- Email: [email protected]
Characterization of a monoclonal antibody to human alpha-fetoprotein and its use in affinity chromatography
Journal o f Immunological Methods, 46 (1981) 337--345 Elsevier/North-Holland Biomedical Press
337
CHARACTERIZATION OF A MONOCLONAL ANTIBODY TO HUMAN ALPHA-FETOPROTEIN AND ITS USE IN AFFINITY CHROMATOGRAPHY
ULF-HAKAN STENMAN 1, MARJA-LIISA SUTINEN 1, RITVA-KAJSA SELANDER 2, KAIJA TONTTI 2 and JIM SCHRODER 1
Folkhi~lsan Institute o f Genetics 1, Medix Laboratories Ltd. 2, PB 819, SF-O0101 Helsinki 1 O, Finland (Received 27 February 1981, accepted 14 July 1981)
A hybrid cell line secreting monoclonal antibodies to human alpha-fetoprotein (AFP) was produced by fusion of a mouse myeloma cell line with spleen cells from a BALB/c mouse immunized with human AFP. The affinity constant of the antibody was about 1.7 × 109 l/tool. When clones were grown in vitro, the highest concentration of specific antibody in the culture medium was 25/lg/ml. A clone was transplanted intraperitoneally into pristine-primed BALB/c recipients. Ascites developed within 3--4 weeks of transplantation, and the maximal antibody concentration in the ascitic fluid was 2.5 mg/ml. The immunoglobulin fraction of ascitic fluid was coupled to cyanogen bromideactivated Sepharose and used for affinity chromatography of AFP. AFP containing less than 1% contaminating proteins was obtained by passing amniotic fluid through the column and eluting the adsorbed AFP with 4 mol/l urea. The monoclonal antibody was used for radioimmunoassay (RIA). The sensitivity obtained was 50/~g/1, which is adequate for certain clinical applications.
INTRODUCTION
Conventionally produced antisera are heterogeneous, containing antibodies of different specificities, affinities and classes. Therefore the production of homogeneous standardized antisera has been an almost impossible task. However, KShler and Milstein (1975) have shown that antibodies of predefined specificity can be produced by fusion of spleen cells from immunized mice with a mouse myeloma cell line. When cloned, such antibodysecreting hybrid cell lines produce homogeneous so-called monoclonal antibodies. During the past few years, monoclonal antibodies have been produced against a variety of antigens (Melchers et al., 1978). In this paper we describe the production and characterization of a monoclonal antibody to human alpha-fetoprotein (AFP), and its use in radioimmunoassay and in purification of AFP from amniotic fluid by affinity chromatography.
0022-1759/81/0000--0000/$02.50 © 1981 Elsevier/North-Holland Biomedical Press
338
MATERIAL AND METHODS
Immunization and fusion BALB/c mice were immunized and boosted intraperitoneally with 100 pg human AFP (a gift from Dr. E. Ruoslahti) by injection in Freund's complete adjuvant at 2-weekly intervals. After 2 months, one mouse had antibody titers of 1 : 15,000 and was then b o o s t e d twice intravenously with 100 pg of the antigen in saline. The spleen was removed 4 days after the last booster, and 1.6 × 108 spleen cells were fused in polyethylene glycol 1500 (PEG, Baker Chemicals) with 3 × 1 0 7 P3-NS1/1-Ag-4 m y e l o m a cells (BALB/c) in exponential growth. After fusion, cells were plated into U-bott o m e d microplates in Dulbecco's modified Eagle's medium (DME) with either 15% horse serum (Gibco) or 15% fetal calf serum (Gibco) at a density of 1 X 106 cells/well. Each well contained 0.1 ml of culture medium. After 1 day, 0.1 ml o f selective HAT medium (100 pM hypoxanthine, 10 phi aminopterin, 30 pM thymidine in DME + 15% horse or fetal calf serum) was added to the cultures. After this, half o f the medium was removed every other day and replaced with fresh H A T medium. After 2 weeks of culture the cells were transferred to HT culture medium (HAT w i t h o u t aminopterin), which was changed 3 times a week, and finally 2 weeks later back to DME + 15% serum. Vigorously growing antibodyproducing cultures were transferred to tissue culture tubes and expanded, and then cloned b y limiting dilution in flat-bottomed microtiter plates on human fibroblast feeders. One antibody-producing clone was recloned in 200 microtiter wells. T w e n t y subclones were expanded in test tubes and later in Falcon tissue culture flasks, and the antibody in the supernatants o f these cultures was characterized.
