Production and purification of murine monoclonal antibodies: Aberrant elution from protein A-Sepharose 4B

ANALYTICAL

BIOCHEMISTRY

142, 189-195 (1984)

Production and Purification of Murine Monoclonal Aberrant Elution from Protein A-Sepharose JOHN R. STEPHENSON,

Antibodies: 48

JOHN M. LEE, AND PETER D. WILTON-SMITH

Vaccine Research and Production Laboratory, PHLS Centre for Applied Microbiology and Research, Porton Down, Salkbury, Wiltshire, United Kingdom Received February 6, 1984 A rapid, one-step method for the efficient purification of murine monoclonal antibodies from tissue culture supematants is described. This process is based on affinity chromatography on protein A-Sepharose columns. It was found that murine monoclonal antibodies raised against tick-borne encephalitis virus frequently eluted at more than one pH value and these pH values did not always correspond to those of antibodies of the same subclass from polyclonal mouse sera. The two populations of antibody molecule eluting at different pH values showed no variation in molecular weight, isoelectric profiles, specific enzyme-linked immunosorbent assaytiter, or antibody subclass. 0 1984 academic mess h. KEY WORDS: monoclonal antibodies; protein A; hybridoma; flaviviruses; tick-borne encephalitis; athnity chromatography.

During the last decade, monoclonal antibodies, produced from hybridoma cell lines, have found a wide and varied use in many areas of biomedical research (reviewed in (1,2)). Although monoclonal antibodies can be readily produced against most antigens, their production in large amounts involves either their growth as ascitic tumors in mice or the production of large amounts of tissue culture fluid. In the former case, the amount of antibody produced is high, but is frequently contaminated with other mouse immunoglobulins. In the latter case, a large volume of dilute antibody solution is produced, free of other mouse proteins, but contaminated with bovine serum proteins. In either case, estimation of the amount of specific antibody is difficult, and thus it is hard to compare results obtained with different antibodies or different preparations of the same antibody clone. The purification of murine immunoglobulins by binding to protein A-Sepharose has been previously described by Ey et al. (3), and this methodology has been adapted to purify

murine monoclonal antibodies from large volumes of tissue-culture fluid. Unlike these authors, however, we have found that murine monoclonal antibodies do not always elute from protein A under the same conditions as their polyclonal counterparts. This phenomenon has been studied by the methods described in this report, and a method is described for the efficient largescale preparation of highly purified monoclonal antibodies, avoiding the use of a large number of animals. MATERIALS

AND METHODS

Preparation of monoclonal antibodies. A library of monoclonal antibodies was raised against an inactivated preparation of tickborne encephalitis virus (TBEV)’ as described by Stephenson et al. (4), and their polypeptide ’ Abbreviations used: TBEV, tick-borne encephalitis virus, DMSO, dimethyl sulfoxide, ELISA, enzymslinked immunosorbent assay; RIP, radioimmune precipitation; IgG, immunoglobulin G. 189

0003-2697184 $3.00 Copyright Q 1984 by Academic Press. Inc. All right3 of reproduction in any form reserved.

190

STEPHENSON,

LEE, AND WILTON-SMITH

specificities, biological activities, type, and subclass were also described. The hybridoma lines were stored at - 196°C in RPM1 containing 10% fetal calf serum and 10% DMSO and had a viability rate >50% under these conditions. To maximize the recovery rate from storage the cells were recloned immediately after resuscitation as follows. Cultures were rapidly warmed to 37”C, resuspended and washed in fresh growth medium, and seeded into culture Basks at 2-4 X 10’ cells/ml. Mouse peritoneal macrophages were used as feeder cells and added to a concentration of 4 X lo4 cells/ ml. Once logarithmic growth was established, the cultures were expanded to maintain a cell density between 2 and 8 X 10’ cells/ml until the required volume of culture had been reached. Best yields were obtained by allowing the cells to overgrow and die as antibody was released from the lysed cells. However, some monoclonal antibodies de-

FIG. 1. Elution of IgG2, monoclonal antibodies from protein A-Sepharose columns. The columns were loaded and eluted as under Materials and Methods while monitoring the eluate at ODz,. (A) T6; (B) T7; (C) T9.

