolling, K., and Sjoquist, J. (1983) Binding of immunoglobu-
lins to protein A and immunoglobulin levels in mammalian sera. J. immunol. Mefh.
62,1-13.
7. Bonnard, C., Papermaster, D. S., and Kraehenbuhl, J. -I’. (1984) The streptavidin-bio-
tin bridge technique: Application in light and electron microscopic immunocyto-
chemistry, idmmunolabellingfov Electron Microscopy (Polak, J. M. and Varndell, I. M.,
eds.), Elsevier, Amsterdam, pp. 95-111.
8. Khalfan, H., Abuknesha, R., Rand-Weaver, M., Price, R. and Robinson, D. (1986)
Aminomethyl coumarin acetic acid: A new fluorescent labellingagent for proteins.
Hisfochem. J l&497-499.
9. Hayman, E. G., Pierschbacher, M. D., Ohgren, Y., and Ruoslahti, E. (1983) Serum
spreading factor (Vitronectin) is present at the cell surface and in tissues. Proc. Nut.
Acad. Sci. USA SO, 4003-4007.
1oa Small, J, V. (1981) Organisation of actin in the leading edge of cultured cells: Influ-
ence of osmium tetroxide and dehydration on the ultrastructure of actin mesh-
works. 1, Cell Biol. 91,695-705.
Immunofluorescent Labeling 485

12. Heath, J. P. (1983) Direct evidence for microfilament-mediated capping of surface re-
ceptors on crawling fibroblasts. Nature 302, 532-534.
12. Spiegel, S., Schlessinger, J., and Fishman, P. H. (1984) Incorporation of fluorescent
gangliosides into human fibroblasts: Mobility, fate, and interaction with fibro-
nectin. J. CeZI Biol. 99,699-704.
13. Weber, K., Rathke, I’. C., Osbom, M., and Franke, W. W. (1976) Distribution of actin
and tubulin in cells and in glycerinated cell models after treatment with cytochala-
sin B. Exp. Cell Res. 102,285-297.
14. Ash, J. F., Louvard, D., and Singer, S. J. (1977) Antibody-induced linkages of plasma
membrane proteins to intracellular actomyosin-containing filaments in cultured
fibroblasts. Proc. Natl. Acad. Sci. USA 74,5584-5588.
15. Mangeat, P. H. and Burridge, K. (1984) Immunoprecipitation of nonerythrocyte
spectrin within live cells following microinjection of specific antibodies: Relation to
cytoskeletal structures. J. Cell Bid. 98,1363-1377.
26. Avnur, Z. and Geiger, B. (1981) Substrate-attached membranes of cultured cells: Iso-
lation and characterization of ventral membranes and the associated cytoskeleton.
J Mol. Biol. 133,361-379.
17. Osborn, M., Born, T., Koitzsch, H. J., and Weber, K. (1978) Stereo immunofluores-
cence microscopy: I. Three dimensional arrangement of microfilaments, microtu-
bules, and tonofilaments. CelI 14,477-488.
18. Wright, S. D. and Silverstein, S. C. (1984) Phagocytosing macrophages exclude pro-
teins from the zones of contact with opsonised targets. Nature 309,359-361.
19. Chen, W.-T., Olden, K., Bernard, B. A., and Chu, F. F. (19841 Expression of transfor-
mation-sensitive proteases that degrade fibronectin at cell contact sites. J. Cell Biol.
98,1546-1555.
20. Avnur, Z. and Geiger, B. (1985) Spatial interrelationships between proteoglycans
and extracellular matrix proteins in cell cultures. Exp. Cell Res. 158,321-332.
21. Chen, W.-T. and Singer, S. J. (1982) Immuno-electron microscopic studies of the sites
of cell-substratum and cell-cell contacts in cultured fibroblasts. J. Ceil BioZ. 95,205-
222.
22. Wulf, E., Deboben, A., Bautz, F. A., Faulstich, H., and Wieland, T. (1979) Fluorescent
phallotoxin, a tool for the visualisation of cellular actin. Proc. Nutl. Ad. Sci. USA 76,
4498-4502.
23. Burridge, K. and Feramisco, J. R. (19801 Microinjection and localization of a 130K
protein in living fibroblasts: A relationship to actin and fibronectin. Cell 19,587-595.
24. Scherson, T., Kreis, T. E., Schlessinger, J., Littauer, U. Z., Borisy, C. G., and Geiger,
B. (1984) Dynamic interactions of fluorescently labelled microtubule-associated
proteins in living cells. J. Cell Biol. 99,425-434.
 Chapter 36

Gene Mapping
to Chromosomes
by Hybridization In Situ

John R. Gosden
1. Introduction
Since the technique of hybridizing labeled nucleic acid sequences to
cytological preparations (hybridization in situ) was first described in 1969
by Pardue and Gall (I), it has undergone considerable refinement. In those
early days, the technique was only capable of detecting and locating highly
repeated target sequences such as satellite DNAs (2) or the genes for ribo-
somal DNA (3).
The introduction of methods that combined positive identification of
every chromosome in each cell with a quantitative analysis of the distribu-
tion of autoradiographic grains (4) opened the way to more sensitive uses
of the technique and ultimately to the present situation, in which hybridi-
zation in situ is routinely used to map the precise chromosomal location of
DNA sequences as little as 1000 base pairs (one kilobase, 1 kb) in length.
The technique described here can be broken down into six stages:
1. Chromosome preparation. Chromosomes are prepared from cell cul-
tures, which may be of permanent cell lines (fibroblasts, lymphoblast-
oid cells, and so on), or primary cultures. The cell type most frequently

487
488 Gosden

used for mapping human DNA sequences is the phytohemagglutinin

(PHA) stimulated peripheral blood lymphocyte. The chromosome
preparations are spread onto clean glass slides. Since it is essential to
be able to identify each chromosome, some form of banding proce-
dure is needed. The method I use, which produces bands on the chro-
mosomes after hybridization in situ, requires specific treatment dur-
ing culture and preparation. The method used must produce a high
mitotic index with well-spread metaphase plates and chromosomes of
adequate length for good band resolution. It is obviously important
that the method minimizes any loss of DNA from the chromosomes,
while ensuring that the DNA is readily accessible to the probe.
2. Labeling the probe. The label of choice is tritium, because this pro-
vides the best combination of sensitivity and resolution. Alternative
labels are discussed in the notes. The maximum specific radioactivity
that can be obtained without compromising the integrity of the probe
is the ideal. The method of labeling may be either nick translation or
random oligonucleotide primed synthesis.
3. Hybridization. This comprises two steps: denaturing the chromo-
somal and probe DNA, and annealing them together. The denaturing
steps must minimize loss or degradation of the DNA while maxi-
mizing the separation of the DNA duplex, yet still retain chromosome
morphology. The annealing conditions are selected to reduce the
chance of producing a poorly matched hybrid duplex, yet obtain the
greatest possible amount of specific hybridization. In practice, the hy-
bridization conditions are standard, and the posthybridization wash
conditions are varied to increase or reduce the stringency of the reac-
tion as required.
4. Autoradiography. This involves coating the slides with sensitive
emulsion, exposing them for a suitable length of time, and developing
the autoradiograph.
5. Staining-banding chromosomes after hybridization.
6. Analysis-of grain distribution.

2. Materials
2.1. Chromosome Preparation
from Human Peripheral Blood
1. Medium: RPM1 1640 (store at 4OC).
2. Phytohemagglutinin (PHA): dissolved in 5 mL sterile H20 (store at
4°C).
Gene Mapping to Chromosomes 489

3. Fetal calf serum (store at -2OOC).

4. 5-Bromo-2’-deoxyuridine (BUdR): 10 mg/mL in sterile water, filtered
through a sterile 0.2~pm millipore filter (store at 4OC).
5. Tris-EDTA (TE): 1OmM Tris-HCI pH 7.5; 1mM EDTA, (autoclaved).
6. Thymidine: 1O3M in TE, filtered through sterile millipore filter (store
at 4OC).
7. Colcemid: 10 pg/mL (store at 4OC).
8. Methanol, absolute.
9. Glacial acetic acid.
10. 0.075M KCl: 0.56 g/100 mL 30.

2.2. Labeling Probe DNA by Random

Oligonucleotide Primed Synthesis
1. Deoxy(1’,2’,2,8-3H) adenosine 5’-triphosphate (50-100 mCi/mmol)
(dATl?) (store at -2OOC).
2. Deoxy(1’,2’,5-3H) cytidine 5’-triphosphate (50-85 mCi/mmol) (dCTP)
(store at -20°C).
3. (MethyZ,l’,2’-3H) thymidine 5’-triphosphate (90-130 mCi/mmol) (dT”TE)
(store at -2OOC).
4. 2’-deoxyguanosine 5’-triphosphate (dGTP): 0.1 mM in TE (store at
-2OOC).
5. Mixed sequence hexadeoxynucleotides: 1 mg/mL in TE (store at
-20°C).
6. DNA Polymerase 1: large fragment, Klenow enzyme (store at -2OOC).
7. 10 x concentration salt solution: 0.5M Tris-HCl pH 7.8; 50 mM MgCl,;
10 mM B-mercaptoethanol; 500 pg/mL bovine serum albumin (BSA)
(store at -2OOC).
8. TNE column buffer: 02MNaCl; O.OlM Tris-HCl pH8.0; 1 .OmM EDTA
(store at room temperature (RT).
9. Water-saturated, distilled phenol (store at RT).
10. Chloroform: octan-2-ol(24:l) (store at RT).
11. Sonicated salmon sperm DNA: 10 mg/mL in TE (store at 4OC).
12. 3M ammonium acetate (store at 4OC).
13. Scintillation fluid made as follows: 2(4’- t-Butylphenyl)-5-(4’-biphenyl)-
1,3,4-oxadiazole (Butyl-PBD), 8 g, dissolved in 1000 mL of toluene, to
which is added 600 mL 2-ethoxyethanol (Scintillation grade).
14. Nick column (Pharmacia).

2.3. Hybridization
1. 20 x SSC: 3M NaCl; 0.3M tri-sodium citrate.
490 Gosden

2. Ribonuclease A (10 mg/mL in 2 x SSC, boiled 5 min to inactivate

DNAse) (store at 4OC).
3. Formamide (Analar).
4. Dextran sulfate.
5. 10 x SSCB: 1.5M NaCl; 0.15M tri-sodium citrate; 0.25M Tris-HCl pH
7.4; 0.5mM EDTA (store at 4°C).
6. E. coli tRNA: 4 mg/mL in TE (store at 4OC).
7. Rubber solution (e.g., Pang Super Solution, Fritz Hesselbein Chemis-
chefabrik, Nurderstedt).

2.4. Autoradiography
1. Dark room with safelight (Kodak S902filter), water bath at 45”C, light-
tight cabinet.
2. Ilford L4 Nuclear Emulsion (KodakNTB2 emulsion can besubstituted)
(store in dark at 4°C).
3. Dipping vessel (flat Pyrex tubing, 30 x 7 mm inside measurements, 80
mm long, sealed and rounded at one end).
4. Light-tight slide containers.
5. Kodak D19 developer.
6. Kodafix.

2.5. Staining-Post-hybridization Banding and Analysis

1. Bisbenzimide H33258 fluorochrome (H33258): 500 pg/mL in HO
(store in dark at 4OC).
2. Giemsa stain: 2% in Gurr’s buffer pH 6.8, diluted fresh on each occa-
sion.
3. 2 x SSC: 0.3M NaCl; 0.03M tri-sodium citrate (store at RT).
4. D.P.X. mountant.
5. Printed chromosome idograms for grain distribution analysis (seeFig.
1).
3. Methods
3.1. Chromosome Preparation from Human
Peripheral Blood Igmphocytes
All solutions used prior to harvest must be sterile.
1. Medium: To a sterile medical flat or other closed glass vessel contain-
ing 200 mL sterile RPMI 1640 medium, add 2mL PHA (dissolved in
Gene Mapping to Chromosomes 491

sterile distilled water as directed on container) and 30 mL sterile fetal

calf serum. For culturing, this is then dispensed with a sterile pipet in
9-mL aliquots into sterile glass vessels of 20-25 mL capacity.
2. Culture: Blood is obtained by venipuncture with a sterile syringe and
needle and placed in a vial containing lithium heparin as an antico-
agulant. Of this fresh whole blood 1 mL is added to 9 mL of medium,
gently mixed, and placed in a water bath or incubator at 37°C for 72 h.
It will be found convenient to do this at about 3 PM on Friday. After
72 h (3 PM Monday) 200 ~1 of BUdR solution is added, gently mixed,
and incubation is continued for a further 18 h (seeNote 1).
3. At 8 AM the following day (Tuesday), transfer the cultures to sterile
capped centrifuge tubes and spin them in a bench centrifuge at 1800
rpm for 5 min.
4. Gently decant the supernatant, without disturbing the pellet, and
gently resuspend the cultures in RPM1 1640 (without PHA or serum).
5. Repeat the centrifugation and resuspension, and then centrifuge for a
third time. On this occasion, the cultures are resuspended in complete
medium and transferred to fresh, sterile glass culture vessels.
6. To each, add 100 PL of sterile 1PM thymidine. Incubate the cultures
for a further 5 h 15 min at 37”C, at which time add 100 PL colcemid sol-
ution to each culture and continue incubation for a final 40 min.
7. Transfer the cultures to fresh, sterile capped centrifuge tubes, and spin
at 1800 rpm for 5 min.
8. Gently remove the supernatant by pipet, and add 5 mL freshly made
hypotonic (0.075M) KC1 solution at RT drop by drop while mixing on
a vortex mixer.
9. Leave the cultures at RT for 10 min, and then centrifuge again.
10. Carefully remove the supernatant and add (drop by drop on a vortex
mixer as before) 3 mL of a freshly made mix of methanol: glacial acetic
acid (3:l).
11. Top the tube up to the 10 mL level with fix, and mix well. Leave in this
first fix for up to 30 min, then spin again, decant the supernatant and
add a further 10 mL of 3:l fix, mixing well.
12. Centrifuge again, decant the supernatant, and add 10 mL of fix.
13. At this stage, a test slide can be made by dropping one drop of the re-
suspended chromosome preparation onto a clean glass slide and ex-
amining it under phase contrast. If phase optics are not available, an
approximation can be made by removing the condenser from a stan-
dard microscope. This will let the observer see how many cells are in
metaphase, how well the chromosomes are spread, and how much cy-
492 Gosden

toplasm is still attached to the chromosome spreads. It is worth stain-

ing one slide (2 min in 2% giemsa) to check that nuclear envelopes are
removed to permit access of probes to chromosomes. This cannot be
readily determined under phase contrast.
14. A further two or three changes of fix will be necessary in most cases.
After the final change, resuspend the pellet in 1 mL of fix.
15. Double-frosted glass slides should be used. The slides are cleaned
prior to use by soaking in absolute ethanol to which 1% HCl has been
added, and polishing with fine muslin. (The same treatment is used
for coverslips. These will be termed clean slides and coverslips.)
16. One drop of cell suspension is dropped into the center of a clean slide.
The height from which it is dropped, and whether the slide should be
warm or cold, wet, dry, or damp (from being breathed on) will vary
from day to day according to the local climatic conditions. Each work-
er will find the best treatment for the local conditions by trial.
17. The slides should be dried at room temperature and then stored in
light-tight containers in desiccators under vacuum. Under these con-
ditions, a batch of slides should remain useful for at least 2 mo, and
may remain so for much longer.

3.2. Labeling Probe DNA by Random Priming

(See Note 2)
1. A minimum of 0.3 pg of linear DNA is needed, and the maximum to
be labeled at any one time is 1 pg.
2. In a sterile plastic microcentrifuge tube, place 0.5 nmol of each of 3H-
labeled dATP, dCTl?, and dTTP. Dry down by blowing a gentle
stream of air into the tube (just sufficient to ruffle the surface of the
fluid) from a glass Pasteur pipet connected to an aquarium aeration
pump. This procedure takes about 3 h, and may conveniently be done
overnight.
3. Mix the probe DNA, 2 PLof hexanucleotide solution and sterile, dis-
tilled Hz0 to a total volume of 16 PL in a second sterile 1.5 mL plastic
microcentrifuge tube. Spin briefly to collect, pierce a fine hole in the
lid with a heated syringe needle, and place in a boiling water bath for
3 min. Remove from the bath, cool to room temperature, and spin
again to collect.
4. Transfer the contents of the second tube to that in which the labeled
nucleotides have been dried. Add 2 PL of 10 x salt solution and 1 PL
of unlabeled dGTl?. Mix well (ensuring that all the dried nucleotides
Gene Mapping to Chromosomes 493

are redissolved), and add 1 PL of Klenow enzyme. Mix gently and in-
cubate in a water bath at 37OC for 90 min to 2 h.
5. Meanwhile, prepare a Nick column by washing with three reservoirs
full of TNE to replace storage buffer with column buffer, and ensure
column bed is properly equilibrated.
6. Add 80 PL of TNE and 50 PL of water-saturated phenol to the incuba-
tion mix. Mix well and incubate for a further 10 min at 37°C. Add 50
~.LLof chloroform: octan-2-01, mix again, and spin for 2 min to separate
the aqueous and phenol phases.
7. With a glass Pasteur pipet, carefully transfer the upper (aqueous)
phase to the Nick column being scrupulously careful to avoid picking
up any of the lower (phenol) phase. Having allowed the aqueous
phase to run into the column bed, add 400 JJL TNE to the column and
let this run through. Add a further 400 uL of TNE and collect the drops
from the column (total volume should be approximately 400 pL) in a
clean, sterile plastic microcentrifuge tube.
8. Take a 4-VL aliquot from this, place in a scintillation vial containing 3
mL scintillation fluid, mix well, and count in a liquid scintillation
counter. There should be no less than 50,000 cpm (about 100,000 dpm)
in this sample. If there are fewer counts than this, incorporation of
label may well be insufficient to detect in the hybridization.
9. If the counts are adequate, add 10 uL of sonicated salmon sperm DNA
(to act as coprecipitant) and 50 yL of 3M ammonium acetate to the
microcentrifuge tube, mix well, and add 1 mL of absoluteethanol. Mix
again and place the tube at -2OOC for at least 1 h.
10. Centrifuge for 5 min, decant the supernatant (into a new tube) and
place the sample in a lyophilizer (Freeze-drier) to evaporate the resid-
ual alcohol.

3.3. Hybridization (See Note 3)

1. Make a stock solution of 20% dextran sulfate in formamide. Since this
takes a long time to dissolve and seems stable at room temperature for
several months, it is worth making a large volume such as 10 mL.
2. Treat the slides with ribonuclease A to remove any endogenous RNA
that might otherwise compete for the probe. Make a solution of 100
pg/mL RNAse A in 2 x SSC by adding 1% of the stock solution to the
appropriate volume of 2 x SSC. For a normal experiment, using six
slides, 50 mL final volume in a Coplin jar is sufficient. Place the jar in
a water bath at 37°C and incubate the slides in it for 1 h. Remove the
494 Gosden

slides and pass them through 70%, 90%, and 100% ethanol (5 min
each). Dry the slides under vacuum.
3. Denature the chromosomal DNA by placing the slides in a solution of
0.6 x SSC; 70% formamide at 70°C for 2 min, transferring them in-
stantly to 70% ethanol after this time, and passing through the alcohol
series as above, before drying under vacuum. At this stage, provided
the slides are kept dry, they are stable for at least several hours, if not
days.
4. Meanwhile, make up the hybridization mix as follows: For each slide,
allow 20 PL of mix, plus an extra 40 pL, i.e., for six slides (6 x 20) + 40
= 160 pL. Of this total, 20% is the labeled probe dissolved in sterile dis-
tilled water, 20% is 10 x SSCB, 10% is E. co2i tRNA solution, and 50%
is 20% dextran sulfate in formamide. For the six slides, dissolve the
dry, labeled probe in 33 u.L sterile distilled water, leave 5 min to allow
complete solution, and take 1 PL and count in a scintillation counter
(leaving the 32 PL needed for the hybridization mix). Add 32 PL 10 x
SSCB, 16 PL tRNA and 80 PL dextran/formamide. This last is veryvis-
cous: it is best to cut off the end of a micropipet tip to dispense this
material, and use the same tip to mix it with the remainder of the hy-
bridization mix by pipeting up and down. Mix very thoroughly and
place in a water bath at 70°C for 5 min. At the end of this time, trans-
fer to an ice bath.
5. Prepare six 20 x 40 mm clean coverslips. Place 20 PL of hybridization
mix on each coverslip, and invert a slide over each. Allow the mix to
spread, ensuring that there are no air bubbles; when the area under the
coverslip is covered, seal with rubber solution to prevent evaporation
of hybridization mix.
6. When the rubber solution is dry, transfer the slides to a thin metal ves-
sel floating in a 37OCwater bath with a lid, or a closed container in a
37OC incubator. Leave overnight.
7. Meanwhile, prepare the washes. In 1000 mL reagent bottles or flasks,
make (a) 50% formamide; 1 x SSC and (b) 1 x SSC. Place in a water bath
at 42OC.
8. The following morning, carefully peel the rubber solution away from
the coverslips, and remove them from the slides. (They should come
away easily if the rubber solution provided a good seal.)
9. Place the slides in a glass stain rack, and place this in a 250-mL capacity
stain dish in the 42OC water bath. Add 50% formamide; 1 x SSC, agi-
tate gently, and leave for 5 min.
Gene Mapping to Chromosomes 495

10. Repeat three times, and then give four washes of 5 min each in 1 x SSC.
11. Transfer to 70% ethanol and pass through the series to absolute alco-
hol as before. Air dry or dry under vacuum.

3.4. Autoradiography
1. In a darkroom under a safelight, heat the water bath to 45OC. In the
water bath place: (a) a 25-mL measuring cylinder containing 5 mL dis-
tilled water and (b) a lOO-mL beaker full of water, in which the dipping
vessel is supported. Place the slides to be dipped face up on a hot plate
that is just warm to the touch. (This helps the emulsion to spread
thinly and uniformly.)
2. With a pair of sterile plastic forceps, transfer emulsion from the con-
tainer to the measuring cylinder until the volume in the latter is 10 mL.
Allow to stand 5 min to reach melting temperature. Mix very gently,
avoiding air bubbles, and pour the diluted emulsion into the dipping
vessel. Take two slides, back to back, and dip into the emulsion, so that
the frosted area is just clear of the emulsion surface.
3. Remove gently (without scraping the sides of the vessel), separate the
slides, and place in a rack to dry. (A standard test-tube rack, with the
slides diagonally across the apertures will do.)
4. When all the slides have been dipped, place the rack in a locked, light-
tight cabinet, and leave for 2.5-3 h to dry.
5. When dry (still under safelight), transfer the slides to a light-tight box
containing a small amount of silica gel separated from the slides by a
piece of tissue paper. Seal the box and place at 4°C in a refrigerator free
from all sources of radiation.
6. After 6 d, bring the slides to room temperature (this takes about an
hour) and, in the dark room, under safelight, remove one slide from
the box.
7. Develop for 5 min in Kodak D19, agitating once a minute, stop in
water, and fix for 5 min in Kodafix, diluted 1:3 with water, again agi-
tating once a minute.
8. Wash under gently running cold water for 6 min.
9. Allow to dry, and stain in 2% Giemsa in Gurr’s buffer for 5 min.
10. Air dry, and examine under a microscope. If there are more than 2 or
3 autoradiographic grains/metaphase, develop the remainder of the
slides. If not, leave a further 6 or 7 d, by which time, if the experiment
has worked, there should be enough grains.
496 Gbsden

3.5. Staining-Banding
1. After the slides have been developed and washed, allow them to dry
thoroughly-at least overnight-to ensure that the emulsion is hard
again.
2. Stain slides for 15 min inH33528 diluted to 50 pg/mL in 2 x SSC. Rinse
in2xSSC.
3. Mount in 2 x SSC under a 20 x 40 mm coverslip and, in a darkroom,
place under UV lamp, set to long wave, at a distance of lo-12 cm, for
30-60 min.
4. Place slides in a Coplin jar containing 2 x SSC for 35-45 min at room
temperature.
5. Dry slides, and stain for 15 min in freshly diluted 2% Giemsa. It is
probably wise to treat only three or four slides at a time. The chromo-
somes should show clearly defined G-bands.
3.6. Analysis (see Note 4)
1. When completely dry, dip the slides in xylene for 5 min before cover-
ing with a 20 x 40 mm coverslip mounted in D.P.X. When they are
completely dry, examine under microscope.
2. A suitable form of idiogram is shown in Fig. 1. It is important that the
resolution implied by the bands on the idiogram is equivalent to that
seen down the microscope!
3. Identify those chromosomes on which grains can be seen, and mark
the grains in the correct location on the idiogram. Only grains in con-
tact with a chromosome should be scored. In this way, the total dis-
tribution of grains is accumulated, and it is possible to distinguish
specific signals from the background.
4. The number of cells needed to obtain a significant result varies from
probe to probe. In general, the highest proportion of grains seen on the
specific chromosome is about 20%, i.e., only one grain in five is actu-
ally found on the correct chromosome, and of these between 60 and
90% may be located at the correct band. It is therefore rarely possible
to get a significant result without scoring at least 20 cells, and it may
be necessary to score many more-perhaps as many as 80 or 100. It
should be possible to resolve the location of a single copy sequence to
within a major band, and in most cases to within one or two minor
bands.
4. Notes
1. Chromosome preparation: It is important to remember that the BUdR
treatment makes the chromosomal DNA sensitive to light, and there-
Gene Mapping to Chromosomes 497

Fig. 1. Hybridization in situ of ?&ll to human metaphase chromosomes. Mcll is a

cloned genomic probe containing 5.3 kb of human DNA. It was labeled with tritium by
random priming and hybridized as described in the text. The autoradiographic exposure
was for 7d. Forty-nine cells were analyzed, with a total of 100grains on the chromosomes.
Of this total, 17 (17%) were on chromosome 6, and of these,7 grains (41%) were on band
6~12.

fore to minimize exposure to light, the slides should be stored in light-

tight containers. Storage of the slides in dessicators under vacuum
seems to slow down the aging process which eventually reduces their
hybridization efficiency.
Other methods of banding can be used that do not require the BUdR
treatment. These include methods of banding the chromosomes be-
fore hybridization, either by Lipsol-Giemsa (5) or Trypsin-Giemsa (6).
These have several disadvantages: First, selected cells must be photo-
graphed to record the bands, since these are lost during the hybridi-
zation process. The photographed cells are relocated after autoradio-
graphy, and the labeled chromosomes identified by reference to the
photograph. Only a proportion of the selected cells prove suitable for
analysis, which means some labor and time has been wasted. Sec-
ondly, the staining treatment results in some loss of DNA from the
chromosomes. Should the lost sequences include that of interest, the
efficiency of hybridization will be reduced. A method of banding after
hybridization using Wright’s stain on cells synchronized with meth-
otrexate has been described (7), but in my hands, this does not give the
498 Gosden

consistency or quality of bands produced by the method I have de-

scribed. Some methods for producing elongated chromosomes have
depended on the incorporation of ethidium bromide into the DNA.
This seems to reduce the efficiency of hybridization, probably because
the dye is intercalated into the DNA, and the use of such methods is
not encouraged.
2. Labeling the probe: In order to obtain the highest specific activity with
“H, three labeled nucleotides are used. Since the incorporation of
dGTP is less efficient, this base is not normally labeled. Of the two pos-
sible methods of labeling DNA, nick translation (8) and random oligo-
nucleotide primed synthesis (9), the latter usually gives higher speci-
fic activity, but in some circumstances (seebelow), nick translation may
be preferable.
Random priming requires the substrate probe to be linearized, by
cutting with a restriction endonuclease with a single recognition site.
This method gives longer stretches of intact labeled DNA than nick
translation. Most experiments with irz situ hybridization are attempts
to map the location of a sequence that exists in a single copy at one site
in the haploid genome. However, if the probe is derived from a cloned
genomic DNA sequence it may, in addition to the single copy se-
quence of interest, include sequences that are repeated and dispersed
throughout the genome. Even if the probe is derived from a cDNA, it
may still include sequences that are repeated elsewherein the genome.
In either of these cases, it may be difficult or impossible to distinguish
the site of interest from the other labeled sites and background. In this
case, hybridization of the repeated elements can be reduced or even
eliminated by preannealing the probe with an excess of unlabeled
sheared total DNA of the same species (20). This process (known as
“stripping”) requires that the fragments of labeled probe containing
the repeats can form duplexes with the unlabeled total DNA (and
thereby be prevented from taking part in the hybridization reaction)
without compromising the unique sequence of interest. For this
purpose, the lower mol wt product of nick translation may be prefer-
able.
Alternative radioisotopes are ‘%I and %. These have the advantage
over 3H that higher specific activities can be obtained using only one
labeled nucleotide. However, because of the greater energy of emis-
sion from these isotopes, the resolution is poorer and background
higher. Nevertheless, some groups have success, particularly with *?I
(11). Nonradioactive systems, most of which are based on the use of
biotin incorporation into the probe, and its detection by immunologi-
Gene Mapping to Chromosomes 499

cal methods, hold hope for the future (12). Despite encouraging re-
ports (23), however, they are not yet sufficiently sensitive for routine
use.
3. Hybridization: All glass pipets and containers, micropipet tips, mi-
crocentrifuge tubes, and solutions must be sterilized by autoclaving or
filtration to reduce the risk of contamination with DNase. The num-
ber of slides suggested for an average experiment (six) is not an arbi-
trary figure. It provides one or two slides for flat staining (unbanded)
to check the progress of the autoradiographic exposure, plus four
slides for banding and full analysis. This should give an adequate
number of analyzable metaphase chromosomes to obtain a result (see
above).
4. Analysis: It is possible to carry out analysis of grain distribution by
photographing suitably banded and labeled cells, identifying the la-
beled chromosomes on the photograph, and transferring the accu-
mulated data to an idiogram. This has the advantage of providing a
permanent photographic record of the analyzed cells, but the disad-
vantage that the autoradiographic grains are frequently in a different
focal plane from that of the chromosomes. In half-tone photographs,
these are not always readily distinguished from stain debris. The
method that I use, however, stores the permanent data in the form of
the slides themselves, with microscope vernier references to identify
the analyzed cells.

Acknowledgments
I would like to thank Professor H. J. Evans for his continued support
during the course of the development of the techniques described here,
Derek Rout for his excellent technical assistance, and Alison Brown and
Derek Rout for their critical reading of the manuscript.

References
1. Pardue, M. L. and Gall, J. G. (1969) Molecular hybridization of radioactive DNA to
the DNA of cytological preparations. Pm. Nafl. Acud. Sci. USA 64,6oo-604.
2. Pardue, M. L. and Gall J. G. (1970) Chromosomal location of mouse satellite DNA.
Science 168,1356-1358.
3. Henderson, A. S., Warburton, D., and Atwood, K. C. (1972) Location of ribosomal
DNA in the human chromosome complement. Pm. N&l. Ad. Sci. USA 69,3394-
3398.
4. Gosden, J. R., Mitchell, A. R., Buckland, R. A., Clayton, R. I’., and Evans, H. J. (1975)
The location of four human satellite DNAs on human chromosomes. Eqd. Cell Res.
95148-158.
500 Gosden

5. Gosden, J*R., Middleton, P. G., Rout, D.,andDe Angelis, C. (1986) Chromosomal lo-
calization of the human oncogene ERBA2. Cytogenet. Ceil Genet. 43,150-153.
6. Garson, J. A., van den Burghe, L. J. A., and Kernshead, J. T. (1987) High-resolution
hybridization technique using biotinylated NMYC oncogene probe reveals periodic
structure of HSRs in human neuroblastoma. Cytogenet. Cell Genet. 45,X&15.
7. Harper, M. E. and Saunders, G. M. (1981) Localization of single-copy DNA sequen-
ces on G-banded human chromosomes by in situ hybridization. Chromosoma 83,
431-439.
8. Rigby, I’. W. J., Diekman, M., Rhodes, C., and Berg, I’. (1977) Labelingdeoxyribonu-
cleic acid to high specific activity in vitro by nick translation with DNA polymerase
I. J. Mol. Biol. 113,237-251.
9. Feinberg, A. I’. andvogelstein, B. (1983) A technique for radiolabeling DNA restric-
tion endonuclease fragments to high specific activity. And. Biochem. 136,613.
10. Porteous, D. J., Bickmore, W., Christie, S., Boyd, I’. A., Cranston, G., Fletcher, J. M.,
Gosden, J. R., Rout, D., Seawright, A., Simola, K. 0. J., Hastie, N. D., and van
Heyningen,V. (1987) HEASI-selected chromosome transfer generates markers that
colocalize aniridia- and genitourinary dysplasia-associated translocation break-
points and the Wilms tumor gene within band 11~13. Proc. Nufl. Ad. Sci. USA 84,
5355-5359.
22. Buckle, V. J. and Craig, I. W. (1986) In situ hybridization, in Human Genetic Disease:
II Practical Approach (Davies, K. E., ed.), IRL Press, Oxford, pp. 85-101.
12. Gosden, J. R. andPorteous,D. J. (1987) HRASI-selected,chromosomemediated gene
transfer; in situ hybridization with combined biotin and tritium label localizes the
oncogene and reveals duplications of the human transgenome. Cyfogenet. CeEZGenet.
45,44-51.
23. Landegent, J. E., Jansen in de Wal, N., van Ommen, G. -J. B., Baas, F., de Vijlder, J. J.
M., vanDujin, P., and van der Ploeg, M. (1985) Chromosomallocalizationof a unique
gene by nonautoradiographic in situ hybridization. Nature 317,175-177.
 Chapter 37

In Situ Hybridization with

Radiolabeled cRNA Probes,
Using Tissue Sections
and Smears

Carolin Adams, Qutayla A. Hamid,

and Julia M. Polak

1. Introduction
Immunocytochemistry., together with biochemical procedures, has
ensured rapid and accurate identification of tissue components. This has
led to a better appreciation of cellular events, particularly storage and se-
cretion of products. However, such techniques have certain drawbacks, in
particular the impossibility of monitoring the intracellular processes con-
cerned with protein synthesis. Thus, there was a need for a method that
would provide more detailed information about the functional morphol-
ogy and gene expression in tissues at cellular level. Advances in molecular
biology allowed the development of in situ hybridization, a procedure that
localizes specific nucleotide sequences (DNA or RNA) in tissue prepara-
tions using labeled complementary probes (DNA or RNA).

501
502 Adams, Hamid, and Polak

Initial in situ studies (1,2,3) involved the visualization of ribosomal

DNA, using radiolabeled cRNA probes. Since then, the technique has ex-
panded to include the hybridization of both DNA and cytoplasmic RNA
with a variety of nucleic acid probes (4,5).
For localization of RNA, there are four types of nucleic acid probes:
double-stranded DNAprobes (5), single-stranded cDNA probes (6), cRNA
probes (7), and finally, synthetic oligonucleotide probes (5). Double-
stranded probes require “melting” and may reanneal during hybridiza-
tions, thus reducing the amount of probe available for reaction with the
mRNA. The advantages of cRNA probes include the high thermal stabil-
ity and affinity of RNA-RNA hybrids, a constant probe size, no vector se-
quences, and the availability of RNase to remove unhybridized probes (8).
In addition, a probe with an identical sequence to the mRNA can be pre-
pared and used as a sense probe.
The probes can be labeled with radioactive or nonradioactive tags.
The most successful of the nonradioactivemethods uses biotin-substituted
nucleotides. However, at present, autoradiography appears to be the most
sensitive detection method available for hybridization histochemistry. In
the following pages, we will outline the method we follow using radioac-
tive labeled riboprobes and give some examples of in situ hybridization for
the investigation of the diffuse neuroendocrine system (9,10,11).

2. Materials
2.1. Tkanscription
1. 5 x Transcription Buffer: 0.2M Tris-HCl, pH 7.5,30 mM Mg Cl, 10 mM
Spermidine.
2. 100 mM dithiothreitol (DTT).
3. RNasin (Human Placental Ribonuclease Inhibitor): 25 U/pL.
4. Nucleotide Mixture: 2.5 mM each of ATP, GTP, and UTP.
5. 100 cln/l cytidine triphosphate (CTP).
6. 1 mg/mL linearized plasmid template DNA in water or Tris-EDTA
Buffer.
7. Cytidine (~-~~l? triphosphate) 10 mCi/mL.
8. SP6 RNA Polymerase, T7 RNA Polymerase, or T3 RNA Polymerase:
10 U/ pL. These enzymes are very labile and should be out of the -20°C
deep freeze for a minimal time.
9. DNase (RNase free): 1 pg/pL.
10. t RNA: 10 pg/j,tL.
In Situ Hybridization 503

11. 4MNaCl.
12. Phenol: Melt the solid at 68”C, add 8-hydroxy-quinoline (antioxid-
ant) to a final concentration of 0.1%. Extract several times with an
equal volume of buffer (1 .OM Tris HCl, pH 8.0, followed by O.lM Tris
HCl, pH 8.0 with 0.2% p mercaptoethanol) until the pH of the aqueous
phase is 7.6.
13. Chloroform:isoamyl alcohol (24:l).
14. 7M Ammonium Acetate.
15. Absolute Ethanol.
16. 10% TCA (Trichloroacetic acid).
17. Bovine Serum Albumin: 10 pg/pL.

Store solutions l-10,15,17at -2O”C, and 12 at 4°C. All other solutions

may be stored at room temperature.

2.2. Fixation
1. O.lM phosphate-buffered saline (PBS): Dissolve 87.9 g NaCl, 2.72 g
KH,l?O, and 11.35 g N%Hl?O, (or 23.9 g Na, HPO,* 1230) in 10 L of
distilled water.
2. 4% Paraformaldehyde: Dissolve 4 g of paraformaldehyde in 100 mL
of hot PBS (50-6OOC) with stirring. Continue stirring until the solu-
tion is clear. If necessary, add 1ONNaOH dropwise until the solution
clears. Cool, check pH (7.2), and use immediately.
3. 15% Sucrose in phosphate-buffered saline.
4. Polyz-lysine-coated glass slides: Soak slides in detergent overnight.
Wash in running tap water for 4-6 h. Rinse in several changes of
double-distilled %O. Bake at 250°C for 4 h. Coat slides with 0.01%
poly+lysine (Sigma, mol wt 300,000) (stored at -20°C). Air dry.

2.3. In Situ Hybridization with cRNA Probes

1. Phosphate-buffered saline,
2. O.lM Glycine in phosphate-buffered saline.
3. 0.3% Triton X 100 in phosphate-buffered saline.
4. Proteinase K solution: 1 ug/mL, O.lM Tris HCl pH 8.0,500 mM EDTA
pH 8.0 (Stock proteinase K: 500 pg/mL, store at -20°C).
5. 4% Paraformaldehyde (freshly prepared).
6. Acetylation solution: 0.25% (v/v) Acetic Anhydride, O.lM Trietha-
nolamine pH 8.0. Use immediately.
7. Formamide.
504 Adams, Hamid, and Polak

8. lOxstandardsodiumcitrate(lOxSSC)stock: 1.5MNaCl,O.l5MTriso-
dium citrate. Dilute as required.
9. Hybridization solution: 50% deionized formamide, 12.5 x Denhardts,
10% dextran sulfate, 250 mMTris HCl pH 7.5,0.5% sodium pyrophos-
phate, 0.5% sodium dodecyl sulfate (SDS) and 250 pg/mL denatured
salmon sperm DNA. This is prepared freshly from the following stock
solutions, which are stored at -2OOC:
a. 100% deionized formamide.
b. 100x Denhardts: 2% bovine serum albumin, 2% polyvinyl-
pyrrolidone (WI?-360), 2% Ficoll400.
c. 5M Tris-HCl, pH 7.5.
d. 30x standard sodium citrate (SSC).
e. 20% sodium dodecyl sulfate (SDS).
f. 50% dextran sulphate.
g. Salmon sperm DNA: 20 mg/mL.
10. RNAase A solution: 20 pg/mL in 0.5MNaC1,lO mM Tris-HCl pH 8.0,
1 mM EDTA pH 8.0 (Stock RNase A: 10 mg/mL, store at -2OOC).
11. Dimethyl-dichlorosilane-coated coverslips: Dip coverslips in 5% di-
methyl-dichlorosilane in chloroform. Rinse several times with ddT0.
Dry.
12. 70,90, and 100% ethanol containing 0.3M ammonium acetate.
13. Autoradiography emulsion: Kodak NTB-2 or Ilford K-5 diluted 1:l
with double-distilled HZO.
14. Kodak D19 developer.

3. Methods
3.1. Synthesis of High Activity Single-Stranded cRNA
Probes (lkanscription)
The following protocol is a modification of that given by Promega Bio-
tee for synthesis of RNA probes (seeNotes 1-4; Figs. 1 and 2).
1. To a sterile microfuge tube, at room temperature, add in the following
order: 4.0 PL 5 x Transcription Buffer, 2.0 PL 100 mM dithiothreitol
(DTT), 0.8 PL RNasin, 4.0 PL Nucleotide Mixture, 2.4 PL 100 pm cyti-
dine triphosphate (CTP), 1.0 PL linearized plasmid template DNA (1
pg), 5.0 PL w~~I?-CTP (50 PCi), 0.5-0.8 PL SP6 RNA polymerase, T7
RNA Polymerase or T3 RNA Polymerase, and DEPC-treated water to
20 PL final vol.
In Situ Hybridization

Hind =,Pst I (568)*

Hint 11(259)

Y Eco RI

-rr
Pvu II
\ t-

*NUMBERING REFERS TO RAT ANP

cDNA SEQUENCE
Fig. 1. Schematic diagram of the SP6 plasmid (ANP-cDNA) used for in vitro tran-
scription.

2. Incubate for l-1-1/2 h at 37-4O”C.

3. To terminate transcription, add 1 PL of RNase-free DNase and 1 FL of
RNasin. Incubate for 10 min at 37OC.
4. Add: 1 PL of tRNA, 175 PL of DEPC-treated water, and 5 JJL of 4M
NaCl. Extract with an equal vol(200 pL> of phenol/chloroform (1:l v/
v). Mixby vortexing. Separate the phases by centrifugationin amicro-
fuge (5 min).
5. Remove the upper aqueous phase (200 I.~.L>and extract this with an
equal vol(200 pL) of chloroform. Mix and spin as above.
6. To the upper aqueous layer add 100 PL of 7M ammonium acetate
(2.5M final concentration), 750 l.tL of absolute ethanol (2.5 vol, -2OOC).
Mix and leave at -2OOC overnight.
7. Spin in a microfuge for 30 min. Discard the supernatant. Dry the RNA
pellet under vacuum. When dry, dissolve the pellet in 20 PL of DEPC-
506 Adams, Humid, and Polak

TCGATCG C-DNA
Vector
Transcription SP6 T 7 T a polymerase

Biotin

Fig. 2. Diagrammatic representation of cRNA synthesis.

treated water. Remove 1 PL for assessment of incorporation of radio-

activity. Store SE?probes at -7O”C, 32Pprobes at -2OOC. The maximum
storage time will depend on the radioisotope used. However, back-
ground increases with storage time.
8. Incorporation of radioactivity is estimated by determination of tri-
chloroacetic acid (TCA)-precipitable counts. Mix 1 PL labeled RNA
probe, 50 PL Bovine Serum Albumin, and 100 PL 10% trichloroacetic
acid. Vacuum filter on GF/C (Glass microfiber paper, Whatman).
Wash the filter twice with 10% TCA and twice with absolute ethanol.
Dry the filters. Count the radioactivity on the filters and from this de-
termine the percent of incorporation of radioactivity.

3.2. Fisation of Material (See Note 5)

3.2.1. Tissue
We have evaluated the use of various fixatives, including 10% for-
malin, Bouin’s solution, 2.5% glutaradehyde/paraformaldehyde mixtures,
In Situ Hybridization 507

acetic acid/alcohol, and 4% paraformaldehyde. Formalin and Bouin’s

gave the best morphology but poor RNA retention, whereas glutaralde-
hyde fixation retained more RNA, but the resultant morphology was poor.
Thus, we compromised with 4% paraformaldehyde, which gave the best
results when considering both RNA retention and cellular morphology.
1. Tissues for hybridization must be collected as fresh as possible. Fix
small pieces of tissue (1 x 1 x 0.5 cm) by immersion in freshly prepared
4% paraformaldehyde for 4 h at 4OC(prolonged fixation reduces the
hybridization signal). To avoid RNase contamination, wear gloves,
use sterile equipment, and use DEPC treated water.
2. Transfer the tissue to phosphate-buffered saline containing 15% su-
crose and store at 4°C (maximum 2 mo).
3. After washing the tissue, prepare cryostat blocks, cut sections (15 pm),
thaw-mount on poly-L-lysine (PII,)-coated glass slides, and allow to
dry at 37OCovernight before processing for hybridization. Use these
slides as soon as possible; otherwise store in a container with desiccant
at -7OOC.
3.2.2. Culture
1. Cultured cells can be grown on coverslips, slides, or in suspension.
Rinse coverslips or slides, with attached cultures, in cold phosphate-
buffered saline, and fix by immersion in 4% paraformaldehyde for 60
min at 4°C.
2. Rinse several times in phosphate-buffered saline and finally in double-
distilled H20.
3. Dry the cultures for 6 h at 37°C and store in a dessicated box at -70°C.
4. Cell lines, grown in suspension, are cytospun onto poly-L-lysine-
coated slides, which are then air-dried for 5-10 min before being fixed
and stored as above.

3.3. In Situ Hybridization with cRNA Probes

Precautions should be taken to avoid RNase contamination until hy-
bridization is complete. Coverslips with cultured cells grown on them
should be attached to slides with paper clips (12).
3.3.1. Tissue Preparation
1. Rehydrate in phosphate-buffered saline (PBS) (pH 7.2) for 5 min.
2. Immerse in O.lM glycine/PBS (5 min).
508 Adams, Hamid, and Polak

3. Permeabilize by immersion in 0.3% Triton X 100 in PBS for 15 min.

4. Wash with PBS (2x 3 min).
5. Deproteinize by incubation with 1 pg/mL Proteinase Ksolution for 20
min at 37OC.
6. Stop deproteinization by immersion in 4% paraformaldehyde/PBS
(5 min).
7. Immerse in freshly prepared acetylation solution for 10 min to reduce
nonspecific binding.
8. Prehybridize in 50% formamide, 2x SSC (37OC at least 15 min) to
enhance signal to noise ratio.
3.3.2. Hybridization
1. Drain the slides briefly (do not dry).
2. Apply 20 PL of hybridization mixture preheated to 37OC, containing
2-3 ng radiolabeled cRNA probe (5 x lo5 cpm/section) diluted in hy-
bridization buffer.
3. Cover the sections with suitably sized dimethyl dichlorosilane-coated
coverslips.
4. Incubate at 37-43OC for 16 h.
3.3.3. Posthybridization Washing
1. Remove the coverslips by immersion in 4 x standard sodium citrate
(SSC).
2. Wash the slides with 4 x SSC (37”C, 3 x 20 min) with gentle shaking.
3. Remove unhybridized single-stranded cRNA probe by treating prep-
arations with RNase A solution for 30 min at 37°C.
4. Wash the slides with 2 x SSC (37OC, 30 min), with gentle shaking.
5. Wash the slides with 0.1 x SSC (37OC,30 min), with gentle shaking.
6. Dehydrate in 70,90, and 2 x 100% ethanol containing 0.3M ammonium
acetate (10 min each at room temperature).
7. Air dry (60-90 min).
3.3.4. Autoradiography
1, Dip the slides in emulsion.
2. Air dry for l-2 h.
3q Store in light box for 2-5 d, depending on the radioisotopes used (2-3
d for 321?,5 d for =S>.
4. Develop in Kodak D19 developer prepared and used according to
manufacturer’s instructions. Fix as appropriate.
5. Wash well with water.
In Situ Hybridization 509

3.3.5. Counterstaining
1. Preparations are usually lightly counterstained with hematoxylin or
hematoxylin/eosin. However, eosin may cover fine grains. Other
counterstains may be used where appropriate (e.g., Coomassie blue
for myocytes, Toluidine blue for brain preparations).
2. Dehydrate, clear, and mount with DPX. Examples of in situ hybridi-
zation results are shown in Figs. 3-6.