Transplantation of hybrid clones Antibody-producing hybrid cells (2 × 106/animal) were injected intraperitoneally into BALB/c recipients primed with pristine (2, 6, 10, 14-tetramethylpentadecane, Aldrich, Europe). Ascites developed within 3--4 weeks. The ascitic fluid was withdrawn and the antibody in the fluid tested.
Screening for antibody production The supernatants from the cultures were screened for antibody production at weekly intervals from hybridization or cloning. Titrations o f the antibodies were performed 5 weeks after hybridization o f the uncloned cultures and at different times after cloning. Unless otherwise stated, these tests were always performed on supernatants from 4-day-old cultures. Titration o f antibodies was also carried o u t on antisera from immunized mice and on ascites fluid.
Iodination AFP was iodinated with 12sI b y the lactoperoxidase m e t h o d (Thorell and
339 Johansson, 1971), and purified by gel filtration on Sephadex G-25 in PBS-E buffer containing 0.1% bovine serum albumin.
Radioassay of antibody to AFP A rapid [~2SI]AFP binding assay was used to demonstrate anti-AFP in culture supernatants. The radioassay buffer (PBS-E) contained 50 mmol/1 sodium phosphate, pH 7.4, 150 mmol/1 NaC1 and 10 mmol/1 Na2-EDTA. Fifty pJ of supernatant was incubated with 200 pl of [12SI]AFP containing about 10,000 cpm. After 3 h, bound [~2SI]AFP was precipitated by addition of 100 pl normal human serum and 1 ml 12% polyethylene glycol 6000 (Fluka) in PBS-E. After centrifugation at 1500 × g for 10 min, the supernatant was aspirated and the radioactivity of the precipitate counted. Duplicate tubes containing growth medium instead of culture supernatant served as background controls. About 15% of the added radioactivity was precipitated non-specifically in these tubes. A sample was considered positive for anti-AFP when the activity precipitated was significantly greater (about 20%) than the background. The titer was defined as the dilution binding 50% of the maximal amount of [ ~2sI]AFP specifically bound by a large excess of antibody.
Radioimmunoassay (RIA ) The ~ f i n i t y of the antibody and its concentration were determined by RIA. One hundred ~1 of [12SI]AFP, 100 pl standard or sample and 100 #l antibody solution were incubated overnight at room temperature. The antibody solution was diluted to bind 30--40% of the ['2SI]AFP; it contained 0.25 pl normal mouse serum per tube. The bound fraction was precipitated by addition of 10 #l rabbit anti-mouse immunoglobulin (Dako Immunoglobulins, code 2109) and 90 #l 10% PEG 6000 in PBS-E. The precipitate was spun down after further incubation for 4 h. AFP standards were prepared by diluting pure AFP with serum containing less than 2 #g/1 of AFP. The ratio of bound/free antigen and the concentration of bound antigen were calculated and used to prepare a Scatchard plot. From these data the affinity constant and concentration of antibody were calculated. Because of the partial sequence homology between albumin and AFP (Ruoslahti and Terry, 19.76) we also tested the effect of human serum albumin in the RIA.
Typing of antibodies The Ig isotype produced by the cell lines was studied by immunodiffusion with rabbit antisera specific for IgGl, igG2a and IgG3 (obtained from Miles Laboratories).
Isoelectric focusing Ascitic fluid and AFP prepared by affinity chromatography were studied by isoelectric focusing in thin-layer polyacrylamide gel with a pH gradient of 3.5--9.5 (Ampholine PAG plate, LKB).
340
Gel electrophoresis Gel electrophoresis in polyacrylamide containing sodium dodecylsulfate (SDS) was p e r f o r m e d according to t he m e t h o d o f Laemmli (1970).
Preparation of immunoadsorbent Two ml o f mouse ascitic fluid containing 4.5 mg anti-AFP was precipit a t ed with an equal volume o f saturated a m m o n i u m sulfate. The precipitate was washed twice with half-saturated a m m o n i u m sulfate and dialyzed against 0.1 mol/1 NaHCO 3, pH 8.3. This immunoglobulin preparation was coupled t o 2 ml o f cyanogen bromide-activated Sepharose (Porath et al., 1967).
