A note about scale:

The methods described here are most suited to the 1-10g production range of a normal rat or
mouse hybridoma. We would recommend the following methods as a rough guide for different
scales of operation:
Less than 10mg:
use normal spent culture supernatant and protein A (or G) purification.
10-100mg:
use 1 litre roller bottle cultures and protein A (or G) purification after concentration.
100mg-1g:
use "dialysis cartridge" roller culture system such as the "MiniPerm" bioreactor now
available from Vivascience Ltd. Alternatively, we now use the CELLine System from
Integra available from BD BioSciences, that is a simple to use, two chamber flask.
1g-10g:
use our hollow fibre system (or a commercial equivalent).
10g-10Kg:
now you are into giant air lift fermenters and custom built factories!

Why the Therapeutic Immunology Group need LARGE quantities of antibodies

We are studying transplantation tolerance induction and treatment of autoimmunity with CD4,
CD8 and other T-cell specific monoclonal antibodies in vivo. Experimental studies use 110mg/experiment while pre-clinical studies use 100-1000mg/course. Therefore we need to
produce 10-40 g each of 6-8 different monoclonal antibodies per year (preferably free of
endotoxin), and it is this that has led us to develop our own methods for large scale production.
In addition, a number of genetically engineered humanized mAbs and structural variants are
required from transfected cells in sufficient quantitiy for in vitro functional and structural studies,
in some cases leading to pre-clinical studies. Therefore, in total we require 1-10 g of each of a
number of mAb variants to allow selection of those which can be moved to pilot scale production
in the Therapeutic Antibody Centre (TAC) for clinical trials.
For laboratory experimental purposes we do not need the high levels of quality assurance
required by the TAC, and we do not wish to have to optimise growth conditions for a whole
range of different cell lines. We therefore chose a hollow fibre based method that has the
following advantages:

high antibody concentrations that minimise and simplify purification

no need to develop serum free lines/optimal medium combinations

minimal serum usage

simple to keep sterile and endotoxin free

low capital and consumable costs

relatively compact and convenient

Total experimental antibody production target is 150 g per year, but total budget allowed for
consumables is 30,000 pounds/year. Therefore we must keep costs below an average of: 200
pounds per gm.

Strategies to minimize mAb production costs

Ascitic fluid must be phased out, but we need antibody produced at high concentration/small
volume to minimize handling and purification costs.
Equipment must have low cost consumables: we use hollow fibre based on cheap kidney dialysis
cartridge.
Run system in air in available 37deg C warm room to cut costs of gassed incubators, and for
convenience (we also use trolleys carrying both the bioreactor and medium bags so that sterile
connections etc can be made in a hood).
Select high producing cell lines/clones where possible, but definitely reclone until all clones are
positive for antibody production over 3 month period to avoid outgrowth of negative variants.
Major cost component is medium:

negotiate bulk purchase in 20L disposable bags: we use a special formulation of IMDM
w/o HEPES from Gibco/Life Technologies (Cat No: 041/91344R).

use serum at 5% only on cell (EC) side.

use HEPES (the major cost component) free medium .

use high glucose, rich medium (eg. IMDM) likely to give best yields for a range of
different cell lines (with minor modifications can be made suitable for Glutamine
Sythetase selection).

Purification of Endotoxin Free mAbs

The following method is recommended for purification of monoclonal antibodies produced

by tissue culture (hollow fibre cartridges or roller cultures) for laboratory use.

General notes:

1) Endotoxin - you must assume that all standard laboratory glassware will be coated with loads
of endotoxin. Wherever possible use sterile, specifically endotoxin-free (clinical grade)
disposable plastics. If you need to use glassware or other materials that are not guaranteed
endotoxin free you should soak them in 0.5M NaOH and then rinse thoroughly with endotoxin
free water. Likewise, assume any source of water or buffers are contaminated unless you have
checked them using the standard LAL assay or similar (all starting reagents, including medium
and buffers should be less than 1EU per ml). Generally speaking, once you have endotoxin in
any of your reagents or samples it is alomost impossible to remove it (even though you will find
adverts that claim various columns can remove endotoxin, we have not found those we have tried
to be very efficient).
2) Concentration - if convenient volumes (eg. less than 1 litre) are being processed you can
proceed straight to the ammonium sulphate precipitation step. If you have larger volumes, or
your antibody is present at low concentrations/poor yields, we find it useful to concentrate the
supernatants by circulating them through the same type of hollow fibre dialysis cartridges we use
for cell growth, allowing the slight pressure of circulation to force water out through the
membrane, and rapidly concentrating the antibody/proteins up to 100 fold. The advantage of this
method is very large volumes can be processed using sterile and endotoxin free components.
Ammonium Sulphate Precipitation
Prepare saturated (NH4)2SO4, using Analar Grade reagent, 1kg per litre endotoxin free water,
and autoclave in NaOH treated glass bottles. Store at 4 deg C. Mix supernatant above excess
crystals before use to ensure even saturation.
Add equal volume of saturated (NH4)2SO4 to your (concentrated) antibody sample. Mix well
and leave to stand at 4 deg C (overnight if not previously concentrated). Spin out antibody
precipitate (eg 30 mins 3500rpm using 50ml Falcon tubes in bench centrifuge). Discard
supernatant, and redissolve pellet in endotoxin free water. Repeat saturated (NH4)2SO4
precipitation.
Redissolve Ig fraction pellet in endotoxin free water and dialyse against PBS using at least 3
changes of 100x volume. (Dialysis can be performed either in hollow fibre cartridges, with
antibody circulating through the fibres and buffer circulating in the opposite direction outside the
fibres, for very large volumes, or using standard dialysis tubing that has been boiled in endotoxin
free water containing 10mM EDTA and subsequently stored sterile in endotoxin free water
containing 0.1% sodium azide).
Finally, centrifuge (at least 3500rpm for 30 min in bench centrifuge) and 0.2 micron sterile filter
to remove any grossly aggregated material, and store frozen at -20 deg C.
If your supernant was produced in a hollow fibre cartridge with good yields (ie more than
1mg/ml with less than 5% FCS) then you should expect that more than 90% of the protein visible
on a native gel electrophoresis will be the monoclonal antibody. In this case we normally
estimate antibody concentration simply by the OD280/1.4 value. This material is suitable for
most purposes, including in vitro conjugation to FITC or biotin, or for experimental use in vivo.