4. Notes
1. Control experiments are very important to assess the specificity of the
hybridization and should include the following:
a. Sense probes: Probes identical to the coding strand of the
mRNA under investigation are transcribed and hybridized as
above.
b. Ribonuclease treatment: Sections or cultures are treated with
RNase A (20 pg/mL, 37OC, 30 min) before the prehybridiza-
tion step. A remnant of the ribonuclease could result in probe
degradation and thus invalidates the results.
c. Inappropriate probe for the tissue in question.
d. Inappropriate tissue for the probe in question.
e. Northern Blot Analysis: The presence of the particular mRNA
in the tissue may be confirmed by Northern Blot hybridiza-
tion.
f. Several probes, coding for different regions of the same gene.
g. Immunocytochemistry: The correlation of immunocytochem-
istry results with thoseobtained by in situ hybridization is one
of the most useful indications of the specificity of the signal.
2, Signal/ Background Ratio: Although probes labeled with % give
better subcellular resolution than those labeled with 3?l?,there is an in-
crease in background. The background may be reduced by the follow-
ing manipulations:
a. Decreased autoradiography time.
b. Minimize the amount of probe used for hybridization.
c. The inclusion of dithiothreitol(50 mM) in one or all the follow-
ing: prehybridization solution, hybridization mixture or post-
hybridization washings. This is particularly true with the use
of ?S-labeled probe.
d. The addition of “cold” cytidine triphosphate to the prehy-
bridization buffer.
510 Adams, Hamid, and Polak

Fig. 3. In situ hybridization of ANP-mRNA in cultured myocytes of rat atrium and

ventricle using ?P-labeled ANP-cRNA probes.

Fig. 4. Autoradiographic preparations of cultures, fixed in4% paraformaldehyde, and

counterstained with hematoxylin x 2!?10.
In Situ Hybridization 511

Fig. 5. In situ hybridization of CGRP-mRNA in rat colon (fixed by perfusion with 4%

paraformaldehyde) using a T-labeled cFNA probe. Positive hybridization signal in sub-
mucous plexus (SMP) x 300. MM = muscularis mucosa, IG = intestinal glands.

Fig. 6. The localization of bombesin-mRNA in culture of small cell carcinoma of the

lung, cytospun onto slides, using 9 bombesin-cRNA probe.
512 Adams, Hamid, and Polak

3. In situ hybridization:
a. Combined immunocytochemistry and in situ hybridization.
Immunocytochemistry may be done subsequent to in situ hy-
bridization (a), the difference being the ommission of dextran
sulfate from the hybridization buffer. After the final post-
hybridization wash, slides are processed as normal for immu-
nocytochemistry. On completion of these protocols, they are
dehydrated and dipped for autoradiography as described
above.
b. Quantitation of autoradiographic signal (12). Before attempt-
ing to quantify the autoradiographic preparations, one should
take into consideration many factors that could affect hybrid-
ization signal such as the thickness of tissue or emulsion, loss
of mRNA in tissue processing, and the efficiency of in situ
procedure. Densitometry or computer imaging analysis give
semi-quantitative estimates.
4. Transcription: A major problem when working with mRNA prepar-
ations is RNase contamination. Thus, gloves should be worn through-
out the transcription and hybridization protocols. Glassware should
be baked at 250°C for 4 h; batches of plasticware should be set aside
exclusively for RNA work and autoclaved where appropriate before
use. All solutions should be prepared with DEPC-treated water.
DEPC (O.l%, final concentration) is added to distilled water and left at
room temperature for 12 h. Residual DEPC is destroyed by autoclav-
ing this water for 15 min. Solutions prepared for transcription with
this water are then aliquoted into sterile tubes and stored at -20°C.
Other labeled rNTPs can be used (10 mCi/mL, 400 Ci/mmol). This
reaction can be run in the absence of unlabeled cytidine triphosphate.
For a 20 PL reaction, 100 PCi of 400 Ci/mmol CX-~~P-CTPis 12 pm.
However, the yield of full-length transcripts drops as the concentra-
tion of limiting nucleotide cytidine triphosphate falls below 12 l.t.m.
The size of the probes may be reduced by alkaline hydrolysis (7).
5. Fixation: In animal experiments, perfusion with 4% paraformalde-
hyde gives the best results: Anesthetize rats with ether and perfuse in-
tracardially with PBS followed by 4% paraformaldehyde. Remove the
appropriate tissue and continue fixation in 4% paraformaldehyde for
1 h as above. Human tissue could also be perfused, for example, per-
fusion of the bowel through mesenteric vessels and the brain through
the Circle of Willis.
In Situ Hybridization 513

Acknowledgments
The authors are grateful to J. Dixon, Purdue University, West La-
fayette, USA; E. Spindel, Harvard Medical School, USA; and S. Amara,
Yale University, New Haven, USA, for supplying ANP, bombesin, and
CGRl?, cDNA, respectively. This work was supported by Amsersham
International and the Upjohn Company.

References
1. Gall, J. and Pardue, M. (1969) Formation and detection of RNA-DNA hybrid mole-
culesin cytological preparations. Proc.Nutl. Acad. Sci. USA 63,378-383.
2. John, H. A., Birnstiel, M. L., and Jones, K. W. (1969) RNA-DNA hybrids at the cyto-
logical level. Nature 223,582-587.
3. Buongiorno-Nardelli, S. and Amaldi, F. (1970) Autoradiographic detection of mo-
lecular hybrids between RNA and DNA in tissue sections. Nature 225,946-948.
4. Coghlan, J. I’., Aldred, I’., Haralambidis, J.,Niall, H. D., Penschow, J. D., and Tregear,
G. W. (1985) Hybridization histochemistry. Anulyf. Biochem. 149,1-28.
5. Penschow, J. D.,Haralambidis, J., Darling, P. E.,Darby, I. A., Wintour,E. M.,Tregear,
G. W., and Coghlan, J. P. (1987) Hybridization histochemistry. Experienfiu 43,
741-750.
6. Vamdell, I. M., Polak, J. M., Sikri, K. L., Minth, C. D., Bloom, S. R., and Dixon, J. E.
(1984)VisualisationofmessengerRNAdirectingpeptidesynthesisbyinsifuhybrid-
ization using a novel single-stranded cDNA probe. Hisfochemisfry 81,597-601.
7. Cox, K. H., De Leon, D. V., Angerer, L. M., and Angerer, R. C. (1984) Detection of
mRNAs in sea urchin embryos by in situ hybridization using asymmetric RNA
probes. Deal. Biol. 101,485-502.
8. Hoefler, H., Childers, H., Montminy, M. R., Lechan, R. M., Goodman, R. H., and
Wolfe, H. J. (1986) In situ hybridization methods for the detection of somatostatin
mRNA in tissue sections using antisera RNA probes. Hisfocbrn. J. l&597-604.
9, Terenghi, G., Polak, J. M., Hamid, Q. A., O’Brien, E., Denny, I’., Legon, S., Dixon, J.,
Minth, C., Palay, S. L., Yasargil, G., and Chan-Palay, V. (1987) Localisation of
neuropeptide Y-mRNA in neurones of the human brain cortex using in situ hy-
bridization and cRNA probes. Pm. Nufl. Acud. Sci. USA 84,7315-7318.
10. Hamid, Q. A., Wharton, J., Terenghi, G., Hassalle, C. J., Aimi, H., Taylor, K. M.,
Nakazato, H., Dixon, J. E., Bumstock, G., and Polak, J. M. (1987) Localization of atria1
natriuretic peptide mRNA and immunoreactivity in rat heart and human atria1
appendage. Proc. Nufl. Acud. Sci. USA 84,6760-6764.
11. Hamid, Q. A., Bishop, A. E., Springall, D. R., Adams, C.,Giaid, A., Denny,P., Ghatei,
M., Legon, S., Cuttitta, F., Rode, J., Spindel, E., Bloom, S. R., and Polak, J. M. (1989)
Detection of human probombesin mRNA neuroendocrine (in small) cell carcinoma
of the lung: in situ hybridization with cRNA probe. Cancer 63,266271.
12. McCabe, J, T., Morrell, J. I., and Pfaff, D. W. (1986) In situ hybridization as a quan-
titative autoradiographic method: Vasopressin and oxytocin gene transcription in
the Brattleboro rat, in In situ Hybridization in Bruin (Uhl, G. R., ed.), Plenum, New
York, Ch. 5, pp. 73-97.
 Chapter 38

Immunogold Labeling for

Electron Microscopy

John Pacy
1. Introduction
Colloidal gold immunocytochemistry was first introduced by Faulk
and Taylor (I), and has rapidly become a major technique in electron
microscopy, covering many aspects of biological research. The rapid
expansion of this technique in electron microscopy is first the result of its
simplicity compared with other labeling techniques, and secondly, the
result of the properties of the gold probes themselves. These probes are
usually 3-15 nm in diameter and coated with immunologically active
proteins. They have the advantages of being very electron dense, giving a
characteristic appearance that cannot be confused with other biological
structures; they are highly sensitive, producing very specific labeling of
both monoclonal and polyclonal antibodies; they are also permanent,
nonhazardous, and can be easily quantified.
Several techniques now exist for obtaining different types of infor-
mation from the sample. The successof these techniques depends on us-
ing very specific antibodies that have a high affinity or binding capacity to
their antigen. Any unwanted antibodies present in the preparation should
be of a lower affinity than the specific antibody. The choice of antibody,
whether monoclonal or polyclonal, is still mainly determined by availabil-
ity. Fixation reduces antigenicity, and polyclonal antibodies are less
515
516 Pacy

affected than monoclonals by the fixation procedures. However, greater

specificity can be obtained with monclonal antibodies.
The choice of using secondary antibody/gold or protein A/gold is
still a complex one, especially as more secondary antibody/gold probes
are becoming commercially available. In general, secondary antibody/
gold produces less background than protein A/gold when used in higher
dilutions. Some signal amplification is found using the secondary anti-
body/gold probes because of the binding of more than one probe to one
primary antibody. This results in clusters of gold particles that can be help-
ful when the marker density is low.
The following techniques are now well established and will be dealt
with in this chapter:
1. The preembedding technique for the localization of external antigens
on whole cells.
2. The postembedding technique for the localization of both internal and
external antigens.
3. Cryo-techniques for sensitive antigens that would be destroyed by
conventional processing.
4. The immunonegative staining technique for localization of surface
antigens on small specimens, such as viruses and bacteria, which lend
themselves to negative staining.
5. The double-labeling technique for localizing more than one species of
antigen by using different sized probes. Thesetechniques listed above
are the most common, but other techniques exist such as those in-
volving freeze-fracture and immunoreplica.
There is now a wide range of commercially available gold probes,
and consequently, methods for producing these probes will be dealt with
only briefly after the immunolabeling techniques.

2. Materials
1. Fixatives:
a. 1% glutaraldehyde in O.lM sodium cacodylate buffer at pH
7.2 for polyclonal antibodies.
b. 4% paraformaldehyde plus 0.05% glutaraldehyde in O.lM so-
dium cacodylate buffer at pH 7.2 for monoclonal antibodies.
c. 2.5% glutaraldehyde in O.lM sodium cacodylate buffer at pH
7.2. (For use in the preembedding and cryo-techniques.)
2. 2% osmium tetroxide in distilled water.
3. Ethanol solutions: 30,50,70,90,100% by volume.
Immunogold Labeling for Microscopy 517

4. Embedding media:
a. L. R. White embedding resin.
b. Spurr embedding resin: NSA (nonyl succinic anhydride) 26
mL, ERL 4206 (vinyl cyclohexane dioxide) 10 mL, DER
(Diglycidyl ether of polypropylene glycol) 6 mL, Sl (dimeth-
ylaminoethanol) 0.4 mL.
c. 10% gelatin in O.lM sodium cacodylate buffer pH 7.2 (dis-
solve the gelatin powder by warming and store in micro-
centrifuge tubes at 4OC).
5. Stains:
a. 2% aqueous uranyl acetate.
b. 2% neutral uranyl acetate (2). Prepare by mixing 4% aqueous
UA (O.lM) and 0.3M oxalic acid in equal proportion. Add a
small amount of 10% ammonium hydroxide to adjust the pH
to 7.2-7.4.
c. 2% aqueous ammonium molybdate.
d. Reynolds lead citrate (3). To prepare: Add 1.33 g lead nitrate
and 1.76 g sodium citrate to 30 mL boiled distilled water in a
50 mL volumetric flask. Shake the suspension vigorously for
1 min, and then allow to stand for 30 min with intermittant
shaking. Add 8 mL of 1 NaOH (freshly prepared) and dilute
to 50 mL with boiled distilled water. Mix by inversion.
6. Phosphate immunobuffer: The phosphate immunobuffer used for
incubations is PBS plus 1% BSA. Phosphate buffered saline (PBS):
0.14M NaCl, 3 mM KCl, 10 mM Na,HPO,, 3 mM KH2POI, 0.1% NaN,
(sodium azide) can be added as a preservative.
7. Tris immunobuffer:
Stock buffer: 6 g Tris (tris-hydroxymethyl-aminomethane), 9g
NaCl, 1 g gelatin (dissolved in 10 mL of distilled water by warming),
990 mL distilled water, pH to 7.4 with cont. HCl. 1 g of NaN, can be
added as a preservative.
Final buffer: Add 1% (w/v) ovalbumin and 0.01% (v/v) Tween 20
to the stock buffer and pH to 8.2. (1% Bovine serum albumin (BSA) can
be used instead of ovalbumin.) TheTris immunobuffer used for incu-
bations is therefore: 0.05M Tris, 0.01% Tween 20,0.1% gelatin, 1.0%
ovalbumin, made up in 0.9% NaCl solution.
8. 1% (w/v) gelatin in PBS.
9. 0.02M glycine in PBS.
10. 2.3M sucrose.
11. 0.5M NH,Cl.
518 Pacy

12. Primary antibody prepared in serial dilutions of immunobuffer. Start

with 1 in 10.
13. Gold conjugate prepared in serial dilutions with immunobuffer. Start
with 1 in 4 (e.g., for rabbit primary antibody use Goat Anti Rabbit or
Protein A; for mouse primary antibody, use Goat Anti Mouse or
Protein A). The gold conjugate can be obtained commercially or made
in the laboratory.
14. 2% aqueous uranyl acetate plus 1.5% methyl cellulose (1:l mixture).
Store at 4OC.
15. 0.01% chloroauric acid (gold chloride) in distilled water.
16. 1% (w/v) trisodium citrate in distilled water.
17. 0.2M potassium carbonate.
18. Ether solution of white phosphorus. Small pieces of white phospho-
rus sticks are cut under water and quickly transferred to 20 mL diethyl
ether. After stirring for 2 h, the supernatant is centrifuged for 20 min
at 4800 g. One part of the supernatant is added to 4 parts &ethyl ether.
19. 1% (w/v) polyethylene glycol.

3. Methods
The labeling methods dealt with in this chapter are shown in the form
of a schematic diagram (seeFig. 1).
3.1. The Preembedding Technique
This method is for the localization of external antigens with the
advantage that conventional electron microscopy (E. M.) can be used to
give good preservation and contrast to the sample.
1. Prepare antigen for immunolabeling.
2. Wash sample in PBSpH 7.2 plus 1% Bovine Serum Albumin (BSA) for
5 min.
3. Incubate with a primary antibody of suitable dilution in PBSplus 1%
BSA for 1 h. (SeeChapter 35 for details of dilutions.)
4. Wash in PBS plus 1% BSA, five changes of 1 min each.
5. Incubate with a gold probe of suitable dilution with PBSplus 1% BSA
for 1 h. (Start with a 10 nM sized probe.)
6. Rinse in PBS, three changes of 1 min each.
7. Fix the sample in 2.5% glutaraldehyde in PBS for 5 min.
8. Wash well in distilled water, five changes of 1 min each.
9. Transfer to 2.5% glutaraldehyde in O.lM sodium cacodylate buffer pH
7.2 for 20 min.
Immunogold Labeling for Microscopy 519

Fzig.1. Schematic diagram of gold labeling methods.

10. Wash in O.lM cacodylate buffer, three changes of 5 min each.
11. Postfix in 1% osmium. tetroxide in O.lM sodium cacodylate buffer
for 1 h.
12. Wash in 30% ethanol for 5 min.
13. Dehydrate through a series of ethanols, 50-70-90-lOO-lOO-100% for
15 min each.
14. Clear in propylene oxide, two changes of 5 min each.
15. Infiltrate with Spurr resin by increasing the resin to propylene diox-
ide ratio slowly over a few hours.
16. Infiltrate with 100% Spurr resin for 2 h. (This is an arbitrary time
dependent on the sample.)
520 Pacy

17. Embed in Beem capsules or equivalent. The sample can be spun down
through the resin to form a pellet once inside the capsule. Polymerize
in a 60°C oven for 2 d.
18. Section and stain as for conventional E. M. (See Chapter 34, this vol.
for details.)
Apart from the polymerization of resin, all the other procedures are
carried out at room temperature.

3.2. The Postembedding Technique

This method is used for the localization of both external and internal
antigens that can survive the fixation and embedding procedures. Os-
mium tetroxide is usually omitted resulting in the problem of obtaining
good contrast in the specimen. (SeeFigures 2a and 2b.)
3.2.1. Fixation and Embedding
1. Prepare antigen for immunolabeling.
2. Fixation: For monoclonal antibodies, fix the sample in 4% parafor-
maldehyde plus 0.05% glutaraldehyde in O.lM sodium cacodylate
buffer at pH 7.2 for as long as required. For polyclonal antibodies, fix
the sample in 1% glutaraldehyde in O+lM sodium cacodylate buffer at
pH 7.2 for half an hour.
3. Wash in cacodylate buffer, three changes of 5 min each.
4. Postfix in 1% osmium tetroxide for 1 h only if labeling is not affected.
This stage is usually omitted.
5. Dehydrate in an ethanol series, 10 min each step: 50-70-90-100-
100-100%.
6. Transfer to L. R. White resin (4). Four changes of half an hour each.
NOTE: pour out enough resin from stock to cover the infiltration and
embedding steps. Be prepared to throw away any excess that is left.
7. Embed in airtight molds, preferably oven-dried gelatin capsules.
Leave in a 60°C oven for 22 h.
8. Section onto gold or nickel grids. Copper grids react with Tris buffer.
Use carbon/formvar coated grids if the sections are delicate.
3.2.2. Antibody Incubation
During this procedure the grids are floated, sections downwards on
droplets. Allow two drops/grid at each stage. The time in the first drop
need only be brief enough to wash off the previous solution.
With Tris Immunobuffer:
Immunogold Labeling for Microscopy 521

Fig. 2. (a) Localization of myosin antigens on B granules of Islet of Langerhans using

the postembedding technique (70,OtXIx). fb) Localization of 4 protein antigens in Pseudo-
monasputida using the postembedding technique (30,000~). (c) Localizationof G specific
antigens on the surface of G organisms of B-hemolytic streptococci using the cryo-labeling
technique (50,000x) (Courtesy of Dr. J. E. Beesley, Wellcome Research Laboratories,
Beckenham, Kent.). (d) Simultaneous localization of Ca and HMFG 2 antigens in breast
tissue using the double-labeling technique. Detection of Ca is with a 20 nM probe and
HMFG 2 is with a 5 nM probe (30,000x) (Courtesy of Dr. J. E. Bee&y, Wellcome Research
Labs.) Published in Proc. Roy. Mic. Sac. 20 (41,187-196 (permission granted).

1. Float grids on 0.5M NH,Cl for 10 min to neutralize free aldehyde

groups.
2. Wash sections with Tris immunobuffer (No. 7 in materials section),
two changes of 2 min each.
3. Incubate sections on drops of primary antibody diluted with im-
munobuffer, for 1 h at room temperature.
522 Pacy

4. Wash with immunobuffer, five changes of 1 min each. This is to

remove unbound primary antibody.
5. Transfer grids to drops of gold conjugate (either secondary antibody
gold, e.g., GAR. or protein A) diluted with immunobuffer, for 1 h at
room temperature.
6. Wash on drops of distilled water, five changes of I min each. This is
to remove unbound gold conjugate.
7. Obtain contrast with conventional E. M. staining procedures.
With PBS Immunobuffer:
1. Float grids on PBS plus 0.02M glycine for 5 min to neutralize any free
aldehyde sites.
2. Transfer to PBS plus 1% gelatin for 10 min to neutralize “sticky” sites
on the sample, and thus reducing nonspecific binding.
3. Wash in PBS plus 1% BSA, two changes of 3 min each. BSA attaches
to nonspecific sites, thus competing with the primary antibody and
reducing nonspecific background labeling.
4. Incubate grids in a suitable dilution of antibody in PBS plus 1% BSA
for 1 h.
5. Wash in PBS plus 1% BSA, five changes of 1 min each.
6. Incubate in a suitable dilution of gold conjugate in PBS plus 1% BSA
for 1 h.
7. Wash in distilled water, five changes of 1 min each.
8. Obtain contrast as for conventional E. M. sections.

3.3. The Cryo-Technique

This method is for antigens that would not survive conventional pro-
cessing and presents us with the inherent difficulties of cryo-sectioning, It
does, however, give the best preservation of antigens (seeFig. 2~).
1. Prepare antigen for immunolabeling.
2. Fix sample, depending on the antibody. See previous method.
3. Wash in cacodylate buffer, three changes of 5 min each.
4. Embed in 10% gelatin in cacodylate buffer. Heat the gelatin (stored at
4°C) to 37OC before use. Warm pipets and tubes before use. Re-
suspend the sample in the warm gelatin and spin down for 2 min into
a pellet. Place in ice to set the gelatin.
5. Cut out the pellet and refix small pieces of it in the original fixative for
10 min to cross-link the gelatin.
Inamunogold Labeling for Microscopy

6. Cryo-protect in 2.3M sucrose for 1 h.

7. Freeze in nitrogen slush. Prepare slush by placing a wide-mouthed
polystyrene container with about l-in depth of liquid nitrogen into a
vacuum chamber. Evacuate for a few minutes until the liquid N, turns
to slush. Return the chamber quickly to atmospheric pressure, and use
the slush immediately.
8. Section at about -80°C, with the knife slightly colder at about -9OOC.
9. Collect sections on a droplet of saturated sucrose solution, and mount
onto a carbon/formvar coated grid (3).
10. Transfer the grid to a droplet of PBS plus 1% BSA to remove sucrose
(3).
NOTE: It is advisable to check that there is a sample in the sections.
Take preliminary sections, mount on a grid, wash with three droplets of
distilled water and negative stain with 2% ammonium molybdate.
3.3.1. Incubation
l-6. Follow steps 14 of the postembedding technique with phosphate
immunobuffer.
7. Rinse in PBS, two changes of 2 min each.
8. Fix in 2.5% glutaraldehyde in PBS for 2-10 min.
9. Rinse in distilled water, five changes of 1 min each to remove phos-
phate.
10. Stain in uranyl acetate oxalate for 5 min (3).
11. Rinse in distilled water, three changes of 20 s each.
12. Stabilize and enhance the contrast of the sections by floating them on
a 1:1 mixture of 2% aqueous uranyl acetate and 1.5% methyl cellulose
for 2-3 min. This has to be done on ice to keep the methyl cellulose
from setting.
3.4. The Immunonegative Staining Technique
This method is for the localization of surface antigens on small
specimens such as bacteria and viruses. Fixation and embedding pro-
cedures are not involved and therefore the maximum labeling potential of
the untreated antigens is retained. This means that low levels of antigen
can be labeled. Negative staining is a fast and simple technique to per-
form, giving high contrast and high resolution.
1. Prepare antigen for immunolabeling.
2. Dry the sample down on to a carbon/formvar coated 400 mesh grid.
524 Pacy

3. Rinse in PBSplus 1% BSA, 2x 1 min.

4. Incubate grids with primary antibody diluted with PBS plus 1% BSA
for 15 min.
5. Wash with PBS plus 1% BSA, 4x 1 min.
6. Incubate grids with gold probe diluted with PBS plus 1% BSA for
15 min.
7. Wash with PBS, 4x 1 min.
8. Negative stain with 2% ammonium molybdate.

3.5. The Double-Labeling Technique

This method is for localizing more than one antigen by using different
sized gold probes (seeFig. 2d). Double-labeling can be carried out using
any of the previous techniques. The incubations can be simultaneous,
providing there is no cross-reaction of reagents, or sequential, with a
blocking step if there is cross-reaction. In general, the smaller probe is
incubated first. Sections on uncoated grids are incubated on one side and
then carbon coated. The second incubation can then be carried out on the
reverse side without risk of contamination.

3.6. Preparation of Gold Probes

The production of gold probes is in two distinct steps. The first is the
production of colloidal gold. This involves the reduction of gold salts into
metallic gold spheres that can vary in size depending on the method used
for their preparation. The second step is the coating of the spheres with the
chosen protein/antibody. This procedure stabilizes the colloidal gold and
is more complicated than the reduction process. The preparation of
protein A-gold complex will be used as an example.
3.6.1. Preparation of Colloidal Gold
All glassware must be thoroughly cleaned, and double-distilled water
should be used in all solutions.
For 5 nM spheres (5):
1. Add 1.5 mL of 1% HAuCl, to 120 mL distilled water. To this add 2.7
mL 0.2M potassium carbonate.
2. Add 1 mL of ether-saturated white phosphorus slowly and shake for
15 min at room temperature. The mixture becomes brownish-red.
3. Gently heat over an open flame until the color changes to a wine-red
color.
Immunogold Labeling for Microscopy 525

For 10 nM spheres (6):

1. Add 15 mL of 1% trisodium citrate to 247.5 mL boiling water. Con-
tinue boiling for 5 min.
2. Quickly add 2.5 mL of 1% HAuCl, (chloroauric acid) while stirring.
The solution changes from blue to red; continue boiling for a further
10 min to complete the reduction.
3. Allow to cool and pH to 6.9 with 0.2M potassium carbonate.
For 15,30, and 40 nM spheres (6):
1. Heat 100 mL 0.01% HAuCl, to boiling.
2. Add 3.0 mL, 1.5 mL, and 1.2 mL of 1% sodium citrate, respectively, for
15,30, and 40 nM spheres.
3. Continue boiling for 10 min, cool, and pH to 6.9 with 0.2M potassium
carbonate.
3.62. Preparation of Protein A-Gold Complex (7)
1. Dissolve 0.5 mg purified protein A in 0.1 mL distilled water and mix
with 10 mL colloidal gold.
2. After 2-3 min, 1 mL of 1% aqueous polyethylene glycol is added, and
centrifuged for 1 h at 20,OOOg.
3. The dark red, loose pellet is resuspended in 6 mL of 0.2 mg/mL of
polyethylene glycol in phosphate buffered saline (pH 7.4) and stored
at 4°C.

4. Notes
1. It is imperative to include appropriate controls in all immunolabeling
techniques. Negative controls are straightforward to carry out. They
could include replacing the primary antibody with an alternative
antibody from the same species raised against an antigen known to be
absent in the sample, replacing the primary antibody with non-
immune serum, or replacing the primary antibody with buffer only.
Positive controls, where possible, should also be included.
2. Longer incubation with higher dilutions of primary antibody reduces
nonspecific binding and, thus, produces more specific binding. Higher
dilutions of gold conjugate can reduce nonspecific binding to the
highly charged components of the sample (e.g., necrotic cells, nucleic
acids, and so on).
3. Clustering of gold particles can occur. This may result from the
natural amplification of the gold conjugate when using IgG gold
526 Pacy

probes, and does not occur with protein A gold. Clustering may also
result from clumping of the primary antibody because of age or
storage conditions.
4. Trouble Shooting
a. No label
(1) The antigen is absent or destroyed by the preparative
procedures. Change procedure.
(2) The ant’b1 o d y is not working. Could be the result of age,
poor storage, wrong dilution, wrong antibody, excessive
freezing and thawing. Change the antibody and run a
positive control.
(3) The antigen is present, but in very low amounts. Extend
the incubation times and use more concentrated primary
antibody.
(4) The gold probe is not working. Could be the result of its
poor condition or even a wrong probe. Repeat the incu-
bation to check.
(5) The pH of the solutions is wrong. Check this.
b. Excessive background labeling
(1) Concentration of primary antibody and/or gold conju-
gate is too high. Dilute at least 10 times.
(2) Insufficient washing between incubations. Increase the
washing times substantially.
(3) Nonspecific binding is occurring. Add ovalbumin or
BSA or increase their concentrations if already present.
(4) Nonspecific charge attraction of the antibody is occur-
ing. Use a detergent in all solutions (e.g., 1% Tween 20).
Osmium tetroxide fixation can introduce excess charge
into the sample. Omit the osmium.
(5) Binding to free aldehyde groups in fixed tissue might
have occurred. Neutralize these with glycine or NH&l
before incubation.
(6) Bad fixation can sometimes produce background label-
ing in dead or damaged cells. Improve fixation.
5. Low temperature embedding can be achieved using Lowicryl K4M
resin (8). The fixation and embedding procedures are carried out at
4°C with the polymerization of resin completed with UV radiation at
the same temperature. Low temperatures enhance preservation of
antigens (9), and proteins are less likely toleachoutof the tissue. There
Immunogold Labeling for Microscopy 527

are some doubts, however, that because of a rise in temperature

during polymerization (10) true low temperature embedding is
achieved.
6. A number of useful general references (1Z-1 8) are included at the end
of this chapter.

Acknowledgments
I wish to thank Trish Dopping-Hepenstal for her assistance in prepar-
ing this chapter. I would also like to thank Dr. Julian Beesley for his micro-
graphs and for his help in the past.

References
1. Faulk, W. P. and Taylor, G. M. (1971) An immunocolloid method for the electron
microscope. immunochemistry 8,1081-1083.
2. Tokuyasu, K. T. (1978)A study of positive staining of ultrathin frozen sections. J. of
Ultrastructure Res. 63,287-307.
3. Reynolds, E. S. (1963) The use of lead citrate at high pH as an electron opaque stain
in electron microscopy. J. Cell. Biol. 17,208.
4. Timms, B. G. (1986) Postembedding immunogold labeling for electron microscopy
using “L. R. White” resin. Amer. J, of Anat. 175,267-275.
5. Zsigmondy, R. (1905) Zur Erkenntnisse der Kolloide. Z. Physik. Chem. 56,65.
6. Frens, G. (1973) Controlled nucleation for the regulation of particle size in mono-
disperse gold suspensions. Nature Phys. Sci. 241,20-22.
7. Roth, J. and Binder, M. (1978) Colloidal gold, ferritin and peroxidase as markers for
electron microscopic double labeling lectin techniques. I, Histochem. Cyfochem. 26,
163-169.
8. Valentino, K. L., Crumrine, D. A., and Reichardt, L. F. (1985) Lowicryl K4M embed-
ding of brain tissue for immunogold electron microscopy. J.Histochem. Cyfochem. 33,
969-973.
9. Armbruster, B. L., Garavito, R. M., and Kellenberger, E. (1983) Dehydration and
embedding temperatures affect the antigenic specificity of tubulin and immunola-
belling by the protein A-colloidal gold technique. 1, Hisfochem. Cytochem. 31,
1380-1384.
10. Ashford, A. E., Allaway, W. G., Gubler, F., Lennon, A., and Sleegers, J. (1986)
Temperature control in Lowicryl K4M and glycol methacrylate during polymer-
isation: is there a low temperature embedding method? 1.Microscopy 144,107-126.
11. Beesley,J. E. (1987) Colloidal gold electron immunocytochemistry: its potential in
medical microbiology. Seriodiug.and lrnmuno therapy1,239-252.
12. Roth, J. (1982) The protein A-gold technique. Tech. in Immunocytochem. 1,108133.
13. Polak, J. M. and Van Noorden, S.(1984) An introduction to immunocytochemistry;
current techniquesand problems, inM. S.Handbook22(Royal Microscopical Society,
Oxford, England).
528 Pacy

14. Bullock, G. and Petrusz, P. (1983,1984,1985) Techniques in immunocyfochmisty,

~01s.1,2,3 (Academic, London: New York).
25. Tokuyasu, K. T. (1986) Immunocryoultramicrotomy. J. Microsc. 143,139-149.
26. Polak, J.M. and Varndell, I. M, eds. (1984) Immunolabeling for electronmicroscopy.
Elsevier, Amsterdam, BV., pp. 129-142.
17. Beesley, J. E. (1986) The use of gold markers in immunocytochemical studies of
microbiological organisms: a review. 1. Microsc. 143,177-186.
18. Polak, J. and Van Noorden, S. (1986) hnmunocytochemisfry: Modern Methods and
Applications. J. Wright and Sons.
 Chapter 39

Human Chromosome
Analysis and Sorting

Judith A. Fantes
and Da@ K. Green

I. Introduction
Flow cytometry has provided the cytogeneticist with a fast and
accurate method of measuring the quantity of DNA in each human
chromosome (I). Almost all the chromosomes in the human complement
can now be resolved and abnormal chromosomes and aneuploidies (13,21,
and X) recognized. A flow karyotype shows apattern of peaks and troughs
that is unique for each individual or cell line because of the variation in
heterochromatic regions of the chromosomes (2). When combined with
family studies, flow cytometry has been able to resolve homologues dif-
fering in DNA content by as little as l/2000 of the human genome (3,4), less
than a metaphase band. In addition, the sorting capabilities of most flow
machines have provided a method for the purification of small but useful
quantities of individual chromosomes, for example, 2x lo6 average sized
human chromosomes are equivalent to 500 ng of DNA. Using recombin-
ant DNA techniques, this material can be used to generate a large number
of DNA probes to produce a chromosome-specific library, which can be

529
530 Fantes and Green

used for the molecular analysis of genetic disease (5,6). More recently,
molecular biologists have experimented with gene mapping by sorting
small quantities of individual chromosomes onto filters for spot-blot
hybridization with DNA probes (7).
The sample preparation and flow machine techniques relating to all
the above biological objectives will be discussed here. The art of producing
an enriched sample of a particular group of human chromosomes by flow
cytometry lies in bringing a clean, well-separated chromosome suspension
to a clean, sterile, and well-adjusted flow machine. Debris, unbalanced
stain/chromosome concentration or clumps of aggregated chromosomes
in the suspension or noise in the flow machine in the form of obstruction
to the flow, optical misalignment, or electronic noise all contribute to a
reduction in purity of the sorted sample. The methods described here are
aimed at producing the best possible sample obtainable from a starting
material of peripheral blood lymphocytes or lymphoblastoid cells to the
end point of verifying the identity and purity of the chosen sorted chromo-
some group.

2. Materials
1. Complete culture medium: RPM1 1640,10% fetal calf serum, 12.5 mM
MOPS, lOOU/mL penicillin, 100 pg/mL streptomycin,
2. Phyto-hemagglutinin (R-IA): reagent grade from Wellcome.
3. Lymphoprep from Nyegaard.
4. Polyamine buffer (Bl): 15 mM Tris, 0.2 mM spermine, 0.5 mM spermi-
dine, 2 mM EDTA, 0.5 mM EGTA, 80 mM KCI, 20 mM NaCl, 14 mM
(0.1% v/v) B-mercaptoethanol. Adjust to pH 7.2 with 1N HCl before
adding B-mercaptoethanol. Prepare fresh every week; store at 4OC.
5. Polyamine buffer plus digitonin (B2): 0.1% solution of digitonin in Bl.
Prepare a saturated solution by heating to 37OC and filtering through
a 0.2 pm filter to remove any undissolved digitonin. The best source
of digitonin is Fluka, since some batches of digitonin from other
sources are difficult to dissolve.
6. Hoechst 33258: 100 w in distilled water. Store in dark at 4OC.
7. Ethidium bromide: 1 mg/mL in distilled water. Store in dark at 4OC.
8. Chromomycin A3: 1 mg/mL in distilled water. Leave overnight in
cold to dissolve. Store in dark at 4OC.
9. Colcemid: 0.01 mg/mL in distilled water. Filter sterilize and store at
4OC.
Chromosomes, Analysis and Sorting 531

10. DAPI (4,6-Diamidino-2-phenylindole l 2HCl): 50 pg /mL in distilled

water. Store in dark at 4°C.
11. Spermidine Es-acridine (8): Dissolve 5 mg in 2 mL methanol. Make
up to 100 mL with 10 mM disodium orthophosphate, adjust to pH 6.5
with 0.2M HCl. Store in dark at 4*C.
12. Decon or 7X detergent diluted 1:25with sterile filtered distilled water.
13. Activated glutaraldehyde (Cidex).
14. Distilled water sterilized by filtration through 0.2 pm filter.
15. Phosphate buffered saline (Dulbecco A); filter sterilized.

3. Methods
The method of isolating chromosomes described here uses poly-
amines to stabilize the chromosomes and detergent treatment to lyse the
cells (9). Since the lysis of interphase nuclei is minimal, this method can be
used successfully with all types of cell lines, including suspension cultures
with a significant proportion of interphase cells. Although the isolated
chromosomes are highly condensed, they maintain most of their in vivo
structure, and it is possible to prepare high mol wt DNA from them.

3.1. Cell Culture

Human chromosomes are usually prepared from lymphoblastoid cell
lines (EBV transformed B lymphocytes, seeChapter 5, this volume) or from
PHA-stimulated peripheral blood lymphocytes. The best chromosome
suspensions and hence the flow karyotypes with the greatest resolution are
prepared from cultures with a reasonable proportion of mitotic cells
(>20%, seeNote 1).
1. Lymphoblastoid cells: set up cells from a stationary phase culture at
3 x 105viable cells/mL. Thirty hours later, add colcemid to a final con-
centration of 0.1 pg/mL to block the cells at metaphase. Sixteen to
eighteen hours later harvest cultures and disperse any cell clumps by
gentle pipeting.
2. Lymphocytes: Defibrinate a 40-mL sample of peripheral blood with
orange sticks. Spin down the cells at 4008 for 10 min, and remove se
rum (=20 mL). Replace with an equal volume of medium without FCS.
Layer 10 mL of suspension onto 7 mL lymphoprep in a sterile plastic
centrifuge tube. Spin for 15 min at SOOg.The lymphocytes collect at
the interface and can be removed with a pastette. They are washed
532 Fantes and Green

twice by centrifugation with excess complete medium and counted

before setting up at 0.5 x W cells/ml in complete medium containing
15% fetal calf serum and 1% PHA. Colcemid is added to 0.1 pg/mL
after 46 or 66 h of incubation, and the cultures are harvested 17 h later.

3.2. Chromosome Preparation

1. Take an aliquot of cell suspension for a cell count. Centrifuge cells at
180s for 10 min. Pour off supernatant and resuspend cells in fresh ice-
cold complete medium. Centrifuge cells at 18Ogfor 10 min; this wash-
ing step removes some dead cells and debris.
2. Pour off the supematant and resuspend cells in hypotonic 0.075M KC1
solution to swell the cells. Use 10 mL of hypotonic for every lo7 cells.
3. Incubate the lymphoblastoid cells for 20 min at 37OC;peripheral blood
lymphocytes require 10 min incubation at room temperature at this
stage.
4, Remove 0.25 mL for mitotic index determination: add 5 mL of 3:l
methanol:acetic acid and allow to stand for 10 min. Centrifuge at 3008
for 5 min, pour off the supernatant, and resuspend the pellet in a small
volume of fixative (see Note 2). Drop onto a clean slide, and dry
quickly in air. Stain in 2% Giemsa in pH 6.8 buffer and air dry. Count
the number of divisions in 500 cells.
5. After incubation in hypotonic solution, centrifuge the cell suspension
at 1808 for 5 min. All further steps should be carried out at 4OC.
6. Pour off the supernatant and resuspend the pellet in cold polyamine
buffer (Bl), 1 mL/ 10’ cells. Centrifuge at 1808 for 5 min.
7. Pour off the supernatant. Resuspend the pellet in cold polyamine buf-
fer plus digitonin (B2), 1 mL/ lo7 cells (seeNote 3).
8. Vortex vigorously for 30-60 s to break the cell walls. Monitor cell lysis
by phase-contrast microscopy or by fluorescence microscopy. In this
case, drop the chromosome suspension onto a slide previously spread
with a drop of fluorochrome such as Hoechst 33258 or ethidium
bromide. Place a coverslip in position, seal with rubber solution, and
examine. Most of the chromosomes should be free and in suspension
after 60 s vortexing; further vortexing will only cause an increase in
chromosome degradation and stickiness.
9. Nuclei should be removed from the chromosome suspension before
flow analysis and sorting, since they can contaminate sorted frac-
tions. Spin down the nuclei at 1808 for 10 min, and transfer the
Chromosomes, Analysis and Sorting 533

supernatant carefully to another tube. Add I mL of B2 buffer to the

pellet and resuspend by a 5s vortex. Centrifuge at 180s for 5 min,
remove supernatant, and add to the first supernatant. Check for the
presence of nuclei as described above in stage 8 (seeNote 4).
10. This chromosome suspension can be stored for 24 wk at 4OCwith
little loss of resolution when analyzed by flow cytometry.

3.3. Staining
1. If the chromosome suspension was prepared from cells with a high
mitotic index (> 30%), dilute I:1 with fresh B2 buffer before staining
(seeNote 5).
2. For single fluorochrome analysis, add Hoechst 33258 to 0.5 pg/mL or
ethidium bromide to 50 pg/mL.
3. For dual fluorochrome analysis (seeNote 6), add chromomycin A3 (see
Note 7) to 20 pg/mL, magnesium chloride to 1mMand Hoechst 33258
to 0.5 pg/mL from stock solutions. Leave for at least 1 h for the fluoro-
chromes to equilibrate before analyzing or sorting the chromosomes.

3.4. Chromosome Identification

Although chromosomes prepared in polyamine buffer (seeNotes 8
and 9) after exposure to 16 h colcemid are condensed, it is possible to band
and identify them if they are swollen and elongated slightly by prior ex-
posure to phosphate buffered saline. Not all the chromosomes on a slide
will be sufficiently decondensed to give adequate banding for identi-
fication, but over 50% should have sufficient bands and examples are
shown in Fig. 1.
1. Sort 60,000 chromosomes into a cold Eppendorf tube containing 0.25
mL of buffer B2. On our machine, this quantity of chromosomes will
be sorted in 0.25 mL of PBS sheath fluid so the sorted chromosomes
will finally be exposed to 1:l buffer B2:PBS.
2. Fix chromosomes by adding 40% formaldehyde to give a final concen-
tration of 4%. Leave for 10 min on ice.
3. Spin chromosomes onto alcohol-cleaned slides using a Shandon cyto-
centrifuge at 150s for 7min. 60,000 chromosomes can be split between
two slides.
4. Allow slides to air dry. Wash briefly in deionized water and air dry.
5. Fix in 3:l methanohacetic acid for 5 min and air dry.
534 Fantes and Green

Fig. 1. Selected photomicrographs of sorted chromosomes 1,2,11,13 and Y. The chrom-

osomes were sorted by flow cytometry, and deposited on microscope slides according to
the method described in the text. Dapi staining was used for chromosomes 1,2,11, and 13
and spermidine his-acridine for chromosome Y. The distinctive banding of the chromo-
somes associated with these dyes can be seen.

6. Either: (a) Stain in 0.5 PgJmLDAPIin distilled water for 10 min, wash,
and air dry. Mount in citifluor/glycerol; or (b) Stain in 0.005% sper-
midine his-acridine for 10 min, wash, and air dry. Mount in deionized
water.

