Journal of Immunological Methods 278 (2003) 283 – 292

www.elsevier.com/locate/jim

Protocol
Simple and cost-effective isolation of monocytes from buffy coats
Urška Repnik a,*,1, Miomir Knezevic a,b, Matjaz Jeras a,*,1
a
Blood Transfusion Centre of Slovenia, Tissue Typing Centre, Slajmerjeva 6, 1000 Ljubljana, Slovenia
b
Transcell d.o.o., Teslova 30, 1000 Ljubljana, Slovenia
Received 16 March 2003; accepted 8 May 2003

Abstract

We have optimized the procedure of monocyte isolation on a Percoll density gradient. The new procedure consists of three
steps: (1) the isolation of MNC on a Ficoll density gradient; (2) the separation of monocytes from lymphocytes on a high-
density hyper-osmotic Percoll density gradient; and (3) the separation of monocytes from platelets and dead cells on a low-
density iso-osmotic Percoll density gradient. The procedure is simple and cost-effective. Monocyte purity and recovery are both
about 75% and platelet contamination is low. The isolated monocytes retain their capacity to differentiate into dendritic cells in
vitro.
D 2003 Elsevier B.V. All rights reserved.

Keywords: Monocytes; Isolation; Percoll

1. Background (1981) showed that monocyte purity can be improved

by using a hyper-osmotic density gradient medium.
Monocytes are widely used for in vitro generation The effect of the osmolarity changes was studied
of dendritic cells. Several methods for monocyte systematically by Bøyum (1983). Using Nycodenz,
isolation from peripheral blood exist such as adhesion an iodinated gradient medium, he prepared separation
and negative and positive immunoselection. From the fluids with densities ranging from 1.061 to 1.096 g/ml
late 1960s through to the 1980s, many protocols for and, by increasing osmolarity together with density, he
monocyte isolation were devised based on density obtained a 95 –98% pure monocyte suspension. The
gradient centrifugation, which is still an attractive principle is that as osmolarity increases, the cells expel
alternative (Almeida et al., 2000; Lehner and Holter, water, thereby increasing their density and thus are
2002). However, the densities of lymphocytes and able to pass a density barrier that otherwise they could
monocytes overlap so that efficient separation proved not. Separation of the two cell types is feasible because
difficult on the basis of density difference alone. Fluks lymphocytes are more sensitive to an increase in
osmolarity than monocytes. He showed that the higher
the osmolarity, the higher the purity but the lower the
* Corresponding authors. U. Repnik is to be contacted at Tel.: yield of monocytes although the viability of the cells
+386-1-54-38-223; fax: +386-1-43-12-304, M. Jeras is to be recovered at higher osmolarities was not affected.
contacted at Tel.: +386-1-23-02-313.
E-mail addresses: [email protected] (U. Repnik),
In our laboratory, we were faced with a problem of
[email protected] (M. Jeras). high platelet contamination of monocyte and dendritic
1
These authors contributed equally to this work. cell suspensions destined for electrofusion in immu-

0022-1759/03/$ - see front matter D 2003 Elsevier B.V. All rights reserved.
doi:10.1016/S0022-1759(03)00231-X
284 U. Repnik et al. / Journal of Immunological Methods 278 (2003) 283–292

nohybridoma generation. As pilot experiments indi-  Percoll, product number: P1644 or P4937 (Sigma).
cated that platelets could impede this procedure, we  1.6 M NaCl solution.
sought a cost-effective method for obtaining highly  1.5 M NaCl solution.
enriched, platelet-free monocyte suspensions.  Distilled or deionized water.