FIG. 2. Elution of monoclonal antibodies from protein A-Sepharose columns. Running conditions were as in Fig. 1. (A) T4; (B) T33/1; (C) T12.

teriorate significantly in these terminal stages due to the low pH and the activity of cellular proteases. Where this is known to occur, the antibody is harvested as soon as the culture reaches the desired volume and density. Pwijication of monoclonal antibodies on protein A-Sepharose columns. Up to 1 liter of cloned hybridoma cell culture was clarified by centrifugation at 2000g for 10 min and then at 1OOOOg for 10 min. The resultant supernatant was mixed with a one-half volume of sodium phosphate buffer (0.1 M, pH 8.0) and the pH of the final solution adjusted to 8.0 with Tris base. The tissue culture fluid was then passed at 0.5 ml/min over a column of protein A-Sepharose 4B ( 1.5 ml of swollen gel-total binding capacity 37.5 mg IgG) which had been previously equilibrated in phosphate buffer. When all the tissue culture fluid had been applied to the column, it was washed with phosphate buffer until the ODZSO of the eluate had reached background levels. The column was then washed with sodium

CHROMATOGRAPHIC

MONOCLONAL TABLE

ANTIBODY

191

PURIFICATION

1

SUMMARYOFTHEELUTIONPROFTLESOFVARIOUS ANTIBODYCLONES ONPROTEIN A-SEPHAROSE~B Clone

Subclass

Polype.ptide specificity

Percentage antibody@ at pH 6.0

T4

w%

51K

T6

W2A

Tl

I62A

T9

Iti2A

Tll

I&2*

Igc2il

58K 58K 58K 58K 58K 51K 51K 58K 58K 51K 51K

66.6 0 0 45 61 38 60 1.5 29 12 20.5 6

W2e

58K

23

51K 51K

0 26

T13 T17 T18 T35/1 T35/3 T15 T33/1 T33/2 T33/3 T12

W2A I@2A W2A I&2A W2A w2s

w2e W3

Percentage antibody at pH 4.5

Percentage antibody at pH 3.5

33.3

0

100 56 55 39 62 40 92.5 71 88 71.5 28 77 85.2 74

0 44 0 0 0 0 0 0 0 0 66 0 14.8 0

‘The amount of antibody was estimated by monitoring the eluate at OD 2t0, calculating the area under each peak, and expressing each area as a percentage of the total uv-absorbing material eluting from the column.

phenson et al. (6). Proteins were visualized by Bio-Rad silver stain. Isoelectric focusing. The isoelectric points of the monoclonal antibody solutions were measured as follows. Approximately 2 ml of each solution was dialyzed at 4°C against a 25 mM solution of ammonium bicarbonate and lyophilized. The dried protein was taken up in 50 ~1 of the ampholyte solution and 20 ~1 applied to a prerun gel. Electrophoresis was for 2 h under the conditions described by LZs et al. (7) and the gel was stained and fured in a solution of Coomasie blue (0.25%), Enzyme-linked immunosorbent assays methanol (40%), and Formalin ( 12%). (ELISA). Tissue culture supematants and Reagents. Protein A-Sepharose 4B and purified antibody solutions were analyzed in ampholytes were obtained from Pharmacia. an ELISA system using purified inactivated Peroxidase-linked antisera were purchased antigen from TBEV-infected suckling mouse from Dako. brain as previously described (4). RESULTS Radioimmune precipitation (RIP). The polypeptide specificities of the antibody prep- Purljication of Monoclonal Antibodies on arations were determined by RIP as described Protein A-Sepharose previously (5). Polyacrylamide gel electrophoresis(PAGE). When monoclonal antibodies were eluted Antibody preparations were analyzed by from protein A-Sepharose columns by sePAGE on 15% discontinuous SDS gels under quential washing with citrate buffers, as dedenaturing conditions as described by Ste- scribed under Materials and Methods, anticitrate buffer (0.1 M) at pH 6.0 and the eluate collected until the ODzso had again reached background levels. The column was then washed sequentially in a similar manner with citrate buffers at pH 4.5 and 3.5. All eluates were adjusted to pH 8.0 with T&chloride buffer immediately upon elution. The column was finally washed in phosphate buffer and stored at 4’C. All procedures described above were performed at 4°C. The eluates from the columns were either made 0.001% w.r.t. sodium azide and stored at 4°C or stored at -20°C.