3.5. Flow Cytometry

The following methods do not relate to any particular commercial
flow cytometer, but given that a machine consists of a light source (usually
a laser beam), an optical train, a liquid flow arrangement to deliver the
chromosome suspension, and a signal detection system, they are generally
applicable to all machines. A general text dealing with the basic principles
and a wide variety of applications of flow cytometry is recommended for
a newcomer to the subject (10). A typical bivariate flow karyotype of nor-
mal human chromosomes is shown in Fig. 2.
The analysis of human chromosomes requires the flow cytometer to
be performing as well as or better than specification. It is not good enough
to have the instrument roughly tuned to performance, which often suf-
Chromosomes, Analysis and Sorting 535
Human Chromosomes

CHROMOMYCIN A3

Fig. 2. Human chromosome fluorescence distribution for two DNA-specific fluores-

cent dyes Hoechst 33258 and Chromomycin A3. Differential amounts of AT- and GC-rich
DNA on each chromosome, which determines the intensity of Hoechst and Chromomy-
tin fluorescence, respectively, produce further separation of the chromosome peaks than
is seen with a single DNA-specific fluorochrome.

fices for cell analysis, and attention to detail will be rewarded with a
reduced coefficient of variation (CV) of the chromosome peaks, provided
of course that the suspension of chromosomes is well prepared. It will be
assumed that the reader has some knowledge of a particular flow cytom-
etry machine, and is able to adjust the signal detection and display system,
clean all lens and signal detection surfaces, and optimize the output of the
lasers or other light sources. Those parts of the system that are most likely
to give rise to performance loss when they are not optimized are dealt with
in detail below.
536 Fantes and Green

3.5.1. Sample Delivery

The arrangement of nozzles and pressurized sheath and sample
stream is designed to deliver chromosomes in single file through a well-
defined position in the cytometer, where a beam of light can be focused. All
tubes, nozzles, and reservoirs therefore must be clean, aligned, and purged
before injecting a series of samples. This will ensure an unobstructed lam-
inar flow, which leads to precise positioning of chromosomes and a stable
droplet formation. Any liquids introduced into the cytometer must be
filtered through a 0.2 pm filter.
1. Cleaning (see Note 10): Flush the following solutions through the
machine for 30 min each. This should include backflushing through
the sample stream.
a. Warm Dilute 7-X (or equivalent) solution to remove chromo-
some material from previous experiments.
b. Cidex to sterilize.
c. Use distilled water to rinse out unwanted detergent or Cidex.
d. Sheath buffer chosen for the experiment.
2. Alignment: In many cases, the sample and sheath nozzles have a fixed
geometry, but in cases where adjustment is provided, make sure that
the sample injection nozzle is placed centrally inside the sheath nozzle
and terminates where the inside diameter of the sheath nozzle is wid-
est. This can be checked by injecting a concentrated fluorescent dye
(preferably the dye to be used in subsequent experiments) through the
sample stream and focusing the laser beam somewhere near the sig-
nal detection point such that fluorescent light piping occurs right up
to the sample nozzle tip. Figure 3 shows an example of sheath and
sample nozzles both in and out of correct adjustment. Take care with
plastic sheath nozzles not to melt the plastic when looking for a light
piping effect.
3. Purging: Every flow cytometer is equipped with a liquid exhaust port
well above the sample injection nozzle, which can be opened to release
trapped air bubbles. These bubbles must be exhausted before a stable
flow and droplet formation can be established.
4. Sample pressure: Too high a sample pressure can cause the sample
stream to balloon, as shown in Fig. 3, and leads to a broad final sam-
ple stream and consequently a high chromosome peak CV. Most com-
mercial flow cytometers are equipped with a simple pressurized
sample delivery system, which is adequate but sometimes oversensi-
tive to adjustment when low sample to sheath pressure differentials
Chromosomes, Analysis and Sorting 537

SHEATH LIQUID
I I

. Widest 4
. Sheath Diameter ’

Fig. 3. The example shown is for a glass nozzle system for “in air” flow where the laser
beam is focused on the edge of the sheath nozzle to produce light piping in the concen-
trated dye sample. On the left, the sample nozzle is shown misaligned, and a high sample
pressure is also shown to be causing ballooning of the emerging sample stream. On the
right, the dye sample emerges at an acceptable relative pressure and passes centrally
down to the “in air” detection point. Alignment and sample pressure are equally import-
ant for cuvette detection systems.

are needed. A more satisfactory way of controlling sample flow is to

use a motorized syringe driver capable of delivering less than 1 mL
from a l-mL syringe. A cooling jacket around the syringe will main-
tain the chromosomes in good condition throughout an experiment.
5. Sample stream coating: A stable flow karyotype will be achieved
more rapidly when, prior to the introduction of a chromosome sus-
 Fantes and Green

pension, a “dummy” sample of buffer and dye (at twice the final con-
centration) is injected through the sample stream for approximately 5
min. Using a “dummy” sample between different chromosome sus-
pensions also helps to flush out remaining chromosomes from pre-
vious samples.
Generally, flow cytometers have a sheath liquid pressure of about
one atmosphere. Under these conditions, a chromosome suspension
containing lO’chromosomes/mL flowing at 0.4 mL/h, which leads to
a flow rate of about 1000 chromosomes/s, should result in a well-
resolved flow karyotype. A slower sample flow rate may produce
even better resolution, and depending on the quality of cytometer ad-
justment and the intensity of the laser beam(s), a faster flow rate may
not necessarily spoil the resolution. As high a flow rate as possible
should be achieved for chromosome sorting experiments.
3.5.2. Beam Alignment
1. Single beam experiments: It is important that the laser beam passes
precisely along the optical axis of the focusing lens system. Check this
by first removing the forward light scatter detector and then adjust the
laser beam to pass through an aperture at the geometric center of the
final focusing lens, and at the same time check that the beam is incident
on the center of a screen some distance away from the stream axis.
Swing the liquid stream to one side for this adjustment. A mark,
known to be on the optical axis, on an adjacent wall often provides a
useful target for beam alignment checking. Swing back the liquid
stream and adjust the beam focus at the stream to be at its narrowest
by observation through the stream viewing optics.
2. Dual beam experiments: Here, the lower wavelength beam, which in
our example is the ultraviolet beam, is aligned in a similar way to the
single beam experiments. The higher wavelength beam is aligned suf-
ficiently off axis to allow the spherical and chromatic aberrations to
compensate and bring the two beams into focus along the liquid
stream axis. Figure 4 shows how the longer focal length of the higher
wavelength beam and the shorter focal length of off axis rays produce
the desired alignment of focal points.

3.5.3. Fluorescence Detection

To optimize the detection of each fluorescent color, it is important to
select suitable long and short pass filters. The following combinations
Chromosomes, Analysis and Sorting 539

Liquid Stream

Fig. 4. The focusing arrangement of an ultraviolet (U.V.) and a blue 458 nM wavelength
beam is shown. A converging lens that could be cylindrical or convex focuses the U.V.
beam on axis and the 458 nM beam at a slight angle and off axis. The scale of distances
and beam widths are somewhat exaggerated in the figure, as are the chromosomes in the
liquid stream.

were used for the specific fluorescent dyes described here: (a) Hoechst
33258475 nM long pass + 550 nM short pass, (b) Chromomycin A3-515 nM
long pass.
3.5.4. Final Adjustment
The final adjustment of laser beam, stream position, flow rate, and
photomultiplier detector positioning needs an actual sample of flowing
objects. Standard practice is to use a suspension of fluorescent micro-
beads. This is not recommended prior to chromosome analyses and sort-
ing, since there is a risk of chromosome aggregation around stray micro-
beads left behind after the tuning process. Use instead a portion of the
chromosome suspension. The distribution of signal pulses appearing on
the oscilloscope will soon become familiar, particularly the prominent
cluster of signals, all with the same peak height, arising from the human
chromosome groups 9 to 12. Optimize the peak height and minimize the
peak width of the signal pulses by adjusting the optical components and
stream position in an ordered fashion. Readjustment of the stream posi-
tion between adjustment of each optical component will usually ensure a
steadily improving signal height and width.
540 Fantes and Green

3.5.5. Sorting
A working day will, as a rule, produce lo6 sorted chromosomes.
Reaching this target depends on a clean chromosome suspension, a steady
flow rate of 1000 to 2000 chromosomes/s, and a degree of stability of the
“live parts” of the flow machine. The purity of each sorted fraction will de-
pend on the resolution of chromosome peaks seen in the flow karyotype
and on prior identification of the chromosome or chromosomes appearing
in particular peaks of interest. Identification requires a sorted fraction of
about 60000 chromosomes into 0.25 mL of sheath buffer followed by the
procedure described in the chromosome identification section. Contami-
nation of a sorted fraction with chromosomes from adjacent peaks can be
minimized to a limited extent by drawing narrow sort windows, but the re-
searcher must be sure that the extra purity gained is justified by a possible
further day’s sort to accumulate the required quantity of chromosomes. It
is usually possible to be occupied on another task while each sample of
chromosome suspension is sorted, but it must be possible to check fre-
quently on the relative positioning of sorting windows, chromosome
peaks, and the droplet stability. No amount of care and attention to the
machine will improve sorted fraction purity if there are large numbers of
undividing nuclei or large quantities of debris in the original suspension.

4. Notes
1. The starting point for a good chromosome preparation must be a rap-
idly dividing cell culture with very few dead cells and free from
bacterial or mycoplasma contamination. Mycoplasma contamination
will cause the chromosomes in the final suspension to stick together
and form large clumps.
2. It is important to treat the cells gently during the preparation; cen-
trifuge at low speeds, resuspend the cell pellet by tapping the tube not
by vortexing, and ensure that all the cells are uniformly exposed to hy-
potonic and buffer solutions by maintaining a single-cell suspension.
3. The concentrationof cells to buffer 82inmethods,stage7is important;
if the amount of B2 is decreased, cell breakage will be incomplete.
4. The differential centrifugation step described in stage 9 will remove
most of the contaminating nuclei. Centrifugation of the chromo-
somes at higher speeds to concentrate them or remove debris will
increase the number of clumps and degraded chromosomes giving
Chromosomes, Analysis and Sorting 541

poor resolution. It is better to start again from a culture with a higher

mitotic index.
5. Do not attempt to analyze very concentrated chromosome suspen-
sions. Staining irregularities will occur as well as excessive signal
coincidences and nozzle blockages. Dilute with buffer B2.
6. Addition of 10 mM sodium citrate and 10 mM sodium sulfite (11) to
stained chromosomes has been reported to increase the resolution of
dual beam flow karyotypes.
7. Chromomycin A3 intercalates into the DNA, and this has sometimes
reduced the efficiency of subsequent DNA manipulations. This
problem has been overcome by dialysis of the sorted chromosomes
against two changes of BI to remove the stain.
8. Chromosomes stabilized by polyamines are not suitable for chromo-
some-mediated gene transfer into other cells. An alternative buffer
containing 15 mM Tris, 3 mM calcium chloride, and digitonin should
be used (12).
9. Chromosomes stabilized with polyamines are also not suitable for use
with antibodies against chromosomal proteins; other buffers are
recommended (23 ).
10. We have found that the best routine for cleaning the tubes of the flow
machine involves washing first with warm detergent solution to
remove residual chromosomes without fixing them to the tube walls,
and then sterilizing with activated glutaraldehyde (Cidex). Extensive
washing with sterile distilled water is then necessary to remove all
traces of Cidex. Using 70% ethanol, which is commonly used by flow
cytometer operators for flushing and sterilizing, will fix residual
chromosomes to the tube walls. The method recommended here
avoids this problem.

References
1. Langlois, R. G., Yu, L. C., Gray, J. W., and Carrano, A. V. (1982) Quantitative karyo-
typing of human chromosomes by dual beam flow cytometry. Proc. N&Z. Ad. SC!.
USA 79,7876-7880.
2. Green, D. K., Fantes, J. A., Buckton, K. E., Elder, J. K., Malloy, P., Carothers, A., and
Evans, H. J. (1984) Karyotypingand identification of human chromosome polymor-
phisms by single fluorochrome flow cytometry. Hum. Genef. 66,143-146.
3. Harris, P., Boyd, E., Young, B. D., and Ferguson-Smith, M. A. (1986) Determination
of the DNA content of human chromosomes by flow cytometry. Cyfogenet. Cell
Genet. 41,14-21.
542 Fantes and Green

4. Harris, I’., Cooke, A., Boyd, II., Young, B. D., and Ferguson-Smith, M. A. (1987) The
potential of family flow karyotyping for the detection of chromosome abnormali-
ties. Hum. Genet. 76,129-133.
5. Krumlauf, R.,Jeanpierre, M., and Young, B.D. (1982)Construction and characterisa-
tion of genomic libraries from specific human chromosomes. Proc. NutZ. Ad. Sci.
USA 79,2971-2975.
6. Cooke, H. J.,Fantes,J.A., and Green, D. K. (19831Structure and evolution of human
Y chromosome DNA. Differentiation 23, S48-S55.
7. Lebo, R. V., Gorin, F., Fletterick, R. J.,Kao, F., Cheung, M., Bruce, B. D., and Kan, Y.
W. (1984) High resolution chromosome sorting and DNA spot-blot analysis assign
McArdle’s syndrome to chromosome 11. Science 225,57-59.
8. Van de Sande, V. H., Lin, C. C., and Dengau, K. V. (1979) Clearly differentiated and
stable chromosome bands produced by a spermine bis-a&dine; a bifunctional
analogue of quinacrine. Exp. Cell Res. 120,439-444.
9. Sillars, R. and Young, B. D. (1981) A new method for the preparation of metaphase
chromosomes for flow analysis. J Histochem. Cytuchem. 29,74-78.
10. Melamed, M. R., Mullaney, P. F., and Mendelsohn, M. L. teds.) (1979) Flow Cytom-
etry and Sorting (Wiley, New York).
II. Van den Engh, G., Trask, B.,and Gray, J.W. (1987) Improved resolution of bivariate
flow karyotypes by manipulation of staining conditions. Cytometry Suppl. 1,3.
12. Porteous, D. J., Morten, J. E. N., Cranston, G., Fletcher, J. M., Mitchell, A., van
Heyningen, V., Fantes, J. A., Boyd, P. A., and Hastie, N. D. (1986) Molecular and
physical arrangements of human DNA in HRASl-selected, chromosome mediated
transfectants. Mol. Cellular Biol. 6,2223-2232.
13. Trask, B., Van den Engh, G., Gray, J. W., Vanderlaan, M., and Turner, B. (1984) Im-
munofluorescent detection of histone 2B on metaphase chromosomes using flow cy-
tometry. Chromosoma 90,295-302.
 Chapter 40

Flow Cytometry

Michael G. Ormerod
and Patrick R. Imrie

1. Introduction
There are several commercial flow cytometers on the market. They all
operate on the same basic principle, but there are important differences in
their design, and the methods for alignment, and so on, depend on the type
of instrument. In this chapter, therefore, we describe procedures for pre-
paring samples and limit description of the flow cytometry to comments
about important features of the analysis. One of the exciting applications
of the techniques is the ability to sort cells. Good separation of cells de-
mands good preparation of the sample and adequate analysis. Although
procedures for sorting cells are not given for reasons given above (see
Chapter 41, this vol.), the methods described in this chapter and the com-
ments about analysis are applicable. A description of the basic principles
involved in the design and operation of a flow cytometer can be found in
references 1 and2. An associated microcomputer is important, since multi-
parametric data cannot be analyzed adequately without this accessory.
The comments on analysis assume that there is an adequate facility for
computing.
Flow cytometry measures properties of cells in a flow system. It is
therefore a prerequisite of the method that the particles (cells or nuclei) are
in suspension and free of clumps. The quality of the information obtained
will largely depend on the quality of the preparation of the sample. This
comment applies to all the methods described.
543
544 Ormerod and Imrie

2. Materials
1. Phospate buffered saline (PBS). 8.5 g NaCl, 1.07 g Na,HPO, (or 2.7 g
Na,Hl?O,~lZ~O), 0.39 g NaH,PO, in 1 L of water.
2. Ribonuclease A (RNase). Type III-A. Stock solution: 1 mg/mL in PBS.
Store at -2OOC.
3. Collagenase. Type 1A. Stocksolutionl mg/mLinPBS. Storeat-20°C.
4. Pepsin. Lyophilized powder from stomach mucosa. Prepare solu-
tions as needed by dissolving the enzyme in 0.9% NaCl and adjusting
the pH to 1.5 with HCl.
5. Propidium iodide (PI). Stock solution: 100 pg/mL in PBS. Store at
4°C.
6. Flourescein diacetate (FDA). Stock solution: 1 mg/mL in acetone.
Store at -20°C.
7. Stain-detergent solution for preparing nuclei from fresh tissue: 1 g tri-
sodium citrate, 564 mg NaCl, 300 PL Nonidet P-40,10 mg ethidium
bromide in 1 L distilled water. Just before use, dissolve 1 mg of RN-
ase in 100 mL solution.
8. Solution for diluting antibodies in the Bromodeoxyuridine method:
1% BSA, 5% normal goat serum, 0.5% Tween 80, and 20 mM EDTA in
PBS, pH 7.5.
All the figures shown in this chapter were recorded using an Ortho
Cytofluorograf 50H equipped with a 50 mw Lexel argon-ion laser and a
5W Coherent laser driving a dye laser. The instrument was interfaced to
an Ortho 2150 computer system, and the figures were taken directly from
the computer’s graphics screen.

3. Methods
3.1. Labeling Cells with Antibodies
3.1.1. Analysis of a Single Antigen (See Notes l-3)
The basic method for cell labeling is straightforward-a typical proce-
dure is included in the method for studying DNA labeled with bromo-
deoxyuridine. The method of analysis given below is suitable for cells
labeled with a fluorescein conjugated antibody. It is assumed that propi-
dium iodide (PI) has been added in order to distinguish dead cells. In cell
mixtures, light scatter can be used to distinguish between different types
of cell (seeFig. 1.).
Flow Cytometry 545

Fig. 1. Human keratinocytes labeled with a mouse monoclonal antibody (23.10) that
reacts with basal cells (3) followed by fluoresceinated goat anti-mouse immunoglobulin;
PI has been added. A shows a histogram of red fluorescence; a region 1 has been set to in-
clude cells that have taken up PI and are therefore fluorescing red. These are excluded
from further analysis. B is a cytogram of light scattered orthogonally against that scat-
tered over a narrow forward angle. Each dot represented a cell recorded by the instru-
ment. Region 1 has been set to include the smaller cells and region 2 to include the larger.
The green (antibody) fluorescence of the small cells is shown in histogram C and that from
the large in histogram D. Cells supplied by Dr. J. Burman.

1. With a flow cytometer employing an argon-ion laser, tune the laser to

488 nm and set up the instrument to measure four parameters-light
scattered orthogonally, and in a forward direction, green (fluorescein)
and red (PI) fluorescent light. The green fluorescence should be sel-
ected using a band pass filter centered on 520 nm and the red with a
long pass filter with a cutoff close to 600 nm.
2. Display red fluorescence and set a gate to exclude labeled cells from
further analysis.
3. Display a cytogram of orthogonal vs forward light scatter, and set a re-
gion to include single cells of the desired type and to exclude, as far as
possible, clumps and debris.
4. Display a histogram of green fluorescence of the single cells of interest.
546 Ormerod and Imrie

RED FLUOR. RED FLUOR.

Fig. 2. Single cells prepared from the human mammary gland and labeled with rab-
bit antiepithelial membrane antigen (6) and mouse anti-common acute lymphoblastic
leukemia antigen (7) followed by fluoresceinated goat antirabbit and phycoerythrin
conjugated sheep antimouse Ig. The single cells have been selected using light scatter.
A is a cytogram showing green vs orange fluorescence; inB the computer has been used
to correct for overlap between the green and orange channels. Cells prepared by Dr. M.
O’Hare.

3.1.2. Simultaneous Analysis of Two Antigens

(See Notes 4 and 5)
If two antigens are to be visualized, it is customary to label one anti-
body with fluorescein and the other with either phycoerythrin (PE) (4) or
Texas red (a derivative of rhodamine) (5). PE can be excited by the same
laser line as fluorescein. There is some overlap between the emission spec-
tra of these two dyes (that is, fluorescein will be detected in the PE channel
and vice versa), and a correction must be applied either electronically or by
the computer. Figure 2 shows an example of this. The procedure for the
simultaneous analysis of fluorescein and phycoerythrin is given below.
1. Tune the argon-ion laser to 488 nm, and set up the instrument to meas-
ure light scatter (orthogonal and forward), green fluorescence at 520
nm and orange fluorescence at 580 nm.
2. Display a cytogram of the light scatter, and set a suitable region to in-
clude the cells of interest.
3, Display a cytogram of green vs red fluorescence of the selected cells.
4. Display a cytogram of corrected green vs orange fluorescence.
Flow Cytometry

3.2. Analysis of DNA

A variety of methods have been published for measuring the DNA
content of cells (8). Those used in this laboratory are given below. The im-
portant features of the analysis are to ensure that the profiles are as sharp
as possible (as measured by the coefficient of variation acrosss the peak),
and that two or more cells stuck together are excluded from the analysis.
The width of the peaks arising from cells in Gl or G2 of the cell cycle
will depend on the flow rate, and this should be kept aslow as possible. For
this reason, the concentration of cells should be between 5.105and 106/mL.
Two cells or nuclei stuck together will have the same DNA content as a sin-
gle cell in G2 of the cell cycle, and the two must be distinguished if the DNA
histogram is to reflect accurately the state of the cell cycle. In instruments
that focus the laser beam down to an ellipse with a cross-section of 5 pm or
less, the difference can be resolved by analyzing the shape of the signal
generated as the particle crosses the laser beam. The width of the signal
generated is given by the sum of the width of the laser beam and nuclear
diameter. As a cell progresses through its cycle, while the content of DNA
doubles, the nuclear diameter will not increase by more than 25%. Conse-
quently, there is only a small increase in the signal width. Because of the
nature of the flow system, two particles stuck together will align one be-
hind the other; their combined diameter will be at least two nuclear diam-
eters. This is illustrated in Fig. 3. Using the Ortho Cytofluorograph, the
width of the signal is best measured by comparing the signal height to its
area. This can be displayed on a cytogram on which the clumps can be dis-
tinguished from single cells. A region can be selected to include only the
latter (Fig. 4). On instruments in which the laser beam is focused to a cir-
cular spot of diameter 60 PM or more (for example the FACS series made
by Becton-Dickinson), this type of analysis is not possible.
Whenmeasuring ploidy, the concentration of cells from one sample to
another should be kept the same so that the dye/DNA ratio does not vary
too much. A standard sample is also needed. Some workers use either
chicken or trout erythrocytes (9). These are not entirely satisfactory, since
their DNA content is considerably lower than that from mammalian cells.
We use ethanol-fixed sheep lymphocytes stained in parallel, and run a
sample before and after the unknown. If there is any doubt or a small
change in ploidy is suspected, the standard and the unknown are mixed
and rerun. It is often possible to distinguish the sheep lymphoctyes from
other cells on the basis of light scatter so that the two populations can be
548 Ormerod and Imrie

Iarer

“‘Ln In In Gl
limo

G2 2xQl

Fig. 3. The signal shape generated at the photomultiplier by the fluorescence froma nu-
cleus in Gl, a nucleus in G2, and two nuclei inG1 stuck together. The area under the curve
generated by a nucleus in G2 and that from two nuclei in Gl are the same and equal to
twice that of the curve from a single nucleus in Gl. However, the peak height of the G2
signal is twice that from the clumped nuclei.

selected from the mixture, the histograms of their DNA content displayed
separately, and their peak channels compared. When samples of cancerous
tissue are studied, there are usually some normal diploid cells present, and
these can act as an internal standard.
Having recorded a DNA histogram, a computer program is often used
to deconvolute the curve into the separate contributions from the different
phases of the cell cycle (GO/Gl, S, and GZ). A variety of programs have
been written; we have recently published an aligorithm that is particularly
fast and handles perturbed histograms (20).
3.2.1.
Measurement of the DNA Content of Cultured Cells
and Cells in Suspension (See Notes 6-8)
1. Prepare a suspension of single cells in 200 PL of PBS.
2. Add vigorously 2 mL of ice-cold 70% ethanol, 30% PBS. Leave for at
least 30 min in the cold.
Flow Cytometry 549

Fig. 4. Sheep lymphocytes fixed in ethanol, treated with RNAase and labeled with PI.
The cytogram in A shows the signal peak displayed against the signal area. The region
was drawn to include single cells and exclude clumps. Note the appreciable contribution
madebytwocellsstuck together(arrowedincytogram1. ThehistogramB shows theDNA
histogram (signal area) of the single cells only, and C shows the histogram that resulted
when all the cells were included.

3. Harvest the cells by centrifugation. Resuspend in 800 clr, of PBS.

4. Add 100 PL of RNase (1 mg/mL) and 100 PL of PI (400 pg/mL). In-
cubate at 37OC for 30 min.
5. Analyze.
3.2.2. Measurement of the DNA Content of Fresh Tissue
Several methods for extracting nuclei from fresh tissue have been pub-
lished. We use one adapted from that of Petersen (II).
1. Take 200 mg of wet weight or more of tissue. Mince with scalpels in
tissue culture medium.
2. Filter through a stainless-steel tea-strainer.
3. Centrifuge at 8OOgfor 5 min.
4. Aspirate and discard the supernatant, and resuspend the pellet in 10
mL of stain-detergent solution.
550 Ormerod and Imrie

5. Mix on a rotator at 4OCfor 1-4 h. The exact time will depend on the tis-
sue used, and must be found by trial and error.
6. Filter through 60qM pore nylon mesh.
7. Analyze.
3.2.3. Measurement of the DNA Content
of Fixed Tissue Embedded in Parafin Wax
(See Notes 9 and 10)
Tissue removed during an operation or an autopsy is often fixed in for-
malin and embedded in paraffin wax prior to cutting tissue sections for
staining and histopathological examination. These blocks are stored for
many years, and the development of a method for extracting nuclei from
them has given flow cytometrists access to much archival material (12).
1. Cut three or four 40-pm sections from each block. Store in xylene in
glass containers. (Check the lids. Xylene will dissolve the rubber
washers often found in screw-top lids.)
2. Harvest tissue by centrifugation in glass tubes. Reextract twice with
xylene to complete removal of paraffin wax.
3. Wash the tissue twice in ethanol, once in 50% (v/v) ethanol and twice
in PBS.
4. If the tissue is particularly fibrous, incubate for 1 h at 37OCin 1 mg/mL
collagenase in PBS.
5. Centrifuge and resuspend in 0.9% NaCl, 0.5% pepsin adjusted to pH
1.5 with HCl. Incubate at 37OC for 1 h.
6. Pass through a 23-gage needle to break up clumps. Wash once in IM
phosphate buffer, pH 7.0 and once in PBS.
7. Finally, resuspend in PBS, containing 100 pg/mL RNase, 10 pg/mL
PI, and incubate for 30 min at 37°C.
8. Analyze, measuring orthogonal and forward light scatter and peak
and area of the red (PI) fluorescent signal.
3.2.4. Combined Analysis of a Surface Antigen and DNA
This can be used to observe any cell cycle dependence of expression of
an antigen.
1. Prepare a suspension of 1-2 x lo6 single cells.
2. Incubate in an appropriate dilution of primary antibody (for example,
a murine monoclonal antibody) in tissue culture medium (TCM) plus
10% fetal calf serum (FCS) at 0°C for 1 h.
3. Wash the cells by centrifugation twice in TCM plus FCS.
Flow Cytometry 551

4. Incubate the cells in an appropriate dilution of fluorescein conjugated

second antibody (for example, goat antimouse Ig) in TCM plus 10%
FCS for 1 h at OOC.
5. Wash the cells once in TCM plus 10% FCS; once in PBS.
6. Resuspend the cells in 200 FL PBS, and take up and down through a
pipet tip to ensure a good suspension of single cells.
7. Add vigorously 2 mL of ice-cold 70% ethanol, 30% PBS. Leave at least
30 min in the cold.
8. Harvest the cells by centrifugation. Resuspend in 800 FL of PBS.
9. Add 100 FL of RNase (1 mg/mL), 100 PL of PI (400 pg/mL). Incubate
at 37°C for 30 min.
10. If necessary, pass the cells through a 25-gage needle.
11. Analyze on the flow cytometer using an argon-ion laser tuned to 488
nm and recording light scattered in a forward angle, green (antibody)
fluorescence (520 nm) and red (PI) (>630 nm) fluorescence on peak and
area (seeabove). On an instrument capable of recording 5 parameters,
also measure orthogonal light scatter.
12. On analysis, gate out any debris and large clumps using light scatter,
select single cells using peak vs area for the red fluorescence, and then
display green vs red fluorescence. If the antibody staining is weak, it
may be necessary to correct for red fluorescence in the green channel.
An example of this analysis is shown in Fig. 5.
3.2.5. Combined Analysis of Cytoplasmic
or Nuclear Antigen and DNA
If the antigen is inside the cell, the cell has to be fixed before incubat-
ing with the first antibody The fixative used should be chosen to avoid
damage to the antigen being studied. The DNA in a fixed cell is suscepti-
ble to attack by nucleases, and this can cause problems. It can be avoided
by a clean technique, adding EDTA to the buffers, and incubating the cells
on ice.
A procedure for observing progesterone receptors is given (worked
out in collaboration with Dr. J. Mansi).
1. Prepare a suspension of l-2 lo6 single cells in 910 PL of PBS.
2. Add 90 PL of 40% formaldehyde solution. Leave at room tempera-
ture for 15 min.
3. Harvest the cells by centrifugation, and wash once in PBS.
4. To the cell pellet, add 2 mL of absolute methanol at -2OOC. Leave for
4 min.
552 Ornaerod and Imrie

Fig. 5. Cultured cells from the human mammary gland labeled with a murine mono-
clonal antibody against CALLA (71 followedby fluoresceinated sheep antimouse Ig, fixed
in 70% ethanol, and incubated with RNase and PI. Cytogram A shows the signal peak
against area for red (DNA-PI) fluorescence. The single cells have been included in Region
1 and all other particles havebeenexcluded from further analysis. CytogramB shows the
orthogonal vs forward light scatter. The larger cells in the elliptical fluorescence dis-
played against red fluorescence in cytogram C. Cells prepared by Mrs. S. Davies.

5. Centrifuge, remove the methanol, and add 2 mL of acetone at -2OOC.

Leave for 2 min.
6. Centrifuge, remove the acetone, and wash the cells twice in PBS,
finally resuspending in 100 PL of PBS, 10 mM EDTA. Take the cells up
and down through a pipet tip to ensure a suspension of single cells.
7. Add 200 PL of 5% normal goat serum in PBS/EDTA. Incubate on ice
for 15 min.
8. Add 200 PL of monoclonal rat antibody (antiprogesterone receptor,
Abbott Diagnostic). Add an irrevelant rat antibody to the control. In-
cubate for 1 h at 0°C.
9. Wash the cells twice in PBS/EDTA/O.l% bovine serum albumin
(BSA).
Flow Cytometry 553

10. Resuspend final pellet in a fluoresceinated goat antirat Ig diluted ap-

propriately in PBS/EDTA/BSA. Incubate for 1 h at OOC.
11. Wash the cells twice in PBS/EDTA, finally resuspending in 800 of yL
PBS.
12. Add 100 PL of PI solution (100 ug/mL), and 100 PL of RNase solution
(1 mg/mL).
13. Incubate for 15 min at room temperature before analysis by the same
procedure described in the preceding section.
3.2.6. Cells with DNA Labeled
with Bromodeoxyuridine (Brd U)
This is another example of the method given above. It is also described
because of its importance (seeNotes 11 and 12).
If BrdU is injected into an animal or added to a cell culture, it will be
incorporated into the DNA of cells synthesizing DNA. The BrdU can be
visualized subsequently by reaction with an antibody directed against
bromouracil. This is a powerful technique for studying the cell cycle and
perturbations therein caused by radiation and cytotoxic drugs. It can also
be used to estimate the growth fraction of a culture.
Most antibodies available will not react with BrdU in native DNA. It
is necessary to disrupt the structure of the DNA; one of several published
methods can be used (seereference 13).
1. Prepare a suspension of about 2 x lo6 cells.
2. Resuspend in 200 PL of PBS pipeting up and down to ensure a good
suspension of single cells.
3. Vigorously add 2 mL of ice-cold 70% ethanol.
4. Stand on ice for at least 15 min. (The cells can be stored at this stage
for several days at 4OC.)
5. Harvest the cells by centrifugation and wash once in PBS.
6. Resuspend the cell pellet in 2 mL of 2M HCl. Incubate at 37OC for 30
min.
7. Harvest the cells; wash once in 5 mL of 1Mphosphate buffer, pH 7, and
once in PBS.
8. Resuspend cells in 200 PL of incubation buffer (seeMaterials section).
Add 10 PL of rat anti-BU (ICR2, Seralab, Ltd., Crawley). Incubate on
ice for 1 h.
9. Wash the cells twice in PBS, 0.5% Tween 80.
10. Resuspend in 200 PL of fluoresceinated, goat antirat Ig diluted appro-
priately in incubation buffer. Incubate on ice for 1 h.
11. Wash once in PBS/Tween and once in PBS.
554 Ormerod and Imrie
12. Resuspend in 450 PL of PBS. Add 50 PL of PI solution (100 pg/mL).
13. Incubate at 0°C for 15 min.
14. Analyze according to protocol described above for combined analy-
sis of surface antigen and DNA.
A typical result is shown in Fig. 6.

3.3. Monitoring the Ebctropermeabilization of Cells

Electroporation is the permeabilization of cells by a high voltage
pulse. On subsequent incubation, if the correct conditions are used, the
cells will reseal their plasma membranes. The method is an efficient way
of introducing foreign DNA into cells.
A variety of parameters require optimization. For effective trans-
fection, it is necessary to know whether holes have been punched in the
plasma membrane, for how long they stay open, how efficiently they re-
seal, and the effects of this treatment on cell viability. Different types of cell
require different conditions for electroporation in order to optimize these
conditions (24,X). The method is based on the observation that propidium
iodide is excluded by an intact plasma membrane. If it is added imme-
diately after electropermeabilization, cells whose membrane has been
ruptured will fluoresce red. After the cells have been incubated in warm
medium, fluorescein diacetate is added. This is taken up by live cells and
hydrolyzed to fluorescein, which fluoresces green and will be retained by
cells with an intact membrane.
The details of the flow cytometric method are given. For information
about the method of electroporation, seeref. 15.
1. Electroporate a precooled suspension of single cells at 0°C.
2. Add PI to a concentration of 10 pg/mL. Incubate at 0°C for 10 min.
3. Dilute the cells in culture medium at 37OC. Incubate for 10 min.
4. Wash the cells in PBS and resuspend in 1 mL of PBS.
5. Add FDA to a final concentration of 10 ng/mL. Incubate at room tem-
perature for 10 min.
6. Analyze, recording orthogonal and forward light scatter, red and
green fluorescence. Gate on the light scatter from single cells and dis-
play green vs red fluorescence.
7. In the final cytogram, three classes of cells are distinguished: fluores-
cein positive, PI negative (unpermeabilized); fluorescein positive, PI
positive (permeabilized and resealed membranes); fluorescein nega-
tive, PI positive (permeabilized and failed to reseal) (Fig. 7 and Note
13).
Flow Cytometry 555

Fig. 6. Chinese hamster V79 cells labeled in culture for 30 min with 10 @I BrdU. The
cells were labeledwith a rat anti-BU antibody and PI and analyzed as in Fig. 5. An iso-
metric display of the final histogram of green vs red fluorescence is shown. The cells in
S phase of the cell cycle have incorporated BrdU and are labeled with the antibody. Cells
prepared by Mrs. P. Loverock.

4. Notes
1. Dead cells (including fixed cells) tend to take up antibodies nonspeci-
fically; this can be reduced by adding an extraneous protein to the
solutions of antibody and taking particular care in washing the cells.
2. When antigens on the surface of the cell are studied, it is usual to label
live cells. In such cases, the cells should be kept cold; all incubations
should be at 0°C and all buffers kept on ice. At higher temperatures,
the antigens in the plasma membrane may become crosslinked by
antibody (called “capping”) and then be pinocytosed.
3. Since clumps of cells and dead cells may label nonspecifically, they
should be excluded as far as possible from the analysis by measuring
scattered light. Dead cells can also be distinguished from their live
counterparts by adding propidium iodide (PI) at a typical concentra-
tionof 10 pg/mL. This dyeis excluded by theintact plasma membrane
in live cells and fluoresces red when bound to nucleic acid. For some
types of cells, such as lymphocytes, the addition of PI is unnecessary
556 Ormerod and Imrie

t FLUORESCEIN1
l PROPIDIUH IODIDE I

Fig. 7. A fibroblastic rat cell line electoporated at 5 kV/cmin low conductivity medium
(15. The cytogram shows green (fluorescein) vs red (PI) fluorescence. Three groups of
cells can be identified: fluorescein +ve, PI -ve; fluorescein +ve, PI +ve; and fluorescein
-ve, PI +ve. Cells supplied by Drs. T. McMillan and M. J. O’Hare.

since dead cells can be distinguished by their reduced light scatter.

4. Texas red is excited by orange light and fluoresces red (peak emission
at 620 nm). A second laser must be used, either a krypton laser tuned
to 568 nm or a dye laser using rhodamine 6G and tuned to 580 nm. This
has the disadvantage of using additional expensive hardware. The
advantage is that the two laser beams are separated spatially (usually
set between 30 and 130 pm apart), so that the signals from them are
resolved both optically and in time; “cross-talk” between the red and
green channels is usually undetectable. With a dual laser cytometer,
with the combination of either fluorescein, phycoerythrin and either
Texas red or allophycocyanin, three antibodies can be detected simul-
taneously (26,Z7).
5. When measuring two antibodies, PI may be added to exclude dead
cells (28). An anticipated difficulty in correcting for overlap between
PI fluorescence and that from PE or Texas red should not be a problem
since only the excluded (dead) cells are stained with PI.
6. If there are an excessive number of clumps, these may be reduced by
passing the sample through a 25-gage hypodermic needle.
7. If PI or ethidium bromide is used to label DNA, the profiles may be
broadened because of the presence of residual RNA, particularly in
Flow Cytometry 557

larger cells. This can usually be resolved by replacing the stock solu-
tion of RNase. In particularly difficult cases, it may be necessary to
add mithramycin (final concentration: 50 pg/mL, 7.5 mM MgCl,) and
excite at 457 nm (mithramycin binds to DNA but not RNA) (19).
8. Another problem is the degradation of DNA by contaminating nucle-
ases. Attention should be paid to the cleanliness of all equipment
used. Storage of prepared samples is not generally advisable.
9. Occasionally it is found that the DNA in the extracted nuclei is badly
degraded. This appears to be related to the treatment of the tissue
prior to fixation, so little can be done about it. Sometimes light scat-
ter can be used to exclude the more badly degraded material.
10. The amount of PI bound is affected by the conditions of fixation of the
original tissue (20) and account should be taken of this when deter-
mining ploidy.
11. Failure to stain the BrdU-labeled DNA usually relates to a failure to
disrupt the DNA sufficiently.
12. Sometimes there are problems with background stain observed by
staining cells that have not been labeled with BrdU. Attention should
be paid to procedures for washing the cells-increasing the number of
washes and/or the protein concentration in the buffer used.
13. In Fig. 7, the unpermeabilized cells were present because a small num-
ber of cells were retained in the tip of the pipet and were not subjected
to the electric field. They served as a useful built-in control.

Acknowledgments
This_ work
- was supported by a joint program
_ - grant from the Cancer
Research Campaign and the Medical Research Council.

References
2. Shapiro, H. M. (1985) PracticalFlow Cytometry (Liss, NY, New York).
2. Dilla, A. V. D., Dean, P. N., Laerum, 0. D., and Melamed, M. R. feds.1 (1985) Flow
Cytometry: hsfrumentution and Dafa Analysis (Academic, London).
3. Gusterson, B. A., McIlhinnney, R. A. J., Patel, S., Knight, J., Monaghan, P., and
Ormerod, M. G. (1985) The biochemical characterisation and immunocytochemical
characterisation of an antigen on the membrane of basal cells of the epidermis.
Difierenfidion 30,102-110.
4. Oi, V. T., Glazer, A. N., and Stryer, L. (1982) Fluorescent phycobiloprotein conju-
gates for analyses of cells and molecules. J. Cell Biol. 93,981-986.
5. Titus, J. A., Haugland, R., Sharrow, S.O., and Segal, D. M. (1982) Texas red, a hydro-
558 Ormerod and Imrie

philic red-emitting fluorophore for use with fluorescein in dual parameter flow
microfluorometric and fluorescence microscopic studies. J. ImmunoI. Methods 50,
193-204.
6. Sloane, J.P. and Ormerod, M. G. (1981) Distribution of epithelial membrane antigen
in normal and neoplastic tissuesand its value in diagnostic tumor pathology. Cancer
47,1786-l 795.
7. Gusterson, B. A., Monaghan, P., Mahendron, R., Ellis, J., and O’Hare, M. J. (1986)
Identification of myoepithelial cells in human and rat breasts by anti-commonacute
lymphoblastic leukemia antigen antibody A12. J. NUB CancerInst. 77,343349.
8. Taylor, I. W. and Milthorpe, B.K. (1980)An evaluation of DNA fluorochrome, stain-
ing techniques and analysis for flow cytometry. J. Hisfochem. Cytochem. Z&1224-
1232.
9. Vindelov, L. L., Christensen, I. J., and Nissen, N. I. (1983) Standardisation of high-
resolution flow cytometric DNA analysis by the simultaneous use of chicken and
trout red blood cells as internal reference standards. Cytometry 3,321-328.
20. Ormerod, M. G., Payne, A. W. R., and Watson, J.V. (1987) Improved program for the
analysis of DNA histograms. Cyfomefy 8,637-641.
11, Petersen,S. E. (1985) Flow cytometry of human colorectal tumors: nuclear isolation
by detergent technique. Cytomefy 6,452-460.
12. Hedley, D. W., Friedlander, M. L., Taylor, I. W., Rugg, C. A., and Musgrove, C. A.
(1983)Method for analysisof cellular DNA content of paraffin-embedded patholog-
ical material using flow cytometry. J. Histochem.Cytochem.31,1333-1335.
23. Gray, J. W. and Mayall, B. H. (eds.) (1985) Monoclonul Antibodies Against Bromode
Oxyuridine (Liss, New York).
14. O’Hare, M. J., Asche W., Williams, L., and Ormerod, M. G. (1989) Optimisation of
transfection of breast epithelial cells by electroporation. DNA, in press.
15. O’Hare, M. J.,Ormerod, M. G., Imrie,P. I., Peacock,J. H., and Asche,W. (1989)Elec-
tropermeabilisation and electrosensitivity of different types of mammalian cells, in
Electroporufion and Elecfrojksion in Cell Biology (Neumann, E., Sowers, A. E., and
Jordan, C., eds.), Plenum, New York, in press.
16. Parks, D. R., Hardy, R. R., and Herzenberg, L. A. (1984) Three-color immunofluores-
cence analysis of mouse B-lymphocyte poplulations. Cytometry 5,159-W.
17. Loken, N. R. and Lanier, L. L. (1984) Three-color immunofluorescence analysis of
Leu antigens on human periphereal blood using two lasers on a fluorescence acti-
vated cell sorter. Cytometry 5,151-158.
18. Sasaki,D. T., Dumas, S.E., and Engelman, E. G. (1987) Discrimination of viable and
non-viable cells using propidium iodide in two color immunofluorescence. Cyto-
mety 8,413-420.
19. Barlogie, B.,Spitzer, G., Hart, J.S.,Johnston, D. A., Buchner, T., Schumann, J.H., and
Drewinko, B. (1976) DNA histogram analysis of human hemopoietic cells. Blood 48,
245-258.
20. Schutte, B., Reynders, M. M. J., Bosman, F. T., and Blijham, G. H. (1985) Flow cyto-
metric determination of DNA ploidy level in nuclei isolated from paraffin-embed-
ded tissue. Cyfomehy 6,26-30.
 Chapter 41
Fluorescence-Activated Cell
Sorting According
to Receptor Density
Application for Isolating
Transfected Cell Lines

Stella Clark
1. Introduction
Isolation of cells expressing transfected cDNAcan be extremely diffi-
cult and tedious if the efficiency of expression is low (less than l-5%). If the
receptor being selected for is a cell surface or integral membrane protein at
least partly exposed to the exterior, then it is possible to use the extremely
powerful technique of fluorescence-activated cell sorting (FACS) to select
these cells from a receptor-negative background.
The technique depends on being able to tag the receptor-expressing
cells with a fluorescent label. A detector identifies fluorescent cells in a
stream of single cells passing through a laser beam. An electric charge is
applied to this stream and positive cells can be deflected into a collecting
tube. The entire procedure is carried out under sterile conditions, so that
the positive cells can be expanded for further sorting and ultimately exam-
ination of the expressed receptor. Theoretically, this method could select
out a single fluorescent cell from a population of nonexpressing cells; in
practice one is looking at a level of detection of l:l@-104.

559
560 Clark

There are two types of fluorescent probes that are used to label recep-
tors on cells. First, one can use a fluorescently labeled ligand, and indeed,
several groups have used the FACS to measure receptor numbers and to
identify subsets within a cell population based on cell size and receptor ex-
pression using this approach (Z-5). However, ligand receptor binding is
usually reversible and thus the continued presence of the ligand in solu-
tion during a sort would be necessary. As yet, using the ligand to detect
transfected receptor expression by FACS has not been reported.
Secondly, antireceptor antibodies can be used in conjunction with a
fluorescent second antibody. In general, antibodies bind to receptors with
higher affinity than ligands, so dissociation of the label during the sort is
unlikely. By using a double antibody system, a much higher sensitivity can
be achieved. For example, from one transfection, we observed no positive
cells above background at the first sort, but by the fifth sort, a large percen-
tage of the cells were expressing the desired antigen (Fig. 1). This method
has been used both to isolate cells expressing transfected receptors and to
select high expressers for the transferrin receptor (6) and the insulin recep-
tor (7). We have also used this approach to select for EGF-receptor expres-
sion (8).
This chapter describes the preparation of cells for FACS using a fluo-
rescent double-antibody method.

2. Materials
1. Complete cell medium, e.g., alpha minimal essential media plus 10%
v/v fetal calf serum.
2. Phosphate Buffered Saline (PBS): 0.14M NaCl, 2.7 mM KCl, 1.5 mM
KH2P04, 8.1 mM Na,Hl?O,. Sterilize.
3. Trypsin solution: 0.05% Trypsin in PBS/O.5 mM EDTA.
4. Earle’s Balanced Salt Solution (10 x EBSS): Make up as a 10x stock,
sterilize, and store at room temperature. 1.2M NaCl, 54 mM KCl, 12
mM NaH$‘O,, 8 mM MgSO,, 18 mM CaCl,. If using a commercial
source, specify without bicarbonate and phenol red.
5. FACS Buffer A: EBSS (1:lO dilution of stock) containing 10 mM Hepes
pH 7.4, filter sterilize, and add 2% (v/v) heat inactivated (30 min/
50°C) fetal calf serum.
6. FACS Buffer B: FACS Buffer A without serum. NOTE: For sorting one
cell population, about 100 mL of buffers A and B are needed. Make up
on day of sort and cool to 4°C.
Fluorescence-Activated Sorting 561

SORT NUMBER

--
SCATTER

Fig. 1. FACS Sorting of Chinese Hamster Ovary (CHO) Cells Transfected with EGF-
Receptor cDNA. G418 resistant CHO cells from one transfection were pooled for the
first sort. The cells were prepared for the sort as described in the Method section.
Approximately lo7 cells were used in each sort (+Rl) and 10s cells for the negative
control (-Rl). A431 cells labeled with Rl were used to align the machine. The data is
presented as a series of dot plots with forward scatter on the horizontal axis (allows one
to gate out dead cells and cell aggregates) and fluorescence intensity on the vertical axis
(measures level of receptor expression). Sort 1 showed no difference between k first
antibody. The top 5% of cells were collected and expanded for resorting. From sort 2
onwards, an increasing fraction of the cells showed fluorescence levels above the
control, i.e., sort 3‘8% positive; sort 5,56% positive.

7. Antireceptor Antibody (first antibody): In general, use a high titer

antireceptor antibody directed against the extracellular domain of the
receptor. Use at a concentration of 100'M. For EGF-receptor, the
monoclonal Rl (9) is used at a 1:50 dilution.
8. Second Antibody Conjugated to a Fluorescent Probe. Choose an ap-
propriate second antibody. When using Rl as the first antibody, we
use goat antimouse IgG conjugated to fluorescein isothiocyanate
(FITC). The dilution of second antibody required for best results
needs to be empirically determined. We routinely use a 1:20 dilution.
562 Clark

3. Method

The Method describes sorting cells for EGF-receptor expression.