4.2. Special equipment

2. Type of research
 Centrifuge (preferably Heraeus, Megafuge 1.0).
1. Generation of dendritic cells from monocytes.  Safety cabinet (optional, for isolation under aseptic
2. Research on monocytes and macrophages. conditions).
 Flow cytometer (optional, for evaluating the
composition of the cell suspension).
3. Time required

1. Isolation of peripheral blood mononuclear cells 5. Detailed procedure

(PBMC) from a buffy coat on a Ficoll density
gradient: 2 h. 5.1. Starting material
2. Cell counting: 10 min.
3. Preparation of a cell suspension with the desired The following protocol is most suited to the
cell concentration: 15 min. isolation of monocytes from buffy coats (BC) al-
4. Preparation of the hyper- and iso-osmotic Percoll though it has also been applied for lymphopheresis
solutions: 15 min (can be done during the products and can easily be adapted for smaller vol-
centrifugation or several days in advance). umes of whole blood (see Discussion).
5. Separation of monocytes from lymphocytes on the
hyper-osmotic Percoll: 1 h. 5.2. PBMC isolation
6. Cell counting (optional): 10 min.
7. Separation of monocytes from platelets and dead 1. Dilute the BC with up to twice its volume of
cells on the iso-osmotic Percoll: 45 min. RPMI medium, most frequently to 200 ml.
8. Cell counting: 10 min. 2. Distribute 12.5 ml of Ficoll into each 50 ml
centrifuge tube. The number of tubes depends on
The overall time required is approximately 5 h. the final volume of the diluted BC.
3. Carefully overlay the Ficoll with 25 ml of the
diluted BC.
4. Materials 4. Centrifuge at 950 g for 15 min with the brake off.
5. Collect the PBMC at the interface between the
If the monocytes are to be cultivated, then the Ficoll and the plasma-medium layers with a
material used throughout the procedure must be sterile Pasteur pipette and transfer the cells from two
and endotoxin-free. tubes into one new tube.
6. Add RPMI medium to a final volume of 45 ml.
 50 ml centrifuge tubes. 7. Centrifuge at 350 g for 7 min. Decant the
 Pasteur pipettes. supernatant carefully and resuspend the cells in
 plastic serological pipettes (1, 5, 10 and 25 ml). fresh RPMI medium. Combine the cells from
 stripettor. two tubes and add RPMI medium to a final
volume of 45 ml.
4.1. Chemicals and reagents 8. Centrifuge at 350 g for 7 min. Discard the
supernatant and resuspend the cells in RPMI
 RPMI medium. medium. Combine the cells from two tubes and
 Ficoll. add RPMI medium to a final volume of 25 ml.
 U. Repnik et al. / Journal of Immunological Methods 278 (2003) 283–292 285

9. Mix 20 Al of this cell suspension with 180 Al of 5.4.1. Monocyte fraction

Turk’s solution. Count the cells in a standard
hemocytometer. 3. Collect the cells at the interface with a Pasteur
10. Centrifuge the suspension from Step 8 at 350 g for pipette. Do not collect the last 4 ml.
7 min. Discard the supernatant and resuspend the 4. Combine the cells from up to three tubes. Add
cells at a concentration of 50 – 70  106 cells/ml in RPMI medium to a final volume of 45 ml.
RPMI medium to give a final volume which is a 5. Centrifuge at 350 g for 7 min.
multiple of 3 (i.e. 3/6/9/12/. . . ml). 6. Discard the supernatant and resuspend the cells
in 3 ml of RPMI medium. The suspension is
5.3. Preparation of the Percoll solutions later referred to as a monocyte-enriched suspen-
sion.
The Percoll solutions can be prepared in advance 7. Mix 20 Al of the cell suspension with 180 Al of
and kept at 4 jC until required. trypan blue solution and count the cells in a
standard hemocytometer (optional).
5.3.1. The hyper-osmotic Percoll solution
5.4.2. Lymphocyte fraction
1. For 100 ml of solution, mix 48.5 ml of Percoll,
41.5 ml of water and 10.0 ml of 1.6 M NaCl and 8. After collecting the monocyte fraction at the
shake vigorously. interface, combine the residual sediments in one
centrifuge tube. Add RPMI medium to a final
5.3.2. The iso-osmotic Percoll solution volume of 45 ml.
9. Centrifuge at 350 g for 7 min. Discard the
1. For 20 ml of solution, mix 8.3 ml of Percoll, 9.7 ml supernatant and resuspend the cells in culture
of water and 2.0 ml of 1.5 M NaCl and shake medium to allow them to recover.
vigorously. 10. Count the cells.