192

STEPHENSON,

LEE, AND WILTON-SMITH

TABLE 2 EFFICIENCYOF MONOCONAL ANTIBODY PURIFICATION BY AFFINITY CHROMOTOGRAPHY ON PROTEIN A-SEPHAROSE Antibody preparation T4-TCF” T4-TCFD ’ T4-pH 6.0’ T4-pH 4.5 T4-pH 3.5 T17-TCF T17-TCFD T17-pH 6.0 T17-pH 4.5 T17-pH 3.5 T33/1-TCF T33/1-TCFD T33/1-pH 6.0 T33/1-pH 4.5 T33/1-pH 3.5 Tl2-TCF TlZ-TCFD T12-pH 6.0 TlZ-pH 4.5 TlZ-pH 3.5

Total ELISA unitsd 2.12 x 2.40 x 1.51 x 3.76 x
10’ 10’ lo3 lo4 108 10’ 10’ 10’ 10’ 10’ lo6 10’ 10’ 10’ 103 10’ lo4

Yield 6) 100 11.3 71.2 17.4 loo 0.4 42.1 57.1 100 0.3 5.3 28 57 100 1.4 19.1 76.0 -

a TCF = clarified tissue culture fluid. ’ TCFD = clarified tissue culture fluid after absorption with protein A-Sepharose. cColumn eluate fractions as described in Fig. 1. d One ELISA “unit” is that ELISA titer given by 0.1 ml of an antibody solution when assayed as described under Materials and Methods.

terial elute at pH 3.5 and all four clones showed different elution pro&s. Although the only clone available in subclass 1 eluted at pH 6.0, as expected from that found with polyclonal sera, a @ikant proportion eluted at pH 4.5. The only available clone from subclass 3 did not elute at all at pH 3.5, and this particular clone demonstrated an elution profile more reminiscent of an IgGzA type of molecule. Eficiency of PuriJication by Protein A A@nity Chromatography To examine the efficiency of the purification procedure, the ELISA titers of the various column eluates were compared to those of the crude tissue culture supernatants (Table 2). With all the clones examined, recoveries were between 89 and 99% or higher, regardless of their IgG subclass. There was no evidence of any significant amount of material being retained on the protein A-Sepharose column.

1234 -2OOK -925K -69K 09 we

bodies could be found in all three eluates (Figs. 1 and 2). Although the majority of IgGzA antibody molecules in polyclonal mouse serum elute at pH 4.5 (3), the monoclonal antibodies of this subclass used here did not elute exclusively at this pH, and did not demonstrate an elution pattern which was specific for this antibody subclass (Fig. 1 and Table 1). Of all the monoclonal antibodies used in this study, only T6 demonstrated an elution profile similar to that of IgG% antibodies from polyclonal mouse sera (Fig. 1). When the clones belonging to the other subclasses were eluted from the columns, they too demonstrated abnormal elution protiles (Fig. 2 and Table 1). With only one of the 2B subclass clones did most of the ma-

-46K -3OK

-14*3K

FIG. 3. RIP of [‘5S]methionine-labeled infected cell lysates with monoclonal antibody solutions eluted at various pHs from protein A-Sepharose columns. Track 1, TlS-pH 6.0; Track 2, Tl5-pH 4.5; Track 3, Tl lpH 6.0; Track 4, Tl I-pH 4.5. The positions of the M, markers are shown on the right-hand side.