The entire procedure is carried out under sterile conditions.
1. Following marker selection of transfected cells (seeChapter 42, this
vol.), the resultant colonies are removed from the plates and replated
on 150-mm dishes.
2. When the cells become confluent, themedia is removed and the adher-
ent cells are washed once, on the dish, with 10 mL PBS.
3. The cells are removed with trypsin/EDTA (seeNote 1) and washed 3x
in 10 mL Buffer A by centrifuging at 600g for 5 min at 4OC. This and
all subsequent procedures are carried out on ice.
4. On resuspending the cells (seeNote 2) for the third wash, an aliquot of
the suspension is removed for a cell count. The cells from the final
wash are resuspended at lO’/mL in Buffer A. At least lo7 cells are re-
quired for a satisfactory sort.
5. From the sample to be sorted, remove lo6 cells (100 pL) for a negative
control (no first antibody), and to the rest of the sample add EGF-Rl
(1:50) for 30 min at 4OC.
6. Wash cells twice in 10 mL Buffer A.
7. Make up a small volume of second antibody (1:20 FITC-goat anti-
mouse IgG) in Buffer A, and filter through a 0.22~pm membrane. Re-
suspend the cells (sample and negative control) at 107/mL in the
second antibody containing solution. Incubate 30 min at 4OC.
8. Wash cells 3x in 10 mL Buffer B, and resuspend at 107/mL in this
buffer.
9. Once the FACS analysis has been performed on the cells (seeNote 3)
and the difference (if any) between f first antibody observed, a deci-
sion has to be made about what fraction to collect in the sort (seeNote
4). If the cells are being sorted for the first time, both positively stain-
ing and negative cells should be collected. As a rough guide, collect
separately the brightest 5% of cells and the bottom 10%. The cells are
collected into a tube(s) of relevant culture medium.
10. Following sorting, the cells should immediately be plated onto a dish
of suitable size for the number of cells recovered. The culture medium
will have been substantially diluted by the buffer the cells were sorted
in, so after 1 h (or when the cells have reattached), the medium should
be changed.
Fluorescence-Activated Sorting

4. Notes
1. The EGF-receptor in intact cells is not susceptible to digestion by
trypsin; however, other receptor proteins may be. If this is the case, it
is possible to trypsinize the cells the day before the sort and replate
them on bacteriological dishes. Cells do not attach well to this surface
and can be removed simply by washing. It may also be possible to
remove cells by using EDTA alone.
2. It is very important to make sure the cells are in a single-cell suspen-
sion. Cell aggregates cannot be sorted properly and may in fact clog
up the machine. This can be a serious problem with some cells and
may require passing the cell suspenion gently through a needle.
3. A positive control is required by the FACS operators to align the laser
and photomultiplier tubes to get optimum sensitivity from the ma-
chine prior to sorting. In general, use cells expressing moderate to
high levels of the receptor under study. To sort cells expressing EGF-
receptors, we use A431 cells (106cells in 100 pL), processed with Rl
exactly as described.
4. The number of rounds of FACS sorting required to obtain a popula-
tion of cells expressing the transfected receptor varies depending
upon the efficiency of the initial transfection. We keep sorting until
over 50% of the cells are positive. In the case illustrated in Fig. 1, this
may require five sorts, but more usually two or three. To obtain pure
cell lines, the cells are then cloned by limiting dilution.
Do not be disheartened if there is no difference between + 1st
antibody on the initial sort (seeFig. 1 bottom set); collect the top 5% of
cells and expand for resorting. However, if no difference is seen on the
second sort, then it is not worth proceeding further.
5. Although the scope of this article (and author) does not cover a
detailed description of the FACS machine and the actual sort, these
being carried out by a separate cell-sorting laboratory, a brief descrip-
tion of the equipment used to sort the author’s samples is presented
(seealso Chapter 40 of this volume).
All analysis and cell sorts were performed with a modified FACS I
(Becton Dickinson, Mt. View, CA), which operates with an argon laser
used at a wavelength of 488 nM. For the fluorescence signal a 530 nm
filter was used, and for the scatter signal, a neutral density filter (1%).
Ten thousand cells were analyzed at 500-2000 cells/s to determine the
window settings for the sort. The nozzle tip was 100 pm in diameter.
564 Clark

The windows were set from the dot plots of the analysis, with all
dead cells and cell aggregates excluded. Approximately 10’ cells were
sorted over a 90-min period.

Acknowledgments
The author holds a C. J. Martin traveling fellowship from the National
Health and Medical Research Council of Australia.
With many thanks to Dorothy Cheng for maintaining all the cell lines
and assisting with the sorts and to David Capillaro and the FACS “team”
at the Imperial Cancer Research Fund for carrying out the sorts.

References
1. Sklar, L. A. and Finney, D. A. (1982)Analysis of l&and-receptor interactions with the
fluorescence activated cell sorter. Cytometry 3,161-165.
2. Macinnes, D. G., Green, D. K., Harmer, A., Nairn, E. G., and Fink, G. (1983) Neuro-
endocrine receptor -l&and binding using quantitative video-intensification mi-
croscopy and fluorescence-activatedcell sorting. Quart. J. Exp. Physiol. 68,463-474.
3. Due, C., Linnet, K., Langeland Johansen,N., and Olsson, L. (1985) Analysis of insulin
receptorson heterogeneous eukaryotic cell populations with fluorochrome-conju-
gated insulin and fluorescence-activated cell sorter. Diubetologia 28,749-755.
4. Chatelier, R. C., Ashcroft, R. C., Lloyd, C. J.,Nice, E. C., Whitehead, R. H., Sawyer,
W. H., and Burgess,A. W. (1986)Binding of fluoresceinated epidermal growth factor
to A431 cell sub-populations studied using a model-independent analysis of flow
cytometric fluorescence data. EMBO J. 5,1181-1186.
5. Tubiana, N., Mishal, Z., Le Caer, F.,Seigneurin, J. M., Berthiox, Y., Martin, P. M., and
Carcassonne,Y. (1986) Quantification of oestradiol binding at the surface of human
lymphocytes by flow cytometry. Br. J. Cancer54,501~504.
6. Newman, R., Domingo, D., Trotter, J., and Trowbridge, I. (1986) Selection and
properties of a mouse L-cell transformant expressing human transferrin receptor.
Nu ture 304,643-&5.
7. Ellis, L., Clauser, E., Morgan, D. O., Edery, M., Roth, R. A., and Rutter, W. J. (1986)
Replacement of insulin receptor tyrosine residues (1162 and 1163) by site-directed
mutagenesis compromises insulin-stimulated kinase activity and uptake of 2-De-
oxyglucose. Cell 45,721-732.
8. Clark, S, Cheng, D. J.,Hsuan, J.J.,Haley, J.D., and Waterfield, M. D. (1988) Loss of
three major auto phosphorylation sites in the EGF-receptor does not block the
mitogenic action of EGF. J. Cell Physiol. 133,421X25.
Waterfield, M. D., Mayes, E. L. V., Stroobant, l?., Bennett, I’. L. l?., Young, S.,
Goodfellow, P. N., Banting, G. S.,and Ozanne, B. (1982) A monoclonal antibody to
the human epidermal growth factor receptor. 1. Cell Biochem. 20,149-161.
 Chapter 42

Transfection
and Transformation
of Human Thyroid
Epithelial Cells

Nicholas Robert Lemoine

and David Wynford-Thomas
1. Introduction
High efficiency gene transfer into mammalian fibroblasts is a routine
procedure that can be performed by a variety of methods (I), but trans-
fection of epithelial cells has been more difficult to achieve. Each of the
standard techniques has associated problems; special equipment is re-
quired for electroporation (2) and also for direct microinjection (3). Proto-
plast fusion (4) is suitable only for cloned genes. Retroviral transduction
of genes (5) is a method of high efficiency for animal cells, but again is
applicableonly to cloned genes of restricted size, and safety considerations
are likely to restrict greatly the use of retroviral vectors suitable for human
cells. Calcium phosphate coprecipitation has been used with success in
several epithelial lines (6,7,8) and, more recently, strontium phosphate
coprecipitation (9). These techniques are very simple to perform, require
no special equipment, and can be applied to genomic DNA transfection.

565
566 Lemoine and Wynford-Thomas

The human thyroid follicular cell appears to be a particularly suitable

epithelial cell for gene transfection by coprecipitation, in which insoluble
complexes of DNA with calcium or strontium phosphate are taken up by
recipient cells, because phagocytosis can be specifically increased by
thryoid-stimulating hormone (TSH). We have recently shown that this
maneuver enhances the efficiency of gene transfer, with a fivefold increase
in the frequency of cells transiently expressing SV40 large T antigen as
assessed by immunocytochemical assay 48 h after transfection (10). In
addition, these cells appear remarkably resistant to the toxic effects of
calcium phosphate, with plating efficiencies of 295% even after 16 h
exposure to coprecipitate.

2. Materials
1. Plasmid DNA: closed circular plasmid DNA (prepared by cleared
lysate method with banding on a CsCl, gradient) (II) is dissolved in
1.0 mM Tris-HCl, 0.1 mM EDTA pH 8.0 at a concentration of 0.1 pg/
NJ*
2. Carrier DNA: high mol wt (seeNote 1) carrier DNA (prepared by a
method of caesium chloride banding) (12) is dissolved in 1.OmM Tris-
HCl, 0.1 mM EDTA pH 8.0 at a concentration of 0.5 pg/pL.
3. 25 rnM Hepes: dilute 2.5 mL of sterile 1M Hepes to 100 mL with sterile
water. Store at 4*C for up to 3 mo.
4. 2x Hepes-buffered saline (2 x HBS): dissolve 1.64 g NaCl plus 0.023 g
Na,HPO,*2H,O in 90 mL of water. Add 5 mL of 1M Hepes, then: (a)
for CaCl, coprecipitation, adjust pH to 7.1 with 0.5M NaOH. (b) for
SrCl,coprecipitation, adjust pH to 7.8 with 0.5MNaOH. Adjust vol to
100 mL with water (seeNote 2). Filter sterilize (seeNote 3), and store
at 4*C for up to 3 mo.
5. 2.5M CaCl,: 10.8 g CaCl,~G~O in 15 mL of water. Adjust vol to 20 mL
with water. Filter sterilize and store in aliquots at -2OOC.
6. 2M SrCl,: 10.665 g Sr Cl, in 15 mL of water. Adjust vol to 20 mL with
water. Filter sterilize and store in aliquots at -2OOC.
7. 1.OmM Tris-HCl, 0.1 mMEDTA, pH 8.0: 100 PL l.OM Tris-HCI pH 8.0
plus 20 I.~L0.5M EDTA, pH 8.0 in 100 mL water. Filter sterilize and
store at 4*C.
8. SF-12 medium: This must contain the appropriate concentration of
sodium bicarbonate for the conditions of incubation: for a pH of 7.4
at 37°C in 5% CO, add 2.9 mL of 7.5% NaHCO, solution/100 mL of
medium (seeNote 4).
Human Thyroid Epithelial Cells

9. RPM1 medium: prepared with sodium bicarbonate concentrations as

above.
10. 11.8% (v/v) Glycerol: 11.8 mL Glycerol, diluted to 100 mL in Hanks’
balanced salt solution, (HBSS). Filter sterilize and store at 4°C.
11. TSH: thytropar (Armour Pharmaceuticals, Tarrytown, NY) dissolved
in HBSS at 1 U/mL as stock solution, filter sterilized and stored in 100
PL aliquots at -2OOC.
12. 2MNH,Cl: 1.06 g of NH&l in 8 mL of water. Adjust pH to 7.1 and ad-
just vol to 10 mL with water. Filter sterilize and store at 4OCfor up to
3 mo.
3. Methods
1. Thaw human thyroid epithelial cells (seeChapter 13, this vol.) from
frozen stock, wash once with RPM1 medium and plate out as a mono-
layer at 5 x lo5 cells /60-mm dish in RPM1 medium supplemented with
10% fetal calf serum (FCS).
2. Transfection is performed 4 d after plating (seeNote 5). Six hours
before transfection, replace the medium with 5 mL of warm SF-12 sup-
plemented with 10% FCS (seeNote 6).
3.1. Cakium Phosphate Coprecipitation
1. Prepare DNA for transfection by mixing x PL of plasmid DNA (con-
taining 0.5-10 pg/DNA) solution with (200 -x> PL of 25 mM Hepes in
a sterile plastic bijou. If carrier DNA is used, then this is included to
bring the final concentration of DNA to 10 pg/O.5 mL of coprecipita-
tion suspension. Add 50 PL of CaCl, and swirl to mix.
2. Add this solution dropwise through a syringe and needle to 250 PL of
2 x HBS in another sterile plastic bijou with continuous mixing by a
stream of air bubbles via a plugged sterile pipet with plastic tip (see
Note 7).
3. Leave the mixture to stand for 30 min, after which time the fully
formed coprecipitate will have settled to the bottom of the container.
Coprecipitates containing high MW carrier DNA have a coarser con-
sistency than those containing plasmid DNA only.
4. Five minutes before application of coprecipitate, add 50 PL of 1 U/mL
TSH solution to the medium and swirl gently to mix (seeNote 8).
5. Gently take up the suspension in a 1-mL Gilson pipet a few times, so
that the coprecipitate is uniformly suspended through the solution.
6. Add 100 PL of 2M NH&l solution to the medium and swirl gently to
mix (seeNote 8).
 Lemoine and Wynford-Thomas

7. Sprinkle the suspension over the medium of the epithelial monolayer.

8. Incubate at 37OC for 16 h to allow uptake of the coprecipitate.
9. Remove the medium containing coprecipitate and wash the cells
twice with warm HBSS. Refeed with warm RPM1 + 10% FCS.
3.2. Strontium Phosphate Coprecipitation
1. Prepare DNAfor transfection by consecutively mixingx PLof plasmid
DNA (containing 0.5-10 pg DNA), 30 PL of 2M SrCl,, and (220-x) PL
of sterile water in a sterile plastic bijou. Seeabove for details of carrier
DNA.
2. Add this solution dropwise through a syringe and 23-gage needle to
250 PL of 2 x HBS in another sterile plastic bijou with continuous
mixing by a stream of air bubbles (via a plugged sterile pipet with
plastic tip).
3. Leave the mixture to stand undisturbed at room temperature and
observe the development of the coprecipitate. After 5-10 min, the
precipitate is seen as a fine dust and is then ready for application to the
cells.
4. Gently take up the suspension in a Gilson pipet a few times, so that the
coprecipitate is uniformly suspended throughout the solution.
5. Sprinkle the suspension over the medium of the epithelial monolayer
in a 60-mm dish.
6. Incubate at 37OC for 90 min to allow adsorption of the coprecipitate.
7. Remove the medium containing coprecipitate, and wash the cells
twice with warmed HBSS. Apply 1.5 mL of 15% glycerol in HBSS and
incubate at room temperature for 30 s, then wash three times with
warm HBSS, and refeed with warm RPM1 + 10% FCS.

3.3. Transient Expression Assay

The transient expression of genes that have been successfully trans-
fected is maximal at 48-72 h after transfection in these cells. A convenient
assay system that allows calculation of transfection frequency is im-
munocytochemical demonstration of SV40 large T expression (Fig. 1, see
Chapter 44, this vol. for alternative transient assay systems) following the
transfection of a plasmid such as pSVorP(13).

3.4. Stable Expression Assay

The transfection of a dominant selective marker such as the neo gene
that codes for resistance to geneticin in eukaryotic cells allows assay of
stable transfection frequency. Primary human thyroid epithelial cells have
Human Thyroid Epithelial Cells 569

Fig. 1. Pair of recently divided thyroid epithelial cells showing stong nuclear immu-
noreactivity with antiSV40 large T antibody pAB 419,48 h after transfection with a
plasmid containing the whole SV40 sequence. (1mmunocytochemistry was performed on
acetone-fixed monolayers using pAB 419 followed by rabbit antimouse immunoglobulin
conjugated to horseradishperoxidase. Positivityisshownbydepositionof thebrownper-
oxidasecatalyzed polymer of diaminobenzidine).

poor cloning efficiency and require the use of a feeder layer (of geneticin-
resistant fibroblasts) in order to develop viable clones from the low density
required for geneticin selection.
Transfection of plasmids containing SV40 leads to the outgrowth of
clones with extended life span that continue to proliferate after the un-
transformed cells undergo senescence (after an average of 3-6 doublings).
These clones can be shown to express SV40 large T by the immuno-
cytochemical assay mentioned above.

4. Notes
1. DNA prepared by methods that involve phenol/chloroform extrac-
tion of proteinase K digests for example will be of lower mol wt, and
this will reduce the efficiency of transfection.
570 hmoine and Wynford-Thomas

Fig. 2. Relationship between time of addition of TSH (10 mU/mL) to monolayer cul-
ture and efficiency of transfection (as assessed by proportion of cells showing transient
expression of SV40 T antigen 48 h after transfection). q lno TSH (addition of NH&l had no
effect on this result). 4 TSH (10 mU/mL) added at indicated times. A TSH (10 mU/mL)
added at indicated times plus NH&l (20 m&f) added at time 0. All points show mean +/
- SE derived from 6 separate experiments.

2. The effect of pH is critical, and it is recommended that pH of stock so-

lutions should be checked at intervals after preparation, and if incor-
rect, new reagents should be prepared.
3. 0.22 pm filters should be flushed with sterile water before use. This has
been observed to reduce subsequent aggregation of the DNA-cation
coprecipitates.
4. This medium has been found to be ideal for the culture of cells during
transfection by coprecipitation. Some other media (including RPMI)
are unsuitable because the product of divalent cation -phosphate con-
centrations is too high, causing excessive coprecipitation.
5. The growth rate of these cells (as determined by thymidine labeling)
shows a peak on d 4 after stimulation with serum. The highest
efficiency of transfection is at this time (10).
6. Media and wash solutions should be prewarmed to enhance growth
and transfectability of cells.
7. Even mixing of the solutions is best achieved by a gentle stream of
bubbles, and the pipet should have a plastic tip since the precipitate
adheres to glassware.
Human Thyroid Epithelial Cells 571

8. The addition of TSH to stimulate phagocytosis significantly improves

the efficiency of transfection by calcium phosphate coprecipitation
when given within a critical time-window (seeFig. 2): no effect is seen
if it is given more than 20 min before (or at any time after) addition of
the coprecipitate. Lysosomal function inhibitors, such as ammonium
chloride, have no influence when used alone, but have a synergistic
effect when used in concert with TSH in this system. Interestingly, no
effect of TSH has been demonstrable on the efficiency of strontium
phosphate coprecipitation.

Acknowledgments
We are grateful to the Cancer Research Campaign and to the Welsh
Scheme for the development of Health and Social Research for grant
support.

References
1. Pollard, J. W., Luqmani, Y., Bateson,A., and Chotai, K. (1984) DNA transformation
of mammalian cells, in Methods in Molecular Biology, vol. 2 (Walker, J. M., ed.),
Humana Fress, NJ, pp. 321-332.
2. Tur-I&spa, R.,Teicher, L., Levine, B.J.,Skoultchi, A. I., and Shafritz, D. A. (1986) Use
of electroporation to introduce biologically active genes into primary rat hepato-
cytes. Mol. Cell Biol. 6,716-718.
3. Garcia, I., Sordat, B., Rauccio-Farinon, E., Dunand, M., Kraehenbuhl, J. -P., and
Diggelmann, H. (1986) Establishment of two rabbit mammary epithelial cell lines
with distinct oncogenic potential and differentiated phenotype after microinjection
of transforming genes. Mol. Cell Biol. 6,1974-1982.
4. Yoakum, G. H., Lechner, J. F., Gabrielson, E. W., Korba, B. E., Malan-Shibley, L.,
Willey, J. C.,Valerio, M. G.,Shamsuddin, A. M., Trump, B.F., and Harris, C. C. (1985)
TransformationofhumanbronchialepithialcellstransfectedbyHarveyrasoncogene.
Science 22, 1174-1179.
5. Wolff, J. A.,Yee J. -K., Skelly, H. F., Moores, J. C., Respess,J. G., Friedmann, T., and
Leffert, H. (1987)Expression of retrovirally transduced genes in primary cultures of
adult rat hepatocytes. Proc. N&l. Acad. Sci. USA 84,3344-3348.
6. Hynes,N. E.,Jaggi,R., Kozma, S.C., Ball, R., Muellener, D., Wetherall, N. T., bavis,
B. W., and Groner, B. (1985) New acceptor cell for transfected genomic DNA:
oncogene transfer into a mouse mammary epithelial cell line. Mol. Cell Biol. 5,
268-272.
7. Storer, R. D., Stein, R. B., Sina, J. F., DeLuca, J. G., Allen, H. L., and Bradley, M. 0.
(1986) Malignant transformation of a preneoplastic hamster epidermal cell line by
the EJ c-Ha-ras oncogene. Cancer Res. 46,1458-1464.
8. Summerhayes, I. C., Malone, P., and Visvanathan, K. (1986) Altered growth prop-
erties and cell surface changes in ras transformed mouse bladder epithelium. Int. J.
Cancer37,233-240.
9. Brash, D. E., Reddel, R. R., Quanrud, M., Yang, K., Farrell, M. P., and Harris, C. C.
572 Lemoine and Wynford-Thomas

(1987) Strontium phosphate transfection of human cells in primary culture: stable

expression of the simian virus 40 large-T-antigen gene in primary human bronchial
epithelial cells. Mol. Cell Bid. 7,2031-2034.
10. Lemoine, N. R. and Wynford-Thomas, D. (1987) The thyroid follicular cell as a
recepient for DNA transfection. Br. J. Cancer 55,342.
11. Boffey, S. A. (1984) Plasmid DNA isolation by the cleared lysate method, inMethods
in Molecular Biology, vol. 2 (Walker, J. M., ed.), Humana Press, pp. 177-183.
12. Kaiser, K. and Murray, N. E. (1985) The use of phage lambda replacement vectors in
the construction of representative genomic DNA libraries, in DNA Cloning, vol. 1
(Glover, D. M., ed.), IRL, pp. l-74.
13. Gluzman, Y., Sambrook, J. F., and Frisque, R. J. (1980) Expression of early genes of
origin-defective mutants of SV40. Pm. Nufl. Ad. Sci. USA 77,3898-3902.
 Chapter 43

Expression of Foreign Genes

in Cultured Insect Cells
Using a Recombinant
Baculovirus Vector

Martin J. Page
and Vivienne F. Murphy
1. Introduction
An important consideration for the expression of cloned genes in
recombinant expression systems is the ability of the foreign host to produce
the protein faithfully in a form that is similar or identical to that found in
the cell type from which the gene was cloned. For eukaryotic proteins, this
frequently involves many posttranslational modifications of the protein,
such as glycosylation, phosphorylation, processing, and secretory events.
Additionally, very precise interactions are essential for the correct folding
of the polypeptide to achieve the final tertiary structure. If the folding is
incorrect, then the molecule will often be biologically inactive.

573
 Page and Murphy

These considerations have led to the increasing use of recombinant

eukaryotic expression systems to express cloned genes accurately. In par-
ticular, recombinant mammalian expression systems have been used ex-
tensively to achieve this aim, with considerable success. More recently, an
alternative higher eukaryotic expression system involving the use of a re-
combinant baculovirus and insect tissue culture cells has demonstrated
considerable advantages. These include the speed of obtaining expres-
sion of the recombinant product, the potential high yield, and the option of
expressing proteins that would otherwise be toxic at high levels in mam-
malian cells (e.g., c-myc, seeref. 1). Examples of recombinant proteins pro-
duced using the insect expression system have so far shown them to be
appropriately modified, processed, secreted, and correctly folded to give
high yields of biologically active proteins, such as human p interferon (2),
P-galactosidase (3), c-myc (4), interleukin-2 (5), and influenza hemag-
glutinin (6,7).
The insect baculovirus used for these experiments is called Auto-
gvapha Ca2ifornica Nuclear Polyhedrosis Virus (AcNPV), and it has a 128
kdaltons double-stranded circular DNA genome, which replicates within
the nuclei of lepidopteran species with a biphasic life cycle. The first phase
(W-24 h postinfection) involves the formation of mature nucleocapsids
that bud through the cellular membrane to produce extracellular virus
(EV). The second phase (18-24 h postinfection) involves high-level expres-
sion of the AcNPV late gene called polyhedrin. This protein can account
for up to 50% of the protein mass of an infected insect cell and is responsible
for embedding mature virus particles within the nuclei, generating ex-
tremely large viral occlusion bodies (seeFig. 1). These occluded viruses
(OV) are released only after cell death and are essential for lateral transmis-
sion of the virus.
Several features of this biphasic life cycle make the insect baculovirus
system amenable for geneticmanipulation. First, the polyhedrin gene pro-
duct is totally unnecessary for the EV form of the life cycle, which is the only
infectious form in insect cell cultures. Therefore, the polyhedrin gene can
be replaced with a foreign gene, so that it is suitably positioned under con-
trol of the powerful polyhedrin promoter. Second, whereas wild-type
AcNPV will make polyhedrin within infected cells and give rise to viral
plaques easily identified visually as occlusion positive, the recombinant
AcNPV lacks the polyhedrin gene and gives rise to occlusion negative viral
plaques. This forms the basis for rapidly screening and identifying rare re-
combinant viruses that contain and likely express a cloned foreign gene.
Foreign Gene Expression in Insect Cells 575

Fig. 1. The contrast between uninfected Sf insect cells (top) and cells 48 h postinfection
with wild-type AcNPV (bottom). The polyhedrin occlusion bodies are easily visible
within the nuclei of the infected cells and give rise to occlusion-positive plaques that have
a characteristic silver-gray sheen when examined against a light source.

This chapter is aimed at introducing this technology to researchers

who may have little or no previous experience of the insect baculovirus
expression system. Throughout, it is assumed that some knowledge of
basic recombinant DNA and tissue culture techniques are known, al-
though important aspects of each will be emphasized.
576 Page and Murphy

2. Materials
1. The insect cell line Spo&rWu frugiperda Clone 9 (S. f. 9) may be ob-
tained from the American Type Culture Collection (12301 Parklawn
Drive, Rockville, MD 20852, USA) or alternatively from any research
group working in this area.
2. Insect cell stocks are maintained at room temperature in modified
BML TC-10 medium (8) supplemented with 10% fetal calf serum and
antibiotics as log-phase monolayer cultures in glass tissue culture
flasks of approximately 150 cm* surface area. The modifications to the
culture medium are as follows. The concentrations of L-arginine, L-
histidine, and calcium pantothenate are 550 mg/L, 3.4 g/L, and 0.11
mg/L, respectively. Choline chloride is replaced with cyanocobalbu-
men at 0.01 mg/L.
3. lx TE buffer: Prepare a solution containing 10 mM Tris HCI, pH 7.4,
1 mM EDTA, autoclave, and store at room temperature.
4. Transfection buffer: Prepare a solution containing 25 mM Hepes, pH
7.1,150 mMNaC1,125 mM CaCl, autoclave, and store at 4°C. The pH
of this solution is critical and should be 7.1+ 0.05.
5. Neutral red stock solution is obtained from BDH.
6. Low melting temperature ultrapure Seaplaque agarose.

3. Methods
To obtain successful expression of cloned genes using the insect ex-
pression system, it is critical that the principles underlying each exper-
imental step are clearly understood. For this reason, each of the steps
necessary to engineer the baculovirus DNA are shown schematically in
Fig. 2, and described in detail below.

3.1. Cloning of Foreign Genes into an AcNPV

mansfer Vector
To obtain expression of recombinant proteins using this system, it is
necessary to position the foreign gene under control of the strong poly-
hedrin promoter. Owing to the very large size of the AcNPV genome
(about 128 kdaltons), and lack of a suitable restriction site at the desired
position, this cannot be achieved by a direct cloning route. Instead, the for-
eign gene is first subcloned into an AcNPV transfer vector. This consists
Foreign Gene Expression in Insect Cells 577

TRANSFER FOREIGN BACULOVIRUS

VECTOR - GENE

RECOMBINANT
TRANSFER
VECTOR POLYHEDRIN
GENE *
IRAL
NA

HOMOLOGOUS
RECOMBINATION
IN WV0 -INSECT CELL

r 1
\ /
t t + t
PROGENY VIRUS
RELEASED FROM
INFECTED CELL
0 l-l% CONTAIN
FOREIGN GENE
Co

c -LIQUID OVERLAY
PETRI DISH - - - - - - - - - - - - - - -i, AGAROSE OVERLAY
CELL MONOLAYER (INFECTED
27% 3 days WITH PROGENY VIRUS)
NEUTRAL RED STAIN

Fig. 2. Generation of a recombinant baculovirus for expression of a foreign gene. (1)

Construction of the recombinant transfer vector. (2) Extraction of wild-type baculovirus
DNA. (3) Cotransfectionof recombinant transfer vector and baculovirus DNA. (4) Plaque
assay for selection of recombinant baculoviruses. Steps l-4 are detailed in Methods sec-
tions 3.1,3.4., respectively.
578 Page and Murphy

0 / Hind III

3.96 EcoR+
\ Kpn I 4.43
4.0 BamH I’ I
-8 / +177 Hind III 4.1

Fig. 3. The baculovirus transfer vector pAc373. The map units shown are in kilobase
pairs relative to the Hind III site at map unit 0. The insertion site for cloning foreign genes
is at the BamHI site, which is map unit 4.0, and this represents fusion of coordinates-8 and
+177 relative to the original polyhedrin ATG translation codon via a BamHI linker. The
remainder of the polyhedrin gene is shown as a hatched box, whereas flanking bacu-
lovirus DNA is shown as stippled boxes.

of a cloned region of the AcNPV genome encompassing the polyhedrin

gene, in which a unique restriction site has been engineered immediately
upstream of the polyhedrin coding body. A suitable and widely used
AcNPV transfer is pAc373 (51,which has a unique BamHl restriction site
engineered at position -8 relative to the polyhedrin ATG translation ini-
tiation codon (seeFig. 3). The following steps should be performed using
standard cloning techniques (9, and vol. 2 and 4 of this series) to insert the
foreign gene into pAc373:
1. Digest pAc373 with the restriction enzyme BamHl.
2. Phosphatase the protruding 5’ overhang with calf intestinal phos-
phatase. This step is necessary to reduce the high frequency of vector
self-ligation that would otherwise occur.
3. The foreign gene to be cloned must be prepared or manipulated so that
it has homologous cloning ends for the phosphatased pAc373 BamHl
Foreign Gene Expression in Insect Cells 579

digested transfer vector. Ideally, the gene should have BamHl, BglII,
or BclI ends to facilitate “sticky end” cloning. If these sites are not pres-
ent, then there are two options. First, the ends could be altered by
blunting and cloning suitable linkers onto the ends (making sure that
there are no internal sites within the gene), or secondly, the ends could
be left as blunted ends and cloned into a pAc373/BamHl digested
transfer vector that has also had the BamHl ends blunted. Ligate the
foreign gene into the phosphatased pAc373 transfer vector using T4
DNA ligase.
4. Transform competent E. cdi to ampicillin resistance using an aliquot
of the ligation mix.
5. Prepare mini-prep plasmid DNAs from picked colonies and identify
by restriction enzyme analysis recombinants that have the foreign
gene inserted in the correct orientation.
6. Prepare full-scale DNA preparations from an identified positive clone
and band through a cesium chloride density gradient.
7. Perform a series of restriction enzyme digests to confirm that the re-
combinant transfer vector is correct. The foreign gene is now correctly
inserted in the proper orientation immediately downstream of the
polyhedrin promoter within a recombinant AcNPV transfer vector.

3.2. Preparation of High Molecular Weight Infec-

tious DNA fkom Wild-5Qpe AcNPV
The foreign gene contained within the AcNPV recombinant transfer
vector now has to be introduced precisely into the AcNPV genome by hom-
ologous recombination in vivo. This procedure requires the use of good-
quality, intact AcNPV DNA, which should be capable of initiating aninfec-
tion when transfected into insect cells.
1. Seed a 100-mL spinner flask with insect cells at 1 x 106/mL and incu-
bate overnight at 27OC.
2. Centrifuge at 2,500 rpm for 10 min to pellet the cells and debris using
a bench centrifuge. Discard the cell pellet.
3. Pellet the extracellular virus (EV) from the supernatant at 100,000 x g
in a swing-out rotor (e.g., 27,500 rpm in a Beckman SW40 rotor) for 30
min at room temperature.
4. Resuspend the viral pellet in 10 mL of 0.1 x TE, and pellet the EV as
before.
5. Resuspend the viral pellet in 4.5 mL of 1 x TE containing 0.2M NaCl,
580 Page and Murphy

and then add 200 pg of proteinase K. Incubate at 50°C for 1-2 h.

6. Add 0.5 mL of 10% sarkosyl, and incubate for a further 2 h or
overnight.
7. Add an equal vol of 1:l phenohchloroform, and mix the phases very
gently, to prevent shearing of the DNA, for five min.
8. Repeat step 7, and then reextract in the same way with chloroform
alone. Transfer upper aqueous phase to a clean tube. Centrifuge at 100
rpm for 5 min; then transfer upper aqueous phase to a clean tube us-
ing a wide bore pipet.
9. Add 10 mL of cold (-20°C) ethanol and mix gently. A cotton-wool like
precipitate should form at this stage. Centrifuge at 1000 rpm for five
min to pellet the precipitate, discard the supernatant, and wash the
pellet in 5 mL 80% ethanol.
10. Vacuum dry the DNA pellet, and dissolve it in 0.5 mL of 0.1 x TE. The
DNA is now ready for use. The yield should be around 200-500 pg.
Store at 4OC. The quality of the purified AcNPV DNA can be moni-
tored by its ability to initiate an infection after transfection into insect
cells (seebe2ow).

3.3. Cotransfection of Wild-Type AcNPV DNA

and Recombinant AcNPV l’kansfer Vector
into Insect Cells
This step is necessary to introduce into cultured insect cells both DNA
molecules consisting of the wild-type AcNPV DNA and the smaller recom-
binant transfer vector containing the foreign gene to be expressed. Once
inside the cell, a small proportion of the molecules (approximately 1 in
100-1000) will undergo double-reciprocal homologous recombination in
which the polyhedrin gene of the wild-type AcNl?V DNA is replaced by the
foreign gene.
1. Seed 2 x lo6 insect cells into a 25 cm2 tissue culture flask, and leave at
27OC for 1 h to attach.
2. Into a sterile 1.5 mL Sarstedt tube, add 1 pg of high mol wt wild-type
AcNPV DNA and 2 pg of recombinant transfer vector. Add 0.75 mL
of transfection buffer, and mix by pipeting up and down.
3. Remove all of the medium from the flask of insect cells, and replace
with exactly 0.75 mL fresh medium. Be sure to cover the entire
monolayer with a film of medium. Slowly add the 0.75 mL of DNA
solution directly onto the cell monolayer dropwise using a I-mL
Foreign Gene Expression in Insect Cells

pipet. The DNA will form a calcium phosphate DNA precipitate in

situ because of the phosphates in the medium.
4. Leave at 27OCfor 4 h, then remove the transfection solution and care-
fully wash the cell monolayer twice with 5 mL of culture medium.
Finally, add 5 mL fresh medium and incubate (with cap tightly closed)
in a 27°C incubator for 3 d.
5. At the end of this period, a significant proportion (usually about 10%)
of the cells should be showing positive signs of viral infection as evi-
denced by large granular occlusions of polyhedrin in the nuclei of in-
fected cells. By this stage, the culture medium will contain virus parti-
cles at a concentration of about 2 x 10’ pfu/mL. Only 0.1-l% of these
will consist of recombinant viruses, and these have to be identified by
visual inspection following a plaque assay.

3.4. Plaque Assay

Recombinant viruses present in the medium harvested following
cotransfection can be identified and isolated by plaque assay. This meth-
od can also be used to determine the titer of an unknown virus stock
solution. The principle of the method is to obtain well-isolated viral
plaques in a confluent cell monolayer, which can then be screened for
recombinants or simply counted for determination of virus titer.
1. Prepare a series of lo-fold serial dilutions in culture medium of the
virus stock. An appropriate range for culture supernatant harvested
from cotransfections is 1 in 100-l in 10,000, and for determination of
virus titer, 1 in 100-l in l,OOO,OOO.
2. Seed 35-mm plastic Petri dishes with 2 x lo6 insect cells, and incubate
at 27OCfor 1 h to allow the cells to attach. Seed enough for 5-10 plates/
dilution for screening recombinant plaques and two plates/dilution
for determination of virus titer.
3. Remove most of the culture medium from the dishes, leaving a small
amount to prevent drying out of the monolayer. Gently drip 100 PL
of the diluted virus onto the cells in the center of the dish, removing
any bubbles that form. Incubate for 1 h at room temperature to allow
adsorption of the virus.
4. Prepare a 1% agarose/culture medium overlay solution as follows:
Autoclave an appropriate volume of 3% Seaplaque agarose solutionin
distilled water (0.5 ml/plate) and cool to 37OC. Warm twice this vol-
ume of culturemedium to 37°C. Mix the agarose and the medium, and
maintain at 37OC.
582 Page and Murphy

5. Remove the inoculum from the Petri dishes using a sterile Pasteur
pipet. Add 1.5 mL of the overlay mix to the center of each plate, allow-
ing it to spread evenly over the monolayer. Allow it to set for 20 min
at room temperature, and then overlay with 1 mL of culture medium.
Incubate undisturbed in a humid environment at 27°C for 3-4 d.
Movement of the plates during incubation can cause smearing of the
plaques.
6. Add 1 mL of a neutral red stock solution, diluted 1 in 10 in PBS, to the
liquid overlay. Incubate at 27OC for 1-2 h. Tip off the liquid overlay
containing stain, invert the plates, and leave for several hours or over-
night at 4°C for the plaques to clear. Plaques appear as circular regions
of weak staining about 1-3 mm in diameter in a darker stained
monolayer, and may show retarded cell growth. Cells in the center of
wild-type plaques will contain nuclear occlusion bodies (seeFig. 1)
clearly visible under low-power magnification. The rare recombin-
ants generated following cotransfection will lack occlusionbodies and
can be selected visually by this phenotype. Each plaque represents
one virus or pfu in the inoculum; therefore, the titer of the original
stock can be determined allowing for the dilutions made.
7. To screen for recombinants, select plates with well-isolated plaques at
a density of 20-200 plaques/plate. With practice, it is possible to pick
out occlusion-negative plaques by eye, and then confirm these by
microscopic analysis. Wild-type plaques containing occlusion bod-
ies have a characteristic refractile appearance, giving a slight silvery
sheen when viewed at an angle against a light source. Plaques lack-
ing this sheen can be marked for closer inspection under 400x magnifi-
cation with an inverted microscope. Check every cell in the plaque
carefully for occlusion bodies, and if none are seen, mark the plaque
for plaque purification.

3.5. Plaque Purification

Occlusion-negative plaques may still contain some wild-type virus
contamination because of diffusion from neighboring plaques. Thus, to
isolate a pure viral stock, recombinants are purified by successive rounds
of plaque purification.
1. Using a sterile Pasteur pipet, pick a plug of agarose from directly over
the potential recombinant occlusion-negative plaque into 0.5 mL of
culture medium. Vortex briefly, and leave at room temperature for 30
Foreign Gene Expression in Insect Cells 583

min or longer to allow the virus to diffuse from the plug. Keep an ali-
quot at 4°C for long-term storage.
2. Carry out a plaque assay as in section 3.4., using the plaque suspen-
sion and lo-fold dilutions of it down to 1 in 1000.
3. Stain and screen as described in the previous section.
4. Repeat until the plaques generated are 100% occlusion negative. Gen-
erally, three rounds of plaque purification are carried out to ensure
elimination of wild-type virus, but the presence of the foreign gene can
be confirmed by Southern Blotting or product analysis at an earlier
stage if required.

3.6. Derivation of High-Titer Medium

and Determination of pj%lmL
Once an occlusion-negative plaque has been successfully identified
and plaque purified, it is necessary to obtain culture medium containing
high titers of the recombinant virus for further infections. This is achieved
by a series of “scale-up” infections and the resulting high-titer culture me-
dium is then titerd by plaque dilution assays to determine the pfu/mL of
infectious recombinant virus.
1, The isolated pure recombinant AcNPV plaque is picked and transfer-
red into a sterile Bijoux bottle (5 mL size) containing 1 mL of culture
medium, and left overnight at room temperature to allow diffusion of
the virus particles into the medium.
2. The next day, 0.5 x lo6 insect cells are seeded onto a 35-mm tissue
culture dish and left for 1 h at 27OCto attach. The medium is removed
and replaced with 0.8 mL of the culture medium containing the virus
plaque. The cells are left for a further hour at 27OCto allow infection
to occur. Then the medium is removed and replaced with 2 mL fresh
medium. The cells are left at 27OCfor 4 d. The remaining 0.2 mL of
culture medium containing the original plaque is stored at 4OCas a re-
serve in case of contamination at this stage.
3. After 4 d, the medium is removed and spun at 1500 rpm in abench-top
centrifuge for 5 min and the culture medium transferred to a labeled
sterile tube. Of this, 1.8 mL is used to infect a T-25 flask seeded with
5 x lo6 insect cells, as described above, then replaced with 5 mL of cul-
ture medium, and left for a further 3 d at 27°C.
4. The medium is then removed and spun as described above. Of this,
4.8 mL is used to infect a T150 flask seeded with 30 x lo6 insect cells,
 Page and Murphy

as described above, then replaced with 50 mL culture medium and left

for a further 3 d at 27°C.
5. The medium is removed and spun as described above. At this stage,
one has 50 mL of culture medium containing virus at a titer in the
range of 1-4 x lo8 pfu/mL. Of this, 1 mL is frozen at -8OOC for long-
term storage, the remainder is kept at 4OCand should retain its high
infectivity for at least 1 yr.
6. It is usual to determine the titer of the virus in the culture medium at
this stage, and this is done by a dilution plaque assay. Assuming that
the titer is about lo8 pfu/mL, carry out the necessary dilutions to give
10-100 plaques/35-mm plate using a standard plaque assay as de-
tailed in section 3.4.

3.7. Isolation of Recombinant AcNPV DNA

for Characterization by Southern Blotting
It is good practice to verify that the recombinant virus, identified by
virtue of producing occlusion-negative plaques, does indeed contain the
inserted foreign gene, in the correct orientation, and without rearrange-
ment or deletion. Fortunately, it is not necessary to isolate pure viral DNA
as described in section 3.2. for this purpose. Instead, since about 25% of the
total nucleic acid from cells late in infection consists of viral DNA, this can
be used to isolate a total DNA preparation, which is very suitable for re-
striction enzyme and Southern Blot analysis.
1. Infect 30 x lo6 cells in a T-150 flask at a multiplicity of infection of 5 pfu/
cell, and leave for 3 d.
2. Smack the flask hard a few times to dislodge the insect cells, spin at
1000 rpm for 5 min, and then wash the cell pellet twice with 25 mL of
PBS. Do not vortex or pipet the cells vigorously; otherwise, consider-
able cell lysis will occur. Finally, allow the cell pellet to drain for a few
minutes.
3. Resuspend the cell pellet in 5 mL of TE buffer. Then add 0.1 mL of pro-
teinase K solution (1 mg/mL in TE), mix quickly, rapidly add 0.25 mL
of 10% SDS, and mix rapidly by inverting the tube several times. Leave
the mixture at 37OC for 4 h.
4. Add 5 mL of phenol:chloroform (l:l), mix by inverting and rolling the
tube (do not vortex), spin in a 30-mL glass sterile Corex tube at 10,000
rpm in a Sorvall centrifuge for 10 min, then remove the lower organic
layer with a pipet, and discard. Repeat this procedure once more with
phenol:chloroform and then once with chloroform alone.
Foreign Gene Expression in Insect Cells

5. Remove the top aqueous layer this time, into a new labeled sterile Cor-
ex tube. Add 0.15 mL of 5M NaCl and 12.5 mL of ethanol, and swill
or invert to mix. A fluffy white precipitate of DNA should form at this
stage. This is spun, washed in 70% ethanol, and then air dried.
6. The pellet will contain RNA and DNA at this stage. Restriction en-
zyme digests can be performed on this material, but must include
RNase A (preheated at 70°C for 10 min to inactivate any DNases pre-
sent) at a final concentration of 10 pg/digest. Alternatively, the total
nucleic acid preparation can be dissolved in TE buffer, incubated with
RNase A (50 pg/mL) for 3 h at 37OC,then deproteinized again by add-
ing NaCl to O.l5M, SDS to 0.1% and proteinase K to 20 pg/mL, and
leaving for a further 2 h at 37°C. The solution is then extracted once
with phenol:chloroform and once with chloroform before precipitat-
ing with ethanol. The DNAobtained after this step will be “purer” and
more likely to digest completely with restriction enzymes. The final
DNA pellet is dissolved in 1 mL sterile distilled water and quantitated
by UV absorbance at 260 nM.
7. 2 t.tg aliquots of DNA are digested with restriction enzymes, electro-
phoresed, and blotted using standard procedures. The blot is hybrid-
ized with a radioactive probe consisting of the cloned foreign gene.
The signals obtained are usually very strong because of the high pro-
portion of pure viral DNA within the total DNA preparation. Control
digests using the cloned foreign DNA in the original recombinant
transfer vector should be included to identify correct co-migrating
bands.

4. Notes
1. Subculturing of the insect cell monolayer stocks is required at approx-
imately 1-wk intervals, while cells are still in log phase growth (i.e.,
subconfluent). Cells are detached from the flask by sharply smacking
the flask several times or by scraping the cells off with a sterile rubber
policeman. Trypsin/versene should not be used. Viability counts
should be done at this stage using Trypan blue stain. The resuspended
cells are normally subcultured 1/lO for weekly maintenance.
2. Spinner cultures can be used for virus preparations or for large scale
production of recombinant protein. These shouldbe grown at 27-28”C,
which is the permissive temperature for virus replication. Standard
glass spinner flasks are seeded at 5 x 104-1 x 106 cells/mL, stirred at
50-100 rpm. Aeration is not required for volumes up to 500 mL, but
586 Page and Murphy

for larger scale cultures, refer to reference 20 and Chapter 6 of this vol.
for methodology.
3. We have noticed considerable variation in the growth and plaquing
properties of several Spodopteva frugiperda cell lines from different
sources. If problems are encountered, then a different cell isolate from
another research group should be tried.
4. Some protocols recommend the use of TClOO or Grace’s insect culture
medium. However, in our experience, commercial sources of these
support only poor growth of the Spodopteru fiugiperda cell line. The
culture medium recipe given in the Materials section (and in detail in
the Appendix) has, in our experience, proven to be the most reliable
for good cell growth.
5. Although the insect cells grow optimally at 27OC,we prefer to restrain
the growth of the cells by maintaining the stock cultures at room tem-
perature. These slower growing cells seem to perform consistently
better in plaque assays (carried out at the optimal temperature of
27°C).
6. The insect cells are considerably more fragile than most mammalian
cells. For this reason, the cells should not be subjected to rapid pipet-
ing or vortexing. Otherwise, substantial cell death could occur. It is
recommended that viability counts be performed regularly on the
cells during passaging.
7. If difficulties are consistently found in attempting to identify the rare
occlusion-negative recombinant plaques among the wild-type occlu-
sion-positive plaques, then it is advisable to obtain a pure recombinant
virus stock from a research group and plaque this just to familiarize
oneself with plaques totally lacking polyhedrin. This could be further
reinforced by mixing together wild-type and recombinant virus in a
1O:l ratio, carrying out a plaque assay, and then practice identifying
the 1 occlusion-negative plaques in every 10 occlusion positive plaques.
8. The condition and number of cells used in a plaque assay is critical for
obtaining well-formed, easily identifiable occlusion-positive or neg-
ative plaques. The cells must be subconfluent on the first day of the
assay. If a confluent monolayer is formed on the first day (Le., by seed-
ing too many cells), then the plaques will be small and poorly formed.
If the cells fail to reach confluence by the end of the assay, identifica-
tion of plaques is often very difficult. This is usually the result of un-
healthy stock cells, or obtaining high numbers of dead cells upon har-
vesting with which to set up the assay.
Foreign Gene Expression in Insect Cells 587

9. Occasionally, occlusion-negative plaques can be generated by events

other than inactivation of the polyhedrin gene by insertion with the
foreign gene. It is strongly recommended that Southern Blot analysis
be performed on final plaque purified recombinants to ensure identi-
fication of true recombinants.
10. The factors affecting the levels of foreign gene expression with the ba-
culovirus/insect cell system are still poorly understood. Although
polyhedrin can be expressed from the wild-type virus at levels ap-
proaching 1 mg/mL, the levels of foreign gene expression are usually
in the range of l-50 pg/mL using the transfer vector pAc373.
11. pAc373 can only be used to express mature foreign proteins (i.e., from
genes that contain their own ATG translation start codon followed by
the uninterrupted coding open reading frame). However, there are
vectors available, such as pAc360 (II), that use the polyhedrin gene
ATG translational start followed by a unique cloning site shortly
downstream of theN terminus. The foreign genes are cloned in frame
into this site and will now be expressed as polyhedrin/foreign gene
fusion proteins. The levels of these proteins can often be considerably
greater than that obtained with mature proteins generated from
pAc373 and can easily provide large quantities of material suitable for
antigenicity evaluation. The pAc373 vector is most suitable for ex-
pression of therapeutic proteins where biological activity is more
important.