5.4. Separation of monocytes from lymphocytes 5.5. Separation of monocytes from platelets and dead
cells
1. Carefully lay 3 ml of the PBMC suspension
(containing 150 –200  106 cells) onto 10 ml of 1. In each centrifuge tube, overlay 10 ml of iso-
the hyper-osmotic Percoll solution in each cen- osmotic Percoll solution with 3 ml of the mono-
trifuge tube so that the interface between the layers cyte-enriched suspension (refer to Section 5.4.1,
is preserved. Step 6) (up to 200  106 cells).
2. Centrifuge the tubes at 580 g for 15 min with the 2. Centrifuge at 350 g for 15 min with the brake
brake off. off.

Table 1
The average monocyte-to-lymphocyte ratio and the average monocyte yield and recovery obtained with the isolation of monocytes on a Percoll
density gradient
Starting MNC suspension Final monocyte-enriched suspension MNC yield Monocyte
Monocytes Lymphocytes Monocytes Lymphocytes (  106) Yield Recovery
(%) (%) (%) % (  106) (%MNC)
Mean value 13 87 75 25 603 77 13.5
SD (n = 41) 6 6 10 10 238 36 6
The values were obtained from the analysis of 41 isolations using buffy coats as starting material. Cell suspensions were analysed by flow
cytometry and the monocyte-to-lymphocyte ratio determined on forward scatter (FSC) vs. side scatter (SSC) dot plots (see Fig. 1). Absolute cell
numbers (in millions) were obtained by cell counting in a standard hemocytometer.
286 U. Repnik et al. / Journal of Immunological Methods 278 (2003) 283–292

3. Collect and discard the supernatant with a Pasteur 6. Results and discussion
pipette, or alternatively, save it for analysis of its
composition. In Table 1, the average monocyte-to-lymphocyte
4. Resuspend the sediment in less than 1 ml of RPMI ratios in the starting and the final cell suspensions
medium and transfer it into a new centrifuge tube. are presented together with the average monocyte
Add RPMI medium to a final volume of 45 ml. yield and recovery. The values were obtained from
5. Centrifuge at 350 g for 7 min. Discard the 41 isolations performed over a period of one year.
supernatant and resuspend the cells in 5 ml of The cell suspension composition was evaluated by
culture medium. flow cytometry using forward scatter (FSC) vs. side
6. Mix 20 Al of the cell suspension with 180 Al of scatter (SSC) dot plots gated on peripheral blood
trypan blue solution and count the cells. MNC. In Fig. 1, dot plots of successive cell sus-

Fig. 1. Forward scatter (FSC) vs. side scatter (SSC) dot plots of cell suspensions at different stages during the monocyte isolation procedure.
Phenotypes of the different cell populations are shown in Figs. 2 and 3. Monocytes in the final suspension are shown in a rectangle. The dot
plots relate to three separate isolations (in columns), which yielded different final monocyte purities: 91%, 84% and 68% (from the left). (A) The
starting monocyte suspension; (B) the monocyte-enriched suspension after the hyper-osmotic Percoll step; (C) the lymphocyte-enriched
suspension after the hyper-osmotic Percoll step; (D) the suspension of platelets, debris and dead cells after the iso-osmotic Percoll step; and (E)
the final monocyte-enriched suspension after the iso-osmotic Percoll step.
 U. Repnik et al. / Journal of Immunological Methods 278 (2003) 283–292 287