Amniotic fluid This was obtained f r om therapeutic amniocentesis p e r f o r m e d between the 15th and 18th weeks o f pregnancy. RESULTS
Hybridization and testing of antibody production Two weeks after hybridization, 20 o u t o f 100 wells showed a n t i b o d y p r o d u c t i o n . This n u m b e r o f positive cultures rose to 25 within 5 weeks o f hybridization. One such culture showed an a n t i b o d y titer o f 1 : 5 0 0 (Table 1). Five h y b r id cultures were cloned, and t h e yield was 10 antibodyproducing hybrid clones. Three weeks after cloning, t h e binding o f AFP t o culture supernatants was 5 times as high as t he background. After recloning o f on e clone, 110 a nt i body- pr oduc i ng subclones were recovered and 20 o f these were f u r t h e r expanded. One week after recloning the a n t i b o d y titer in t h e supernatant was onl y 1 : 50, but it rose t o a b o u t 1 : 5000 after continuous culture (Table 1). The highest c o n c e n t r a t i o n o f specific a n t i b o d y in such cultures was 25 gg/ml. T h e a n t i b o d y titers in t he supernatans correlated with t he n u m b e r o f viable cells and th e time o f culture w i t h o u t change o f medium. After t he culture med iu m had been changed, a substantial rise in a n t i b o d y titer could be
TABLE 1 Titers of antibody to AFP produced by mice and hybrid cells in culture and in ascites from mice bearing intraperitoneally transplanted hybrid cells. = dilution giving half-maximal binding
Source
Titer
Mouse antiserum Culture supernatant from uncloned hybrid Culture supernatant from clone Ascitic fluid
1 1 1 1
: 15 000 : 500 : 5 000 : 200 000--1 : 800000
341
1:800
w
--
O
0.8x106
I%400
0,4
0
~
• 1:200
O
0
; 0.2
/
o~* I
I
I
'
o 2 4 6 Days Fig.1. Correlationbetweennumberof viablecellsper m] (o
o) and antibody titer ; ) in a lymphocyte hybrid culture producing monoclonal antibodies to AFP.
(=
demonstrated with each additional day of culture. However, the number of viable cells began to decrease 3--4 days after the change of medium (Fig. 1). Ascitic fluid developed within 3--4 weeks in all mice that had received an intraperitoneal injection of hybrid cells. The antibody titers in the ascitic fluid of these mice varied between 1 : 200,000 and 1 : 800,000 (Table 1), and the maximal concentration of specific antibody was 2.5 mg/ml. RIA The antibody produced by one clone was chosen for closer study. The inhibition curve obtained with RIA using this anti-AFP is shown in Fig, 2. Significant binding inhibition was obtained with an AFP concentration of 50 B/T
3020:10
-
-
~
.....
I 25
I 50
I I I00 200
Bg
I |, 500 pg/]O00
Fig. 2. Standard curve for radioimmunoassay of AFP. The abscissa gives the concentration o f AFP in the standards.
342
pg/1. This corresponds to a concentration of 6/~g/1 in the incubation mixture. This sensitivity is about one-tenth o f that obtained with our standard AFP RIA using rabbit anti-AFP and goat anti-rabbit IgG but otherwise identical conditions. Analysis o f serum samples showed no systematic difference between results with the two assays. Nor was any difference observed when AFP in amniotic fluid was studied. Human serum albumin did not cause inhibition of binding even in a final concentration o f 50 mg/ml. The affinity constant determined from the Scatchard plot was 1.7 X 109 1/mol. A straight line was obtained, indicating homogeneity of the antibody.
A
....
i ¸ ! i!ii! i i! ii ~
!: i il
:! ..... i~ I~ ~ !ii~I~:
1
2
3
4
5
~
3
1
i
:
6
Fig. 3. A: isoelectric focusing in polyacrylamide gel, pH range 3.5--9.5. High pH is at top and low at the b o t t o m of the figure. Sample 1 contains amniotic fluid, 2 contains 10/~g and 3, 20 pg of purified AFP. Samples 4 and 5 contain ascitic fluid from two different mice transplanted with the same anti-AFP producing monoclonal hybrid cell line. The strong bands at the top correspond to the monoclonal antibodies B: SDS-polyacrylamide electrophoresis in a 10% gel. Anode at the bottom. Sample numbers are the same as in A except for No. 6, which contains molecular weight standards. The molecular weights (x 10 -3) of the standards are marked on the right side of the figure.