References
1. Wurm, E M., Gwinn, K. A., and Kingston, R. E. (1986) Inducible overproduction of
the mouse c-myc protein in mammalian cells. Proc.Nutl. Ad. Sci. 83,5414-5418.
2. Smith, G. E., Summers, M. D., and Fraser, M. J. (1983) Production of human J3inter-
feron in insect cells infected with a baculovirus expression vector. Mol. Cd. Bid. 3,
2156-2165.
3. Pennock, G. D., Shoemaker, C., and Miller, L. K. (1984) Strong and regulated ex-
pression of Escherichia coli P-galactosidasein insect cells with a baculovirus vector.
Mol. Cell. Biol. 4,388X%.
4. Miyamoto, C., Smith,G. E.,Farrell-Towt, J.,Chizzonite,R.,Summers, M. D.,and Ju,
G. (1985) Production of human c-myc protein in insect cells infected with a bacu-
lovirus expression vector. Mol. Cell. Biol. 5,2860-2865.
5. Smith, G. E., Ju, G., Ericson, B. L., Moschera, J., Lahm, H. W., Chizzonite, R., and
Summers, M. D. (1985) Modification and secretion of human interleukin-2 pro-
duced in insect cells by a baculovirus expression vector. Proc. N&l. Ad. Sci. 82,
8404-8408.
588 Page and Murphy

6. Possee, R. D. (1986) Cell surface expression of influenza virus hemagglutinin in

insect cells using a baculovirus vector. Virus Research 5,43-59.
7. Kuroda, K., Hauser, C., Rott, R., Klenk, H. D., and Doerfler, W. (1986) Expression of
the influenza virus hemagglutinin in insect cells by a baculovirus vector. EMBO. J.
5,1359-1365.
8. Gardiner, G. R. and Stockdale, H. (1975) Two tissue culture media for production of
lepidopteran cells and nuclear polyhedrosis viruses. 1. Znvertebr. Path. 25,363-370.
9. Maniatis, T., Fritsch, E., and Sambrook, J. feds.) (1982) Molecdur Cloning-A Lab-
ovatovy Manual (Cold Spring Harbor Laboratory, New York).
10. Weiss, S. A. and Vaughn, J. L. (1986) Cell culture methods for large-scale propa-
gationofbaculoviruses,inTheBiologyofBaculoviruses(Granados,R.andShapiro,M.,
eds.), CRC Press, Florida.
22. Summers, M. D. and Smith, G. E. (1987) A manual of methods for baculovirus
vectors and insect culture procedures. Texas Agriculture Experiment Station,
bulletin No.: 1555.
 Chapter 44

Chloramphenicol
Acetyltransferase
as a Reporter
in Mammalian Gene Transfer

Christopher D, Corsica
and Bruce H. Howard
1. Introduction
Reporter genes can be used to advantage in a variety of different kinds
of gene transfer experiments. One of the most common applications is the
optimization of transfection methods. Owing to the large number of vari-
ables that influence the uptake and expression of exogenously introduced
genes, it is extremely helpful to have available rapid, sensitive methods for
determining transfection efficiency. A second frequent application of re-
porter genes is to study promoters and transcriptional enhancers. To rig-
orously map functional domains in transcriptional control elements it is
often necessary to generate large numbers of deletions, insertions, and
point mutations, and then carry out functional assays on each construct. As
in optimization studies, an appropriate reporter gene can greatly facilitate
the rate with which information is accumulated. Another gene regulation
application involves measuring activity of transacting transcriptional
regulatory proteins. In such experiments, one plasmid codes for the
transacting factor, while the second cotransfected plasmid carries a target

589
590 Corsica and Howard

&-acting element coupled to a reporter gene. Examples of more novel

reporter gene uses include introducing translational stop codons into a
reporter coding sequence to measure suppressor function in mammalian
cells (1,2), and measuring transcriptional competence of UV-irradiated
reporter plasmid DNA (3). Finally, reporter genes can be used to normal-
ize expression where other transfected genes are measured at the RNA
level by Sl nuclease, primer extension, or Northern analysis (4). Given this
multiplicity of applications, there has been a steady increase in the num-
ber of studies in which reporter genes are used.
Of the many different reporter genes available for gene transfer
studies in mammalian cells, the chloramphenicol acetyltransferase (CAT)
gene (5) has probably been the most widely used. Another reporter gene
that is frequently employed is the E. coli B-galactosidase gene (6). The E. coli
xanthine-guanine phosphoribosyltransferase (7,8) and galactokinase (9)
genes, and recently, the firefly luciferase gene (20) represent alternative
reporter systems.
There are several variables in the choice of an enzymatic activity as a
reporter system. Perhaps the first consideration is whether there is a
corresponding endogenous enzyme activity in recipient cells. The ab-
sence of endogenous activity eliminates the possibility that the investi-
gator will encounter interference in the detection of the expression of
exogenously introduced genes. An additional concern is whether the
assay is appropriate for the experiments contemplated, e.g., whether cost
or sensitivity is of primary importance. Finally, an ancillary but often
crucial point is whether a given reporter system provides functions in
addition to an assay for soluble enzyme activity.
With respect to these considerations, each of the above-mentioned
reporter systems has attendant advantages and disadvantages. For ex-
ample, the E. coli P-galactosidase must be distinguished from endogenous
P-galactosidase activity and is generally less sensitive than CAT as a re-
porter (Padmanabhan, R. and Howard, B., unpublished results); on the
other hand, an excellent in situ detection system for P-galactosidase based
on the X-gal substrate is available (12). E. coli xanthine-guanine phos-
phoribosyltransferase is also somewhat less sensitive than CAT, but can be
used simultaneously as a dominant selectable marker (12). Firefly lu-
ciferase is a novel activity in mammalian cells and offers a more sensitive
assay than CAT, but at present requires relatively expensive equipment to
obtain maximum sensitivity.
This chapter will present details concerning the use of the CAT
reporter system in mammalian gene transfer applications (seeNote 1). The
Chloramphenicol Acetyltransferase 591

focus will be on the commonly used CAT enzyme assay developed by

Shaw (13) and subsequently adapted by Cohen and coworkers (14). This
assay measures acetylation of chloramphenicol at the l- and 3-positions of
the carbon side chain. CAT initially catalyzes acetylation at the 3-position.
The 3-acetoxy chloramphenicol can then undergo a nonenzymatic rear-
rangement that results in the formation of l-acetoxy chloramphenicol. In
a second CAT catalyzed reaction, the l-acetoxy chloramphenicol can be re-
acetylated at the 3-position yielding 1,3-diacetoxy chloramphenicol(Z5,Z 6).
CAT
Chloramphenicol + AcCoA - 3-acetoxy chloramphenicol + CoA

3-acetoxy chloramphenicol t-------3 1-acetoxy chloramphenicol

CAT
lacetoxy chloramphenicol + AcCoA - 1,ddiacetoxy chloramphenicol + CoA

2. Materials
1. Acetyl &enzyme A, Lithium. Store at -7OOC.
2. 14C-chloramphenico140-60 mCi/mmol. Store at -70°C.
3. Ethyl acetate.
4. Baker flex silica gel 1B TLC plates.
5. Chloroform.
6. Methanol.
7. Phosphate Buffered Saline (PBS) (seeAppendix, this vol.)-without
Mg2+ or Ca2+.
8. TEN buffer: 0.04.M Tris HCl, pH 7.4/l mM EDTA/O.lSM NaCl.
9. 0.25M Tris HCl, pH 7.5.
10. 4 mM Acetyl Coenzyme A-Prepare in 0.25M Tris HCl, pH 7.5. It is
preferable to prepare a fresh solution each time the assay is done;
however, storage of the solution at -7OOCfor up to 2 wk is possible.

3. Methods
1. Wash transfected cells in a T-25 tissue culture flask three times with 5
mL of PBS.
2. Add 1 mL of TEN, and let the flask sit at room temperature for 5 min.
3. Gently scrape the cells off the bottom of the flask with a tissue culture
scraper, and transfer them to an appropriately labeled Eppendorf tube
(seeNote 2). Immediately centrifuge the cells at 4OC, aspirate the
supernatant, and resuspend the cell pellet in 0.10 mL of 0.25M Tris
592 Corsica and Howard

HCl, pH 7.5. At this point, the suspension can be stored at -2OOC for
later use.
4. For cells grown in suspension, the washing can be done in centrifuge
tubes. After the final wash, resuspend the cells in 1 mL of 0.25M Tris
HCl, pH 7.5, and transfer to an Eppendorf tube.
5. Lyse the cells by freeze-thawing them three times. To ensure com-
plete lysis, cells can be sonicated briefly. A cuphorn apparatus is
preferable to a sonication tip; if a tip is to be used, it is imperative that
it be washed thoroughly in between each sample to prevent cross-
contamination.
6. Pellet the cellular debris by centrifugation at 12000 x g at 4OC for 10
min.
7. Transfer the supernatant, being careful not to disrupt the pellet, to a
new Eppendorf tube. At this point, the supernatant can be stored at
-20°C.
8. The supernatants can now be assayed radiochemically. Place a new
Eppendorf tube on ice and add the following:
l-50 I.LL Extract
78-28 /.tL 0.25M Tris HCl, pH 7.5
W-J ‘4C-chloramphenicol
20 /.tL 4 mM Acetyl Coenzyme A
The final reaction vol is 100 pL; acetyl CoA is the last reagent that
should be added to the reaction mixture.
9. Generally, it is only necessary to assay 1 PL of an E. coli (pRSVcat)
extract from a positive control sample (seeNote 5); 1 PL of extract can
usually convert 80-90% of the ‘4C-chloramphenicol to its acetylated
product in less than 1 h.
10. Centrifuge the reaction mixture at 12000 x g for 15 s.
11. Incubate the reaction mixture at 37°C for 30 min. Incubation times
range from as short as 10 min to as long as 18 h. To enhance the
accuracy and the sensitivity of the assay, time points can be taken
during the course of the incubation. By plotting the formation of
acetylated 14C-chloramphenicol as a function of time, one can min-
imize random variation and ensure that all the assays are within the
linear range. When taking time points, stop the reaction by adding 1
mL of ethyl acetate to the reaction mixture. Process the samples as
outlined below.
12. The reaction is terminated by adding 1 mL of ethyl acetate to the re-
action mixture (seeNote 4).
Chlorarnphenicol Acetyltransferase 593

13. Vortex the reaction mixture vigorously for 30 s, and separate the
phases by centrifugation for 30 s.
14. Remove 950 PL of the top organic layer, and transfer it to a fresh
Eppendorf tube (seeNote 2).
15. Evaporate the ethyl acetate overnight in a fume hood. This step can be
expedited by evaporating the samples in a Speed Vat concentrator
(Savant).
16. Once the sample is thoroughly dried, resuspend it in 30 PL of ethyl
acetate. Vortex vigorously for 30 s, and spot onto a TLC plate using a
capillary pipet. Although not absolutely necessary, a preequilibrated
TLC plate can be used. To prepare it, place a fresh TLC plate into a
chromatography tank containing 95:5 chloroform:methanol. Once
the solvent has run to the top of the plate, remove it from the tank and
let it dry in a fume hood. Spot as indicated above.
17. The TLC plate is placed in a chromatography tank that has been
preequilibrated with 95:s chloroform:methanol (see Note 3). The
ascending chromatography should be completed in less than 3 h. The
time will vary depending on the commercial source of plates used.
18. The plate is dried in a fume hood and then exposed to film overnight.
19. The conversion of ‘4C-chloramphenicol to its acetylated products can
be quantitated by scintillating counting. Using the film as a guide,
mark the acetylated andnonacetylated spots on the TLC plate. Cut the
spots out from the plate, place into separate scintillation counting
vials, and count. The percent conversion can be calculated as follows:
% acetylated Wchloramphenicol = 100 x acetylated counts
total counts (acetylated + nonacetylated
counts)

4. Notes
1. The focus of this chapter on the standard CAT assay notwithstand-
ing, it should be emphasized that overall success in gene transfer
experiments (seeChapter 50; Vol. 2, Chapter 25; and Vol. 4, Chapter 42
of this series) depends not only on the reporter system, but also on
adherence to proper technique with respect to DNA plasmid prep-
aration, tissue culture, and transfection protocols. In particular, the
efficiency of transfection can vary over a wide range depending on the
competence of recipient cells, which in turn is strongly dependent on
tissue culture parameters. Readers interested in details concerning
594 Corsica and Howard

gene transfer efficiency are referred to other discussions on that topic

(l7,18).
2. Streaking of spots on the TLC plates has several causes:
a. Failure to avoid the interface when transferring the organic
layer (Method, step 14);
b. Inadequate drying of sample (Method, step 15); and
c. Use of tubes from certain manufacturers (e.g., Sarstadt) dur-
ing the ethyl acetateextraction (Method, steps 12 and 13)-we
speculate that the polypropylene in some tubes has been
treated to make it more hydrophilic, leading to more protein
contamination after ethyl acetate extraction and subsequent
“gumming” of the TLC.
3. Abnormally slow migration of acetylated chloramphenicol deriva-
tives (manifested by compression of those spots) is usually the result
of inadequate solvent saturation of the atmosphere in the TLC tank.
To prevent this problem, the inside of the TLC tank should be lined
with chromatography paper. Place the paper so that the lower edge
sits in the solvent.
4. Ethyl acetate rapidly attacks polystyrene. This reagent should be
transferred with glass pipets only.
5. It is often useful to prepare a bacterial extract that contains CAT
activity as a positive control for the assay. Bacteria carrying a plasmid
that codes for chloramphenicol resistance (e.g., pBR325 or pRSVcat)
provide a convenient source. Prepare a 5-mL overnight culture, col-
lect in a 1.5-mL Eppendorf tube, lyse by several freeze-thaw cycles and
brief sonication, and clear by centrifugation in a microfuge at 4OCfor
10 min. Aliquot and store at -2OOC.
6. Low CAT activity in cell extracts can result from premature cell lysis
with consequent loss of CAT during scraping of cells from the plate
and centrifugation. One solution to this problem is to remove cells
from the plate by trypsinization. Addition of serum and pelleting of
cells eliminates most trypsin activity; moreover, trypsin concentra-
tions as high as 0.025% have no observable effect on CAT activity
under standard assay conditions (Holter, W., Howard, T., and How-
ard, B. H., unpublished results).
7. CAT has been used as a reporter in extracts from a wide variety of
mammalian tissue culture cells and in extracts from multiple tissues
prepared from transgenic mice (19). In some extracts, however,
inhibitors of CAT activity have been detected. These inhibitors have
Chloramphenicol Acetyltransferase 595

not been characterized, but may degrade acetyl CoA or acetylated

chloramphenicol derivatives. In most cases, such inhibitors can be
inactivated by incubation of extracts at 65OC for 5 min (19,20).
8. Other methods for determining CAT activity that avoid requirements
for Y-chloramphenicol and thin layer chromatography have been
described (21,22).

References
1. Burke, J. F. and Mogg, A. E. (1985) Construction of a vector, pRSVcatamb38, for the
rapid and sensitive assay of amber suppression in human and other mammalian
cells. Nucleic Acids Res. 13,1317-1326.
2. Capone, J. P., Sedivy, J. M., Sharp, P. A., and RajBhandary, U. L. (1986) Introduction
of UAG, UAA, and UGA nonsense mutations at a specific site in the Escherichia coli
chloramphenicol acetyltransferase gene: use in measurement of amber, ochre, and
opal suppression in mammalian cells. Mol. Cell. Bid. 6,3059-3067.
3. Protic-Sabljic, M. and Kraemer, K. H. (1986) Host cell reactivation by human cells of
DNA expression vectors damaged by ultraviolet radiation or by acid-heat treat-
ment. Carcinogenesis 10,1765-1770.
4. Gaynor, R. B., Feldman, L. T., and Berk, A. J. (1985) Transcription of class III genes
activated by viral immediate early proteins. Science 230,447-450.
5. Gorman,C.M.,Moffat,L.F.,andHoward,B.H. (1982)Recombinantgenomes which
express chloramphenicol acetyltransferase in mammalian cells. Mol. Cell. BioZ. 2,
1044-1051.
6. Hall, C. V., Jacob, P. E., Ringold, G. M., and Lee, F. (1983) Expression and regulation
of Escherichia coli 1acZ gene fusions in mammalian cells. J. Molec. Appl. Genef. 2,
101-109.
7. Mulligan, R. and Berg, P. (1980) Expression of a bacterial gene in mammalian cells.
Science 209,1422-1427.
8. Chu, G. and Berg, P. (1985) Rapid assay for detection of Escherichia coli xanthine-
guanine phosphoribosyltransferase activity in transduced cells. Nucleic Acids Res.
13,2921-2930.
9. Schumperli, D., Howard, B. H., and Rosenberg, M. (1982) Efficient expression of
Escherichia coli galactokinase gene in mammalian cells. PYOC.NafI. Acad. Sci. USA 79,
257-261.
lO# de Wet, J. R., Wood, K. V., DeLuca, M., Helinski, D. R., and Subramani, S. (1987)
Firefly luciferase gene: structure and expression in mammalian cells. Mol. CeII. Biol.
7,725-737.
11. Price, J,, Turner, D., and Cepko, C. (1987) Lineage analysis in the vertebrate nervous
system by retrovirus-mediated gene transfer. PYOC.Nafl. Acad, Sci. USA 84,156-160.
12. Mulligan, R. and Berg P. (1981) Selection for animal cells that express the Escherichia
coli gene coding for xanthine-guanine phosphoribosyltransferase. Proc. Nafl. Acad.
Sci. USA 78,2072-2076.
13. Shaw, W. (1967) The enzymatic acetylation of chloramphenicol by extracts of R
factor-resistant Escherichia coli. J. Biol. Chem. 242,687-693.
596 Corsica and Howard

14. Cohen, J. D., Eccleshall, T. R., Needleman, R. B., Federoff, H., Buchferer, B. A., and
Marmur, J. (1980) Functional expression in yeast of the Escherichia coli plasmid gene
coding for chloramphenicol acetyltransferase. Proc. Natl. Acad. Sci. USA 77,1078-1082.
15. Shaw, W. V. (1983) Chloramphenicol acetyltransferase: enzymology and molecular
biology. CRC Crit. Rev. B&hem. 14,146.
16. Kleanthous, C. and Shaw, W. V. (1984) Analysis of the mechanism of chloram-
phenicol acetyltransferase by steady-state kinetics. Evidence for a ternary-complex
mechanism. Biochem. J. 223,211-220.
17. Gorman, C., Padmanabhan, R., and Howard, B. H. (1983) High efficiency DNA-
mediated transformation of primate cells. Science 221,551-553.
18. Fordis, C. M. and Howard, B. H. (1987) Use of the CAT reporter gene for optimiza-
tion of gene transfer into eucaryotic cells. Methods Enzymol, 151,382-397.
19. Overbeek, P. A., Chepelinsky, A. B., Khillan, J. S., Piatigorsky, J., and Westphal, H.
(1985) Lens-specific expression and developmental regulation of the bacterial
chloramphenicol acetyltransferase gene driven by the murine alpha A-crystallin
promoter in transgenic mice. Proc. Natl. Acad. Sci. USA 82,7815-7819.
20. Mercola, M., Goverman, J., Mire& C., and Calame, K. (1985) Immunoglobulin
heavy-chain enhancer requires one or more tissue-specific factors. Science 227,
266-270.
22. Sleigh, M. F. (1986) A nonchromatographic assay for expression of the chlor-
amphenicol acetyltransferase gene in eucaryotic cells. Anal. Biochem. 156,251-256.
22. Young, S. L., Jackson, A. E., Puett, D., and Melner, M. H. (1985) Detection of
chloramphenicol acetyltransferase in transfected cells: a rapid and sensitive HPLC-
based method. DNA 4,469-475.
 Chapter 45

Immunizing Schedules
for Hybridoma Production

Bjiirn Gustafsson

1. Introduction
Immunization protocols for generating activated B-lymphocytes suit-
able for hybridoma production vary widely. The optimal amount of anti-
gen given, the way of antigen presentation, and the timing between
injections must be determined for each antigen. Often, the antigen of in-
terest is not available in a pure form, and the influence of contaminating
antigens in the preparation should be considered when the immuniza-
tion schedule is designed. Generally, a lower immunization dose will give
rise to antibodies with higher specificity as compared to the total antibody
response obtained after immunization with higher doses. Furthermore,
antibodies obtained after one or two injections often show lower affinity
than those obtained when a prolonged immunization schedule is used.
The number of antibody-producing hybridomas obtained after fusing B-
lymphocytes with myeloma cells depends on how well the mouse was
immunized, and it is therefore worthwhile to put some effort into opti-
mizing the immunization schedule.

597
598 Gustafsson

2. Materials
1. Antigen, dissolved in O.OlM phosphate-buffered saline, pH 7.2.
2. Freunds complete adjuvant (FCA).
3. Freunds incomplete adjuvant (FlA).
4. Female BALB/c mice, 6-8 wk of age.

3. Methods
3.1. Soluble Antigens
1. Mix lo-100 pg of antigen in FCA (1:l; vol:vol). The antigen solution
is sucked into a glass syringe and mixed with FCA in a test tube by
pushing the mixture back and forth. Continue until a thick creamy
solution is obtained. If the solution is properly mixed, it should
remain as one phase after a 1 h incubation at 4OC. If two phases are
obtained, the solution was not properly homogenized.
2. Inject 50-100 PL intraperitoneally (ip) to each mouse. Two to four
weeks later the same dose of antigen is given ip, but this time it is
mixed with FIA. Two to four weeks later, the mice are boosted with
100-200 pg ip without adjuvant. Continue to immunize until a
satisfactory immune response is obtained. Alternatively, the animals
could be immunized subcutaneously with 50 yL at each of four lymph
node sites. Immunization in the footpads causes great pain to the
animal and should be avoided. For highly immunogenic antigens, it
is possible to omit adjuvants which is also more lenient to the animal.
Fuse spleen cells with myeloma cells 34 d after the last booster dose.

3.2. Cellular Antigens

1. Inject approximately 2 x lo7 cells ip. Boost the animals at 34 wk inter-
vals with the same or a lo-fold higher dose. In general, the animals
should be rested for 4 wk before the final booster dose.
2. When optimizing the immunization schedule, the development of a
specific immune response should be measured by a method suit-
able for rapid screening of multiple samples. Collect serum samples
by retrobulbar punction (eye-bleeding) or tail-bleeding 1 wk after
each injection, and measure the increase of specific antibodies in
enzyme-linked immunosorbent assay, radio-immunoassay, or some
other method suitable for rapid screening of multiple samples (see
Chapter 48).
Immunization for Hybridoma Production 599

4. Notes
1. Individual animals may sometimes respond differently to the same
immunization schedule. When optimizing the immunization pro-
gram, each schedule should be tested in groups containing at least five
mice/group.
2. Adjuvants are used to enhance the immune response to molecules that
are poorly immunogenic. FCA is a mixture of mineral oils containing
killed and dried Mycobactevium tubercu2c~is bacteria. FIA is devoid of
the bacteria. Both FCA and FIA cause local inflammatory response at
the site of injection. Precautions should be made to avoid accidental
injections into the hand or splashing of droplets into the eye, since this
may cause irreparable damage.
3. Collecting serum by eye-bleeding should be performed only on suffi-
ciently anesthesized mice under the supervision of a person who is
thoroughly skilled in this technique.
 Chapter 46

Fusion Protocol
for the Production
of Mouse Hybridomas

Bj6rn Gustafsson
1. Introduction
Hybridoma technology makes it possible to produce almost un-
limited amounts of monospecific antibodies. Each hybridoma represents
only one of several B-lymphocytes responding to a particular antigen, and
for this reason, monoclonal antibodies can also be produced against im-
pure antigen preparations.
Monoclonal antibodies are produced by hybridomas derived from the
successful fusion of B-lymphocytes and mouse myeloma cells (1). Most
myeloma cell lines used in hybridoma work are mutants lacking the en-
zyme hypoxanthine guanine phosphoribosyl transferase (HGPRT), which
is utilized in the hypoxanthine-aminopterin-thymidine (HAT) selection
system (2). Aminopterin blocks the main biosynthetic pathways for DNA
and RNA synthesis. By providing thymidine and hypoxanthine in the
growth medium, cells can usually continue to multiply by using the sal-
vage pathway. However, a HGPRT- myeloma cell line will not grow,
since the cell cannot utilize hypoxanthine supplied with the medium.
Only hybrids formed between a myeloma cell (HGPRT-), and a B-lympho-
cyte (HGPRT’) will survive and multiply.

601
602 Gustafsson

The production of hybridomas begins by immunizing a mouse with

a selected antigen (seeFig. I). B-lymphocytes from the spleen are prepared
and mixed with myeloma cells in the presence of polyethylene glycol to
promote fusion. The cells are suspended in a growth medium containing
HAT. Unfused myeloma cells and myeloma-myeloma hybrids, which
lack the enzyme HGPRT, cannot survive in the HAT-medium. Unfused
spleen cells and spleen-spleen hybrids die naturally after a few replica-
tions. The only surviving cells are hybrids of myeloma cells and B-lympho-
cytes. These hybridomas are screened for specific antibody production,
and hybrids of interest are cloned. Following recloning, antibody-produc-
ing hybridomas are expanded, and the cells are frozen in liquid nitrogen
for future use. A hybridoma cultivated in tissue culture bottles produce
l-50 pg/mL of monoclonal antibodies. For large-scale production, the
hybridomas are injected into the peritoneal cavity of mice, in which they
produce ascites tumors that generate monoclonal antibodies in concentra-
tions of l-10 mg/mL. The entire process of establishing antibody produc-
ing hybridomas takes 34 mo.
A number of HAT-sensitive mouse myeloma cell lines are available
for hybridoma production. However, many myelomas produce their own
immunoglobulin, and this makes themless suitable as fusion partners. The
optimal myeloma should be a nonimmunoglobulin-producing cell line
with a high capacity of hybrid formation. The cell line Sl?2/0-Ag 14 (3) is
such a myeloma and is commonly used for the generation of hybridomas.
Hybridoma production involves a large number of steps requiring
patience and strict adherence to sterile tissue culture technique. This and
following chapters (46-51) describe the basic methods required for the
successful production and characterization of hybridomas.

2. Materials
1. RPM1 1640 tissue culture medium. Store at 4OC.
2. Fetal calf serum. Store at -20°C.
3. Stock solutions of 100 x L-glutamine: 2.92 g L-glutamine dissolved in
100 mL distilled water.
4. Stock solutions of 100 x sodium pyruvate: 1.l g sodium pyruvate dis-
solved in 100 mL distilled water.
5. Stock solution of 100 x PEST: 10,000 U pencillin and 10 mg strepto-
mycin/ml of distilled water.
Hybridoma Fusion Protocol 603

IMMIJNI~ATI~N MYELOMA CELL CULTURE

SPLEEN CELLS
HGPRT+

FUSION IN PEG

SELECTION OF HYBRID CELLS

IN HAT-MEDIUM

ASSAY FOR ANTIBODY

A

13
CLONE ANTIBODY-PRODUCING
@jj'b z& HYBRIDS

FREEZE
HYBRIDOMA
FOR FUTU-
RE USE

MONOCLONAL ANTIBODY

Fig. 1. The generatioxuof hybridomas and production of monoclonal antibodies.

6. Stock solution of 50 x I-IT solution: Dissolve 68 mg of hypoxanthine

and 19.5 mg of thymidine in 100 mL of distilled water under gentle
heating. If the hypoxanthine does not dissolve readily, add a few
drops of LOM NaOH. Adjust to pH 7.2 with 1M HCl when the hypo-
xanthine is solubilized.
 Gustafsson

7. Stock solution of 50 x HAT solution: Dissolve 0.9 mg of aminopterin

in 100 mL of 50 x HT solution. All stock solutions should be sterilized
by membrane filtration and stored in aliquots at -20°C.
8. Standard medium: 500 mL of RPM1 1640 tissue culture medium is
supplemented with 50 mL of fetal calf serum, 5 mL of L-glutamine, 5
mL of sodium pyruvate, and 5 mL of PEST.
9. HAT medium: Add 2 mL of HAT (50x) stock solution to 100 mL of
standard medium.
10. 50% PEG solution: 1 g polyethylene glycol4000 (PEG) in a test tube is
autoclaved for 20 min. While the PEG is still liquid, add 1 mL of RPM1
1640 medium.

3. Methods
1. Give the immunized mouse a final booster dose without adjuvant, 3-4
d before fusion.
2. Dilute the myeloma cells the day before fusion to a concentration of
2-3 x lo5 cells/mL.
3. Kill the mouse by cervical dislocation or CO, asphyxiation.
4. Swab the mouse with 70% ethanol and remove the spleen, aseptic-
ally. Place the spleen in a Petri dish containing 5-mL of RPM1 1640
medium.
5. Put a sterile sieve on top of a Petri dish containing 5-mL medium. Pre-
pare a single-cell suspension by placing the spleen on the sieve and
gently tease the spleen through the sieve using forceps and spatula.
All handling of cells should be performed in a laminar flow hood.
6. Transfer the spleen cells to a conical tube. Exclude large clumps of tis-
sue. Allow the tube to stand for l-2 min, while clumps settle out.
Transfer the cell suspension to a new tube and centrifuge at 300 xg for
5 min. Wash the cells once in RPM1 1640 medium and count them us-
ing a cell counting chamber. Approximately lo* cells are obtained
from a mouse spleen.
7. Harvest the myeloma cells by centrifugation at 300 x g for 5 min, wash
once in RPM1 1640 medium, and count.
8. Mix the myeloma cells with the spleen cells in a 50-mL centrifuge tube.
The ratio of myeloma cells: spleen cells may range from l:l-1:lO. Cen-
trifuge the mixed cell suspension at 300 x g for 5 min and discard the
supernatant.
Hybridoma Fusion Protocol

9. Add 1 mL of the PEG solution (37OC) to the pellet over 1 min, using a
I-mL pipet, continuously stirring the pellet with the tip. Continue
with gentle stirring for 1 min.
10. Dilute the fusion mixture by adding 1 mL of RPM1 1640 medium
(37°C) dropwise, over 1 min. Stir gently to ensure an even dilution.
11. Repeat step 10.
12. Add 10 mL of medium over 5 min to the mixture. Centrifuge the sus-
pension at 300 x g for 5 min and discard the supernatant. Resuspend
the pellet gently in 30-40 mL of HAT-medium and distribute the cells
into four 96-well tissue culture trays, at a concentration of approxi-
mately 2.5 x lo5 cells/well.
13. Incubate the trays at 37OC,5% CO, 80% humidity in a CO,-incubator.
14. Feed the cells twice a week by removing half of the medium in each
well using sterile Pasteur pipets connected to a filter pump, taking care
not to disturb the cells. Add fresh HAT-medium to the wells by gently
dropping medium from a 10-mL pipet. A color change of the medium
from red to yellow is indicative of cell growth. Visible clones will ap-
pear within 2-3 wk.

4. Notes
1. Stock solutions of medium supplements are commercially available.
2. The myeloma cells must be in exponential growth prior to fusion in
order to obtain high yields of hybrid clones. Keep the cells at a den-
sity of 2-6 x l@ cells/mL. The cells should never be allowed to exceed
8 x 105cells/mL. The cells can be cultured in the same bottles for sev-
eral months under the condition that they are split and fed with fresh
standard medium regularly. The size of a B-lymphocyte is about one-
tenth of a myeloma cell, and the two cell types are easily distinguished
in the microscope. Hybridomas and myeloma cells are of the same
size and appear in the microscope as round, heavily light-scattering
cells (seeFigs. 2 and 3).
3. Most myeloma cells grow as suspension cultures. However, a certain
percent of the cells attach to the plastic of the bottle. By vigorously
shaking the bottle, the adherent cells are released from the plastic sur-
face. The cells are split by removing all the spent medium and replac-
ing it with fresh standard medium. There will still be enough cells left
in the bottle to ensure the culture.
606 Gustafsson

Fig. 2. Hybridoma cells in tissue culture (100 x magnification).

Fig. 3. Hybridoma cells in tissue culture (400 x rnagnificationh

Hy bridoma Fusion Protocol 607

References
2. K&ler, G. and Milstein, C. (1975) Continuous cultures of fused cells secreting
antibody of predefined specificity. Nature 256,495-497.
2. Littlefield, J,W. (1964) Selection of hybrids from matings of fibroblasts in vitro and
their presumed recombinants. Science 145,709.
3. Schulman, M., Wilde, C., D., and Kiihler, G. (1978) A better cell line for making
hybridomas secreting specific antibodies. Nature h-m.lon) 276,269,270.
 Chapter 47

Cloning of Hybridomas

Bjiirn Gustafsson
1. Introduction
Hybrid clones will appear within 2-3 wk after fusion (seeChapter 46).
It is of utmost importance that a newly established hybridoma is cloned
thoroughly to ensure that the cells growing in the tissue culture are of
monoclonal origin and not a mixture of two or more hybridomas. A mix-
ture of cells may result in a gradual decline of specific antibody production
because of overgrowth of contaminating hybridomas, until finally the hy-
bridoma of interest is lost. Culture supernatants from the hybridomas
should be tested for specific antibody production by appropriate methods.
Hybridomas are cloned by limiting dilution using thymocytes as feeder
cells as soon as the cell density in the well reaches one-half to two-third
confluence.
The hybridomas will have previously been cultivated in HAT me-
dium. Before the cells can be switched to standard medium, they must be
cultivated in HT medium for 2 wk. If the HAT constituents are excluded
from the medium, there will be a risk that the hybrids will die out because
of remaining amounts of aminopterin in the cells. By maintaining the cells
on HT medium for a couple of weeks, they can still use the salvage pathway
for nucleic acid synthesis while aminopterin is metabolized. In order to
avoid accidental feeding of cells with HT medium instead of HAT medium
610 Gustafsson

and vice versa, we routinely add HAT medium to all fusion trays, I-IT
medium to all cloning trays, and standard medium to all tissue culture
bottles.

2. Materials
1. Standard medium, HAT medium and sterile tissue culture equip-
ment (seeChapter 46).
2. HT medium: Add 2 mL of HT (50x) stock solution to 100 mL of
standard medium.

3. Methods
1. A mouse (BALB/c), 3-4 wk of age, is killed by CO, asphyxiation.
2. Swab the mouse with 70% ethanol and remove the thymus, asep-
tically. Prepare a single-cell suspension of thyrnocytes as described
for splenocytes (seeChapter 46, points 4-6). Approximately lo8 thy-
mocytes are obtained from a young mouse. Use one thymus for each
hybridoma to be cloned.
3. Suspend the thymocytes in 10 mL of HT medium. Distribute 0.1 mL
of the cell suspension to each well of a 96-well tissue culture tray.
4. Remove the hybridoma cells from the microtiter well with a sterile
Pasteur pipet by gently flushing the medium up and down a couple
of times, and transfer the cells to a sterile test tube. Add fresh HAT
medium to the well, and put the tray back into the CO, incubator.
5. Count and dilute the cells in I-IT medium to a concentration of 50 and
5 cells/mL, respectively. Add 0.1 mL of the suspension containing 50
cells/mL to eight wells in the thymocyte-conditioned tray and 0.1 mL
of the suspension containing 5 cells/mL to the rest of the wells.
Incubate the tray in the CO, incubator.
6. Feed the cells twice a week with HT medium. Visible clones appear
within 1-2 wk. A successful cloning should not show growth in more
than approximately 50% of the wells.
7. The clones are screened for antibody production, and hybridomas of
interest are expanded gradually from the 96-well tissue culture well,
over 24-well trays, to 25- and 80-cm2 tissue culture flasks. The hybrid-
oma cells can be cultivated in the same tissue culture flasks for several
months if they are handled properly. Monoclonal antibodies can be
harvested twice a week from such cultures by simply replacing spent
medium, containing antibodies, with fresh medium. Newly estab-
Cloning of Hybridomas 611

lished hybridomas should be frozen and stored in liquid nitrogen as

soon as possible.

4. Notes
1. Hybridoma cells grow poorly at low cell densities (< lo4 cells/mL),
and growth is supported by the use of feeder cells. Thymocytes are
effective as feeder cells, although peritoneal macrophages or sple-
nocytes may be used equally well. Feeder cells are short-lived in
culture, but provide necessary growth supporting factors during the
early stages of growth of the hybridoma cells.
2. Cloning of hybridomas should routinely be performed twice, and
monoclonality should be verified by different methods, e.g., iso-
electric focusing of monoclonal antibodies, immunoglobulin isotype
determination, and/or titration of antibodies produced during pro-
longed cultivation of hybridomas. The frequency of chromosome loss
decreases with time in culture, but may still occur during long-time
cultivation. Hybridomas that are cultivated over longer periods
should be tested for antibody production and recloned if the titer
drops significantly.
3. Scaling up of hybridomas may occasionally be troublesome, since
some hybridomas do not always transfer well to growth bottles. The
easiest way to get the cells started is to incubate the flask in an upright
position for 24 h and then continue cultivation with the flask lying
down. Feeder cells may help. Keep the cells in log-growth phase (3-6
x lo5 cells/mL). They should never exceed lo6 cells/ml. Overgrowth
may result in a dead culture.
 Chapter 48

Enzyme-Linked
Immunosorbent Assay
for Screening of Antibodies
in Hybridoma Supernatants

Bjiirn Gustafsson
1. Introduction
Enzyme-linked immunosorbent assay (ELISA) is a rapid and con-
venient method for screening of antibody producing hybridomas (1,2).
The method is highly sensitive and can be applied to detect antibodies
directed against soluble antigens as well as cell-bound antigens. Over a
thousand culture supernatants can be tested in one day if the assay is
performed in microtiter trays.
In principle, the soluble antigen (protein, glycoprotein, lipopoly-
saccharide) is attached to the plastic by physical adsorption. Whole-cell
antigens may have to be bound by poly-L-lysine to the plastic followed by
fixation with glutaraldehyde. Remaining binding sites are blocked by
bovine serum albumin (BSA). Culture supernatants are added to the wells
followed by incubation with enzyme-conjugated rabbit antimouse immu-
noglobulin antibodies. Unbound conjugate is washed away, and an
enzyme substrate is added to each well. The amount of colored product

613
614 Gustafsson

Antigen-ELISA,

v C------- Substrate

Enzyme-ab
conjugate

Monoclonal
antibody

-e BSA

Fig. 1. ELBA principle.

formed is proportional to the amount of antibodies produced by the

hybridoma cells (seeFig. 1).

2. Materials
1. Coating buffer: 0.05M carbonate buffer, pH 9.6; 1.56 g Na.&03, 2.93 g
NaHCO, 0.2 g NaN, in 1000 mL of distilled water.
2. PBS (stock): Phosphate-buffered saline (PBS), pH 7.2; 36.04 g
Na,HPO,*Z&O, 9.18 g KHJ?O,, 145.8 g NaCl, 0.6 g NaN, in 3 L of
distilled water.
3. PBS (15 mM; working dilution): 500 mL PBS (stock) + 2.5 L of distilled
water.
4. Washing buffer: 0.05% Tween-20 in PBS.
Enzyme-Linked Immunosorbent Assay 615

5. Blocking buffer: 1% bovine serum albumin (BSA), 0.05% Tween-20 in

PBS.
6. Substrate solution: Dissolve 55 mg of 1,2-phenylenediaminedihy-
drochloride in 100 mL of 40 mM Tris-HCl, pH 7.6. Add 40 yL of 30%
H202 immediately before use.
7. 96-well flat bottom microtiter trays (ELISA grade).
8. Conjugate: Peroxidase-conjugated rabbit anti-mouse immunoglob-
ulin antiserum.
9. Stop solution: 1M H$04.
10. Microplate spectrophotometer.
11. Cell bound antigens:
a. 10 pg/mL poly+lysine in PBS.
b. 0.5% glutaraldehyde in cold PBS. Prepare immediately before
use.
C. 0,lM glycine, 0.1% bovine serum albumin in PBS.

3. Methods

3.1. Soluble Antigens

1. Add 100 JJL of antigen (l-10 pg/mL in coating buffer) to the wells.
Incubate overnight at 22OC. The trays can be stored at 4OC.
2. Wash three times with washing buffer. Add 150 PL of BSA-solution
and incubate for 15 min at 22°C.
3. Wash. Add 100 PL of hybridoma culture supernatant, and incubate
for 1 h at 37OC.
4. Wash. Add 100 JJL of conjugate diluted according to manufacturer’s
instructions. Incubate for 1 h at 37°C.
5. Wash. Add 100 FL of substrate solution. Incubate for 10 min at 22°C.
6. Stop the reaction by adding 50 PL of 1 M H,SO, and measure the
absorbance at 492 nm.

3.2. Cell Bound Antigens

1. Add 100 PL of poly-L-lysine solution to each well. Incubate for 30 min
at 22°C.
2. Aspirate.
3. Prewash the cells in PBS. Add lo6 cells/well in 100 PL of PBS.
4. Centrifuge the trays at 200 x g for 5 min.
616 Gustafsson

5. Add 100 l.tL of cold glutaraldehyde solution to each well. Incubate for
15 min at 22°C.
6. Wash twice with PBS. Add 125 PL of PBS-glycine-BSA solution,
incubate for 30 min at 22”C, and then follow the procedure described
for soluble antigens, steps 3-7.
7. If the trays are to be stored: wash twice with PBS. Add 100 u.L of 0.1%
BSA in PBS, then store at -20°C.

4. Notes
1. Carbonate buffer is the standard buffer used for coating of antigen to
microtiter trays. In our experience however, PBS often works equally
well. Microtiter trays made of polystyrene are commonly used. If the
antigen sticks poorly to the plastic, it may be worthwhile to try flex-
ible polyvinyl trays. Sterilized (irridiated) microtiter trays often show
higher affinity to the antigen, although sometimes accompanied by an
increased background level.
2. The enzyme reaction can be performed using different enzyme-
antibody conjugates. Horseradish peroxidase (HRP) and alkaline
phosphatase (ALP) are the most commonly used enzymes. HRP is
popular because of its rapid formation of colored product. However,
the substrate o-phenylenediamine has been suggested to be a car-
cinogen, and gloves should be used when handling this chemical.
3,3,5,5-tetramethylbenzidine is not a known carcinogen and can re-
place o-phenylenediamine as substrate in the HRP-reaction. When
using this substrate, 10 mg of 3,3,5,5-tetramethylbenzidine is dis-
solved in 1 mL dimethylsulfoxide, and 100 PL of this solution is added
to 10 mL O.lM sodium acetate, pH 6.0. I!-$02 is added to a final concen-
tration of 0.01% immediately before use. The ALP reaction is equally
sensitive, but somewhat slower. None of the enzymes should be used
if endogenous activity of these enzymes in the antigen is expected.
Other enzymes that can be used are glucose oxidase, galactosidase (3),
and urease (4).
3. All ELISAs must be optimized for each antigen, and the instructions
given above should only be regarded as guidelines. The use of poly-
L-lysine and glutaraldehyde fixation in whole cell ELISAs may some-
times be omitted, and the number of cells/well may have to be altered
in order to obtain optimal sensitivity.
Enzyme-Linked Immunosorbent Assay 617

References
1. Engvall, E. and Perlman, l? (1972) Enzyme-linked immunosorbent assay, ELISA. III.
Quantitation of specific antibodies by enzyme-labeled anti-immunoglobulin in
antigen coated tubes. 1. Irnmunol. 109,129-135.
2. Gustafsson, B. (1984) Monoclonal antibody-based enzyme-linked immunosorbent
assays for identification and serotyping of Vibrio cholerae 01. J. Clin. Microbial. 20,
1180-1185.
3. Chandler, H. M., Cox, J. C., Healey, K., McGregor, A., Premier, R. R., and Hurrell, J.
G. R. (1982) An investigation of the use of urease-antibody conjugates in enzyme
immunoassays. J, Immunol. Methods 53,187-194.
4. Kincade, P. W., Lee, G., Sun, L., and Watanabe, T. (1981) Monoclonal rat antibodies
to murine IgM determinants. J. Zmmunol. Methods 4217-26.
 Chapter 49

Cryopreservation
of Hybridomas

Bjiirn Gustafsson
1. Introduction
Hybridomas are exposed to many threats, such as contamination with
bacteria and fungi, loss of chromosomes coding for antibody production,
overgrowth by nonsecreting mutants, and cell death resulting from over-
growth. Therefore, newly established hybridomas should be frozen and
stored in liquid nitrogen at -196OC as soon as possible. Alternatively, they
can be stored in the vapor phase above the liquid nitrogen where the tem-
perature will be between -120’ and -195OC. Generally, eucaryotic cells are
very sensitive to the freezing procedure, and programmable freezing de-
vices are available that lower the temperature from 0’ to -3OOC and from
-30” to-100°C at different time rates. However, in our experience, hybrid-
oma cells will survive by simply incubating the vials at -2OOC for 2 h fol-
lowed by overnight incubation at -70°C. The vials are then transferred to
liquid nitrogen for long-time storage, Cells frozen in this way have sur-
vived for more than 5 yr.
However, sometimes hundreds of hybridomas are obtained after a
successful fusion experiment. Since it is not possible to clone all hybrido-
mas within the time available before the cultures become overgrown, one
can freeze hybridomas directly in the culture trays (1). This may also be
useful when time-consuming screening protocols are necessary. Culture
supernatants can be analyzed, while the cells are preserved frozen.