pensions obtained during the isolation are presented significantly after separation on hyper-osmotic Per-
for three separate isolations with different final coll, with further slight increase after separation on
purity of monocytes. The mean percentage of mono- the iso-osmotic Percoll. The average purity of mo-
cytes in the starting MNC suspensions was 13 F nocytes in the final suspensions was 75 F 20%.
12% (X F2S.D.), which is lower than in some other The contaminating MNC were mainly T lympho-
reports (Fluks, 1981; Almeida et al., 2000). The cytes and NK cells (Fig. 2). The separation on the
suspensions were highly contaminated with plate- iso-osmotic Percoll was crucial for eliminating pla-
lets, irrespective of the number of washing steps. telets and dead cells although it resulted in some
The ratio of monocytes to lymphocytes increased loss of monocytes. The final monocyte-enriched

Fig. 2. The final monocyte-enriched suspension was stained for lymphocyte lineage markers to determine the relative position of monocytes on
a forward scatter (FSC) vs. side scatter (SSC) dot plot. Monocytes have higher FSC values than lymphocytes and span a larger interval of SSC
values. The contaminating lymphocytes are mostly T cells and NK cells. The grey area indicates the expression of the selected CD molecules.
The dotted line indicates the binding of isotype-matched, irrelevant control antibodies. R1 (region 1): monocytes with high SSC values; R2
(region 2): monocytes with low SSC values; R3 (region 3): lymphocytes.
288 U. Repnik et al. / Journal of Immunological Methods 278 (2003) 283–292

suspension stained with Hemacolor and monocytes flow cytometry. The results are presented in Fig. 4. The
at the initiation of a cell culture are shown in Fig. monocyte population was not homogenous with re-
3. spect to IgG2a isotype control binding although the
The viability of monocytes in the final suspension cells did not differ significantly in CD83, CD80, CD86,
was at least 95%. The average final cell yield was CD40, CD54, HLA-DR and CD1a expression. Mono-
13% based on the starting MNC numbers, as assessed cytes that did not significantly bind the IgG2a isotype
by cell counting. This was also the value of the control also did not bind Becton Dickinson’s CD14
average percentage of monocytes in the starting mAb, but did bind DAKO’s CD14 mAb, presumably
MNC suspensions, as assessed by flow cytometry. due to different epitopes recognized by the two mAb.
There was only a weak correlation between the These slightly different monocytes, most probably
individual values of the two parameters, which may immature, had also lower SSC values and could ac-
be due to the different methods used for their assess- count for up to 25% of the total monocytes.
ment. Assuming that the average values of 13% for Percoll can be obtained either as an aseptically
the two parameters are accurate and that the average filled (product number P1644) or a sterile product
final monocyte purity is 75%, the final monocyte (product number P4937). Neither is declared endo-
recovery is 75%. toxin-free. In all our isolations, the aseptically filled
On some occasions, the final suspension was stained Percoll was used. Priming monocytes with LPS
with mAb specific for surface antigens and analysed by impairs their ability to differentiate into dendritic cells
(Palucka et al., 1999). Dendritic cells generated from
monocytes primed with LPS have reduced capacity to
produce IL-12 p70 (Lehner and Holter, 2002). With
LAL tests, we found endotoxin contamination in
Percoll (product number P1644) and its solutions as
well as in media sampled from immature DC cultures
to be below 0.50 EU/ml. In our laboratory, the
monocytes isolated on Percoll are successfully differ-
entiated into dendritic cells if judged by the expres-
sion of cell surface markers. Phenotypes of immature
dendritic cells generated in X-VIVO 15, AIM-V or
RPMI supplemented with 10% FBS are shown in
Fig. 5.
Frequently, monocytes were isolated from two or
more BC in parallel using the same Percoll solutions.
Comparison of the final monocyte purity in such cases
led to the conclusion that the efficiency of monocyte
isolation is largely dependent on preparation of the
Percoll solutions, in addition to the factors highlighted
under General precautions.
A lymphopheresis product was used twice as a
starting material, but the purity of monocytes obtained
with Percoll density gradients was only about 50%. A
subsequent plastic adherence step yielded a highly
pure monocyte culture. In the case of apheresis
products, the proposed protocol can serve to precon-
centrate monocytes and thus reduce the amount of
Fig. 3. (A) A cytospin preparation of the final monocyte-enriched
suspension stained with Hemacolor (Merck) (20  ex-objective). work at the subsequent adherence step.
(B) The photomicrograph of monocytes isolated on Percoll density With smaller volumes of blood, when the PBMC
gradients at the initiation of a cell culture (10  ex-objective). yield does not exceed 60  106 cells, the separation
 U. Repnik et al. / Journal of Immunological Methods 278 (2003) 283–292 289