343
Isotype The isotype o f the antibody determined b y immunodiffusion was IgG1. In isoelectric focusing ascitic fluid from one mouse displayed 3 strong bands in the pH range 8--9, and 3 additional bands were observed in the more concentrated Ig fraction prepared b y ammonium sulfate fractionation (Fig. 3A). Ascites from aI~other mouse displayed 6 bands in the same pH region, b u t the pattern was slightly different from that in the first mouse (Fig. 3A).
Affinity chromatography Sixty ml o f amniotic fluid containing a b o u t 6 mg of protein/ml and 28 #g AFP/ml were applied to a 1-ml affinity chromatography column at a flow rate of 10 ml/h. Two-ml fractions were collected and the contents of AFP and albumin were monitored b y immunodiffusion against specific antisera. No AFP was detected in the first 10 fractions; after this AFP came through (Fig. 4). This indicated that 560 gg o f AFP had been adsorbed. After the amniotic fluid had been passed through, the column was washed with 0.1 mol/1 Tris-HC1 buffer, pH 7.4, containing 1 mol/1 NaC1. When the absorbance at 280 nm fell to less than 0.005 (after 10 fractions) the column was eluted with 4 mol/1 urea in 0.1 mol/l Tris-HC1, pH 7.4. AFP was detected in the first 6 ml eluted. This fraction was concentrated under reduced pressure and dialyzed against 0.1 mol/1 Tris HC1, pH 7.4. F r o m the concentrate, 415 #g of AFP was recovered, i.e., a b o u t 74% of the AFP adsorbed. No additional AFP or protein was eluted with 8 mol/1 urea. In a separate experiment we eluted the column with a gradient o f 30 ml Tris buffer and 30 ml o f 4 mol/1 urea in Tris buffer. A b o u t 90% of the AFP was elute.d at urea concentrations b e t w e e n 1.5 and 2 mol/1. The ascitic fluid used to prepared the immunoadsorbent had a calculated
A280 6
..,
I
NaCI 1M
5
AFP
10
1
20
Urea 4M
25
30
35
F r a c t i o n no.
Fig. 4. P u r i f i c a t i o n o f A F P b y a f f i n i t y c h r o m a t o g r a p h y o n a 1-ml c o l u m n containing about 2 mg m o n o c l o n a l a n t i - A F P c o u p l e d t o S e p h a r o s e 4B. S i x t y m l o f a m n i o t i c fluid was a p p l i e d t o t h e c o l u m n . T h e A F P in the first 30 ml was a d s o r b e d o n t h e c o l u m n a n d e l u t e d w i t h 4 mol/1 urea. A F P was m o n i t o r e d b y i m m u n o d i f f u s i o n .
344
binding capacity o f 3.6 mg AFP. After precipitation of the Ig fraction and coupling to Sepharose, 16% of the calculated capacity was recovered. Portions of 10 and 20 ttg of AFP were submitted to isoelectric focusing in polyacrylamide gel. One strong band isoelectric at pH 4.7 and 3 weaker bands were observed (Fig. 3A). This microheterogeneity is typical of human AFP. Polyacrylamide electrophoresis in the presence of SDS gave only one band (Fig. 3B). These tests show that the AFP preparation probably contained less than 1% of contaminating protein. DISCUSSION
The aim of the present study was to investigate the usefulness of monoclonal antibodies for assay and purification of clinically important proteins. AFP was chosen as a model because it was available in pure form in sufficient amounts. In addition, determination o f AFP is clinically important and difficult to standardize (Sizaret et al., 1979). A continuous supply of homogeneous antibody might help to resolve this problem. In this study the antibody produced b y one monoclonal cell line was characterized. With this antibody we could demonstrate some o f the advantages inherent in the monoclonal technique. For RIA work, however, the antibody did n o t prove superior to conventional antibodies, and this has also been the experience o f groups working with other antigens (Accolla et al., 1980; Miggiano et al., 1980). Because of the only moderate affinity of the antibody, the sensitivity obtained with it in RIA was a b o u t 10 times less than that of our conventional RIA for AFP. In affinity chromatography our monoclonal antibody demonstrated useful properties which m a y be typical of monoclonal antibodies. The antigen could be eluted from the affinity column with mild reagents, i.e., 1.5--2 mol]l urea. This was probably n o t only due to the moderate affinity of the antibody, b u t m a y in part be accounted for by the fact that monoclonal antibody binds monomeric antigen at one site only. High avidity resulting from simultaneous binding to t w o antibodies is thus avoided. With gradient elution the adsorbed AFP was eluted as a fairly sharp peak. This was expected, since monoclonal antibody had a constant affinity, not a spectrum from low to high. Absence o f multipoint binding might also be important in this respect. Rather than attempting to prepare an affinity column with maximal capacity, we maximized coupling o f the limited a m o u n t of antibody available. However, the very high antibody concentration that can be obtained in ascites, in combination with low concentrations of other proteins, makes it easy to prepare concentrated antibody and high capacity columns with low nonspecific adsorption. Monoclonal antibody columns have earlier been used to purify antigenically similar H L A antigens (Parham, 1979). This unique property together with those observed in the present study make monoclonal antibodies ideal for affinity chromatography.