619
 Gustafsson

2. Materials
1. 1.0 mL cryopreservation vials.
2. Liquid nitrogen container .
3. Freezing medium: 10% dimethylsulfoxide (DMSO) in fetal calf serum.

3. Methods

3.1. Freezing Cells in Cryotubes

1. Day 1. Remove half of the cell suspension, and add fresh standard
medium (seeChapter 46).
2. Day 2. Centrifuge the cells at 300 xg for 10 min. Resuspend the cells
in ice-cold freezing medium at a concentration of 2-5 x IO6 cells/mL.
3. Put 1 mL of cell suspension into the cryopreservation vials. Incubate
immediately at -2OOCfor 2-3 h, and then transfer the vials to -7OOCfor
overnight incubation.
4. Day 3. Put the vials in the liquid nitrogen container.
To reconstitute frozen cells, carry out the following procedure:
1. Day 1. Thaw rapidly by placing the vial in a beaker of water at 37OC.
As soon as the cells are liquified, wash the cells in 10 mL of cold stand-
ard medium (seeChapter 46). Resuspend the cells in 5 mL of standard
medium, and transfer the cells to a 50-mL tissue culture flask.
2. Day 2. Check the cells for viability. Split if necessary or transfer to a
larger tissue culture flask.

3.2. Freezing of Hybridomas

in 96-Well Microtiter Tkays
1. Day 1. Remove half of the medium from the wells, and add fresh
medium.
2. Day 2. Remove all medium from the wells. Add 0.1 mL of ice-cold
freezing medium to each well, and wrap the tissue culture tray in plas-
tic. Incubate at -2OOC for 2 h, and then at -7OOC. Cells frozen in trays
have remained viable for at least 2 mo.
To reconstitute the cells, carry out the following steps:
Cryopreservation of Hybridomas 621

1. Add 0.1 mL of RPM1 1640 medium, prewarmed to 37OC.

2. Aspirate the freezing medium as soon as it is liquified. Add HAT-me-
dium to the wells (seeChapter 46). Centrifuge the trays at 300 xg for
5 min.
3. Aspirate all medium from the wells. Add HAT-medium to the wells.
Incubate the trays in the CO, incubator.

4. Notes
1. Freeze only cells that are in mid-log phase (2-5 x lo5 cells/mL).
Rapidly growing cultures can be identified because they contain cells
of varying size.
2. Freeze several vials each time. Reconstitute one vial after a couple of
days, and check for viability (by cultivating the cells for a couple of
days) that the freezing was successful before discharging the cell cul-
tures.
3. Recently thawed cells may sometimes have difficulties in getting
started. Such cultures may improve by increasing the fetal calf serum
content to 20%. Once they are started, they can be switched back to
10%.
4. An alternative to the freezing procedure described above is to put the
vials in a polystyrene box with l-cm thick walls, and put the box over-
night in -70°C before transferring the vials to liquid nitrogen.

Reference
I. Wells, D. E. (1983) Simple rapid method for freezing hybridomas in 96-well micro-
culture plates. 1. Immunol. Melhods 59,49-52.
 Chapter 50

Isotype Determination
of Monoclonal Antibodies
by Immunodiffusion

Bjiirn Gustafsson
1. Introduction
Immunodiffusion is an important qualitative method for the detec-
tion of antigens and antibodies. The technique may be used to determine
the immunoglobulin heavy and light chain class and subclass of hy-
bridoma antibodies. The monoclonal antibodies are allowed to diffuse
toward anti-mouse isotype specific antisera in an agarose gel. In a pos-
itive reaction, a precipitin line will develop in the agarose between two
wells. A monoclonal antibody should show only one heavy chain and
one light chain. Knowledge of the isotype is also useful for purification
purposes.

2. Materials
1. Cell culture supernatants, lo-fold concentrated by ammonium sul-
fate precipitation.
2. Specific antisera against mouse IgM, IgA, IgGl, IgG2a, IgGILb, IgG3,
kappa and lambda light chain.

623
624 Gustafsson

3. 1% agarose in 15 mM phosphate-buffered saline (PBS), pH 7.2.

4. Plastic film (Gel BondTM).
5. Gel pouch (4 mm).
6. PBS.
7. Washing solution: 0.9% NaCl in distilled water.
8. Staining solution: 1.5 g Coomassie Brilliant Blue R250, in 300 mL
destaining solution. Stir for 1 h, 22OC. Filter before use. The solu-
tion can be used several times.
9. Destaining solution: 350 mL 95% ethanol, 100 mL glacial acetic acid.
Make up to 1000 mL with distilled water. The destaining solution
may be reused if the dye is first removed by passing through a funnel
of activated charcoal.

3. Methods
3.1. Immunodiffusion
1. Add 1 g of agarose to 100 mL PBS. Heat to boiling while stirring.
When the agarose is completely dissolved, transfer to a 60°C water-
bath.
2. Cut the plastic film to appropriate size.
3. Place the plastic film on a horizontal table with the hydrophilic side
facing upward (check with a drop of water).
4. Pour the melted agarose solution onto the film (0.15 mL/cm2).
Allow the agarose to solidify (lo-15 min).
5. Stamp and suck out wells, 4-5 mm apart in a circle and with one well
in the center.
6. Fill the center well (approximately 15 FL) with mouse isotype-
specific antisera and the outer wells with concentrated hybridoma
supernatants. Incubate overnight at 22”C, in a moist chamber.
7. Check for precipitin lines on a light box, or stain in Coomassie Blue
as described below.

3.2. Staining of Agarose Gels

Precipitates may sometimes be difficult to discover. However, the
precipitates will be visualized by staining the gels with a suitable
reagent.
1. Flush the gel, and fill the wells with distilled water.
2. Cover the gel with a wet filter paper (Whatman No. 1). Make sure
that no air bubbles are trapped between gel and filter.
Monoclonal Antibodies 625

3. Cover the filter paper with a 2-cm layer of cellulose wadding, and
place a thick glass plate on top of the cellulose wadding. Apply a
lo-20g/cm2 pressure on the plate for 15 min.
4. Remove the cellulose. Flush the filter paper with distilled water,
and remove it carefully from the gel.
5. Wash the gel twice in 0.9% NaCl for 20 min, and then once in distilled
water for 10 min.
6. Repeat steps 24.
7. Dry the gels with a heating fan (hair dryer).
8. Put the gel in the staining solution for 5 min at 22°C.
9. Transfer the gel to the destaining solution. Stir gently. Change the
destaining solution once or twice. Protein will retain the dye,
whereas the agarose will be destained.
10. Dry the gel and check for precipitin lines.

4. Notes
1. Monoclonal antibodies can be concentrated by adding an equal vol
of saturation (NHJ,SO, to the culture supernatant. Allow the anti-
bodies to precipitate for 1 h, followed by centrifugation at 10,000 x
g for 10 min. Dissolve the pellet in distilled water to one-tenth of the
original vol, and apply the sample directly to the agarose gel with-
out prior dialysis.
2. Thin gels, e.g., isoelectric focusing gels, can be washed in 0.9%
saline, overnight, followed by washing in 95% ethanol for 10 min.
When the gel is dry, it can be stained according to steps B-10
described above.
 Chapter 51

Isoelectric Focusing
and Immunofixation
of Monoclonal Antibodies

Bjiirn Gustafsson
1. Introduction
Isoelectric focusing of monoclonal antibodies in agarose gel is one of
several ways to characterize a hybridoma. Isoelectric focusing is per-
formed on each monoclonal antibody, and the mouse immunoglobulin
bands are immobilized by soaking the gel with rabbit antimouse im-
munoglobulin before staining with Coomassie Blue.
A single band obtained after isoelectric focusing and immunopre-
cipitation is a strong indication of monoclonality. However, results from
isoelectric focusing are sometimes difficult to interpret, since amonoclonal
antibody often show multiple bands. Nevertheless, these bands often tend
to show a characteristic grouping that also remains constant after exten-
sive recloning (I).

2. Materials
1. Electrofocusing apparatus.
2. Thermostatic circulating cooling bath (flow capacity > 6 L/min).
3. Agarose, IEF-grade.
627
 Gustafsson

4. Ampholytes, pH 3.5-9.5.
5. pH surface electrode (optional).
6. Glass plates 250 x 125 x 3 mm and 270 x 125 x 3 mm.
7. Spacers, 250 x 6 x 0.5 mm.
8. Plastic film (Gel-BondTM) for agarose gels.
9. Electrode filter strips.
10. Sample application strips; 5 x 10 mm pieces of Whatman No. 1 filter
paper.
11. Rubber roller.
12. Hemoglobin marker: Add 100 PL of mouse whole blood to 1 mL of
O.lM citrate in PBS, pH 7.2. Centrifuge at 2008 for 10 min. Wash the
erythrocytes twice, in 5 mL of PBS, and resuspend the cells in 1 mL of
distilled water. Store at -2OOC.
13. Electrode solutions: 0.5M acetic acid (anode) and 0.5M NaOH (cath-
ode).
14. 0.9% NaCl solution.
15. 99.5% ethanol.
16. Coomassie Brilliant Blue staining and destaining solutions (seeChap-
ter 50). Rabbit antimouse immunoglobulin antiserum.

3. Methods
3.1. Gel Preparation
1. Add 0.18 g of agarose to 16.6 mL of double-distilled water. Heat to
boiling while stirring. When the agarose is dissolved, transfer to a
75OC water bath. Add 1.4 mL of ampholytes pH 3.5-9.5 when the
agarose solution is at 75°C. Add the ampholytes immediately before
casting the gel (steps 6-10).
2. Prior to preparing the mold, wash and dry the glass plates thoroughly.
3. Pour a few mL of water onto the larger glass plate. Place the plastic
film with the hydrophilic side facing upwards on the glass plate, and
squeeze out excess water and air bubbles with the rubber roller. Make
sure that the surface is dry and free of dust.
4. Place the spacers along both sides, and put the shorter glass plate on
top of the spacers, leaving 1 cm free on both sides. Fix the plates to-
gether by placing clamps along the long sides.
5. Place the mold in a heating cabinet at 75°C for 20 min before casting
the gel.
Isoelectric Focusing of Antibodies 629

6. To cast the gel take the mold from the heating cabinet and place it at
an angle of 30°C.
7. Fill a preheated pipet with the agarose solution. Put the tip of the pipet
to the slit between the glass plates, and let the gel solution flow into the
mold. Move the pipet along the slit to ensure an even distribution of
the agarose solution and so that no air bubbles are formed.
8. Lower the mold to a horizontal position, and leave i t for 30 min at 22OC.
9. Remove excess agarose with a spatula. Grasp protruding plastic film
and tear the glass plates gently apart.
10. Place the gel in a moist chamber.

3.2. Isoelectric Focusing and Immunoffxation

1. Connect the isoelectric focusing apparatus to the circulating water

bath (+lO”C).
2. Pour a few mL of water on the cooling plate and place the gel on top.
Avoid trapping air bubbles.
3. Soak the electrode strips with electrode solutions. Apply the elec-
trode strips to the gel. Cut off parts of the strips that protrude beyond
the end of the gel.
4. Apply sample application strips 2 cm from the cathode. Sample vol-
umes of < 5 PL can be applied directly on the gel surface.
5. Apply samples (2-20 PL vol). Hemoglobin can be used as marker.
Add 2 PL of hemoglobin in two separate samples, one at the anode side
and one at the cathode side.
6. Run the gel at 15 W for 15 min, then turn off the power, and remove
the application pieces with forceps. Continue the isoelectric focusing
for approximately 30 min at the same wattage.
7. Turn off the power supply, and remove the electrode strips. Measure
the pH gradient at l-cm intervals across the gel with the surface elec-
trode.
8. Place the gel in a box, and soak it with rabbit antimouse immuno-
globulin antiserum. Incubate for 30 min at 22OC.
9. Wash the gel overnight in 0.9% NaCl solution.
10. Wash the gel in 95% ethanol for 10 min at 22OC, and then dry the gel
in front of a heating fan.
11. Stain by soaking the gel in Coomassie Blue solution for 5 min and de-
stain in destaining solution for 5-15 min (seeChapter 50, steps S-10).
630 Gustafsson

4. Notes
1. Gel casting kits are commercially available.
2. The quality of the agarose and distilled water is essential for success-
ful isoelectric focusing.
3. It is not important to measure the pH gradient unless an exact p1 of the
antibody is of interest, since it is the banding pattern of the antibody
that is important.

Reference
1. Gustafsson, B. and Holme, T. (1983) Monoclonal antibodies against group- and type-
specific lipopolysaccharide antigens of Vibrio cholerue0:l. J. Clin. Microbiol. l&480-
485.
 Chapter 52

Large-Scale Production
and Storage of Monoclonal
Antibodies and Hybridomas

Stuart A. Clark, J. Bryan Grifiths,

and Christopher B. Morris
1. Introduction
The end product of all the selection and cloning procedures described
in Chapters 4548 will be a monoclonal hybridoma culture growing in a
0.2-mL culture well. From this single well, sufficient cells must be grown
for storage in liquid nitrogen. It is advisable that at least six ampules, di-
vided between two liquid nitrogen freezers, be stored. After the security
of the cell line has been assured, then the clone can be further expanded for
bulk antibody production. Large quantities of antibody may be produced
either in ascitic fluid or in vitro in various culture vessels. The following
sections describe the various areas covered in the chapter.

1.1. Freeze Preservation

The maintenance of a cell line in continuous culture is not only im-
practical in terms of resources, but generally undesirable because of the
risk of microbial infection and of genetic drift. The ideal situation is to have
a cell stock that has been fully characterized and frozen at the lowest
passage possible. In view of the strict regulations concerning the manu-

631
632 Clark, Griffiths, and Morris

facture of diagnostic and therapeutic products from animal cell lines, e.g.,
monoclonal antibodies and enzymes, it becomes increasingly important to
ensure that cell lines used in production processes have been fully screened
for microbial contaminants.

1.2. Clonal Expansion

The expansion of the clone from a small well to larger culture vessels
can be difficult. The rate of cell growth, which can vary significantly be-
tween clones, and the cell density are critical factors. Fortunately, there are
a range of plastic vessels that permit a gentle stepwise scale-up of the cul-
tures. The normal route is: 0.2-mL well > l-mL well > 25-cm2 flask (up to
15 mL) > 752 flask (up to 50 mL).
At all expansion steps, it is important that the cells have reached a den-
sity of between 5 x lo5 and 1 x lo6 cells/ml before transfer to the next vessel.
During the initial stages of this progression, especially at the lower end of
this cell density, murine thymocyte feeder cells may be required to ensure
that cells survive the transfer. Although the main aim is to expand the
culture, it is very important to ensure that the cell density does not increase
above 1 x lo6 cell/mL. Above this density, the cells are under stress caused
by the limited availability of nutrients in the medium and may as a result
alter in their secretory or replicative capacity.
1.3. Hybridoma Production in Ascites
The hybridomas are derived from plasmacytoma and spleen cells of
an inbred strain of mouse, and are therefore immunologically compatible
withother mice of the same strain. Moreover, the parental plasmacytomas
were originally induced in the peritoneal cavity of mice treated with
Staphylococcal adjuvants (I). By injecting the hybridoma cells into the
peritoneal cavity of mice pretreated with Pristane or incomplete Freund’s
adjuvant (FIA), rapid cell proliferation occurs generating an ascitic fluid
containing large quantities of cells and high concentrations of antibody
(l-3 mg/mL).
1.4. Production of Monoclonal Antibodies In Vitro
Production in ascitic fluid, although convenient for producing small
quantities of high concentration monoclonal antibody (50-500 mg), does
have the disadvantage of being contaminated with other immunoglobu-
lins (and also possibly infectious agents), lacks reproducibility, may be
unsuitable for therapeutic use, and is unsuitable for human MCAbs. The
method is inconvenient for extensive scale-up because it is costly in man-
power and facilities with no scale-up economics, in addition to the ethical
Large-Scale Hybridoma Production 633

considerations of using animals. Thus, in vitro culture is the scale-up route

for quantities of over 1 g, and certainly for the commercial operations now
producing l-100 kg quantities. This scale of production is becoming neces-
sary now that the use of antibodies has extended from diagnostic to affinity
Durification and therapeutic use.

2. Materials
1. Freezing medium, e.g., growth medium with 25% serum and 7-10%
cryoprotectant; alternatively, whole serum, i.e., newborn calf with
7-10% cryoprotectant (seeNote 1).
2. One Balb/c or Fl (Balb/c x CBA) mouse 4-8 wk old.
3. One sterile stainless-steel mesh (Expanded Metal Co., Hartlepool),
Type 940,200-mm wide or a tea strainer.
4. Pristane (2,6,10,14-Tetramethypentadecane).
5. 20 mM EDTA containing 1% (w/v) sodium azide solution.
6. Mature Balb/c or Fl (Balb/c x CBA) mice.
7. Diethyl ether.

3. Methods
3.1. Freeze Preservation (See Notes l-5)
1. Harvest the cells in the same manner as for routine subculture (see
Note 5).
2. After counting the viable cells, pellet the required number in centri-
fuge tubes. Trypsinized cells must be washed in medium containing
serum to neutralize the enzyme. The number of ampules required is
determined by allowing a final concentration of 4-10 x lo6 cells/
ampule.
3. Prepare the freeze medium while the cells are being centrifuged,
allowing I mL/ampule.
4. Clearly mark your ampules (artline for plastic and ceramic ink for
glass) with an identification code and preferably some designation
that can be identified with that particular freeze run. Always main-
tain full records of freeze runs, with dates and locations.
5. The medium is decanted from cell pellets and freeze medium added.
The pellets can be ampuled individually or pooled. If the number of
ampules is small, i.e., c 20, dispensing can be by plastic pipets.
6. Transfer the ampules to the programmable freezing apparatus or two-
stage freezer (Linde BF-S), and cool at a steady rate of between 1 and
3”C/min. With a programmable freezer, it is possible to reduce the
634 Clark, Griffiths, and Morris

problem of a temperature rise during phase transition, i.e., liquid-solid

state, by temporarily increasing the cooling rate to 15 -2O”C/min, and
then return to the original slow cooling rate.
7. On completion of freezing, the ampules are stored either in the gas or
liquid phase of nitrogen depending on whether the ampules have
been hermetically sealed. For temporary handling use a polystyrene
container with liquid nitrogen to keep the ampules cold, e.g., while
transferring from the freezer to the nitrogen refrigerator. Wear pro-
tective clothing at all times.
8. One ampule should be examined for viability, i.e., Trypan blue ex-
clusion test, and growth capacity. Thaw by placing immediately in a
37OC water bath, preferably in a rack to prevent total submersion.
9. Resuspend the cells by slowly adding 1 mL of growth medium
(dropwise), and then transfer to a flask or tube with 5-10 mL of me-
dium. Take an aliquot, i.e., l-2% of the total volume, and measure cell
viability.
10. If it is necessary to remove the cryoprotectant, pellet the cells at low
speed to prevent damaging them, i.e., 50-70g.
11. Start a new culture by resuspending the viable cells at 40-50% of their
usual saturation density. If the viability is below 70%, new stocks
should be prepared. Attached cells can be medium changed after
24-48 h to remove cell debris.
3.2. Clonal Expansion
1. A mouse (4-8 wk old) is sacrificed by cervical dislocation, pinned ven-
tral surface uppermost to a dissection board, and swabbed in 70%
alcohol.
2. The skin is dissected away from the abdomen and thorax, the skin
flaps pinned back, and the underlying tissue swabbed with 70%
alcohol.
3. The peritoneum is opened and scissors used to cut through the ribs,
which are removed to expose the heart and lungs. Care must be taken
not to puncture the heart. The thymus is a cream colored bilobed or-
gan, found lying above the heart (seeFig. 1).
4. The thymus is removed and immediately placed in 10 mL of medium
containing 20% FCS (seeNote 6). The thymus, with the medium, is
dissociated by pushing the tissue through a sterile tea strainer or steel
mesh placed in a Petri dish, using the barrel of a 5-mL plastic syringe.
5. The cells are diluted to 85 mL in medium with 20% FCS.
6. 100~pL aliquots are dispensed/well of eight g&well plates and in-
cubated at 37°C in 5% CO,. (The outer wells can be filled with distilled
Large-Scale Hybridoma Production 635

Cerwcal kmph node

lngvlnrl lymph node

Fig. 1. The main lymphoid organs of the mouse.

water containing 2x concentration of antibiotics to reduce dehydra-
tion in nonhumidified CO, atmospheres.)
7. Thymocyte feeder cells should be used to expand the positive hybri-
doma clones within 24 h of production.
3.3. Hybridoma Production in Ascites
(See Notes 7-14)
1. The mouse is injected ip with 0.5 mL of Pristane using a 23 g x l-in
needle.
2. These pretreated mice can be used between 4 and 60 d after injection.
3. For each mouse, approximately 20 mL of cell suspension is harvested
and centrifuged at 15Ogfor 10 min. The cells (approximately 1 x 10’
cells) are resuspended in 0.5 mL of PBS and injected ip into the mouse
using a 21 g x l-in needle.
4. Within 7-10 d after injection, the peritoneum of the mouse will be-
come distended and the ascitic fluid can be drained.
5. The mouse is anesthesized with diethyl ether vapor, and a 19 g x l-in
needle attached to a 5-mL syringe is inserted into the peritoneal cav-
ity through the left flank of the mouse. The mouse is gently tilted so
636 Clark, Griffiths, and Morris

that the liquid drains to the site of the needle and the excess fluid is
gently removed.
6. Add 10% (v/v) of EDTA/sodium azide solution to the ascitic fluid to
prevent clot formation.
7. The ascitic fluid is centrifuged for 10 min at 2008 and the supernatant
recovered (seeChapter 53 for the purification of the monoclonal anti-
body from this supernatant) and Notes 15 and 16.

3.4. Production of MonocZonaZ Antibodies In Vitro

3.4.1. Method 1. Small-Scale Culture System
Where less than 100 mg of monoclonal antibody is required, a few
liters of end-culture supernatant may suffice. End-culture supernatant is
that obtained after the cells are grown to the maximum cell density, and
includes antibody secreted from viable and lysed cells. Antibody con-
centrations in such supernatants tend to reach between lo-20 &mL (see
Note 17).
1. A lOOO-mL Spinner culture flask (TechneLtd., Note 18) is seeded with
an initial culture volume of 200 mL containing 1 x 105cells/ml, stirred
at 100 rpm and incubated at 37OC for 5 d.
3.4.2. Method 2. Bioreactors for Scale-Up
(Table 1 and Fig. 2)
The basic cell culture techniques described in this chapter (and Chap-
ter 6, this volume) apply to the production of MCAb in all the culture sys-
tems reviewed. In addition, manufacturers instructions are now very com-
prehensive. The aim of this section is mainly to make one aware of the
range of culture systems and concepts that are available, and to introduce
this range in order that a rational choice can be made (2) (seeNotes 20-22).
The following reactor types are available:
a. Techne BR06 Bioreactor: This retains the spinner culture sim-
plicity of use, but with the option of controlling pH and oxygen
and of using it as a continuous-flow culture. Oxygen supply is
good because of a very low aspect-ratio (i.e., high surface area/
unit vol) and the use of a surface aerator. The surface stirrer
impeller is suitable for most fragile cells and is convenient for
building up the culture volume within the vessel as the cells
grow (from 500 mL-3 L). It is currently only available as a 3-L
system.
b. Stirred Fermentation: There are a wide range of commercial re-
actors available in the range l-30 L and upwards (Applicon,
 Table 1
Scale-Up Culture Systems for the Production of Monoclonal Antibodies

Scale Process intensity Ifs Process Vol (L)

Culture system (L) cells/ml x lo6 Contamination for Ig MCAb

Bioreactor (Techne) B 1-3 2-3 Y lo-100

Fermentors, stirred B l-1000 2-3 Y lo-100
Airlift fermentor B 5-1000 2-3 Y 10-100
Opticell” C 1 50 Y 0.5-5
Ultrafiltration fibers” C 2.5 >lOO’ N 0.1-l
Encapsulation” B/SC 1-4-O >lOO’ N/Y 0.1-l
Membroferm# C l-6 50-100’ N 1

a See also Fig. 3.

b B = batch, C = continuous, SC = semi-continuous processes, Y = yes, N = no.
c Concentration in cell compartment, which is M-30% of total medium volume.
638 Clark, Griffiths, and Morris

Fig. 2. High process intensity culture systems for producing monoclonal antibodies
(further data inTable 1). (A) Opticell ceramic cartridge with coarse surface configura-
tion for trapping cells against the medium flow, (B) hollow fiber cartridge with recir-
culating medium via a reservoir (R). Harvest product through (H), (Cl membroferm
unit showing three-compartment configuration for medium, cells, and product, (D)
encapsulation method in a stirred fcrmentor with the option of containing the product
(a) or allowing the product to diffuse out of the capsule (b).
Large-Scale Hybridoma Production 639

LH Fermentation, Life Science Laboratories, Contact Flow,

New Brunswick, SGI, Chemap). Models specially adapted for
animal cells have a modified marine, or large flat blade, im-
peller, curved (or domed) bottom configuration, slow stirring
rates (10-200 rpm), and water jacket or heat pad temperature
control. Above 10 L, special laboratory facilities are needed for
in situ steam sterilization and so on. These units are not always
suitable for fragile cells.
c. Airlift Fermenter: Models are available from 500-mL (Fischer
Scientific) through 2-, 5-, lo-, 30- to 90-L volumes (LH Fer-
mentation, Chemap). They are very efficient for oxygenation,
provide very gentle mixing, and are relatively easy and trouble-
free to operate. A good air flow (and mixing) control unit is
needed.
d. Opticell (CRBS): These systems are based on a ceramic car-
tridge that is a honeycomb of l-mm square longitudinal chan-
nels with an uneven surface. This provides a large surface area
for the entrapment of cells as the media flows continuously
through the unit (Fig. 2). The system comes in a highly sophis-
ticated computer-controlled package, and the cost is such that
it can only really be considered for commercial operation.
e. Ultrafiltration Fibers: Complete turn-key units (Amicon, Cell-
Pharm CD Medical Inc.), specialized high performance units
(Endotronics), or a wide range of biochemical and kidney di-
alysis cartridges are commercially available. High cell densi-
ties (> 108/mL) can be achieved, and maintained for long
periods (months) allowing a daily harvesting of antibody. Ad-
vantages of all high density systems are the reduced serum
requirement and a high antibody titer in the process fluid.
These cartridges cannot be steam sterilized and need aseptic
washing procedures before use.
f. Encapsulation: The entrapment of cells within gel beads is a
commercially used procedure for producing MCAb’s (e.g.,
Damon Biotechnology, Karyon Technology), but is a skilled
and complicated operation. There is a choice of suitable gels
depending upon whether the product is required to diffuse out
of the capsule containing the concentrated cell suspension (al-
ginate gels), or retained with the cells (poly-L-Lysine). Other
materials that have been used are agarose, collagen, and fibrin.
The gel spheres can be suspended in growth medium in both
stirred fermentors and fluidized beds. A fluidized bed biore-
 Clark, Griffiths, and Morris

actor is available with in situ gel entrapment (Bellco Glass Inc.).

Procedures for entrapping and culturing cells in gels are pub-
lished (e.g., 3).
g. Membroferm Fermenter (MBR Bioreactor Ltd.): This fermen-
ter contains a stack of parallel semi-permeable membranes that
form a series of closed compartments (each of 20-25 mL).
Membranes with two molecular cut off values are available so
there is a choice of keeping the product compartmentalized
separately from the cells (Fig. 2), (3 chamber reactor), or only
having the cells compartmentalized (2 chamber reactor). Fresh
medium continuously diffuses across the membranes form a
recirculating reservoir into the cell compartment, thus main-
taining high cell densities (over 50 million/ml). Units are
available with up to 300 compartments.
3.4.2.1. Procedures
1. Seed cultures at between 1-2 x 105 cells/mL with the expectation of
achieving a yield of 1-3 x lo6 cells/mL after 5 d (mean generation time
will be within the range 11-36 h, with a mean of 17-20 h).
2. pH range tolerated is 6.5-7.2. Initiate the culture at pH 7.0-7.2. A
hepes buffered medium will be beneficial, especially in the absence of
pH control.
3. The specific rate of antibody synthesis increases during the death
phase (Fig. 3) for most hybridomas, thus keep the culture during the
death phase of the cells.
4. Expected antibody yields will be within the range of 5-100 pg/mL,
with a mean of about lo-20 pg/mL. With process control of pH and
oxygen, the yield should increase to 2040 pg/mL.
5. Scale-up increases the need for more sophisticated downstream proc-
essing, especially for initial clarification. Filtration (especially tan-
gential flow ultra-filtration) of the supernatant should be used. With
large volumes, flocculation of the cells (e.g., with polygalacturonic
acid or ammonium sulfate) increases filtration efficiency.

4. Notes
4.1. Freeze Preservation
1. Several cryoprotectants are available, the most commonly used are
dimethyl sufoxide (DMSO), glycerol, and polyvinyl pyrrolidone (PVP).
Only PTFE membranes can be used to filter sterilize DMSO; it must
never be autoclaved because it will explode! Use only the highest
Large-Scale Hybridoma Production 641

Time ---+

Fig. 3. Typical hybridoma cell growth and monoclonal antibody production kinetics.

grades available, e.g., ultra pure and ensure that the newest stocks
available are supplied. DMSO can be stored at -2OOCafter aliquotting.
2. The condition of the cells at the time of freezing is essential to their
survival. Therfore, only cells that are healthy and in log-phase growth
should be frozen.
3. For suspension cultures, the cell density should be 5-8 x l@/mL and
viability > 90%. Attached cell lines should be no more than 90%
confluent.
4. Cells should be in medium free of antibiotics, which might otherwise
mask low levels of infection.
5. Remember that it is advisable to perform all culture work in class 2
cabinets. This reduces the risk of cross-contamination between cell
lines, and also meets the requirements necessary to protect the user
from extraneous pathogens. For further reading, the following texts
are recommended (43).

4.2. CZonuZ Expansion

6. A thymus from a recently weaned mouse provides approximately lo*
thymocytes, which should be used at 5 x 105/mL. Thymocyte feeder
cells have the advantage that they die within 7-10 d in culture.
 Clark, Griffiths, and Morris

4.3. Production in As&tic Fluid

7. Mice treated in this way rapidly deteriorate in health. The ascitic fluid
must be drained regularly (every l-3 d). Frequently, solid tumors are
produced in addition to, or instead of, ascites. Such solid tumor for-
mation tends to be a property of the clone. Mice should be sacrificed
as soon as signs of distress are observed. The volume of ascitic fluid
obtainable from each mouse will depend largely on the biological
properties of the clone, but between 15-30 mL can be expected.
8. Some hybridomas may not produce a great deal of ascitic fluid in the
first instance and may tend toward solid tumor production. Any
liquid harvested from such mice when transferred directly into sec-
ondary pris tane primed mice should enhance the level of liquid accu-
mulating in the peritoneal cavity.
9. When a liquid tumor has developed, cells may be recovered and re-
established in in vitro culture or frozen in aliquots. The cells frozen
directly from the mice can be reintroduced into pristane primed mice
directly from the frozen state.
10. It is also possible to “clean up” clones when they have reached over lo6
total cell number, should they become infected, by passaging them
through pris tane primed mice.
11. Alternatively, Freund’s incomplete adjuvant (FIA) can also be used to
pretreat the mice for ascitic fluid production. In this case, the cells are
injected 24.48 h following the injection of FIA. Although this method
successfully produces ascitic fluid with fewer cells and in less time
than the pristane method, in our experience some monoclonal anti-
bodies lose significant activity in the process.
12, It is essential to monitor the antibody concentration throughout the
process of purification. Monitoring antibody activity at regular inter-
vals, using a suitable screening procedure, is necessary when the puri-
fication techniques may denature the monoclonal protein or when the
preparation is contaminated with high concentrations of other pro-
teins, especially immunoglobulins. Such is the case in supernatants
concentrated from bulk cultures, unless the cells have been adapted
from growth in medium supplemented with little, or no, FCS. How-
ever, the screening procedures are relatively time-consuming and
give no indication of the degree of contamination with other proteins.
13. At high concentrations (above 1 mg/mL), the monoclonal antibody
can be seen as a discrete band, distinct from other contaminating
proteins, after gel electrophoresis. Several electrophoretic systems are
commercially available based on thin agarose gels supported with
Large-Scale Hybridoma Production 643

plastic films. Results are obtainable within 90 min with such systems
and allow rapid monitoring of downstream processes.
14. Figure 4 illustrates the results of the purification of several mono-
clonal antibodies from ascitic fluids, run on a Paragon Electrophoresis
System (Beckman Ltd).
15. It is important to remember that storage conditions under which the
activity of polyclonal antibodies will survive can easily destroy the
activity of monoclonal antibodies. This is because polyclonal anti-
bodies have a range of physical properties; therefore, only a propor-
tion of the antibodies will be denatured by mistreatment, whereas all
of the molecules of a monoclonal antibody have the same properties.
16. Repeated freezing and thawing causes immunoglobulin aggregation
with concommitant loss of antibody activity. It is, therefore, essential
to freeze supernatants, ascites fluid, or purified antibody in small ali-
quots. Such aliquots are best stored at -5OOCor below. This inhibits
enzymic proteolytic activity and helps to stabilize IgM isotypes.
Alternatively, the antibody can be diluted 1:1 in glycerol and stored at
-20°C; at this temperature, the mixture remains liquid. Despite these
precautions, some monoclonal antibodies will still not be recoverable
after freezing. Such antibodies should be diluted in O.lM Tris-HCl
buffer, pH 7.2, and stored at 4°C; Tris has considerable antibacterial
activity (6).
4.4. Production of MCAbs In Vitro
17. By adapting the hybridomas to grow in a serum substitute, such as
CPSR-1 (Sigma Chemicals, Ltd.), cells can be maintained in a serum
protein concentration of 1% or less. End-culture supernatants from
these cultures may then be concentrated (up to 30-fold) using ultra-
filtration prior to further purification.
18. The Techne MCS spinner vessels are very useful and efficient first-step
culture units, which provide a gentle but efficient agitation and are a
preferred alternative to the use of stationary flasks or roller bottles.
However, the next step in scale-up must allow the introduction of en-
vironmental control equipment (02, pH) into the process. In choosing
a suitable system, consideration has to be given to the relative fragil-
ityof the cell (i.e., canit withstand conventionalmarine-typeimpellers
or does it need gentle agitation methods), oxygen, and nutrient re-
quirements.
19. The scale-up of animal cells has been reviewed in Chapter 6, and the
contents that refer to suspension culture are applicable to hybridoma
cells and should be read in conjuction with this chapter.
644 Clark, Gri@ths, and Morris

Fig. 4. Monitoring antibody purification. Track 1. Normal mouse serum (l/5 dil.)
Track 2. IgM monoclonal as&tic fluid. Track 3.lgM monoclonal purified using AcA 22
Column (0.75 mg/mL). Track 4. IgG,, monoclonal ascitic fluid. Track 5. I& mono-
clonal purified using protein-A-Sepharose (1 mg/mL). Track 6. IgG, monoclonal
ascitic fluid. Track 7. IgG, monoclonal purified using DEAE Trisacryl column (0.8 mg/
mL). Track 8. IgGzr,monoclonal ascitic fluid. Track 9. IgG, monoclonal ascitic fluid.
Track 10. Normal mouse serum (a 115 dil.). The isotypes were determined by ELISA
(seeSection 6.1.2.). The position of monoclonal bands cannot be used to determine the
isotype, seetracks 6 and 9.

20. If only milligram quantities of MCAb are required, then the mouse
ascitic fluid method or simple replicate culture vessels (flasks, roller
bottles) are adequate.
21. Once the requirement exceeds 100 mg, then scale-up methods should
be considered. Otherwise, one has to use over 50 mice or 30 roller
bottles.
22. Until the requirement exceeds 0.5 g, the simple versions of fermenters
(stirred or airlift) can be conveniently used, but once 0.5-l g has to be
produced (equivalent to 50-100 L of low process intensity culture),
then the more sophisticated and expensive culture systems listed in
Table 1 will be needed.

Acknowledgments
The authors wish to thank Miss A. Williams for the artwork on Fig. 1
and Mrs. Heather Hogton for typing the manuscript.
Large-Scale Hybridoma Production 645

References
1. Potter, M. and Robertson, C. L. (1960))x~naZ of the National Cancer Institute 25,847.
2. Griffiths, J. B. (1988) Overview of cell culture systems and their scale-up, in Animal
Cell:Biotechnology,vol. 3 (Spier, R. E. and Griffiths, J.B.,eds.), Academic, London, pp.
179-220.
3. Nilsson, K. (1987) Methods for immobilizing animal cells. Trends in Biofechnol. 5,
73-78.
4. Ashwood-Smith, M. J. and Farrant, J. (eds.) (1980) low Temperature Presentation in
Medicine and SioZogy (Pitman Medical).
5. Doyle, A., Morris, C. B., and Armitage, W. J. (1988) Cryupreservation of Animal Cells
in Advunces in Biotechnological Processes, vol. 7 (A. R. Liss, New York), pp. l-17.
6. Newell, D., McBride, B. W., and Clark, S. A. (1988) Making Monoclonals (PHLS,
London).
 Chapter 53

Purification of Murine
Monoclonal Antibodies

Michael G. Baines, Andrew J. H.

Gearing, and Robin Thorpe
1. Introduction
Hybridoma technology has made possible the production of highly
specific, homogeneous antibodies withpredefined binding characteristics,
which can be produced in large amounts, from immortal cell lines. They
probably represent the immunochemist’s ideal as reagents, and mono-
clonal antibodies (MoAbs) are invaluable for the investigation and assay of
the entire antigen spectrum. However, it must be remembered that, al-
though MoAbs can be exquisitely specific, they are far from pure unless
specific procedures are used for their isolation. If produced as ascitic fluid,
they contain proteins derived from the host animal, whereas in vitro pro-
duction as cell culture supernatant produces MoAbs contaminated with
tissue culture additives, nonimmunoglobulin secretion products, and ma-
terial derived from dead disrupted cells.
Preliminary purification procedures (1,2) generally utilize a precipi-
tation technique allowing the easy production of a moderately pure, stable
and concentrated MoAb preparation. Such procedures are the ideal start-
ingpoint for further purification, for storage, and since the precipitates are
particularly stable at room temperature, they are suitable for distribution
or exchange between laboratories. Although precipitation techniques are
647
648 Baines, Gearing, and Thorpe

excellent for producing purified, concentrated MoAb preparations they

rarely produce the highly pure antibodies necessary for many immuno-
chemical procedures, e.g., radiolabeling or enzyme conjugation for use in
immunoassays. For this, it is usually necessary to carry out some type of
chromatographic technique. Both ion-exchange and gel filtration have
been used for this, and there are many adaptations of these for use in
conventional chromatography and HPLC systems (3,4,5,6). Affinity chro-
matography techniques (7) are also extremely valuable for the purification
of MoAbs.
At present, mouse and rat MoAbs are by far the easiest to prepare
because of the availability of inbred strains, good fusion partners, and ease
of immunization. However, mouse and rat immunoglobulins are gener-
ally less stable than those of higher mammals, and can be more difficult to
purify and successfully fragment.

2. Materials
1. PBS: 0.14M NaCl, 2.7 mM KCl, 1.5 mM KH,PO, 8.1 mM Na,HPO,.
Store at 4°C.
2. PBSA: PBS containing 0.015M NaN,.
3. Saturated Ammonium Sulfate Solution: Add excess (NH,),SO, to
distilled water and stir overnight at 4OC. This solution (in contact with
solid salt) is stored at 4°C.
4. PEG Solution: 20% w/vPEG 6000inPBS;forIgMprecipitationuse6%
PEG 6000 in PBS.
5. Silicon Dioxide.
6. Glycerol.
7. Phosphate Buffer: 2 mM Na,HPO,, 2 mM NaH,PO,*2%0, pH 6.0.
Add the dihydrogen salt to the disodium salt until the required pH is
obtained. Phosphate buffer is also used at O.lM, pH 8.0. Store at 4OC.
8. Sephacryl S-300.
9. DEAE - Sephacel.
10. HCl: lOaM, 0.5M, 1M.
11. NaOH: 0.5M, lM, 4M.
12. Tris-HCl Buffer: O.O5M, lM, 2M Tris, pH 8.0. Adjust pH with 1M HCI
or concentrated HCl as appropriate. For ionic gradient separations
use Tris-HCl buffer containing 0.3M NaCl. Store at 4OC.
13. Triethanolamine Buffer: 0.02M triethanolamine, pH 7.7. For ionic
gradient separations by FPLC, use triethanolamine buffer containing
1M NaCl.
14. Sepharose 4B.
Murine Monoclonal Antibodies 649

15. Sodium Carbonate Buffer: 0.5M NaHCO, pH 10.5.

16. Cyanogen Bromide.
17. Sodium Citrate Buffer: O.lM Na&H,O,*2HO, pH 2.5,3.5,6.5. Adjust
pH with citric acid.
18. Ethanolamine Buffer: 2M ethanolamine.
19. Glycine-HCl Buffer: O.lM glycine, pH 2.5. Adjust pH with 1M HCl.
20. Protein A Sepharose.
21. “Mini-Gel” Solution: 1M Tris, 1M Bicine (2.0 mL), 50% w/v acryla-
mide (2.5% w/v his-acrylamide) (4.0 mL), 1.5% w/v (NHJ2 S,O, (0.4
mL), 10% w/v SDS (0.2 mL). Make up to 20 mL with distilled water.
22. “Mini-Gel” Running Buffer: 1M Tris, IM Bicine (2.8 mL), 10% w/v
SDS (1.4 mL). Make up to 140 mL with distilled water. Store at 4OC.
23. “Mini-Gel” Sample Buffer: Sucrose (1 g), 1M Tris, 1M Bicine (0.2 mL),
10% w/v SDS (1 mL), Zmercaptoethanol(O.25 mL). Made up to 3 mL
with distilled water and add 0.001% w/v bromophenol blue. Store at
-2OOC.
24. Coomassie Blue R Stain: Add Coomassie brilliant blue R (0.025 g) to
methanol (50 mL) and stir for 10 min. Add distilled water (45 mL) and
glacial acetic acid (5 mL). Use within 1 mo.
25. Destain Solution: Glacial acetic acid (7.5 mL), methanol (5 mL). Make
up to 100 mL with distilled water. Store at room temperature.
26. TEMED.
27. High and Low Molecular Weight Markers.
All reagents should be of Analar grade.