Fig. 4. The phenotypes of monocytes isolated on the Percoll density gradients. Monocytes exhibiting different side scatter (SSC) values (see Fig.
2: R1 and R2 gates) differ in their IgG2a isotype control binding and CD14 expression, but show similar CD83, CD80, CD86, HLA-DR and
CD1a expression. The grey area indicates the expression of HLA-DR and the selected CD molecules. The dotted line indicates the binding of
isotype-matched, irrelevant control antibodies. R1 (region 1): monocytes with high SSC values; R2 (region 2): monocytes with low SSC values.
The CD14 staining was performed by two different CD14 mAb, provided by DAKO and Becton Dickinson-Pharmingen, and gave different
results.

can be performed in a 15 ml centrifuge tube with 1 ml solutions should be prepared and used at room
of cell suspension being laid over 3 ml of the Percoll temperature.
solution.
The protocol was also tested on thawed PBMC and 6.1.2. Overlaying the Percoll gradient with a cell
the results were comparable to those obtained using suspension
freshly prepared cell suspensions. The cell suspension should be laid over the
Percoll slowly to preserve the interface. It was
6.1. General precautions observed that in the case of a disturbed interface,
monocytes tend to migrate deeper into the Percoll
6.1.1. Preparation of Percoll solutions layer.
Percoll is an extremely viscous liquid and special
attention should be taken when measuring its vol- 6.1.3. Collecting monocytes from the interface with
ume. To transfer accurately a given volume of hyper-osmotic Percoll
Percoll, we use plastic serological pipettes and take In contrast to separation on Ficoll, the distributions
the 1 ml mark as zero point. Percoll and its of monocytes and lymphocytes after the hyper-osmotic
290 U. Repnik et al. / Journal of Immunological Methods 278 (2003) 283–292

Fig. 5. Phenotypes of immature dendritic cells generated from monocytes isolated on the Percoll density gradients. Monocytes were cultured for
5 days either in AIM-V (solid line), X-VIVO 15 (dashed line) or RPMI + 10% FBS (dotted line) medium supplemented with IL-4 (400 U/ml)
and GM-CSF (1.000 U/ml). The histograms show the expression of selected surface markers of the gated immature dendritic cells, as depicted in
forward scatter (FSC) vs. side scatter (SSC) dot plots.

Percoll separation are not completely separate, but collected. To ensure high recovery, the starting numb-
overlap partially. To avoid significant lymphocyte ers of MNC laid over the Percoll solution should not
contamination, the last 4 ml of Percoll should not be greatly exceed 200  106 cells.
 U. Repnik et al. / Journal of Immunological Methods 278 (2003) 283–292 291