345 REFERENCES Accolla, R.S., S. Correl and J.P. Mach, 1980, Proc. Natl. Acad. Sci. U.S.A. 77,563. KShler, G. and C. Milstein, 1975, Nature 256,495. Laemmli, U.K., 1970, Nature 227,680. Melchers, F., M. Potter and N.L. Warner (eds.), 1978, Lymphocyte Hybridomas. Curr. Top. Microbioi. Immunol. 81, 1. Miggiano, V., C. St~/hli, P. H~iring, J. Schmidt, M. Le Doin, T. Glatthaar and T. Staehling, 1980, Protides of the Biological Fluids, Belgium, May, 1980, Abstr. 24. Parham, P., 1979, J. Biol. Chem. 254, 8709. Porath, J., R. Ax6n and S. Ernback, 1967, Nature 215, 1491. Ruoslahti, E. and W.D. Terry, 1976, Nature 260, 804. Sizaret, P., N. Breslow et al., 1979, Clin. Chim. Acta 96, 59. Thorell, J.I. and B.G. Johansson, 1971, Biochim. Biophys. Acta 251,363.
337
CHARACTERIZATION OF A MONOCLONAL ANTIBODY TO HUMAN ALPHA-FETOPROTEIN AND ITS USE IN AFFINITY CHROMATOGRAPHY
ULF-HAKAN STENMAN 1, MARJA-LIISA SUTINEN 1, RITVA-KAJSA SELANDER 2, KAIJA TONTTI 2 and JIM SCHRODER 1
Folkhi~lsan Institute o f Genetics 1, Medix Laboratories Ltd. 2, PB 819, SF-O0101 Helsinki 1 O, Finland (Received 27 February 1981, accepted 14 July 1981)
A hybrid cell line secreting monoclonal antibodies to human alpha-fetoprotein (AFP) was produced by fusion of a mouse myeloma cell line with spleen cells from a BALB/c mouse immunized with human AFP. The affinity constant of the antibody was about 1.7 × 109 l/tool. When clones were grown in vitro, the highest concentration of specific antibody in the culture medium was 25/lg/ml. A clone was transplanted intraperitoneally into pristine-primed BALB/c recipients. Ascites developed within 3--4 weeks of transplantation, and the maximal antibody concentration in the ascitic fluid was 2.5 mg/ml. The immunoglobulin fraction of ascitic fluid was coupled to cyanogen bromideactivated Sepharose and used for affinity chromatography of AFP. AFP containing less than 1% contaminating proteins was obtained by passing amniotic fluid through the column and eluting the adsorbed AFP with 4 mol/l urea. The monoclonal antibody was used for radioimmunoassay (RIA). The sensitivity obtained was 50/~g/1, which is adequate for certain clinical applications.