3. Methods
3.1. Preliminary Purification (Precipitation)
Techniques
Generally, ammoniumsulfateprecipitationis themethodof choicefor
most IgG and some IgM MoAbs; however, euglobulin precipitation is
particularly successful with some, but not all IgM MoAbs. The use of PEG
or less often sodium sulfate can be useful for some MoAbs, and techniques
for these procedures are also described below. Other techniques such as
the use of rivanolor ethanol can be of use in a few cases; however, since they
are not generally applicable to most MoAbs, they will not be considered
further here.
3.1.1. Ammonium Sulphate Precipitation
1. Prepare the ascitic fluid or culture supernatant for fractionation by
centrifugation for 20 min at lOOOOg,,(seeNote 1).
650 Baines, Gearing, and Thorpe

2. Cool the antibody containing fluid on ice (or to 4”C), and stir slowly
using a magnetic stirrer.
3. Add saturated ammonium sulfate solution (equilibrated to 4°C) drop-
wise to give a 35-45% final saturation, i.e., for 45% saturation, add 8.2
mL saturated ammonium sulfate for each 10 mL of antibody contain-
ing fluid. Alternatively, add solid ammonium sulfate to give desired
saturation. (45% saturation = 2.7 g ammonium sulfate/l0 mL fluid.)
Stir in the cold for 2-4 h.
4. Centrifuge for 15-20 min at 20004000gaV at 4OC and discard super-
natant. Drain pellet until dry.
5. Dissolve the precipitate in lo-20% of the original volume of PBS (or
other buffer) by agitation with a spatula or drawing repeatedly into a
wide gage Pasteur pipet. Make up to 25-50% of original volume and
dialyze against 100 vol of PBS (or other buffer) with repeated buffer
changes overnight at 4OC. Alternatively, the precipitate may be stored
at 4 or -20°C if not immediately required (seeNotes 2-4).
3.1.2. Precipitation with PEG (See Note 5)
1. Prepare a 20% w/v solution of PEG 6000 in PBS and cool it to 4OC.
2. Prepare the ascitic fluid or culture supernatant for fractionation.
3. Cool the antibody solution on ice, and add an equal vol of 20% PEG
solution with slow stirring. Stir on ice for 15-30 min. This works well
for most IgG MoAbs; if the preparation is very impure, reduce the
amount of PEG added. For IgM MoAbs, less PEG is needed and a final
concentration of 4-6% PEG works in most cases.
4. Centrifuge for 30 min at 2OOOg,,at 4°C and discard the supernatant.
Drain the pellet until dry and resuspend. Dialyze against PBS (or other
buffer) as described for the ammonium sulfate fractionation.
3.1.3. Euglobulin Precipitation of IgM (See Note 6)
1. Prepare the ascitic fluid or culture supernatant for fractionation.
2. Dialyze MoAb solution against lo-50 vol of 2 mA4 sodium phosphate
pH 6.0 at 4OC, changing the buffer frequently until a precipitate de-
velops.
3. Centrifuge for 10 min at 2000gaVat 4”C, discard the supernatant, and
drain the pellet. Resuspend carefully in PBS or other suitable buffer.
3.2. Chromatography Techniques Based on Size
or Charge Separation
3.2.1. Purification of IgM MoAbs by Gel Filtration
Gel filtration is not particularly effective for the purification of IgG
MoAbs, since they tend to elutein a rather broad peak that is normally quite
Murine Monoclonal Antibodies 651

heavily contaminated with albumin (derived from dimeric albumin, M,

135,000). For this reason, gel filtration for isolation of IgM, which can be a
very useful purification procedure for IgM MoAbs, will be the only gel fil-
tration procedure described in this section (seeNote 7).
1. Carefully pack a column of Sephacryl S-300 (seemanufacturer’s in-
structions and 2). Pump through with column buffer (PBSA) at 20-30
mL/h.
2. Carry out a preliminary purification of approximately 10 mL ascitic
fluid by euglobulin precipitation, PEG precipitation, or ammonium
sulfate fractionation as described above.
3. Pump partially purified MoAb preparation onto the column at 20-30
mL/h and collect approximately 6 mL fractions. Monitor absorbance
at 280 nM. The void peak will contain the IgM (seeFig. la).
4. Check purity of fractions and pool fractions accordingly.
3.2.2. Ion-Exchange Chromatography (See Note 8)
Ion-exchange chromatography has been widely used as a method for
the purification of immunoglobulins from both polyclonal antisera and
preparations containing high concentrations of MoAb. Anion exchange
groups, e.g., diethylaminoethyl (DEAE) and cation exchange groups (5),
e.g., carboxymethyl (CM), have both been used coupled to a cellulose
support for the fractionation of serum proteins, although DEAE cellulose
is the most commonly used matrix for this purpose. DEAE has also been
covalently linked to a beaded crosslinked agarose support together with
the dye cibacron blue F3GA (DEAE Affi-gel blue). In this case,the blue dye
displays a differential affinity for a number of serum proteins and has been
used as a method for the purification of immunoglobulins (4).
3.2.3. Preparation and Equilibration of Ion-Exchanger
1. Gently stir the ion-exchanger into approximately five times its swol-
len volume of 0.5M HCl (anion exchanger) or 0.5M NaOH (cation ex-
changer). Leave for 30 min at room temperature with occasional
swirling.
2. Filter out the resin by suction through a Buchner funnel with a
WhatmanNo. 54 filter paper. Wash the resin cake with distilled water
until the pH of the filtrate is greater than 4 (after acid) or less than 8
(after base).
3. Gently stir the ion-exchanger into 5x its swollen volume of 0.5M
NaOH (anion exchanger) or 0.5M HCl (cation exchanger). Leave for
30 min at room temperature with occasional swirling.
4. Repeat step 2.
5. Add ion-exchanger to an equal volume of 10x concentrated starting
652 Baines, Gearing, and Thorpe

20

0 41 I
34
I
40
1
46
I
52
1
58
I
64
I
70
FRACTION NUMBER
Fig. la. Purification of IgM MoAb by gel filtration using Sephacryl S-300. Ascitic
fluid containing an IgM MoAb was partially purified by ammonium sulfate precipita-
tion. The crude IgM obtained from 9 mL of ascitic fluid was applied to a Sephacryl S-
300 column (85 x 2.8 cm). The IgM elutes near the void volume (fractions 3545).
buffer, e.g., Tris-HCl, pH 8.0,0.05M (or other buffer). Mix thoroughly
and leave at room temperature for 30 min.
6. The ion-exchanger will absorb some buffer ions in exchange for pro-
tons or hydroxyl ions and hence alter the PH. The pH must be restored
to the original value by gently stirring the slurry, and adding the acid
or basic forms of the buffer (HCI for an anion exchanger and IM Tris
for a cation exchanger).
7. Leave the slurry at room temperature for another 30 min and recheck
its PH. Adjust it if necessary. The ion-exchanger is now at the required
pH with the counterion bound but the ionic strength will be too high.
8. Wash the resin with 5x its volume of starting buffer through a
Whatman No. 54 filter paper.
Murine Monoclonal Antibodies 653

9. Degas the slurry using a Buchner side-arm flask under vacuum for 1
h with periodic swirling.
10. Suspend the resin in about 5x its volume of starting buffer and leave
to stand until the majority of beads have settled. Remove the fines by
aspirating the supernatant down to about twice the settled slurry
volume.
11. Carefully pack a clean column by first filling it with 10 mL starting
buffer and then by pouring the resuspended slurry down a glass rod
onto the inside wall until the column is filled. Allow to settle under
gravity.
12. Wash the column with starting buffer at the operating temperature
until the pH and conductivity of the eluant is exactly the same as the
starting buffer.
3.2.4. Sample Application and Elution
1. Apply the precipitate ascitic fluid sample to the column (seeNote 9),
and wash the ion-exchanger with two column volumes of starting
buffer. Collect fractions and monitor eluant for absorbance at 280 nM.
2. Using a commercial or homemade gradient maker, elute the proteins
with a gradient of increasing ionic strength. For example, 200 mL of
Tris-HCl (pH 8.0; 0.05M) in chamber A of the gradient maker and 200
mL of Tris-HCl (pH 8.0; O.OSM>containing 0.3M NaCl in chamber B
of the gradient maker.
3. Collect fractions of about 5-mL volume and monitor for absorbance at
280 nM. The first major peak of the elution profile will contain IgG
antibody if DEAE-resin is used. If cationic exchangers are used, the
position of the MoAb peak must be determined empirically.
4. Pool the protein containing fractions determined by absorbance at 280
nM, and dialyze against PBS (or other buffer).

3.3. HPLC and ReZated Techniques

Fractionation of proteins by HJ?LC (3) utilizes the same principles as
chromatography in standard columns by gel filtration, ion-exchange or
affinity chromatography. Therefore, in HPLC, separation occurs in a
column of small cross-sectional area containing the chromatographic
material in the form of very fine particles (the stationary phase). The
solvent (buffer) or mobile phase is pumped through the column using
medium to high pressure pumps. This enables the sample molecules in
the mobile phase to interact reversibly with the stationary phase in a
continuous fashion.
654 Baines, Gearing, and Thorpe

The advantages of HPLC over conventional chromatographic techni-

ques are speed because of the small high capacity columns, improved re-
producibility because of the sophisticated pumps and accurate timers, and
in many cases increased resolution because of the fine resins and control
systems. The technology of the resins in the stationary phase has advanced
considerably in recent years, and the entire fractionation process can be
performed very quickly indeed. Fast Protein Liquid Chromatography
(FPLC) is a variant of HPLC, which has proved useful in the purification
of mouse MoAbs (6 and Note 10).
3.3.1. FPLC Purification of Mouse IgG MoAbs
1. Prepare ascitic fluid by ammonium sulfate precipitation. In the final
step, use triethanolamine buffer (pH 7.7; 0.02M) to redissolve and dia-
lyze exhaustively overnight at 4OCusing 100 vol of this buffer. Filter
sample before use (0.2 cln/r).
2. Assemble FPLC system according to manufacturer’s (Pharmacia) in-
structions for use with a Mono-Q anion exchange column.
3. Equilibrate the column using the Mono-Q ion-exchange equilibra-
tion method, which is programmed into the liquid chromatography
controller. This takes about 13 min.
4. Wash through the sample injection loop with buffer A (triethanol-
amine, pH 7.7; 0.02M).
5. Load the sample injection loop with the filtered sample. (Usually up
to 500 PL can be loaded using the appropriate loop.)
6. Set the sensitivity on the UV monitor control unit to 0.2 and zero the
chart recorder, (The setting will vary depending on concentration of
sample; try 0.2 to start with and adjust as required.)
7. Run the Mono-Q ion-exchange MoAb purification method, which is
programmed into the liquid chromatography controller. Collect frac-
tions (automatic, seeFig. lb). This takes about 1 h. Elution buffers are
buffer A (triethanolamine, pH 7.7; 0.02M) and buffer B (buffer A con-
taining 1M NaCl).
8. Wash the Mono-Q ion-exchange column with NaOH and NaCl
(methods programmed).
9. Store column in 0.02% NaN, (method programmed).

3.4. Affinity Chromatography

Purification of MoAbs by affinity chromatography is a very powerful
method that requires little specialized equipment. Essentially all proce-
dures require a solid matrix to which ligands can be covalently coupled.
Murine Monoclonal Antibodies 655

-41 ,/ I I 1
20 30 40 50
TIME(mins3

Fig. lb. FPLC purified MoAb elution profile. A typical elution profile for an FPLC
purified MoAb. This profile represents the purification of an antiprolactin MoAb by the
authors. The first major peak is IgG, which was eluted after about 30 min.
Ligands are chosen that display specific and reversible binding to some
portion of the antibody molecule (seeNotes 11-13).
3.4.1. Activation of Sepharose with Cyanogen Bromide
1. Wash 10 mL of Sepharose 4B by vacuum filtration on a sintered fun-
nel with 1 L of water, and then resuspend it immediately in 18 mL of
water in a 50 mL beaker.
2. Add 2 mL of 0.5M sodium carbonate buffer pH 10.5 and a magnetic
flea. Place in a fume cupboard on a magnetic stirrer with a pH elec-
trode in the solution.
3. CAUTION: Dispense 1.5 g cyanogen bromide (CNBr) into an airtight
container. Weigh in fume cupboard, and wear gloves. (CNBr is ex-
 Baines, Gearing, and Thorpe

tremely toxic by inhalation or by skin contact.) Contaminated equip-

ment should be immersed in 1MNaOH overnight in the fume hood to
neutralize the CNBr.
4. Add the CNBr to the gently stirred Sepharose. Maintain the pH
between 10.5-11.0 by dropwise addition of 4M NaOH until all the
CNBr is dissolved and the pH stabilizes (lo-15 min). If the pH rises
above 11.5, the Sepharose must be discarded.
5. Wash themixtureon a sintered glass or Buchner funnel with 2L of cold
O.IM sodium citrate pH 6.5. CAUTION: The filtrate contains CNBr,
so discard it carefully. Do not allow it to dry out. Ifusing commercial
CNBr Sepharose, swell 3 g of powder and wash in 600 mL lO3M HCI
for 15 min on a sintered glass filter.
6. Quickly add the cake of activated Sepharose to the protein solution
(2-10 mg/mL) in O.lM citrate buffer pH 2.5. Mix it gently on a rota-
tor overnight at 4OC. NOTE: This is a relatively inefficient pH to
couple at, but results in much less deformation of the proteins and
hence better affinity ligands.
7. Block remaining active sites with 2M ethanolamine over 1 h.
8. Calculate the percentage of protein unbound by measuring the absor-
banceat280 nMof asampleof thesupernatant. Normally less than5%
will be unbound.

3.4.2. Purification of MoAbs on Afinity Columns

1. Preelute column with dissociating buffer, e.g., O.lM glycine-HCl pH
2.5, followed by PBS.
2. Slowly apply sample containing MoAb, and allow to equilibrate for
15-30 min.
3. Elute unbound proteins with PBS until no protein is detected by ab-
sorbance at 280 nM.
4. Disrupt ligand-MoAb complex with dissociating buffer and collect
peak of protein (MoAb) into IM Tris (120 pL/l mL fraction). Dialyze
against PBS to remove glycine-HCl.
5. Wash column in PBSA. Store at 4OC.

3.4.3. Protein A Column Purifications of IgG

1. Preelute a column containing 1 mL of protein A Sepharose with 5 mL
of O.IM phosphate buffer pH 8.0,2 mL citric acid, and 5 mL of O.lM
phosphate buffer pH 8.0. A 1 mL column should contain about 2 mg
protein A, which can retain 20 mg IgG (2 mL ascites or 500 mL culture
supernatant).
Murine Monoclonal Antibodies 657

2. If ascites fluid, adjust pH to 8.0 with 2M Tris base, add equal volume
of O.lM phosphate buffer pH 8.0, and apply to the column at 5 mL/h.
Collect 0.5 mL fractions and monitor the absorbance at 280 nM.
3. Wash with 5 mL of O.lM phosphate buffer pH 8.0 IgM. IgG, or IgA
antibodies are eluted at this stage.
4. Elute with 5 mL O.lM citrate buffer pH 3.5 to obtain all other IgG sub-
classes. Fractions should be collected into 60 PL 1M Tris. The protein
peak should be pooled and dialyzed against PBS to give the final anti-
body solution.
3.5. Assessment of Purity by Sodium Dodecyl Sulfate
Polyacrylamide Gel Electrophoresis (SDS-PAGE)
After the purification procedure, it is then necessary to obtain some
index of sample purity. The simplest method for assessing the purity of a
MoAb is by SDS-PAGE (seeNote 14). Although “full size” slab gels (see
Figs. 2 and 3) can be used with discontinuous buffer systems and stacking
gels, the use of a “mini-gel” procedure is much easier, quicker and is per-
fectly adequate for assessing purity, monitoring column fractions, and so
on (seeFig. 4).
1. Prepare the sample buffer.
2. Adjust the MoAb preparation to I mg/mL in O.lM Tris, O.lM Bicine.
3. Mix the sample in the ratio 2:l with the sample buffer.
4. Heat at 100°C for 24 min.
5. Prepare gel solution and running buffer as described in Materials.
6. Assemble the gel plates and “mini-gel” apparatus.
7. Add 50 PL TEMED to 10 mL of gel solution, and use this to set the
trough in the “mini-gel” apparatus. Leave for 10 min.
8. Add 30 PL TEMED to the remaining 10 mL of gel solution, pour the gel
into the gel plates, and set the comb. Leave for 10 min.
9. Remove comb and locate gel plates into the electrophoresis tank. Add
running buffer.
10. Load the sample prepared as described in 14 (30-50 l.tL per track) and
run standard mol wt markers (10 PIJtrack of manufacturer’s recom-
mended stock solution) in parallel.
11. Run at 150 V for 1.5 h.
12. Remove gel, and stain in excess Coomassie blue R stain for 2 h (gently
rocking) or overnight (stationary).
13, Pour off the stain and rinse briefly in tap water.
14. Add excess destain and 2-3 pieces of sponge to the gel. Leave until de-
staining is complete (usually overnight gently rocking).
658 Baines, Gearing, and Thorpe

ABC D E

Fig. 2. Polyacrylamide gel electrophoresis of Protein A purified IgG, MoAb. The

separating gel contained a gradient of 5-15% total acrylamide and was run under re-
ducing conditions. Track A shows mol wt markers (from top to bottom: myosin heavy
chain M, 200 kdaltons; b-galactosidase M, 116 kdaltons; phosphorylase b M, 97.4 kdal-
tons; bovine albumin Mr 66 kdaltons; egg albumin Mr 45 kdaltons and carbonic anhy-
drase M, 29 kdaltons). Tracks B and C show purified MoAb and tracks D and E show
hybridoma culture supematant containing MoAb.

4. Notes
1. Crude preparations of MoAbs, whether ascitic fluid or cell culture
supernatant, normally contain insoluble material that must be re-
moved before MoAb purification is commenced. This should be
removed by centrifugation (20-30 min at 1OOOOgJ. Ascitic fluid often
contains a considerable amount of lipid that sometimes forms a pelli-
cule after centrifugation and can be removed with a spatula. Alterna-
Murine Monoclonal Antibodies 659

Fig. 3. Affinity purification of two IgM class MoAb by antigen columns. The figure
is a Coomassie blue stained SDS polyacrylamide gel run under reducing conditions.
Tracks A and C show proteins present in ascitic fluid containing two different IgM
MoAbs directed against human IgM. Tracks Band D show the low pH eluate after ap-
plying the ascitic fluid to a polyclonal human IgM affinity column.
660 Baines, Gearing, and Thorpe

Fig. 4. SDS-PAGE of FPLC purified MoAbs. SDS “mini-gel” electrophoresis analysis

of an FPLC purified mouse MoAb stained with Coomassie blue R. The left-hand track
shows standard mol wt markers (see Fig. 2 for details). The remaining tracks show dif-
ferent loadings of the same MoAb. Characteristic heavy chains (50,000) and light chains
(22,000) are clearly seen free of contamination by other proteins.

tively, and if lipid contamination is a par titular problem, the addition

of silicon dioxide powder (15 mg/mL) followed by ten-trifugation (20
min; 2OOOgJ readily eliminates most of the fatty material present in
such ascitic fluids (2).
2. Crude preparations of MoAbs (a&tic fluid or culture supernatant)
should be stored at -20°C and repeated freeze-thawing avoided. Puri-
fied IgG MoAbs should be aliquoted and stored at -20 or -7OOC. Pure
IgM antibodies can be stored frozen, but some are better stored at 4 or
-2OOC in 50% glycerol, i.e., not frozen. Some MoAbs can be freeze
Murine Monoclonal Antibodies 661

dried with no apparent loss in antigen binding properties. If highly

purified it is often necessary to add carrier protein, e.g., 0.5% bovine
or human serum albumin. It should be remembered that all MoAbs
are different, and so the most efficient storage conditions must be
found empirically-some unpurified MoAb containing culture su-
pernatants can be stored at 4OCfor 1-2 yr with no apparent adverse
effects. If MoAbs are stored at 4”C, then 0.02-0.1% sodium azide or
0.01% thimerosal should be added to prohibit bacterial growth.
3. Precipitation of antibodies with ammonium sulfate is one of the most
useful purification procedures. The technique is particularly ap-
plicable to murine MoAbs since it is fast, can be carried out in the cold,
and is efficient. It produces a fairly pure product, particularly if re-
peat precipitation with 40% saturation is used and leads to minimal
denaturation of antibody in most cases. Dependent upon whether
purity or yield are most important, precipitation is carried out using
35-45% saturation with ammonium sulfate. The former will produce
a relatively pure immunoglobulin preparation but usually leads to
some loss of MoAb, whereas the latter enables almost total recovery of
most antibodies but this will be contaminated with other proteins,
mainly transferrin and some albumin. Since the latter proteins are eas-
ily removed by further purification procedures, precipitation with
45% saturated ammonium sulfateis theideal starting point for further
purification. The desired saturation with ammonium sulfate can be
obtained either by addition of solid salt, or by addition of a saturated
solution. For most MoAbs, the latter is preferable because it generally
causes less denaturation. The saturated solution should be prepared
by adding excess ammonium sulfate to distilled water and stirring
overnight at 4°C. The solution, which should be in contact with solid
salt, must be stored at 4OC.In some cases,adjusting the pH to 7.0 using
sodium hydroxide has been advised, but the authors have not found
this to be advantageous.
4. Precipitation of human and rabbit serum with sodium sulfate pro-
duces a fairly pure IgG preparation in good yield with little denatur-
ation. However, fractionation of mouse or rat ascitic fluid with this
salt is generally less successful, and both yield and purity are almost
always lower than for human or rabbit serum. Since the procedure is
carried out at 25-37OC, some denaturation of the less stable mouse or
rat IgG may also occur. The fractionation is performed as for ammo-
nium sulfate precipitation except that it is carried out at all stages
(including centrifugation) at 25°C and solid sodium sulfate is added
to give 15-18% saturation (18% sat = 1.8 g salt/l0 mL fluid). A pre-
 Baines, Gearing, and Thorpe

cipitation time of 30-60 min is normally used. It is emphasized that a

pilot scale fractionation should be carried out to determine the effec-
tiveness of this procedure for purification of MoAb, and it is essential
to check purity, yield, and whether extensive denaturation has occur-
red. The procedure is not suitable for fractionation of cell culture
supernatants in most cases.
5. Precipitation of antibody with PEG (2) is a procedure well known to
those working with liquid-phase radioimmunoassay. It can be a very
effective purification procedure for MoAbs and is particularly mild
resulting in very little denaturation. Although it usually works well,
a pilot scale experiment should always be carried out to check for un-
expected problems. The procedure is suitable for both IgG and IgM
MoAbs, although the percent of PEG required is different.
6. Although IgM MoAbs can be fractionated by the above techniques,
the preparation obtained is usually impure (ascitic fluid fractionated
in this way will contain much IgG), and is often denatured and/or
difficult to redissolve. Many (but not all) IgM antibodies can be pre-
cipitated by dialysis against very low salt concentrations, which
effectively removes most salts from the solution resulting in precipi-
tation of the macromolecular IgM. This is the basis of euglobulin pre-
cipitation, and again pilot scale studies are necessary to establish the
efficiency and efficacy of this procedure for particular MoAbs. Some
IgM MoAbs are not precipitated by this process, and others are
precipitated but will not redissolve or become denatured. If the pro-
cedure does not produce precipitation, try precipitation with ammon-
ium sulfate (40% saturated) or 6% PEG 6000.
7. The macromolecular size of IgM (Mr about 9 x 105-106)causes MoAbs
of this class to elute in the void volume of many molecular seive gels,
e.g., Sephadex G-200, Bio gel P300, Sephacryl S-300. Since the purity
of the IgM MoAbs will be compromised if the full fractionation range
of most gels is not obtained, it is usually necessary to use well-con-
structed commercially available columns equipped with flow adap-
tors (Pharmacia, LKB, and so on). Although Sephadex G-200 is often
used for the preparation of IgM, this gel is rather fragile and the
authors prefer to use the more robust Sephacryl S-300. Alternatively,
Bio-Gel P300, Ultragel AcA 22, or Sepharose 6B can be used. It is
important not to overload the column or to use too fast a flow rate.
Sepharose 4B or 6B and gels of the Ultragel A series have fractionation
ranges such that IgM is within the exclusion volume and will elute
after the void volume. This is usually not advantageous for purifica-
Murine Monoclonal Antibodies 663
tion of IgM MoAbs, but can be of use in some cases. For a detailed dis-
cussion of gel filtration, see manufacturer’s literature and (I). It is
usual to carry out preliminary purification using a precipitation tech-
nique; euglobulin, PEG, or ammonium sulfate precipitation as de-
scribed are the most appropriate.
The precipitate is redissolved in column buffer at l-3 mg/mL, and
this is pumped directly onto the column. The buffer for fractionation
is relatively unimportant, and the authors have used PBSA success-
fully for several different IgM MoAbs. Buffers should contain at least
100 mA4 sodium chloride or a similar salt to avoid binding of the pro-
teins to the gel matrix. Fractions containing the IgM often appear
opalescent, and it is best to avoid further manipulation of the purified
MoAb, e.g., by concentration or further precipitation. A procedure
using Sephacryl S-300 is described and the profile obtained in the
authors’ laboratory is shown in Fig. la.
8. There are a number of criteria to consider when setting up an ion-
exchange chromatography sys tern, e.g., the choice of ion-exchanger,
matrix support, buffer, and column. The most useful ion-exchange
matrix for the purification of IgG antibody (only MoAbs of this class
can usually be successfully purified by ion-exchange chromatog-
raphy) is DEAE cellulose, of which there are several variants. The
DEAE reactive group was originally coupled to fibrous cellulose
(Whatman, Biorad), which resulted in poor flow rates and clogging of
the matrix. It appears that coupling the DEAE to beaded cellulose
(DEAE-Sephacel, Pharmacia) results in a matrix with a higher cap-
acity and greater resolving potential than the fibrous forms. In gen-
eral, an ion-exchange matrix will need to be regenerated and swollen
before use according to the manufacturer’s instructions. Cellulose ex-
changers should be regenerated before and after use, which will
ensure that they are washed free of residual protein, and protons or
hydroxyl ions left bound weakly to the charged groups of cation or
anion exchangers, respectively. DEAE-Sephacel, however, is sup-
plied preswollen and ready to use.
When choosing a buffer, the pH should be within the operational
range of the ion-exchange matrix (pH 2-12 for DEAE-Sephacel) and
suitable for use with the protein in question, in this case IgG. For a
protein to bind to the ion-exchange matrix, the pH should be approxi-
mately one unit above its isoelectric point for anion exchangers (one
unit below for cation exchangers). The ionic strength of the applica-
tion buffer should be between O.Ol-0.05A4, and for reproducible
 Baines, Gearing, and Thorpe

fractionations, the water usedin the buffers should be as pure as possi-

ble. Furthermore the addition of charged bacterial inhibitors, such as
sodium azide or thimerosal, should be avoided.
When choosing a column for ion-exchange chromatography, there
are a few rules to follow. The column should be of controlled diameter
glass with as low a dead space as possible. Column length should be
20-30 cm and internal diameter 1.5-l .6 cm. Homemade columns are
quite acceptable for most purposes, although commercial columns are
available from several manufacturers. The column must be packed
with the regenerated or ready to use ion-exchanger and equilibrated
fully with the chosen starting buffer before use. Once the column is
ready for use, the sample can be applied after preliminary purifica-
tion, e.g., by ammonium sulfate precipitation and exhaustive dialysis
against the starting buffer. The sample can then be applied to the
column (volume is not an important consideration) and the eluant
monitored for its absorbance at 280 nM, either manually or by passing
the eluant through the flow cell of a UV absorbance monitor. If there
is a high concentration of proteins in the eluant, then the ion-ex-
changer or sample were not fully equilibrated or the absorbing capac-
ity of the matrix has been exceeded. After the sample has been
applied, the ion-exchanger is washed with two column volumes of
starting buffer to ensure complete elution of any unbound protein.
The bound proteins are then eluted by changing the pHof the buffer
or increasing its ionic strength. In general, variation in ionic strength
is easier to control than variation in pH, and hence, this method is
preferable; a final ionic strength of 1.0 is usually enough to elute most
proteins. There are two ways of altering the ionic strength, either by
increasing the concentration of the starting buffer or by increasing the
concentration of other ions, e.g., NaCl. In the latter case, the buffering
capacity and pH is kept constant throughout the separation procedure.
The entire process can be automated by using a commercial gradient
maker and fraction collector, although a laboratory made gradient
maker is perfectly adequate. In general, the total volume of the gradi-
ent should be between five and ten times the column volume and the
size of collected fractions should be about 1/5-l /lO of the column
volume. The elution profile can be plotted by recording the optical
density at 280 nM of the eluted fractions and the first major peak
contains the IgG antibody. IgG of different isotypes will elute at
different times in the gradient, e.g., IgG, elutes later than IgG,,.
9. Although ion-exchange chromatography works very well for rabbit
IgG isolated from serum, it does not seem to be as effective in purify-
Murine Monoclonal Antibodies 665

ing mouse IgG from either serum or samples containing MoAbs. One
problem is that some immunoglobulins are unstable at low ionic
strength, e.g., mouse IgG,, and hence, precipitation occurs during the
sometimes lengthy ion-exchange procedure. Similarly, low ionic
strengths have been reported to be deleterious to antibody activity. A
number of studies have shown that DEAE cellulose chromatography
after ammonium sulphate precipitation does not remove contam-
inating protease and nuclease activity. Furthermore, transferrin is not
well separated by ion-exchange methods, and a further purification
step involving gel filtration often has to be employed (6). The use of
hydroxyapatite (hydroxylapatite) has been described for purification
of MoAbs, but the authors have not found this to be very effective in
most cases. Coupled with the need for long equilibration and recycl-
ing procedures inherent in ion-exchange techniques and the fact that
only certain MoAbs will have a suitable charge for use in this method
(all MoAbs are by definition different), conventional ion-exchange
chromatography cannot be considered a rapid and efficient general
method for the purification of mouse MoAbs. Sometimes (but not
always) the use of an alternative support gel, e.g., DEAE Sepharose
improves the efficiency of the procedure.
10. The FPLC system may be used in a fully automated mode, a manual
mode, or a combination of both. The details of the FPLC hardware and
its use are comprehensively described in the manufacturer’s man-
uals. Briefly, however, the system consists of two high-precision
pumps connected to mixer for gradient formation, an automatic
motor valve for injection of samples (samples are delivered manually
to the valve via a loop with a syringe attachment), a single-path UV
monitor (and control unit) for monitoring samples, and a program-
mable liquid chromatography controller for the full or partial
automation of the FPLC system. A fraction collector and chart re-
corder complete the hardware requirements together with the col-
umn of choice, e.g., ion-exchanger, gel filtration, and so on. Simple
and rapid purification of mouse IgG MoAbs may be achieved using
the above apparatus and the Mono-Q HR5/5 anion exchange column.
Several criteria require further consideration. It is essential that all
buffers and the sample to be loaded are filtered prior to use (0.2 JJM
millipore). All methods for column equilibration, elution of sample,
and so on, are programmed into the liquid chromatography con-
troller. The programs have to be put in by the user and the parameters
of each program can be varied, e.g., flow rate, gradient times, and so
on. In practice, for purification methods, we have found that a flow
 Baines, Gearing, and Thorpe

rate of 0.5 mL/min with a gradient of 35% buffer B for 2 min followed
by a gradient of 100% buffer B for 8 min is adequate for rapid and
single-step purifications of IgG MoAb from ascitic fluid. Although the
method itself takes about 1 h to run, the MoAb is eluted after about 25
min, thus equilibration of column to the elution of the MoAb takes
only about 40 min. A typical elution profile of mouse IgG MoAb is
shown in Fig. lb. It is often useful to run a dilute sample of a&tic fluid
preparation by the method outlined above in order to ascertain the
sensitivity required in the system. Assuming ascitic fluid prepara-
tions contain about 5-10 mg/mL of MoAb, a 1 /lO dilution is appro-
priate.
11. In affinity chromatography techniques, the solid matrix used must
have a loose porous mesh to allow good flow rates, should be chem-
ically and physically stable, and should have chemically derivatized
groups for interaction with ligands. For most protein ligands, agar-
ose beads such as Sepharose 4B meet the requirements for MoAb puri-
fications. Sepharose can be activated by cyanogen bromide (CNBr) at
high pH and subsequently linked to the amino groups of proteins at
neutral pH to yield a stable covalently linked complex. For some small
ligands, it is sometimes better to couple to the gel matrix via a spacer
arm that allows more efficient interaction with the MoAb. A number
of spacer arm derivatized gels are available commercially. Polyacryl-
amide beads may also be used for coupling, but are not as versatile and
will not be discussed further. Sepharose can be activated in the labor-
atory, provided a fume hood is available. The activated Sepharose
formed is unstable and should be used immediately. A more conven-
ient, but more expensive alternative is to use commercially available
CNBr-activated Sepharose, which comes as a fairly stable freeze-dried
powder. Once coupled, the matrix should be packed into simple
columns either made from syringe barrels fitted with a sintered poly-
thene sheet to retain the gel or in commercially available columns.
These procedures can readily be scaled up for large-scale purification.
12. Three basic groups of affinity ligands can be used to purify MoAbs:
class-specific reagents such as antiimmunoglobulin antisera, bac-
terial proteins, such as protein A, C or G, which bind to the Fc portion
of some immunoglobulin classes, and antigens that specifically bind
to individual MoAbs. In addition, certain gel matrices that form
hydrophobic interactions with proteins can be used to purify MoAbs.
The choice of ligand depends on the source of the MoAb, its subclass,
and the degree of purity required. For culture supernatants, if fetal
calf serum or serum-free conditions are used, the MoAb is usually the
Murine Monoclonal Antibodies 667

predominant antibody species present. Less specific methods, such as

protein A or antiimmunoglobulin columns, can then give a very pure
product. For ascitic fluid where other host-derived immunoglobulins
are present, a substantially pure MoAb will only be obtained by anti-
gen-based affinity columns. For many immunochemical applications,
however, it is sufficient to use MoAbs from ascites that have beenpuri-
fied on the basis of class or subclass either by protein A or hydrophobic
chromatography. The immunoglobulin contaminating such prepara-
tions will consist of a diverse range of antigen specificities. The MoAb
would therefore be in great excess over any other antibody species,
and the preparation can be used in practice as a specific reagent.
13. All affinity methods involve a number of basic steps, all of which
should be carried out at 4°C. The column should first be preeluted
with buffer containing the dissociating agent, which will be used to
disrupt the ligand-MoAb complex. For MoAbs, O.lM glycine HCl pH
2.5 is most common, although 3M potassium thiocyanate, high salt,
O.lM sodium borate pH 10, or 0.05M diethylamine pH 11.5 are also
used as dissociating agents. The preelution step purges the column of
any loosely bound ligand that would otherwise contaminate the pu-
rified MoAb. After preelution, the column is washed with running
buffer such as PBS. The sample is then slowly applied and allowed to
incubate for 15-30 min. Unbound proteins are eluted by washing with
several column volumes of buffer until no protein is detected by ab-
sorbance at 280 nM. The MoAb is finally eluted from the column in the
dissociating buffer, and the peak of protein is collected, neutralized,
and dialyzed to remove the dissociating agent. The column should be
washed in buffer containing O.lM sodium azide prior to storage. The
use of a protein A column for the purification of MoAbs of various
subclasses is described. The effect of such a column in purifying an
IgG MoAb from ascitic fluid is also illustrated in Fig. 2, which shows
an SDS polyacrylamide gel separation of proteins in ascitic fluid, and
in the pH 2.5 eluate from the column. Figure 3 shows a similar gel of
an IgM class MoAb directed against human IgM purified from ascitic
fluid on a human polyclonal IgM column. Both of these figures clearly
illustrate the power of affinity chromatography in the purification of
MoAbs.
14. The “mini-gel” is easily and quickly prepared consisting only of a
resolving gel, and takes approximately 1.5 h to run once set up. The
sample (MoAb) is prepared for electrophoresis under reduced con-
ditions and is run in parallel with standard mol wt markers. The sam-
ple is loaded onto the “mini-gel” and run at 150 V for approximately
668 Bakes, Gearing, and Thorpe

1.5 h. The gel is then stained using Coomassie blue R to visualize pro-
tein bands, followed by destaining, after which the gel can be photo-
graphed and/or dried onto filter paper. A typical gel showing three
different loadings of an FPLC purified mouse IgG MoAb is shown in
Figs 3 and 4. Run under the reduced conditions described, the heavy
chains have characteristic mol wt of approximately 50,000 and the
light chains mol wt of approximately 22,000. The gel shows that the
purified MoAbs are free from any detectable, contaminating proteins.
The method described clearly indicates the purity of the MoAb. When
this has been established, the biological activity of the MoAb must
then be considered. Clearly, assays utilizing antibodies are extremely
varied, and researchers will have to devise their own methods for
assessing biological activity of purified MoAbs.

Acknowledgments
This investigation received financial support from the Special Pro-
gram of Research, Development and Research Training in Human Repro-
duction, World Health Organization. We would like to thank Chris Bird
for excellent technical assistance and Miss L. Hudson for typing the manu-
script.

References
1. Johnstone, A. and Thorpe, R. (1987) Immunochemistry in Practice. 2nd Ed. (Blackwell
Scientific Publication, Oxford).
2. Neoh, S. H., Gordon, S., Potter, A., and Zola, H. (1986) The purification of mouse
monoclonal antibodies from ascitic fluid. J. Immunol. Mefhods 91,231-235.
3. Burchiel, S. W., Billman, J. R., and Alber, T. R. (1984) Rapid and efficient purification
of mouse monoclonal antibodies from ascites fluid using high performance liquid
chromatography. J. Immunol. Methods 69,33-42.
4. Bruck, C., Portetelle, D., Glineur, C., and Bollen, A. (1982) One-step purification of
mouse monoclonal antibodies from ascitic fluid by DEAE affi-gel blue chromatog-
raphy. J. Immunol. Methods 53,313-319.
5. Carlsson, M., Hedin, A., Inganas, M., Harfast, B., and Blomberg, F. (1985) Purifica-
tion of in vitro produced mouse monoclonal antibodies. A two-step procedure utili-
zing cation exchange chromatography and gel filtration. J. Immunol. Methods 79,89-
98.
6. Clezardin, P., McGregor, J. L., Manach, M., Boukerche, H., and Dechavanne, M.
(1985) One-step procedure for the rapid isolation of mouse monoclonal antibodies
and their antigen binding fragments by fast protein liquid chromatography on a
mono-Q anion-exchange column. J. Chromatography 319,67-77.
7. Dean, P. D. G., Johnson, W. S., and Middle, F. A. teds.1 (1985) Affinity Chromatogru-
phy u Prucficd Approach (IRL Press Limited, Oxford, Washington, DC).
 Chapter 54

Rat x Rat Hybridomas

Christopher J. Dean
1. Introduction
There is an increasing interest in the preparation of rat x rat hybrid-
omas, because they have been found to be more stable in culture than
mouse hybridomas and they secrete consistently high levels (10 &mL
and above) of monoclonal antibody. In addition, certain subclasses of rat
IgG have been found to interact efficiently with human Fc receptors and
the human complement system. More important perhaps in practical
terms is that, when grown as an ascites, up to 30 mL of fluid containing 1-5
mg/mL specific antibody can be obtained from each rat. Furthermore,
when nude rats are used for ascites production, the levels of endogenous
immunoglobulin can be less than 1 mg/mL.
Two rat myelomas are in general use today, both derived from
ileocecal plasmacytomas of the LOU strain of rats, namely R21O.Y3 Ag
1.2.3, abbreviated Y3 (I), which secretes Kappa light chains, andIR983F (Z),
which is a nonsecretor. Although the IR 983F grows normally in suspen-
sion, the Y3 myeloma is strongly adherent to glass and tissue culture
plastics. To obtain cells that fuse well, it is essential to grow cells in spinner
cultures.
For a general introduction to hybridoma production, the reader is
referred to the articles by Galfre and Milstein (3) and Bazin (4), as well as
the chapters in this volume. Essentially, the procedure is to fuse myeloma
cells that lack hypoxanthine phosphoribosyl transferase (HGPRT) and so

669
670 Dean

cannot grow in media containing hypoxanthine, aminopterin, and thymi-

dine (HAT), with specific antibody producing B-cells that contain the
HGPRT gene. Only stable hybrids containing the HGPRT gene will grow
in HAT selection medium. Perhaps the two most important requirements
for successful hybridoma production are first the ability to produce spe-
cific antibody-producing B-cells in the spleen or lymph node of the im-
mune donor, and second the need for a rapid, sensitive, and specific assay
for the antibody required. The appearance of high titers of antibody in
serum will show that immunization has been successful, but it is not al-
ways possible to elicit at the same time the specific B-cells required for
fusion. In this case, it is better to allow the response to decline (weeks to
months) before reimmunizing just before fusion. Our experience with
proteins and intact cells shows that two immunizations via the Peyer’s
patches 14 d apart followed by cell fusion 3 d later is a useful and quick way
of generating the specific B-cells. We immunize routinely via the Peyer’s
patches because the stimulated mesentericnodes yield ahigher proportion
of IgG-producing hybridomas as well as a proportion of IgA secretors (5).
Rat x rat hybridomas grow rapidly with generation times often sim-
ilar to that of the parental myeloma (lo-12 h), and it is essential to clone cells
from positive wells as soon as possible. We plate fusion cultures into 4 x
24 well plates and pick individual hybridoma colonies from antibody
positive wells by aspiration with a Pasteur pipet. After rescreening the
picked colonies, those producing specific antibodies are cloned when
about lo6 cells have been obtained (equivalent to about two confluent wells
of a 24-well plate). One final point to note is the need to freeze down stocks
of cells as soon as possible. Even the best run laboratory suffers with
contamination at times, and there is nothing more frustrating than loss of
a promising hybridoma.

2. Materials
1. Rats: Any strain may be used, preferably of lo-12 wk of age at com-
mencement of immunization.
2. Immunization: Mammalian cells and membrane vesicles can elicit
good antibody responses without further treatment. Soluble anti-
gens, however, usually require injection as an emulsion in an oil base
(Freund’s adjuvant) to achieve good responses.
a. Live mammalian cells (2-5 x 106/immunization/rat) or mem-
brane vesicles as a suspension in phosphate buffered saline
(PBS) or Dulbecco’s Modified Eagle’s Medium (DMEM).
Rat x Rat Hybridomas 671

b. Macromolecules (e.g., protein, carbohydrates, DNA-methyl-

ated BSA precipitates at XI-100 pg/immunization/rat) in
normal saline or PBSand emulsified just before use, by high-
speed vortexing, blending, or ultrasonication, with an equal
volume of Freund’s adjuvant to give a stable emulsion. Emul-
sion stability can be demonstrated by allowing a drop from a
Pasteur pipet to fall onto water. With a stable emulsion, the
drop contracts and remains floating on the water surface,
whereas an unstable emulsion disperses.
c. Peptides or other low mol wt haptens conjugated to carrier
proteins, such as ovalbumen, keyhole limpet hemocyanin, or
bovine thyroglobulin using glutaraldehyde or carbodiimide.
Solutions or suspensions in PBS should be emulsified with
Freund’s adjuvant before use,
3. Cell fusion:
a. Growth medium: DMEM is prepared from dry powder and
supplemented with 50 IU penicillin, 50 pg streptomycin, 100
u.g neomycin/mL, and then filter sterilized. Medium is
stored at 5-6OC and used within a 2-wk period.
b. PEG solution: Weigh 50 g of polyethyleneglycol(1500 MW)
into a capped 200-mL bottle, add 1 mL of water, and then
autoclave for 30 min at 12OOC.Cool to about 70°C, add 50 mL
of DMEM mix, and after further cooling to ambient temper-
ature, adjust the pH to about 7.2 with NaOH (the mixture
should be colored orange). Store as I-mL aliquots at -20%.
c. HAT selection medium: Prepare 100 x HT by dissolving 136
mg of hypoxanthine and 38.75 mg of thymidine in 100 mL of
0.02N NaOH prewarmed to 60°C. Cool, filter sterilize, and
store at -2OOCin l- and 2-mL aliquots. Prepare 100x A by dis-
solving 1.9 mg aminopterin in O.OlNNaOH, then filter steril-
ize, and store as 2-mL aliquots at -2OOC.HAT medium is pre-
pared by adding 2 mL of HT and 2 mL of A to 200 mL DMEM
containing 20% fetal calf serum.
d. HT medium: Add 1 mL of HT/lOO mL DMEM containing
20% fetal calf serum.
e, Fetal calf serum: Use only batches that have been shown to
support the growth of hybridomas.
f. Feeder cells for fusion cultures (seeNote 1): 3 h to 1 d before
fusion, 2-4 x lo6 rat fibroblasts in 10 mL DMEM and con-
tained in a 30-mL plastic universal are irradiated with about
672 Dean

30 grays (3000 rad) of x or y-rays. Dilute with 90 mL of HT (3

h before fusion) or DMEM-10% FCS (I d before fusion) and
plate l-mL aliquots into four 24-well plates.
g. Y3 Ag 1.2.3 myeloma: About 5 x lo6 cells of Y3, stored frozen
in liquid nitrogen as 1-mL aliquots in 5% dimethyl sulf-
oxide/95% FCS are thawed quickly at 37OC, centrifuged
down (5008 x 2 min), then resuspended in 100 mL DMEM-
10% FCS placed in a 200-mL spinner flask. Stand for 2 d at
37OC to allow cells to attach to base of vessel, and then place
spinner flask on a Bellco magnetic stirrer running at about 160
rpm. Cells can be maintained in exponential growth for up to
1 mo by fourfold dilution daily with DMEM-10% FCS. Mye-
loma cells are ready for use&5 d after commencement of stir-
ring (seeNote 1).
h. IR 983F myeloma: Cells growing exponentially in flasks,
roller bottles, or spinner culture are maintained by daily four-
fold dilution with DMEM-10% FCS. Unlike theY3 myeloma,
these cells do not readily attach to glass or plastic.
i. Freezer medium: Freshly prepared 5% dimethyl sulfoxide/
95% fetal calf serum.

3. Methods
3.1. Immunization for Spleen Cell Fusions
1. Rats are anesthetized and a 0.5-mL sample of blood (preimmune) is
taken from the jugular vein into a 1.5 mL capped microcentrifuge tube.
Subsequently, after clotting at room temperature, the serum is re-
moved, clarified by centrifugation and stored at -20°C.
2. Antigens, either in suspension (cells) or emulsified with complete
Freund’s adjuvant (soluble materials), are injected at five sites (4x sub-
cutaneous and 1 x intraperitoneal) using 0.1 ml/site.
3. Fourteen days later, the rats are again anesthetized, bled (0.5 mL), and
then rechallenged at five sites with antigen using incomplete (no bac-
teria) Freund’s adjuvant where necessary.
4. Repeat immunization at 14 d or longer intervals until test bleeds show
high titers of specific serum antibodies using the preimmune serum as
control.
5. When satisfactory titers of antibody have been achieved, give the rats
an intravenous boost of antigen (no adjuvant), and 3 d later, remove
the spleens for cell fusion.
Rat x Rat Hybridomas 673

3.2. Immunization for Mesenteric Node Cell Fusion

-(See Note 3)
1. Rats are anesthetized, bled from the jugular, and then the lower abdo-
men opened along the central line.
2. Carefully extend the small intestine and locate the 8-16 Peyer’s
patches (strain dependent) that lie along the peritoneal wall of the
small gut.
3. Take up antigen-containing samples (in PBS or emulsified with com-
plete Freundls adjuvant) into a l-mL syringe using a 27-gage needle,
inject between 10 and 15 PL into every other Peyer’s patch to give a
total dose of between 0.05 and 0.1 mL/animal (seeNote 4).
4. To prevent the formation of adhesions, place 2 mL of sterile 0.9% NaCl
(containing antibiotics if necessary) in the peritoneal cavity and close
the abdomen.
5. Two weeks later, test bleed the rats and rechallenge intraperitoneally
with antigen in 0.5 mL PBS or incomplete Freund’s adjuvant.
6. Four weeks after the initial challenge, anesthetize the rats, test bleed,
open the abdomen parallel to the initial incision, and reimmunize us-
ing theunchallenged Peyer’s patches as recipient for antigen. Remove
the mesenteric nodes 3 d later, and use for cell fusion.

3.3. Hybridoma Production

3.3.1. Preparation of Cells for Fusion

1. Centrifuge exponentially growing cells of Y3 (seeNote 1) or IR 983F in
50-mL aliquots for 3 min at 4008, wash twice by resuspension in se-
rum-free DMEM, count in a hemocytometer, and resuspend in this
medium to l-2 x 10’ cells/ml.
2. Kill immune rat by cervical dislocation or CO, inhalation, test bleed,
and open abdominal cavity. Remove mesenteric nodes and spleen by
blunt dissection.
3. Disaggregate nodes or spleen by passing through a fine stainless-steel
sieve (e.g., a fine tea strainer) into 10 mL of serum-free DMEM using
a spoon-head spatula. (Dip it into ethanol and flame to sterilize.)
4. Harvest cells by centrifuging for 5 min at 4008, wash twice in serum-
free DMEM, and resuspend in 10 mL of the same medium.
5. Count viablelymphoid cells in a hemocytometer. Immune rat spleens
generally yield between 3-5 x lo* cells total, whereas the nodes can
yield up to 2 x lo8 lymphocytes.
674 Dean

3.3.2. Fusion Protocol

1. Mix lo8 viable lymphocytes with 5 x 10’ myeloma cells in a lo-mL ster-
ile capped tube, and centrifuge for 5 min at 4008.
2. Pour off the supernatant, and while draining remove any surplus me-
dium carefully with a Pasteur pipet. Release cell pellet by gently tap-
ping the tube on bench.
3. Over a period of 1 min, stir 1 mL of PEG solution, prewarmed to 37OC,
into the pellet. Continue mixing for a further minute by gently rock-
ing the tube.
4. Dilute the fusion mixture with DMEM (2 mL over a period of 2 min,
then 5 mL over 1 min).
5. Centrifuge for 5 min at 4OOg,then resuspend the cells in 200 mL HAT
selection medium, and plate 2-mL aliquots into four 24-well plates
seeded with irradiated fibroblasts.
6. Screen cultures for specific antibody&14 dafter commencement of in-
cubation at 37°C in 5% CO,-95% air (seeNote 5).
7, With a Pasteur pipet, pick Individual colonies into 1 mL HT medium
contained in 24-well plates. Split and feed with HT when good growth
commences. Freeze samples in liquid nitrogen.
8. Rescreen the picked colonies and expand positive cultures. Freeze
samples of these in liquid nitrogen and reclone twice.
3.3.3. Cloning of Rat Hybridomas
1. Prepare a suspension of irradiated rat fibroblasts in DMEM-10% FCS,
and seed at 5 x lo3 cells/well into 96-well plates.
2. The next day, centrifuge down cells from at least two wells of a 24-well
plate that contains a confluent layer of hybridoma cells. Carefully
count the number of cells, and then dilute to give about 50 cells in 20
mL of HT or DMEM-10% FCS.
3. Carefully “flick off” the supernatant medium from the feeder cells,
and plate 0.2-mL aliquots of hybridoma cells into each of the 96 wells.
4. Examine plates 5-10 d later, and screen those wells that contain only
single colonies.
5. Pick cells from positive wells into 24-well plates, expand, and freeze
in liquid nitrogen.
6. Repeat by recloning the best antibody-producing colonies.
3.3.4. Bulk Cultures
Rat hybridomas grow well, but more slowly, in DMEM containing3%
fetal calf serum, and they can be quickly adapted by reduction in concen-
trations of serum over a period of 1 wk. To bulk up cultures in vitro, seed
Rat x Rat Hybridomas 675

about lo6 cells into a 24-cm2 flask containing 10 mL DMEM-10% FCS. Use
cells from this flask to seed an 80-cm2 flask containing 3 mL DMEM and
cells for two 80-cm2flasks to seed an 800-cm2roller bottle or 200-mL spinner
flask containing 100 mL DMEM-10% FCS. Hybridomas produced with
either rat myeloma grow well in both roller and spinner cultures, and anti-
body yields of up to50 pg/mL can be obtained. Cultures are fed daily with
DMEM-10% FCS and on alternate days when DMEM-3% FCS is used.
3.3.5. Ascites
In general, rat x rat hybridomas grow well in the peritoneal cavity of
nude rats, which is an advantage because the levels of endogenous im-
munoglobulin are usually low (<lmg/mL).
1. Inject 5 x 106-lo7 hybridoma cells, taken either from roller/spinner
culture or ob tained from a previously grown ascites and stored frozen,
into the peritoneal cavity of each nude rat.
2. When tumor growth is visible by the swelling of the abdomen, drain
off the ascitic fluid into a syringe containing heparin to prevent clot-
ting. If tumor does not grow as an ascites, use rats that have been in-
jected ip with 1 mL pris tane 7-10 d previously.
3. Centrifuge down cells at 6008 for 10 min, aspirate off the supernatant
(seeNote6), add theproteaseinhibitorphenylmethylsulfonylfluoride
to a concentration of lOaM, and store at -2OOC. Collect the cells into
5% DMSO-95% FCS and freeze in liquid nitrogen.

