IMMUNOAFFINITY PURIFICATION OF MONOCLONAL ANTIBODIES: IN SEARCH OF AN. ELUTION BUFFER OF GENERAL APPLICABILITY. Ben-David,A.
BIOTECHNOLOGY
TECHNIQUES
Volume IO No. 10 (October 1996) pp.7994302 Received as revised 2 September
IMMUNOAFFINITY PURIFICATION OF MONOCLONAL ANTIBODIES: ELUTION BUFFER OF GENERAL APPLICABILITY Ben-David,A.
IN SEARCH OF AN
& Firer,M.A.*
Laboratory of Immunology and Cell Biology, Research Institute, 44837, Israel. College of Judea and Samaria, P.O. Box 3, Ariel,
Academic
SUMMARY Of a number of elution buffers tested, 4MgC12 in 25% ethylene glycol was the most eficient for recovering active monoclonal antibodies from immunoaffinity columns. This solution may be useful as a general purpose eluant for monoclonal antibody purification. INTRODUCTION is commonly used to purify antibodies but Immunoaffinity chromatography only some aspects of the technique have been well studied. Knowledge of the polymeric structure of the solid phase and the chemistry of ligand attachment provided methods for preparing solid phases with improved flow rates and high capacities (Harlow & Lane, 1988; Johnstone & Thorpe, 1990). There remains the need to optimize the recovery of bound antibody to the mobile phase. A variety of elution buffers have been recommended 1988) but some of these cause changes to protein (Harlow & Lane, conformation resulting in the need to compromise between the quantity of protein recovered and retention of antibody activity. Recently, Tsang and Wilkins (1991) compared 13 elutions buffers and found that a combination of MgCL2 and ethylene glycol gave high yields MM of both protein recovery and antibody activity. This study, as well as that of Cachet et al (1994), dealt with polyclonal antibodies but there is also need for more efficient methods to purify monoclonal antibodies (mAbs). We set out to compare the ability of several elution buffers to recover active mAbs from immunoaffinity columns overlayed directly with cell culture supernatants. Our aim was to test if recovery of active mAb was an intrinsic property of the buffer or was mAb dependent.
EXPERIMENTAL WETHW Cells
and mAbs
mAbs
from
two different
mouse hvbridomas
799
were used.
U763 (Prof.
Z.
Eshhar, Weizmann Institute, anti-DNP and Ho-30-12 produces Thy 1.2 (Firer et al, 1995). or RPM1 (for Ho-30-12) with separated supernatants stored Preparation
Rehovot, Israel) secretes monoclonal IgGl monoclonal IgM to the surface iso-antigen Cells were grown in either DMEM (for U763) 10% FCS, 2mM glutamine and antibiotics and at -20°C.
of immunoadsorbant
Rabbit anti-mouse Ig was covalently bound to oxyrane acrylic beads (Eupergit C, diameter 150um, Sigma according to the average Approximately 20mg of antibody were bound manufacturer's instructions. beads were processed without antibody. Antibody to lg of beads. Control orientation was judged qualitatively by testing the ability of coated beads to capture U763 in an enzyme immunoassay. Affinity
Chromatography
6ml of processed beads were packed into a plastic column of 2cm wide and equilibrated with 20 bed volumes of O.OlM PBS, pH 7.2. 60 ml of pooled supernatant were passed over the column at 6ml/hr. The column was washed with PBS until the Optical Density (OD) of the effluent at 280nm was ~5% of baseline. 20ml of eluting buffer (see Table 1) were passed over the lml fractions were collected and their OD measured directly at column, the collection tubes received 200~~1 of 280nm. With TG buffer, so as to minimize the potential neutralizing 2M Tris prior to elution Eluted fractions were tested effects of low buffer pH on mAb activity. for antibody purity by 10% SDS-PAGE using non-reducing conditions. Table
I:
Buffers
used to elute
mAbs from immunoaffinity
column
Code
Composition
PH
MC
4M MgC12 O.lM Hepes
8.0
3M MgC12 0.075M Hepes 25% Ethylene Glycol
7.1
TG
0.2M Glycine
2.7
UT
8M Urea 0.05M Tris
8.0
MCE
ELISA assays: Mouse Ig ELISA: Microplates (NUNC Maxisorp) were coated with 1Opg rabbit anti-mouse Ig/ml in PBS. Wells were washed with PBS/Tween-20 and blocked with PBS/Tween/S% fish gelatin (Sigma). Samples were serially diluted in PBS/Tween/0.75% fish gelatin/0.3%BSA and added to plates for lhr at RT. IgG or washed and affinity-purified anti-mouse Wells were emptied, IgM-horseradish peroxidase added. After incubating and washing, TMB substrate containing 0.06% H202 was added for 10 mins at RT. The 2MH2S04 and ODs at 450nm read on an ELISA reaction was stopped with Ig concentrations were quantitated by reference to an U763 reader. standard pool previously standardised by ELISA against purified mouse I!iJ800