INTRODUCTION
Conventionally produced antisera are heterogeneous, containing antibodies of different specificities, affinities and classes. Therefore the production of homogeneous standardized antisera has been an almost impossible task. However, KShler and Milstein (1975) have shown that antibodies of predefined specificity can be produced by fusion of spleen cells from immunized mice with a mouse myeloma cell line. When cloned, such antibodysecreting hybrid cell lines produce homogeneous so-called monoclonal antibodies. During the past few years, monoclonal antibodies have been produced against a variety of antigens (Melchers et al., 1978). In this paper we describe the production and characterization of a monoclonal antibody to human alpha-fetoprotein (AFP), and its use in radioimmunoassay and in purification of AFP from amniotic fluid by affinity chromatography.
0022-1759/81/0000--0000/$02.50 © 1981 Elsevier/North-Holland Biomedical Press
338
MATERIAL AND METHODS
Immunization and fusion BALB/c mice were immunized and boosted intraperitoneally with 100 pg human AFP (a gift from Dr. E. Ruoslahti) by injection in Freund's complete adjuvant at 2-weekly intervals. After 2 months, one mouse had antibody titers of 1 : 15,000 and was then b o o s t e d twice intravenously with 100 pg of the antigen in saline. The spleen was removed 4 days after the last booster, and 1.6 × 108 spleen cells were fused in polyethylene glycol 1500 (PEG, Baker Chemicals) with 3 × 1 0 7 P3-NS1/1-Ag-4 m y e l o m a cells (BALB/c) in exponential growth. After fusion, cells were plated into U-bott o m e d microplates in Dulbecco's modified Eagle's medium (DME) with either 15% horse serum (Gibco) or 15% fetal calf serum (Gibco) at a density of 1 X 106 cells/well. Each well contained 0.1 ml of culture medium. After 1 day, 0.1 ml o f selective HAT medium (100 pM hypoxanthine, 10 phi aminopterin, 30 pM thymidine in DME + 15% horse or fetal calf serum) was added to the cultures. After this, half o f the medium was removed every other day and replaced with fresh H A T medium. After 2 weeks of culture the cells were transferred to HT culture medium (HAT w i t h o u t aminopterin), which was changed 3 times a week, and finally 2 weeks later back to DME + 15% serum. Vigorously growing antibodyproducing cultures were transferred to tissue culture tubes and expanded, and then cloned b y limiting dilution in flat-bottomed microtiter plates on human fibroblast feeders. One antibody-producing clone was recloned in 200 microtiter wells. T w e n t y subclones were expanded in test tubes and later in Falcon tissue culture flasks, and the antibody in the supernatants o f these cultures was characterized.
Transplantation of hybrid clones Antibody-producing hybrid cells (2 × 106/animal) were injected intraperitoneally into BALB/c recipients primed with pristine (2, 6, 10, 14-tetramethylpentadecane, Aldrich, Europe). Ascites developed within 3--4 weeks. The ascitic fluid was withdrawn and the antibody in the fluid tested.
Screening for antibody production The supernatants from the cultures were screened for antibody production at weekly intervals from hybridization or cloning. Titrations o f the antibodies were performed 5 weeks after hybridization o f the uncloned cultures and at different times after cloning. Unless otherwise stated, these tests were always performed on supernatants from 4-day-old cultures. Titration o f antibodies was also carried o u t on antisera from immunized mice and on ascites fluid.
Iodination AFP was iodinated with 12sI b y the lactoperoxidase m e t h o d (Thorell and
339 Johansson, 1971), and purified by gel filtration on Sephadex G-25 in PBS-E buffer containing 0.1% bovine serum albumin.
Radioassay of antibody to AFP A rapid [~2SI]AFP binding assay was used to demonstrate anti-AFP in culture supernatants. The radioassay buffer (PBS-E) contained 50 mmol/1 sodium phosphate, pH 7.4, 150 mmol/1 NaC1 and 10 mmol/1 Na2-EDTA. Fifty pJ of supernatant was incubated with 200 pl of [12SI]AFP containing about 10,000 cpm. After 3 h, bound [~2SI]AFP was precipitated by addition of 100 pl normal human serum and 1 ml 12% polyethylene glycol 6000 (Fluka) in PBS-E. After centrifugation at 1500 × g for 10 min, the supernatant was aspirated and the radioactivity of the precipitate counted. Duplicate tubes containing growth medium instead of culture supernatant served as background controls. About 15% of the added radioactivity was precipitated non-specifically in these tubes. A sample was considered positive for anti-AFP when the activity precipitated was significantly greater (about 20%) than the background. The titer was defined as the dilution binding 50% of the maximal amount of [ ~2sI]AFP specifically bound by a large excess of antibody.