4. Notes
1. Feeders for fusion culture. A useful source of rat cells that can bemain-
tained in culture can be prepared from the xiphoid cartilage that ter-
minates the xiphisternum. Cartilages from 6-8 adult rats are chopped
into 2-3 mm pieces with a scalpel and stirred for 45 min at room
temperature in 15 mL DMEM containing 0.5% trypsin (bovine pan-
creas type III) and 1% collagenase (type II). Add FCS to 10% and filter
through sterile gauze to remove debris. Wash cells in DMEM-10%
FCS and plate into same medium. Passage in DMEM-10% FCS after
removing cells by incubation for 2-3 min in PBS-0.05% N%EDTA con-
taining 0.25% trypsin. Store cells in liquid nitrogen as aliquots of lo6
cells in 5% DMSO-95% FCS.
2. The Y3 myeloma attaches firmly to glass and plastic surfaces in static
culture, and cells produced in this way do not fuse well. Our exper-
ience suggests that good fusions are obtained only using cells taken
from exponentially growing spinner culture.
676 Dean

3. Surgical and immunization procedures: The mesenteric lymphatic

system can be visualized by injecting a small amount of lymphogra-
phy dye (Guerbet’s patent blue V, May and Baker, Dagenham, Eng-
land) into a Peyer’s patch.
4. Administration of the antigen to the Peyer’s patch is facilitated if the
region of the gut containing the Peyer’s patch is gently squeezed either
with forceps or between thumb and forefinger, so that thehypodermic
needle can be inserted just beneath the capsular membrane of the
Peyer’s patch.
5. The test system selected for screening of hybridoma supernatants,
e.g., RIA or ELISA using cells or antigen-coated multiwell plates or
immunostaining of tissue sections and blots of polyacrylamide gels,
will depend on the antigen used for immunization. We recommend
the use of sheep or rabbit antibodies to polyclonal rat F(ab’>, as the
second stage reagent for the primary screen, because neither protein
A (which binds with high affinity only to rat IgG,) nor the mouse anti-
rat kappa monoclonals (MARK-l and MARK-3) bind to all rat anti-
bodies. If a particular antibody isotype is required, specific antiheavy
chain antisera should be used either at the primary screen or after the
hybridomas have been “picked” from the fusion wells.
6. Purification of rat monoclonals is most easily achieved by affinity
chromatography on antigen linked to Sepharose 4B (seeChapter 53
this volume). Where this is not possible, IgG, and IgGgb antibodies can
beisolated from culturesupernatants and ascitic fluid by precipitation
with (NH,),SO, added to give 40% of saturation and purified by ion-
exchange chromatography on Whatman DE52, Pharmacia Mono-Q,,
or similar product. Alternatively, affinity chromatography can be car-
ried out using Protein A-Sepharose for IgGZCantibodies or Sepharose-
linked antiheavy chain antibodies for other isotypes. Immunoaffinity
columns of MARK-l are a useful general reagent for the isolation of rat
monoclonal antibodies. However, when hybridomas have been made
using the Y3 myeloma, it is important to check the protein prepared in
this way both by SDS-PAGE and for antigen binding because MARK-
1 has a high affinity for the Y3 kappa chain (secreted by most Y3
hybridomas), whereas not all of the rat monoclonal antibodies bind to
MARK-l.

References
2. Galfr6, G., Milstein, C., and Wright, B. (1979) Rat x rat hybrid myelomas and a mono-
clonal anti-Fd portion of mouse IgG. Nature 277,131-133.
Rat x Rat Hybridomas 677

2. Bazin,H. (1982)Production of monoclonal antibodies with the LOU rat nonsecreting

IR 983Fmyeloma cell line, in Profidesoffhe BiologicalFluids, vol. 29 (Peeters,H., ed.),
Pergamon, Oxford, pp. 615-618.
3. Galfre, G. and Milstein, C. (1981) Preparation of monoclonal antibodies: strategies
and procedures, in Methodsin Enzymology,,vol. 73 (Langone, J.J. and van Vunakis,
H., eds.), Academic Press,New York and London, pp. 346.
4. Bazin, H. (1987) Rat x rat hybridoma formation and rat monoclonal antibodies, in
MethodsofHybridoma Formation (Bartal, A. H. and Hirshaut,Y., eds.), Humana Press,
Clifton, New Jersey, pp. 337-378.
5. Dean, C. J.,Styles,J.M., Gyure, L. A., Peppard, J.,Hobbs, S.M., Jackson,E.,and Hall,
J. G. (1984) The production of hybridomas from the gut associated lymphoid tissue
of tumor bearing rats. I. Mesenteric Nodes as a Source of IgG Producing Cells. Clin.
Exp. Immunol. 57‘358-364.
 Chapter 55

Monoclonal Antibodies
Against Glycosphingolipids
(GSLs)-Gangliosides

Bruce Fletcher Clark

1. Introduction
Free GSL molecules are poorly immunogenic using conventional
immunization procedures (1,2). Although immunogenicity has been in-
creased by immunizing with purified glycolipids coated onto the surface
of SaZmone2Za minnesota organisms (3,4), this has not always been success-
ful (5). Theoretically, one can take advantage of GSLs absorbed on silica as
occurs in thin-layer chromatography (t. 1.c.) separation (6), and then to use
the silica-absorbed GSL spots scraped from the plate as immunogen. This
proposition is based on the principle that aggregation and/or presentation
on solid surfaces often leads to increased immune responses (7), and is
supported by the demonstration that silica enhanced the response to
poorly immunogenic proteins in an in vitro system (8).
Most attempts at inducing high-titer antibody responses and mono-
clonal antibodies (mAb’s) of the IgG type involve immunization schedules
consisting of several injections given over a fairly long period. An adjuvant
is often employed, at least for the primary immunization. Spitz et al. (9)
found that intrasplenic immunization had the advantages that a very small
amount of immunogen without adjuvant is required, and that it induced
679
680 Clark

a rapid response, a 34-d period only, being necessary between immuniza-

tion and fusion. Such a short schedule, which is equivalent to the primary
immunization of conventional procedures, is likely to produce antibodies
of the IgM rather than the IgG class. This represents a disadvantage, since
IgG antibodies are often preferred. However, in seeking to improve the
efficiency and speed of producing antibodies against GSLs, this is not con-
sidered a major disadvantage, especially in view of the possibility of class
switching by selection of somatic cell variants that arise spontaneously or
can be induced by mutagens (ZO,Zl). Also, it is possible to administer a
second intrasplenic injection at a later time, thereby increasing the proba-
bility of obtaining antibodies of the IgG class.
A modified version of the technique described by Galfre and Milstein
(12), the principles of which have been outlined earlier in this series (13),
was used to fuse immune mouse spleen cells with the azaguanineresis-
tant, nonantibody secreting, mouse myeloma cell line, NS-1. Hypoxan-
thine-aminopterin-thymidine (HAT) medium was used to select monoclo-
nal antibody-producing hybrid cell lines.
These principles have been successfully applied to the generation of
mAB’s against gangliosides from fetal brain, monosialoganglioside (GMI)
from a commercial source, and total glycolipids from colonic mucosa.

2. Materials
2.1. Stock Solutions
1. CulturemediumRl?MI1640 withantibiotics (penicillinlOOU/mLand
streptomycin 100 pg/mL). Store at 4OC.
2. Fetal Calf Serum (FCS) (Note 1). Inactivated by heating at 56°C for 30
min and dispensed into aliquots of 20 mL (store at -2OOC).
3. Supplements to Media (100x strength):
a. Oxaloacetic Acid (15 mg/mL).
b. Sodium Pyruvate (5 mg/mL).
c. Insulin (100 iu/mL).
d.Tris Base (2&I).
4. For Selection media (100x strength):
a. Hypoxanthine plus Thymidine: Dissolve 0.1361 g and 0.0388 g,
respectively, in double-distilled water at 70-BOOC.
b. Aminopterin: Dissolve 17.6 mg aminopterin in 80 mL water,
adding O.lM NaOH if solution is not readily achieved. Adjust
the pH to 7.2 with HCl and make up to 100 mL.
Gangliosides 681

c. Sterilize solutions of supplements and selection agents by fil-

tration through single-use 0.22~pm membrane filters and dis-
pense into 2 mL aliquots. Store at -2OOC.

2.2. Extraction and Purification

of Glycolipids and Immunization
1. Chloroform: methanol (2:l v/v and 1:2 v/v with 5% water).
2. Aluminium-backed silica-coated high-performance thin-layer chroma-
tography (HPTLC) sheets (Art5553; E. Merck, Darmstadt, West Ger-
many).
3. Authentic GSLs (for use as references) (Supelchem UK, Radley and
Co. Ltd., London, England).
4. Orcinol-ferric chloride reagent.
5. D&isopropyl ether/l-butanol/50 mM aqueous sodium chloride (6:4:5
v/v/v>.
6. Silica gel chromatogram carrying GSLs.
7. 10 mM sodium phosphate buffer, pH 7.2,0.17M NaCl.

2.3. Fusion of NM Cells with Immune Spleen Cell

and Cultivation of Hybridomas
1. P3/NSl/l-Ag4-1 cells (22), and immune spleen cells.
2. Waterbath (made by placing a beaker of water inside another contain-
ing water and kept at 37OC, until needed).
3. 50% solution of polyethylene glycol (PEG, MW, 1540), prepared as fol-
lows: weigh 1 g of PEG in a glass universal bottle and me1t in a micro-
wave oven with the bottle top loose. As the PEG cools, but before it sol-
idifies, add 1 mL RPM1 1640 and mix. Store at 37OCbefore use.

2.4. Cloning, Growing, and Storing Hybrid Cells

1. Glass capillary pipet (made by stretching a 100~PL accupette [DADE
Diagnostics Inc., Aguada, Puerto Rico] in a Bunsen burner flame to a
diameter, estimated visually, about l/4 that of the low power field of
an inverted microscope).
2. 10% analar DMSO (dimethylsulfoxide)/90% FCS. Prepare by adding
DMSO dropwise to FCS with mixing to prevent precipitating pro-
teins. Make up fresh and cool on ice before use.
3. Sterile transfer pipets (Pastette’s, LW4010, Apha Laboratories, East-
leigh, Hants).
682 Clark

3. Methods
3.1. Extraction of Glycolipids
and Separation of Gangliosides
This method is essentially as described by Ladisch and Zi-Liang (14).
1. Homogenize with a Polytron homogenizer 1 g tissue in 1 mL water.
Add 40 mL of chloroform/methanol (2:l v/v>; leave for 2 h, and then
vacuum fiter through sintered glass. Reextract the remaining homo-
genate with 10 mL chloroform/methanol (1:2 v/v plus 5% water).
Pool extracts.
2. Dry extract under N, at 7OOC.
3. Partition the dry extract (30 vol/g tissue) in di-isopropyl-ether/l-bu-
tanoll50 mM aqueous sodium chloride (6:4:5 by vol).
4. Remove the lower phase containing gangliosides, repartition, sepa-
rate, lyophilize, and dissolve in distilled water (1 mL).
5. Remove low mol wt components by Sephadex G-50 filtration (300 x 15
mM column) (15).
6. Lyophilize the void volume, and redissolve the gangliosides in 1.0 mL
chloroform/methanol (I:1 v/v).
7. Centrifuge for 5 min at 10,000 x g at 4°C to remove trace residual
protein.
3.2. Thin Layer Chromatography
of Total GSL’S and Gangliosides
In advance, equilibrate a thin-layer chromatography tank lined with
chromatography paper, overnight at 4OCwith about 100 mL chloroform/
methanol containing 0.2% KC1 (60:35:8, by vol) or, for greater resolution of
gangliosides, with chloroform/methanol/0.2% aqueous calcium chloride
(60:40:9 by vol) at room temperature.
1. Apply chloroform/methanol extracts containing glycolipids (gangli-
osides) equivalent to lo-20 mg wet wt of tissue, or authentic individ-
ual GSLs (l-10 pg), to 0.5-l-cm wide lanes on a chromatogram sheet.
This is best achieved using a short piece of solvent-resistant plastic
tubing attached to a 50-PL Hamilton syringe. (Designate parallel lanes
as reference tracks to be visualized chemically.)
2. Develop chromatogram in appropriate solvent system over a distance
of 8.0 cm. This takes about 45 min.
3. Dry the chromatogram in a fume hood, cut reference lanes from the
sheet, and visualize GSLs with orcinol-ferric chloride reagent as
Gangliosides 683

described on the reagent bottle. (Use an acid-resistant fume extraction

hood, and wear protective gloves and glasses.)

3.3. Immunogen and Immunization Procedure

(Essentially as Described by Spitz et al. 191;
See Notes 2 and 3)
1. Gently scrape silica and associated GSLs from an unstained lane of the
chromatogram with a scalpel blade (Shape 13 x 23), using a parallel
stained lane to locate the region(s) of interest.
2. Collect the silica into a glass Bijou bottle with the aid of a glass funnel
(Note 4).
3. Suspend the silica plus associated glycolipids in sodium phosphate
buffer (10 mg wet wt equivalent of tissue/mL) by using sonication
until homogeneous by visual inspection. Pick up in a 1-mL glass sy-
ringe with a l/2 in x 27-gage needle.
4. Anesthetize adult Balb/c mice by ip injection of 6-7 PL of nembu-
tal/g body wt.
5. Swab the abdomen with 70% ethanol, and place the animal on its right
side.
6. Make a skin incision, and blunt dissect skin from the body wall to ex-
pose the abdominal wall.
7. Incise the body wall to expose the spleen.
8. Lift the lower pole of the spleen and insert the l/2 in x 27-gage needle
deeply into the spleen, and inoculate 50 PL from a I-mL glass syringe
of the suspension of silica and associated GSLs as the l/2 in x 27-gage
needle is pulled out (Note 5).
9, Push back the spleen, suture the body wall with thread, approximate
and close the skin edges with stitches.
lO# Repeat the immunization after 1 wk.

3.4. Production of Hybridomas

In advance, grow NS-1 cells (Note 7) using the following procedure:
1, About 1 wk before fusion, remove and thaw about 4 x 105NSI cells
from liquid nitrogen storage (seebelow).
2, Wash the cells twice by centrifugation in RPM1 1640 for 5 min at 1200
X8*
3. Transfer the cells, in a small volume of medium, to a 750-mL culture
flask with 20 mL RPM1 1640plus 20% FCS and supplements (put 2 mL
684 Clark

of each stock solution of oxaloacetic acid and sodium pyruvate; 40 iu

Insulin and 200 PL 2M Tris Base to 200 mL of medium).
4. Put the flask on its side, with the cap loose, in a humidified CO, incu-
bator (7% CO, in air gas mixture).
5. Feed every 2 d by addition of medium to expand the culture, keeping
the cell density between 104/mL and 2 x 105/mL. (Cultures with den-
sities less than 104/mL grow slowly and lose viability above 2 x 105/
mL.) This should produce about 2 x lo5 cells for fusion, in log-phase
growth at about lo5 cells/mL. Prepare a backup culture in case of mis-
calculation or accidental loss.
Also in advance, prepare and plate out feeder cells (peritoneal macro-
phages) by the following procedure:
1. Twenty-four h before the day of the fusion, sacrifice four, large, ma-
ture, Balb/c mice (enough for eight culture plates), sterilize by immer-
sion in 70% ethanol, blot dry, and put in cell culture hood.
2. Make a mid-line cut along the skin of the abdomen, and then makefur-
ther cuts to allow the skin over the entire abdominal area to be dis-
sected away exposing the body wall.
3. Inject ip (small 23-gage needle) 5.0 mL of RPM1 1640 plus 20% FCS and
supplements (ice-cold to avoid cells sticking to walls of containers),
and gently massage the abdomen for about 1 min.
4. Aspirate the fluid from the abdominal cavity using the same syringe
with the larger IPgage needle pushed through the same hole in the
body wall. This avoids leakage. Take care not to puncture internal or-
gans (visible through the body wall) while manipulating the fluid into
the syringe.
5. Pool the aspirates and make up a volume with cold medium (about 10
mL/plate), so that drops of about 100 PL from a lo-mL plastic dispos-
able pipet can be delivered to each well of 8 x 96-well culture plates.
One mouse provides enough cells for two plates (2-5 x lo4 cells/well).
6. Put the plates in an incubator overnight.

3.5. Fusion of Immune Spleen Cells

with NSl Cells
Three days (Note 8) after the second injection of immunogen animals
are sacrificed and immune splenocytes immortalized by fusion with NSl
cells:
Gangliosides 685

1. Warm supplemented RPM1 1640 with 20% FCS and serum free me-
dium to 37°C.
2. Harvest NSl cells by rubbing the bottom surface of the culture flask
with a bent Pasteur pipet and then repeatedly rinsing gently with the
medium using a lo-mL pipet. Do this slowly to minimize damage to
the cells.
3. Wash twice with serum-free medium by centrifugation in a bench cen-
trifuge at 1200 x g for 5 min, and resuspend in medium.
4. Count viable NSl cells using an improved Neubauer hemocytometer.
View the cells directly using a phase contrast microscope (viable cells
appear surrounded by a bright halo). Alternatively, dilute the cell sus-
pension l/2 with 0.1% Trypan blue in PBS, and examine by light
microscopy (viable cells appear free of stain). The formula for calcu-
lating the number of cells/mL of medium is 2 (if the medium is diluted
with Trypan blue) x number of cells within the outermost triple
boundary lines x 104.
5. Remove the spleen(s) aseptically.
6. Release cells from l-2 spleens by teasing the tissue with forceps in
RPM1 1640 in a bacterial (Note 9) culture plate.
7. Gently transfer the medium to a centrifuge tube, and allow fragments
of tissue to sediment.
8. Transfer supernatant medium containing splenocytes to another tube,
and wash cells 3 x by centrifugation in serum-free medium at about
1200 x g for 5 min.
9. Mix together 2-4 x lo7 NSl myeloma cells and the spleen cells in a 50
mL centrifuge tube, and centrifuge at 400 x g for 10 min.
10. Remove supernatant, and put tube in water bath at 37OC for subse-
quent manipulations.
Il. Add 1 mL of warm 50% PEG over 1 min, gently stirring the cell pellet
with the pipet as the PEG is added. Continue to stir gently for another
minute.
12. Using the same pipet, slowly stir in 1 mL serum-free medium at 37°C
and repeat.
13. Stir in 7 mL of 37OC serum-free medium over 2-3 min.
14. Centrifuge at 400 x g for 10 min at room temperature, and remove the
supernatant.
15. Aim a lo-mL pipet of 37OCsupplemented medium with 20% FCS at
the cell pellet. Release the medium directly at the pellet, and then by
stirring produce a suspension of fine clumps of cells.
686 Clark

16. Make up a volume with more medium such that, using a 10-mL pipet,
drops of about 100 PL can be distributed into each well of 8 x g&well
culture plates containing feeder cells.
17. Put the plates in the incubator at 37OC.

3.6. SeZection of Hybrid Cells

1. On the day after fusion, make a 3 x solution of selection agents (HAT)
in supplemented medium with 20% FCS (6 mL of each stock solution/
200 mL medium).
2. Using a lo-mL disposable plastic pipet, distribute one drop of selec-
tion medium into each culture well (already containing two drops),
thereby diluting the HAT to the required strength.
3. Every 2-3 d later, aspirate the medium and replace with fresh HAT
medium.

3.7. Primary Screening to Identify Cultures

Producing ReZevant Antibodies
1. Commencing 10 d after fusion, examine wells visually with an in-
verted microscope every 3-4 d for up to 30 d.
2. Remove aliquots of supernatant medium, for antibody screening,
from those wells that contain sufficient hybridomas cells to cover at
least half the bottom of the well. The cells can easily be distinguished
from the feeder cells, since they should be round and “bright.”
3. The method chosen for screening will depend on the purpose of the
investigation. A recent rapid, micro-enzyme-linked-immunoassay
(16) has the advantage of high sensitivity, allowing the assay of as little
as 5 yL culture medium. This has been found to work satisfactorily in
detecting antibodies against GSL target antigens absorbed to the plas-
tic wellplates.

3.8. CZonaZ Isolation of Hybridomas

by Manual Single Cell Isolation
1. Remove one drop of hybrid cell suspension from a g&well plate with
a pastette (plastic transfer pipet; seeitem 3 in Materials 2.4.) and dis-
perse into 10 mL of the same medium in a lo-cm bacterial culture plate
(hybridoma cells do not stick to bacterial plates).
Ganghides

2. Examine with an inverted microscope, and dilute into further plates

until only 1-3 cells are seen per low-power microscope field.
3. Using a fine pipet drawn from an acupette (capillary pipet; seeitem 1
under Materials 2.4.) transfer single cells to wells of a 96-well plate
containing feeder cells.
4. Expand the remaining cells in order to have enough to freeze, and
store in liquid nitrogen as a backup stock in case of later failures.
5. Retest the supernatants of the clones for antibody after culture for an
appropriate time, feeding every 2-3 d (Note 10).
6. Transfer at least two of the positive clones into 5-mL flasks with 1 mL
of medium and put in the incubator, standing the flasks on end to
concentrate the cells. When the culture becomes dense (1-2x l@/mL)
add4 mL medium and lay the flask on its side. When the culture again
becomes dense, transfer to a larger flask and continue feeding until
enough cells are available for freezing in aliquots of about l-5 x 10’
cells.

3.9. Freezing and Thawing of Cells

1. Count cells in growth medium- harvest sufficient cells to provide ali-
quots of l-5 x 10’ cells.
2. Spin cells in a bench centrifuge (1500 rpm for 5 min). Discard super-
natant, and resuspend the cell pellet in the small amount of medium
left in tube.
3. Immediately add (dropwise from a pastette with mixing) sufficient
90% FCS/lO% DMSO (at 4OC)to resuspend cells such that 1 mL con-
tains l-5 x 10’ cells.
4. Place diluted cells on ice while aliquoting out.
5. Working quickly, aliquot about 1 mL to liquid nitrogen storage tubes
on ice. Cap tightly.
6. Immediately transfer capped tubes to dry ice (-70°C) for 5-8 min (bury
tubes in dry ice).
7. Transfer tubes to liquid nitrogen storage (in drawers and racks for ease
of access). Store at -196OC.
8. To retrieve frozen cells from liquid nitrogen, remove storage tube and
thaw rapidly by moving the tube(s) from side to side in a water bath
at 37OC(wear protective gloves and glasses). Wash immediately with
medium and resuspend at l-2 x lo6 cells/mL for optimum culture.
688 Clark

4. Notes
1. FCS should be screened and known to support growth of NSl cells. A
good serum will support 70-80% cloning efficiencies of NSl.
2. These procedures were first developed using mixtures of gangliosides
separated by t.1.c. Various checks discounted the unlikely possibility
that the relatively abundant yield of antibodies (9-14% of wells con-
taining hybrid cells were positive) was because of protein(s) in the im-
munogen. Thus, positive supernatants obtained after immunization
with a single ganglioside (monosialoganglioside; GMI), obtained
from a commercial source, showed similar activity when retested on
the GM1 preparation after removing nonlipid material by reverse-
phase chromatography. Also, amino acid analysis indicated negligi-
ble protein in the immunogen. Similar yields of positive wells were
also obtained using chromatograms of total glycolipid extracts.
3. It is important that the procedures using animals comply with legisla-
tive regulations.
4. Successful antibody production has been achieved with a complete
spectrum of gangliosides from fetal brain associated with about 30 mg
silica. Similarly, selected GSLs from colonic mucosa or 25 pg GM1
alone (associated with 15 mg silica) produced a successful yield of pos-
i tive hybridomas.
5. Two intrasplenic injections have usually been given successfully 2 wk
apart.
6. The use of an intermittant Bunsen burner, with instant on/off control,
inside a sterile laminar flow hood is helpful and interferes minimally
with the air flow. However, a solenoid gas safety inlet must be fitted.
Keeping an open container of formaldehyde solution (38% w/v) in the
hood overnight is also felt to be useful, as is the use of plastic sleeve
covers in addition to surgical gloves. Any and every reasonable pre-
caution should be taken, including not sharing the use of the hood
with other workers.
7. Not all NSl cells are equivalent, and care must be taken to ensure that
a clone that is able to fuse well with immune spleen cells is used. NSl
cells are &azaguanineresistant and hencelack hypoxanthine-guanine
phosphoribosyl-transferase. Long-term culture gives rise to genetic
drift, which may affect continued resistance. Growth of cells for a few
days in medium containing azaguanine (20 yg/mL) prior to growing
for a fusion may, therefore, be necessary.
Gangliosides

8. The optimum time for obtaining maximum numbers of fused immune

spleen cells does not coincide with that for getting high serum titers of
antibody (25).
9. Spleen cells do not stick to the surface of bacterial culture plates.
10. It may be appropriate at this time to start weaning the hybrid cells
from HAT to media containing, first, lx HT then, l/2 HT.

References
1. Weigandt, H. (1985) Gangliosides, in Glycolipids (Weigandt, H., ed.), Elsevier, Am-
sterdam, pp. 199-245.
2. Alving, C. R. (1977) Immune reactions of lipids and lipid model membranes, in The
Antigens, 4 (Sela, M., ed.), Academic, London, pp. 3-63.
3. Young, W. W., MacDonald, E. M. S., Novinsky, R. C., and Hakomori, S. I. (1979)
Production of monoclonal antibodies specific for two distinct steric portions of the
glycolipid ganglio-N-tryosylceramide (asialo-GM21 I. Exp. Med. 150,1008-1019.
4. Young, W. W., Portoukalien, T., and Hakomori, S. I. (1981) 2 monoclonal anticarbo-
hydrate antibodies directed to glycosphingolipids with a lacto-N-glycosyl type-11
chain. J, Biol. Chem. 256,10967-10972.
5. Brodin, T., Thurin, J., Stromberg, N., Karlsson, K. A., and Sjogren, H. 0. (1986) Pro-
duction of oligosaccharide-binding monoclonal antibodies of diverse specificities
by immunization with purified tumor-associated glycolipids inserted into lipo-
somes with lipid-A. Eur. J. Immunol. 16,951-956.
6. Magnani, J. L., Brockhaus, M., Smith, D. F., Ginsberg, V., Blaszczyk, M., Mitchell,K.
F., Steplewski, Z., and Koprowski, H. (1981) A monosialoganglioside is a monoclo-
nal antibody-defined antigen of colon-carcinoma. Science 212,55,56.
7. Crumpton, M. J. (1974) Protein Antigens: The molecular basis of antigenicity and
immunogenicity, in The Anfigens (Sela, M., ed.> Academic, London, pp. l-78.
8. Van Ness, J., Laemmli, U. K., and Pettijohn, D. E. (1984) Immunization in vitro and
production of monoclonal antibodies specific to insoluble and weakly immuno-
genic proteins. Proc. NaB Acad. Sci. USA S&7897-7901.
9. Spitz, M., Spitz, L., Thorpe, R., and Eugui, E. (1984) Intrasplenic primary immuniza-
tion for the production of monoclonal antibodies. J. Immunol. Methods 70,3943.
10. Morrison, S. L. and Scharff, M. D. (1981) Mutational events in mouse myeloma cells,
a review. C. R. C. Cmf. Rev. lmmunol. 3,1-22.
11. Yelton, D. E. and Scharff, M. D. (1982) Mutant monoclonal antibodies with altered
biological functions. 1. Exp. Med. 156,1131-1148.
12. Galfre, G. and Milstein, C. (1981) Preparation of Monoclonal Antibodies: Strategies
and Procedures, in Methods in Enzymology 73 part B. Immunochemical Techniques
(Lagone, J. H. and Van Vunakis, H., eds.), Academic, London, pp. 346.
13. Wood, J. N. (1984) Immunization and fusion protocols for hybridoma production,
inMefhods inMoleculav Biology 1: Proteins (Walker, J. M., ed.), Humana Press,Clifton,
New Jersey, pp, 261-270.
14. Ladisch, S. and Zi-Liang, W. (1985) Detection of tumor-associated ganglioside in
plasma of patients with neuroblastoma. Luncef i, 136-138.
690 Clark

15. Ueno,K.,Ando,S.,andYu,R.K. (1978)Gangliosidesof human,cat,andrabbit spinal

cords and cord myelin. 1. Lipid Res.19,863-871.
16. Durbin, H. and Bodmer, W. F. (1987) A sensitive micro-immunoassay using p-gal-
actosidase/anti P-galactosidase complexes. J. ImmunoZ. Methods 97,19-27.
 Appendix

Compiled by

Jeffrey W. Pollard
 Table 1
Balanced Salt Solutions”
Dulbecco’s
Earle’s, g/Lb Hanks’, g/Lb Gey’s, g/Lb Puck’s, g/Lb PBS, g/L”
KC1 0.4 0.4 0.37 0.4 0.2
- 0.06 0.03 0.15 0.2
=w4
MgCl.p6H,O 0.1 0.21 - 0.1
MgSO,*qO 0.2 0.1 0.07 0.154 0.1
CaC1, 0.02 0.14 0.17 0.012 -
NaCl 6.68 8 7 8 8
NaHCO, 2.2 0.35 2.27
NaH,PO,*H,O 0.14 -
Na,HPO, l 7H,O 0.09 0.226 2.31
Glucose 1 1 1 -
Phenol Red 0.01 0.01
Reference 1 2 3 5
‘For detailed description of derivations, see ref. 6.
@Sterilize by filtration.
Sterilize by autoclaving.
d12H,0
 Table 2
Eagle’s Minimum Essential Medium and Derivatives’
tp
Dulbecco’s Iscove’s zi
Eagle‘s modification modified Joklik’s s
MEM, DMEM, CiMEMb DMEM, MEM, E!
Component mg/L mg/L mg/L mg/L mg/L R
Amino acids
L-Alanine 25 25 -
L-ArginineeHCl 126.4 84 126.4 84
L-Asparagine 50
L-Aspartic acid 30 30
L-Cysteine HCl
l 89.74
L-Cysteine l

d&odium salt 28.42 56.78 28.42 56.78 29.60

L-Glutamic acid - 75 75
L-Glutamine 292.3 584 292 584 294
Glycine 30 50 30 -
L-Hi&dine l

HClmqO 42 42 42 42 42
L-Isoleucine 52.5 104.8 52.5 104.8 52
L-Leucine 52.5 104.8 52.5 104.8 52
L-Lysine*HCl 73.06 146.2 73.06 146.2 72.5
L-Methionine 14.9 30 14.9 30 15
L-Phenylalanine 33.02 66 33.02 66 32
L-Proline - - 40 40 -
LSerine 42 25 42 -
L-Threonine 47.64 95.2 47.64 95.2 48
L-Tryptophan 10.2 16 10.2 16 10
L-Tyrosine 36.22 72 36.22 37.8
L-Tyrosine:
disodium salt 104.2 -
L-Valine 46.9 93.6 46.9 93.6 46
 Table 2
Eagle’s Minimum Essential Medium and Derivatives” (cuntm~ed)
Dulbecco’s Iscove’s s
Eagle’s modification modified Joklik’s
M’EM, DMEM, CiMEMb DMEM, MEM,
Component mg/L q/L mg/L mg/L mg/L
Vitamins and lipids
L-Ascorbic acid 50 -
Biotin - - 0.1 0.013
D-Ca pantothenate 1 4 1 4
Choline chloride 1 4 1 4
Folk acid 1 4 1 4
i-inositol 2 7.2 2 7.2
Nicotinamide 1 4 1 4
l’yridoxal HCI 1 4 1 4
Riboflavin 0.1 0.4 0.1 0.4
Thiamine HCl 1 4 1 4
Vitamin B12 - 1.36 0.013
Cholesterol - 0.02
Caqe%0 264.9 264.9 264.9
CaCt, 200 165
Fe(Nb,),*9&0 0.1
KCl 400 400 400 330 4-00
MgScw-40 200 200 200 242.2
MgSO, (anhyd) 97.67
NaCl 6800 6400 6800 4505 6500
NaHCO, 2000 3700 2000 3024 2000
158.3 141.3 158.3 125 1500
0.076
0.0173
Others
Adenosine - 10
 Cytidihe 10
Reoxyadenosine - 10
Deoxycytidine - 10 -
Deoxyguanosine - 10
Dihydrostrepto-
mycin
sulfate - 50
Bovine serum
albumin 0.4
Glucose 1000 4500 1000 4500 2000
Guanosine - 10
HEPRS - 5958
Lipoic acid - 0.2
Penicillin G” - . 75000 ru
Sodium Phenol
Red 17 15 10 15 10
Sodium
Pyruvate 110 110 110 110
Soybean lipid - 0.1 -
Thymidine 10
Transferrin 0.001 -
Uridine - 10
Reference 7 8 9 10 21
“For full lists of tissue cuhure medium and references, see ref. 27. It should be noted that medium formulations vary somewhat
from company to company
baMEM is often supplied without ribo and deoxyribonucleosides.
cAntibiotics are often supplied with the medium or they can be added during preparation (see ako Table 4).
 Table 3
Other Useful Media”
s
RrMl HAMS CMRLb McCoy’s Medium Waymouth’s
1640, F12, 1066, 5A, 199, MB752/1,
Component mg/L mg/L mg/L mg/L mg/L mg/L

L-alanine 8.9 25 13.9 25

L-arginine
free base 200 -
L-arginine HCl
l 211 70 42.1 70 75
L-asparagine 50 - 45 - -
L-asparaginee
w 15.01
L-aspartic acid 20 13.3 30 19.97 30 60
L-cysteine
(free base) - 31.5 61
L-cysteine*
HCI*qO 35.12 260 - -
L-cysteine HCl - 0.099
L-cystine 50 - 20 - 15
L-cysteine,
disodium
salt 23.66 -
L-glutamic acid 20 14.7 75 22.1 66.82 150
L-glutamine 300 146 100 219.2 100 350
Glycine 10 7.5 50 7.5 50 50
L-Hi&dine
free base 15 - 128
L-Hi&line
HClGJO 20.96 20 20.96 21.88
L-Hydroxy-
proline 20 10 - 10 19.67
L-Isoleucine 50 3.94 20 39.36 20 25
 L-Leucine 50 13.1 60 39.36 60 50
L-Lysine HCl
l 40 36.5 70 36.5 70 240
L-Methionine 15 4.48 15 14.9 15 50
L-Phenylala-
nine 15 4.96 25 16.5 25 50
L-Proline 20 34.5 40 17.3 40 50
L-Serine 30 10.5 25 26.3 25
L-Threonine 20 11.9 30 17.9 30 75
L-Tryptophan 5 2.04 10 3.1 10 40
L-Tyrosine 20 5.4 40 18.1 40 40
L-Valine 20 11.7 25 17.6 25 65
Vitamins
L-Ascorbic - 50 0.5 0.05 17.50
Biotin 0.2 0.0073 0.01 0.2 0.01 0.02
D-Ca Pan-
tothenate 0.25 0.48 0.01 0.2 0.01 1
Choline
chloride 3 13.96 0.5 5 0.5 250
Folk acid 1 1.3 0.01 10 0.01 0.4
i-inositol 35 18 0.05 36 0.05 1
Menopthone
sodium
bisulpite
trihydra te - 0.019
Nicotinic acid - - 0.5 0.025
Nicotinamide 0.04 0.025 0.5 0.025 1
p-Amino-
benzoic acid 1 - 0.05 1 0.05
Pyridoxal l HCl 0.062 0.025 0.5 0.025
PyridoxineeHCI 1 0.062 0.025 0.5 0.025 1
Riboflavin 0.2 0.038 0.01 0.2 0.01 1
ThiamineeHCl 1 0.34 0.01 0.2 0.01 10
 Table 3 (Continued)
RPMI HAMS CMRLb McCoy’s Medium Waymouth’s
1640, F12, 1066, 5A 199, MB752/1,
Component mg/L mg/L mg/L mg/L
2
mg/L mg/L

DL-a Toco-
pherol
phosphate,
disodium
salt - - - 0.01 -
Vitamin A
acetate - 0.115 -
Vitamin B12 0.005 1.36 2 - 0.2
Cholesterol - - 0.2 -

Inorganic salts
CaCl,eanhyd. - 200 -
CaC1,*2H,O - 44 - 264.9
CaNO, l 4H,O 100 -
CuSO, .5H,O - 0.0025 -
Fe(NOJ,*9H,O - - - 0.1
FeSO,m7H,O - 0.834 - - -
KC1 400 223.6 400 400 150
- - 60 80
KIvo4
MgCJe6H,O - 122 - 240
MgSO,*7H,O 100 - 200 200 200 200
NaCl 6000 7599 6799 6400 6800 6000
NaHCO, 2200 1176 2200 2200 2200 2240
NaH,PO,eH,O - 140 530 158.3 -
Na,HPO,e7H,O 1512 268 - 566
ZnS04*7H,O - 0.863 -

Other Components
Adenine
sulpha te - 10
5’-AMP - 0.2
ATE’, diso-
dium salt 10
Bactopeptone -
2-deoxyribose - 0.5
Glucose 1802 1000
Glutathione
GuanineaHCl 0.3
Hypoxanthine 0.3 25
Linoleic acid -
Lipoic acid - -
Phenol Red 20 17
Putrescine l

wcl 0.161 -
Ribose - - 0.5
Sodium
acetate - 83 - 36.71 -
Sodium
glucoronate 4.2 - -
Sodium
pyruvate - 110 -
Triphospho-
pyridine
nucleotide
Thymidine
Thymine
Tween 80
Uracil
Uridine tri-
phosphate l

4%*
1
Xanthine
CoCarboxy-
lase -
 RrMl HAM’S CMRLb McCoy’s Medium Waymouth’s
1640, F12, 1066, 5% 199, MES752/1,
Component mg/L mg/L mg/L mg/L mg/L mg/L

CoEnzyme A - 2.5 -
Deoxyaden-
osine - 10 - -
Deoxycyti-
d.ine*HCl 10 -
Deoxyguano-
sine - 10 -
Diphyospho-
pyridine
nucleotidee
- 7 - -
FP
Ethanol (for
lipid com-
ponent) - 16 - -
Flavin adenine
n nucleotide 1 -
5-methyl-deoxy-
cytidine - 0.1 -
Reference 12 13 11 14 15 16
“For a full list of the medium, their modifications and references, seerefs. 11 and 23.
bCan be made with Hanks’ salts rather than Earle’s salts.
Appendix 701

Table 4
Useful Antibiotics for Tissue Culture

Recommended Approximate
concentration stability,
Antibiotic Spectrum of action pg/mL” d

Ampho tericin B Fungi and yeast 1

Ampicillin Gram-positive and
negative bacteria 100
Chloramphenicol Gram-negative bacteria 5
Erythromycin Gram-positive bacteria
and mycoplasma 100
Gentamycin Gram-positive and
gram-negative
bacteria and myco-
plasma 50
Kanamycin Gram-positive and
negative bacteria 100
Nystatin Fungi and yeast 50
Penicillin G Gram-positive bacteria 100
Rifampicin Gram-positive and
gram-negative
bacteria 50 3
Streptomycin Gram-positive and
gram-negative
bacteria 100 3
Tetracycline Gram-positive,
gram-negative bacteria
and mycoplasma 10 4

“The concentration given is sufficient to control a mild infection for the length of time
stated at 37OCwithout undue toxicity to cells (seeref. 27), for greater detail). Most media
contain streptomycin and penicillin G. The use of other antibiotics, particularly clinically
relevant ones such as erythromycin, should not be encouraged unless absolutely neces-
sary. This is because media invariably goes into the drainage system and so increase the
range of drug-resistant “wild” bacteria.
702 Pollard

Table 5
Insect Cell Medium
Grace’s insect tissue culture medium (ref. 18)
Ingredient mg/L
L-Isoleucine 50
L-Phenylalanine 150
L-Tryptophan 100
L-Leucine 75
L-Histidine*HCl*H,O 3378
L-Methionine 50
L-Valine 100
L-Arginine HCl l 700
L-Lysine l HCl 625
L-Threonine 175
L-Asparagine l H,O 397.7
L-Proline 350
L-Glutamine 600
DL-Serine 1100
Glycine 650
L-Alanine 225
I)-Alanine 200
L-Cystine disodium salt 22.69
L-Tyrosine disodium salt 62.15
L-Glutamic acid 600
L-Aspartic acid 350
D-Sucrose 26680
D-Fructose 400
D-Glucose 700
L-malic acid 670
a-Ketoglutaric acid 370
D-Succinic acid 60
Fumaric acid 55
p-Aminobenzoic acid 0.02
Folic acid 0.02
Riboflavin 0.02
Biotin 0.01
Thiamin HCl
l 0.02
D-Calcium pantothenate 0.02
Pyridoxine HCl l 0.02
Nice tinic acid 0.02
I-Inositol 0.02
Choline chloride 0.2
NaH,PO, 2H,O l 1140
CaCl, 2H,O
l 1325
Appendix 703

Table 5 (Continued)
Grace’s insect tissue culture medium
Ingredient mg/L

MgCl, .2H,O 2280

MgSO, l 7H,O 2780
KC1 2240
NaHCO, 350

BML-TC/lO” for insect cell culture and production

of nuclear polyhedrosis virus (ref. 19)

Ingredient Values/L
Fetal calf serum 100 mL
Tryptose broth 2.6 g
KC1 2.87 g
NaH,PO, 1.14 g
CaCl,*2H,O 1.32 g
MgCQ6H20 2.28 g
MgSO, 7H,O
l 2.78 g
NaHCO, 0.35 g
Glucose lg
Tlus Grace’s vitamins and amino acids without j3-alanine and o-serine as above.
704 Pollard

Appendix References
I. Earle, W. R. (1943) Production of malignancy in vitro: IV The mouse fibroblast cul-
tures and changes in the living cell. J. N&l. Cancer Inst. 4,X5-212.
2. Hanks, J. H. and Wallace,R. E. (1949) Relation of oxygenand temperature in the pres-
ervation of tissues by refrigeration. Proc. Sot. Exp. Biol. Med. 71,196-200.
3. Gey, G. 0. and Gey, M. K. (1936) The maintenance of humannormal cells and tumor
cells in continuous culture. 1. Preliminary report: cultivation of mesoblastic tumors
and normal tissues and rates or method of cultivation. Am. 1. Cancer 27,45-76.
4. Puck, T. T., Cieciura, S. J., and Robinson, A. (1958) Genetics of somatic mammalian
cells: III Long-term cultivation of euploid cells from human and animal subjects. J.
Exp. Med. 108,945-955.
5. Dulbecco, R. and Vogt, M. (1954) Plaque formation and isolation of pure lines with
poliomyelitis viruses. J, Exp. Med. 99,167-182.
6. Bashor, M. M. (1979) Dispersion and disruption of tissues. Methods Enzymol. 58,
119-131.
7. Eagle, H. (1955) Amino acid metabolism in mammalian cell cultures. Science 130,
432437.
8. Dulbecco, R. and Freeman, G. (1959) Plaque production by polyoma virus. Virology
8,396-397.
9. Stanners, C. P., Eliceiri, G. L., and Green, H. (1971) Two types of ribosomes in mouse-
hamster hybrid cells. Nufure Nezu Biol. 230,52-54.
10. Iscove, N. N. and Melchers, F. (1978) Complete replacement of serum by albumin,
transferrin, and soybean lipid in cultures of lipopolysaccharide-reactive b-lympho-
cytes. 1. Exp. Med. 147,923-933.
11. Jakoby, W. B. and Pastan, I. H., eds. (1979) Cell culture. Methods Enzymol. 58.
22. Moore, G. E., Gerner, R. E., and Franklin, H. A. (1961) Culture of normal human
leukocytes. JAMA 199,519-524.
13. Ham, R. G. and McKechan, W. L. (1979) Media and growth requirements. Methods
Enzymol. 58,44-93.
14. McCoy, T. A., Maxwell, M., and Kruse, P. F. (1959) Amino acid requirements of the
Novikoff hepatoma in vitro. Proc. Sot. Exp. Biol. Med. 100,115-118.
15. Morgan, J. F., Campbell, M. E., and Morton, H. J. (1955) The nutrition of animal tis-
sues cultivated invifro. I. A survey of natural materials as supplements to synthetic
medium 199. J. Nafl. Cancer Inst. 16,557-567.
16. Waymouth, C. (1959) Rapd proliferation of sublines of NCTC clone 929 (Strain L)
mouse cells in a simple chemically defined medium (MB 752/l). 1, Nafl. Cancer Inst.
22,1003-1015.
17. Perlman, D. (1979) Use of antibiotics in cell culture media. Methods Enzymol. 58,
110-119.
18. Grace, T. D. C. (1962) Establishment of four strains ot cells from insect tissue grown
in vitro. Nature 195,788,789.
19. Gardiner, G. R. and Stockdale, H. (1975) Two tissue culture media for the produc-
tion of Lepidoptcran cells and nuclear polyhedrosis viruses. 1. Invert. Path. 25,
363-370.