CHROMATOGRAPHIC

MONOCLONAL

Comparison of the Antibody Species Eluting at Each pH Although the monoclonal antibodies bound to the protein A-Sepharose column eluted at different pH values, the polypeptide specificity (Fig. 3) and the IgG subclass of each pH fraction was the same as that of the original tissue culture fluid. In addition the specific ELBA titers of each population (expressed as ELBA units per milliliter of a 1 mg/ml solution) were very similar (Table 3). The molecular weights of both the light chains and the heavy chains of each pH eluate from

ANTIBODY

193

PURIPICATION

an individual monoclonal antibody population were similar when analyzed by PAGE (Table 3 and Fig. 4). It is also noted from Fig. 4 that, although the molecular weights of each pH eluate from an individual monoclonal antibody population are the same, the monoclonal antibodies examined in this study have light chains of two distinct size classes (Fig. 4 and Table 3), regardless of subclass or of immunization protocol (data from (4)). Finally, the isoelectric profile of each antibody solution was determined by isoelectric focusing in a polyacrylamide gel. Under the conditions employed in these studies, each in-

TABLE 3 PROPERTIES

OF MON~CLONAL ANTIBODY SOLUTIONS ELUT~NG AT VARIOUS pH VALUES FROM PROTEIN A-SEPHAROSE COLUMNS

Clone

pH of elution

T4 T4 T6 T7 T7 T9 T9 Tll Tll T12 T12 T13 T13 T15 T15 T17 T17 T18 T18 T33/1 T33/1 T33/1 T33/2 T33/2 T33/3 T33/3 T35Jl T35jl T35/3 T35/3

6.0 4.5 4.5 4.5 3.5 6.0 4.5 6.0 4.5 6.0 4.5 6.0 4.5 6.0 4.5 6.0 4.5 6.0 4.5 6.0 4.5 3.5 6.0 4.5 4.5 3.5 6.0 4.5 6.0 4.5

Specific ELISA titer 1.86 x 2.25 x 5.50 x 1.87 x 7.50 x 1.36 x 2.45 x 5.88 x 6.25 x 1.40 x 1.13 x 1.98 x 1.32 x 3.60 x 4.10 x 3.06 x 5.21 x 2.42 x 3.64 x 1.77 x 9.34 x 1.43 x 3.28 x 2.52 x 1.32 x 2.25 x 7.91 x 1.25 x 8.53 x 8.56 x

10’ 10’ IO4 lo4 lo4 IO6 lo6 10’ 10’ IO5 10’ 10’ lo5 10’ 103 10s 10’ 10’ 10’ IO4 10’ lo4 lo3 lo3 lo4 lo4 102 ld 10’ 10’

Molecular mass (light chain)”

Molecular mass (heavy chain)”

25 25 25 25 25 28 28 25 25 28 28 25 25 25 25 25 25 25 25 25 25 25 28 28 25 25 25 25 25 25

62 62 62 54.5 54.5 55 55 54 54 51 51 54 54 68 68 62 62 64 64 60 60 60 58 58 66 66 64 64 64 64

’ The molecular mass in kilodaltons as determined by SDS-PAGE under denaturing conditions.

194

STEPHENSON,

1234567

8

LEE, AND WILTON-SMITH

910 -94K -67K

r’irralt,~ -43K

-2OK

-14*4K PIG. 4. PAGE of momxkmal antibody solutions eluted at various pH valuea tim a protein Adkphame column. Electrophoresis was performed as under Materials and Methods and the gels were stained with Bio-Bad silver stain. The positions of the M, markers are shown on the right-hand side. Track 1, TlZ-pH 6.0 eluate; Track 2, TlZ-pH 4.5 eluate; Track 3, TI I-pH 6.0 eluate; Track 4, Tl I-pH 4.5 eluate; Track 5, T9-pH 6.0 eluate; Track 6, T9-pH 4.5 eluate; Track 7, T33/2pH 6.0 eluate; Track 8, T33/2-pH 4.5 eluate; Track 9, T17-pH 6.0 eluate; Track 10, T17-pH 4.5 eluate.