Radioimmunoassay (RIA ) The ~ f i n i t y of the antibody and its concentration were determined by RIA. One hundred ~1 of [12SI]AFP, 100 pl standard or sample and 100 #l antibody solution were incubated overnight at room temperature. The antibody solution was diluted to bind 30--40% of the ['2SI]AFP; it contained 0.25 pl normal mouse serum per tube. The bound fraction was precipitated by addition of 10 #l rabbit anti-mouse immunoglobulin (Dako Immunoglobulins, code 2109) and 90 #l 10% PEG 6000 in PBS-E. The precipitate was spun down after further incubation for 4 h. AFP standards were prepared by diluting pure AFP with serum containing less than 2 #g/1 of AFP. The ratio of bound/free antigen and the concentration of bound antigen were calculated and used to prepare a Scatchard plot. From these data the affinity constant and concentration of antibody were calculated. Because of the partial sequence homology between albumin and AFP (Ruoslahti and Terry, 19.76) we also tested the effect of human serum albumin in the RIA.
Typing of antibodies The Ig isotype produced by the cell lines was studied by immunodiffusion with rabbit antisera specific for IgGl, igG2a and IgG3 (obtained from Miles Laboratories).
Isoelectric focusing Ascitic fluid and AFP prepared by affinity chromatography were studied by isoelectric focusing in thin-layer polyacrylamide gel with a pH gradient of 3.5--9.5 (Ampholine PAG plate, LKB).
340
Gel electrophoresis Gel electrophoresis in polyacrylamide containing sodium dodecylsulfate (SDS) was p e r f o r m e d according to t he m e t h o d o f Laemmli (1970).
Preparation of immunoadsorbent Two ml o f mouse ascitic fluid containing 4.5 mg anti-AFP was precipit a t ed with an equal volume o f saturated a m m o n i u m sulfate. The precipitate was washed twice with half-saturated a m m o n i u m sulfate and dialyzed against 0.1 mol/1 NaHCO 3, pH 8.3. This immunoglobulin preparation was coupled t o 2 ml o f cyanogen bromide-activated Sepharose (Porath et al., 1967).
Amniotic fluid This was obtained f r om therapeutic amniocentesis p e r f o r m e d between the 15th and 18th weeks o f pregnancy. RESULTS
Hybridization and testing of antibody production Two weeks after hybridization, 20 o u t o f 100 wells showed a n t i b o d y p r o d u c t i o n . This n u m b e r o f positive cultures rose to 25 within 5 weeks o f hybridization. One such culture showed an a n t i b o d y titer o f 1 : 5 0 0 (Table 1). Five h y b r id cultures were cloned, and t h e yield was 10 antibodyproducing hybrid clones. Three weeks after cloning, t h e binding o f AFP t o culture supernatants was 5 times as high as t he background. After recloning o f on e clone, 110 a nt i body- pr oduc i ng subclones were recovered and 20 o f these were f u r t h e r expanded. One week after recloning the a n t i b o d y titer in t h e supernatant was onl y 1 : 50, but it rose t o a b o u t 1 : 5000 after continuous culture (Table 1). The highest c o n c e n t r a t i o n o f specific a n t i b o d y in such cultures was 25 gg/ml. T h e a n t i b o d y titers in t he supernatans correlated with t he n u m b e r o f viable cells and th e time o f culture w i t h o u t change o f medium. After t he culture med iu m had been changed, a substantial rise in a n t i b o d y titer could be
TABLE 1 Titers of antibody to AFP produced by mice and hybrid cells in culture and in ascites from mice bearing intraperitoneally transplanted hybrid cells. = dilution giving half-maximal binding
Source
Titer
Mouse antiserum Culture supernatant from uncloned hybrid Culture supernatant from clone Ascitic fluid
1 1 1 1
: 15 000 : 500 : 5 000 : 200 000--1 : 800000
341
1:800
w
--
O
0.8x106
I%400
0,4
0
~
• 1:200
O
0
; 0.2
/
o~* I
I
I
'
o 2 4 6 Days Fig.1. Correlationbetweennumberof viablecellsper m] (o
o) and antibody titer ; ) in a lymphocyte hybrid culture producing monoclonal antibodies to AFP.