Anti-DNP ELISA: The assay was similar to that described above but plates were coated with 5 g of DNP-BSA/ml (Prof.Z.Eshhar, Weizmann Institute) in 0.06M Na Carbonate, pH 9.8. mAb activity was converted to DNP Units/ml by constructing, from repeated assays, a U763 standard curve and arbitrarily assigning the activity of undiluted supernatant as 1000 DNP Units/ml. The functional activity of the Ho-30-12 mAbs were not determined. RESULTS AND DISCUSSION Table 2 shows data on the quantity and quality of mAbs recovered. The most efficient buffer was MCE. With U763, MCE recovered higher total quantities of mAb (MCE/UT=5.9; MCE/TG=l.2; MCE/MC=1.06) and in a more concentrated form (MCE/UT=17.8; MCE/TG=4.4;MCE/MC=2.33). Antibodyantigen complexes depend on different non-covalent bonds such as ionic, hydrogen bonds, Van der Waal forces and hydrophobic interactions. The ionic strength of 3M MgC12 and low pH of TG appear to overcome enough of these bonds to dissociate the complexes and allow mAb recovery. The addition of ethylene glycol seems to result in faster dissociation, maybe by creating favourable hydrophobic conditions in the mobile phase. Poorer results were obtained with UT. The buffer gave good protein recovery, but it causes disulphide bond reduction that would reduce the antibody's functional capabilities. The advantage of adding ethylene glycol to MC was also seen in total and specific anti-DNP activity ratios (MCE/MC=4.3 and MCE/MC=3.8 respectively). This was an unexpected result as MgClP has been used aa successfully to elute enzymes (Civalleri et a1,1986) and active polyclonal antibodies (Tsang et al, 1991; Boyd and Yakazaki, 1993). Possibly with polyclonal sera any effect of MC on the combining site structure of a particular clone is compensated for by other, more stable clones. Ho-30-12 mAb activity could not be analysed for this study so the effect of MC on other mAbs needs to be assessed further. Anti-DNP activity was also successfully recovered with TG. In the study by Tsang et al (1991) this buffer gave poor qualitative and quantitative results even though it is commonly used in many labs. In our hands, TG gave higher specific activity than MCE (TG/MCE=1.5) but lower recovery of total activity (TG/MCE=0.66). With Ho-30-12 the lower quantities of antibody recovered (Table 2) reflect the reduced level of total antibody secretion by these cells (data not shown). Nevertheless, MCE was again the most efficient buffer. MC and TG did not recover active antibody, although the chromatographs of these elutions showed similar protein recovery levels as for U763 (data not shown). Also in contrast to U763, UT succeeded in recovering antibody. These data highlight an important technical point immunopurification processes for mAbs. Due to possible
801
in the design differences
of
Table
II:
Recovery affinity
and immunological characteristics of mAbs after column elution with different buffers U 763
p~,$l
total ,ug/gG
DNP U/ml
total DNP U
9.6
42
90
396
9
16711
0.3
1.6
22.4
45
800
1600
36
71680
2.9
22.0
TG
5.1
37
179
1289
53
47505
0.4
1.6
UT
1.3
8
28
165
22
1246
3.2
7.0
Buffer MC MCE
1. 2.
Ho-30-12 Activity SpI Total2
pg/ml IgG
specific activity of eluted antibody was determined determined as total DNP U x total pg IgG of eluted
total PgIgG
as DNP U/c(g IgG antibody.
in immunoglobulin isotype and almost certain alteration in combining site configuration, each mAb clone may react differently to the environmental changes induced by various elution buffers. Reactivity of individual clones may not be detected when polyclonal sera are purified, which may explain why buffers like TG have been used with general success in the past (Harlow & Lane, 1988; Tsang & Wilkins, 1991; Boyd & Yamazaki, 1993). Our data suggest that one solution to this problem is the addition of 25% ethylene glycol to MC buffer. This has been used for polyclonal sera (Tsang & Wilkins, 1991) and it was an efficient eluant If this is confirmed with the mAbs of different isotypes tested here. purpose mAb elution for other mAbs, then MCE could be used as a general This will be important not only in the laboratory but al so for buffer. biotechnology as it could reduce process development time and therefore production costs. REFERENCES Boyd,S.
and Yamazaki,H.
Civalleri,L., 8; 387-392.
(1993).
Federico,R.
Hasnaoui,M., Cochet,S., and Bertrand,O. (1994). Firer, gi-.,
MA., Ben-Zion, 31;382-388.
Biotech.
and Pi ni,C.
Letters, (1986).
7; 827-832.
Biotech.
Aool.
Debbia,A., Kroviarski,Y., Lambin,P., J.Chromatog., 663; 175-186. Z.,
Kostikov,
M. and Irlin,Y.
a laboratory Antibodies: Harlow,E. and Lane,D. Harbor Laboratory (Publ.,) 1988. Chp.13.
(1995). manual.
Johnstone, A. and Thorpe, R. (1990). Immunochemistry Scientific Publications, Chp.10. 2nd. Ed., Blackwell Tsang,
VCW. and Wilkins,
PP. (1991).
J.Imnunol.Methods,
802
Biochem., Cartron,J. Isr.
J. Med.
Cold
Spring
in Practice. 138; 291-299.