6.1.4. Discarding the supernatant after the iso- surface. The advantage of isolation by adherence is
osmotic Percoll separation the avoidance of any reagents other than medium and
For efficient removal of platelets, the supernatant serum. Percoll itself is not toxic, nor are the other
above the monocytes should be almost completely reagents used in our protocol although long-term
removed. To avoid contamination with debris when exposure of the cells to hyper-osmotic Percoll might
the sediment is resuspended in a small volume of be noxious and therefore should be avoided. Although
medium and transferred to a new tube, the tube walls Percoll is not declared endotoxin-free, the monocytes
above the sediment level should not be rinsed. isolated following the described procedure retain their
capacity to differentiate into dendritic cells with a
6.1.5. Lymphocyte viability characteristic expression of cell surface markers.
To retain lymphocyte viability, the cells should be With special equipment, an automated and closed
exposed to the hyper-osmotic Percoll for as short a system procedure based on the negative immunose-
time as possible; therefore, they should be washed lection can be performed, which is, without doubt, an
immediately after the separation. advantage in the case of therapeutic applications but
rather expensive for research work and possibly
6.2. Troubleshooting inconvenient because an apheresis product is required
as starting material. Extensive removal of platelets is
If the lymphocyte contamination of the final prod- achievable. In our laboratory, negative immunoselec-
uct is higher than stated here, then modifications can tion was performed manually. The monocyte purity
be introduced at the centrifugation step or at the was similar to that obtained by the Percoll procedure,
hyper-osmotic Percoll solution preparation. but platelets were enriched along with monocytes and
could not be removed efficiently by low-speed cen-
6.2.1. Centrifugation trifugation alone.
Acceleration and deceleration rates were observed The procedure described here is simple, reliable
to significantly affect the efficiency of the separation. and cost-effective for isolating monocytes from buffy
With the Heraeus centrifuge, Megafuge 1.0, the rates coats. The procedure is still rather laborious and time-
are fixed and one can only choose between the brake consuming, but this is true for most monocyte isola-
off or on (slow acceleration and unbraked deceleration tion protocols. The yield of 75% and the purity of
can be achieved using the brake key). Where there is a 75% are comparable to, if not better than, those
wide range of options, slow acceleration and deceler- obtainable by the adherence method, while the platelet
ation rates should be selected to prevent disturbance contamination is extremely low and monocytes are
of the gradients. straightforwardly recovered in suspension.

6.2.2. Percoll solution preparation

Make a series of dilutions of volumes of Percoll, Acknowledgements
differing by 0.5 ml and each containing 10 ml of NaCl
solution, made up to 100 ml with water and then test This work was supported by a grant (no. L3-3368)
for optimal cell separation. and a scholarship (no. S4-311-001/21560/2000) from
the Ministry of Education, Science and Sport of the
In our experience, the adherence method gives Republic of Slovenia. We also thank Educell,
monocyte yields which are variable and inversely Ljubljana, especially Sasa Puhar, for performing
proportional to the purity of the monocytes. To ensure LAL tests.
efficient separation, the concentration of cells and the
cell density per unit surface should be relatively low.
References
The procedure therefore becomes quite laborious with
larger volumes of blood. To concentrate the mono- Almeida, M.C., Silva, A.C., Barral, A., Barral Netto, M., 2000. A
cytes after isolation or recover them in a cell suspen- simple method for human peripheral blood monocyte isolation.
sion, they need to be detached from the adhering Mem. Inst. Oswaldo Cruz 95, 221.
292 U. Repnik et al. / Journal of Immunological Methods 278 (2003) 283–292

Bøyum, A., 1983. Isolation of human blood monocytes with nyco- cytes for dendritic cell generation via discontinuous density
denz, a new non-ionic iodinated gradient medium. Scand. J. gradient centrifugation based on diluted Ficoll – Paque plus.
Immunol. 17, 429. Int. Arch. Allergy Immunol. 128, 73.
Fluks, A.J., 1981. Three-step isolation of human blood monocytes Palucka, K.A., Taquet, N., Sanchez-Chapuis, F., Gluckman, J.C.,
using discontinuous density gradients of Percoll. J. Immunol. 1999. Lipopolysaccharide can block the potential of monocytes
Methods 41, 225. to differentiate into dendritic cells. J. Leukoc. Biol. 65, 232.
Lehner, M., Holter, W., 2002. Endotoxin-free purification of mono-