dividual monoclonal antibody demonstrated a characteristic electrophoretic mobility; however, no difference was observed in these electrophoretic mobilities when the column eluates of the same monoclonal antibody at different pH values were compared. DISCUSSION

The use of murine monoclonal antibodies is now widespread in many areas in the biomedical sciences. Growth of these monoclonal antibody-secreting cell cultures in quantities suitable for routine use requires either the production of ascitic tumors in mice or the production of large volumes of dilute antibody solutions in tissue culture. In this communication we describe a rapid onestep procedure for the purification of murine monoclonal antibodies from tissue-culture

fluid by affmity chromatography on protein A-Sepharose 4B. The use of such columns for the separation of murine antibodies from polyclonal sera was first described by Ey et al. (3) and has been widely used by other workers. When this technique was applied to a library of murine monoclonal antibodies raised against TBEV, they did not elute under the same pH conditions as did the similar polyclonal antibody preparations described by these authors, even though the yields obtained were high (Table 2). Of the 15 clones examined, only one (T6) eluted exclusively at the pH predicted for a subclass 2A molecule. All the remainder eluted at two or more pH values. Although most of the only IgG, clone eluted at the expected pH value, a significant proportion eluted at a different pH. Neither the subclass 2A nor the subclass 2B molecules eluted exclusively at the expected pH values (except T6 and possibly T18) and no one pattern of elution was shown to be subclass specific. The elution of IgG3 molecules cannot be compared to similar molecules from polyclonal sera, as these data have not been published. Several techniques were applied to elucidate any differences between the molecules eluting at different pH values. The subclass of each pH eluate was redetermined, and in every case found to be the same as the original. Similarly the polypeptide specificity, specific ELISA titer, and isoelectric profile of each pH eluate from any single monoclonal was identical. The molecular weights of the heavy chains and light chains of either eluate from any one monoclonal were also similar. It was noted that the light chains fell into two separate size classes, 25K and 28K, but these size clones were not correlated with their subclass or any other property. In conclusion, we have described a rapid and efficient method for purifying monoclonal antibodies from tissue-culture supernatants using protein A affinity chromatography. However, these studies demonstrate that the elution of monoclonal antibodies from protein A columns cannot be assumed

CHROMATOGRAPHIC

MONOCLONAL ANTIBODY PURIFICATION

to be the same as their polyclonal counterparts of an identical subclass.Therefore, it is recommended that, even though the subclass of a monoclonal antibody is known, it be routinely eluted from the protein A columns at pH 3.5. Provided that the pH of the eluates and columns are adjusted to 8.0 as soon as practicable, no degradation should occur. ACKNOWLEDGMENTS We thank Mr. A. C. Walkland for help with the isoelectricfocusing experimentsand Mrs. AngelaTabben for typing the manuscript.

195

REFERENCES 1. M&win, C. (1980) Sci. Amer. 243, 56-64. 2. Fidwards, P. A. W. (1981) Biochem. J. 200, l-10. 3. Ey, P. L., Prowse, S. J., and Jenkin, C. R. (1978) Immunochemistry l&429-436. 4. Stephenson, J. R., Lee,J. I$, and W&on-Smith, P. D. (1984)J. Gen.Virol.65, 81-89. 5. Stephenson, J. R., andter Meulen,V. (1979)Prac NatI.Acad.Sci. USA76,6601-6605. 6. Stephenson, J. R., Hay,A. J., andSkehel,J. J. (1977) J. Gen.Viral. 36, 237-248. 7. L&b, T., Olson, I., and Soderberg, L. (1980)An& B&hem. 101,449-46 1.