Levels of endotoxin at this point should be less than 20EU per ml (or less than 10EU per mg of
protein) using the LAL assay.

Ion Exchange Purification

This step is necessary if your starting yield was lower than 1mg/ml or if higher purity is required.
Ion exchange, rather than affinity purification (eg protein A or G) is recommended, because ion
exchange materials are high capacity, cheap enough to be disposable, and can easily be sanitized
with 0.5M NaOH. The following method is applicable to most rat and mouse IgG antibodies (not
IgM), although individual monoclonal antibodies may require modification of either the pH or
ionic strength for optimal purification. Cation exchange using Fast Flow S-Sepharose (Sigma S
1264) at pH5.4 has the advantage that most antibodies will bind, while albumin and degraded
antibody will not. This allows reasonable yields in a simple, step-wise absorption, wash, elution
method performed entirely in tubes with centrifugation, minimizing the risks of introducing
endotoxin by the use of columns etc. However, for maximal purity, and elimination of all
residual bovine Ig and transferrin, elution and fractionation with a slowly increasing NaCl
concentration would be advantageous.
Dialyse your (NH4)2SO4 precipitated antibody preparation into 1x Malonate buffer (buffer A).
At this point you will probably get quite a large precipitate - this is mostly degraded proteins cellular components, FCS and denatured antibody (NB. you get very little precipitate from
ascitic fluid preparations).
To make a 10x stock of buffer A:
Malonic Acid
104g
NaOH
60g
Betaine
20g
Make to 2 litres with endotoxin free water.
(NB. This gets quite hot)
should be pH 5.2 - pH 5.4 at 25 deg C

(pH very temperature sensitive - always use at 20-25 deg C)

Sterile filter to 0.2micron and store at 4 deg C
(bugs will grow in this very well if not kept sterile!)
To make Buffer B (1x):

Make 1x buffer A, but containing 0.5M NaCl

Filter to 0.2micron if not used immediately

Spin out precipitate and discard. If you dialysed at 4 deg C allow to warm to room temperature.
Ensure all steps from here are with room temperature buffers etc.
Measure OD280 and adjust to between 1-20 mg/ml in buffer A. You will need to use the
minimum amount of Fast Flow S-Sepharose to maximise yield, so estimate the total amount of Ig
in your preparation (ie. approx. purity on gel x OD280/1.4) and use 1ml of packed gel for every
20mg of antibody. If you use excess the eluted antibody will be more dilute.
Prepare Fast Flow S-Sepharose: if this is from a fresh bottle it is usually sufficient just to wash it
three times (by centrifugation for 2 mins at 2000 rpm in bench centrifuge) in buffer A. Otherwise
you should first wash once in 0.5M NaOH, followed by buffer B, and then three washes with
buffer A.
Incubate the antibody previously dialysed into buffer A with correct volume of S-Sepharose for 1
hour at room temperature, with gentle rotation.
Remove supernatant (keep and check that antibody has been adsorbed by running analytical gel
at end), and wash three times with buffer A (each wash at least 5 gel volumes).
Elute antibody by adding 1 gel volume of buffer B, rotating at room temperature for 5 mins,
centrifuging, and collecting supernatant. Repeat with another column volume and pool eluates.
It is a good idea to make certain that no Sepharose remains that could re-adsorb antibody in the
eluate, either by spinning again or passing the antibody through a sterile 0.2 micron filter.
Dialyse against two changes of PBS, check final OD280 and purity on SDS reducing PAGE and
native PAGE. Generally, approx. 95% of the protein should be antibody, appearing as a single
band (or cluster of close bands) on native PAGE. With optimal conditions, yields should exceed
70% and the final Ig concentration should be from 5-15 mg/ml.

Steve Cobbold 31/7/96

Home-made Hollow fiber reactor