(=
demonstrated with each additional day of culture. However, the number of viable cells began to decrease 3--4 days after the change of medium (Fig. 1). Ascitic fluid developed within 3--4 weeks in all mice that had received an intraperitoneal injection of hybrid cells. The antibody titers in the ascitic fluid of these mice varied between 1 : 200,000 and 1 : 800,000 (Table 1), and the maximal concentration of specific antibody was 2.5 mg/ml. RIA The antibody produced by one clone was chosen for closer study. The inhibition curve obtained with RIA using this anti-AFP is shown in Fig, 2. Significant binding inhibition was obtained with an AFP concentration of 50 B/T
3020:10
-
-
~
.....
I 25
I 50
I I I00 200
Bg
I |, 500 pg/]O00
Fig. 2. Standard curve for radioimmunoassay of AFP. The abscissa gives the concentration o f AFP in the standards.
342
pg/1. This corresponds to a concentration of 6/~g/1 in the incubation mixture. This sensitivity is about one-tenth o f that obtained with our standard AFP RIA using rabbit anti-AFP and goat anti-rabbit IgG but otherwise identical conditions. Analysis o f serum samples showed no systematic difference between results with the two assays. Nor was any difference observed when AFP in amniotic fluid was studied. Human serum albumin did not cause inhibition of binding even in a final concentration o f 50 mg/ml. The affinity constant determined from the Scatchard plot was 1.7 X 109 1/mol. A straight line was obtained, indicating homogeneity of the antibody.
A
....
i ¸ ! i!ii! i i! ii ~
!: i il
:! ..... i~ I~ ~ !ii~I~:
1
2
3
4
5
~
3
1
i
:
6
Fig. 3. A: isoelectric focusing in polyacrylamide gel, pH range 3.5--9.5. High pH is at top and low at the b o t t o m of the figure. Sample 1 contains amniotic fluid, 2 contains 10/~g and 3, 20 pg of purified AFP. Samples 4 and 5 contain ascitic fluid from two different mice transplanted with the same anti-AFP producing monoclonal hybrid cell line. The strong bands at the top correspond to the monoclonal antibodies B: SDS-polyacrylamide electrophoresis in a 10% gel. Anode at the bottom. Sample numbers are the same as in A except for No. 6, which contains molecular weight standards. The molecular weights (x 10 -3) of the standards are marked on the right side of the figure.
343
Isotype The isotype o f the antibody determined b y immunodiffusion was IgG1. In isoelectric focusing ascitic fluid from one mouse displayed 3 strong bands in the pH range 8--9, and 3 additional bands were observed in the more concentrated Ig fraction prepared b y ammonium sulfate fractionation (Fig. 3A). Ascites from aI~other mouse displayed 6 bands in the same pH region, b u t the pattern was slightly different from that in the first mouse (Fig. 3A).
Affinity chromatography Sixty ml o f amniotic fluid containing a b o u t 6 mg of protein/ml and 28 #g AFP/ml were applied to a 1-ml affinity chromatography column at a flow rate of 10 ml/h. Two-ml fractions were collected and the contents of AFP and albumin were monitored b y immunodiffusion against specific antisera. No AFP was detected in the first 10 fractions; after this AFP came through (Fig. 4). This indicated that 560 gg o f AFP had been adsorbed. After the amniotic fluid had been passed through, the column was washed with 0.1 mol/1 Tris-HC1 buffer, pH 7.4, containing 1 mol/1 NaC1. When the absorbance at 280 nm fell to less than 0.005 (after 10 fractions) the column was eluted with 4 mol/1 urea in 0.1 mol/l Tris-HC1, pH 7.4. AFP was detected in the first 6 ml eluted. This fraction was concentrated under reduced pressure and dialyzed against 0.1 mol/1 Tris HC1, pH 7.4. F r o m the concentrate, 415 #g of AFP was recovered, i.e., a b o u t 74% of the AFP adsorbed. No additional AFP or protein was eluted with 8 mol/1 urea. In a separate experiment we eluted the column with a gradient o f 30 ml Tris buffer and 30 ml o f 4 mol/1 urea in Tris buffer. A b o u t 90% of the AFP was elute.d at urea concentrations b e t w e e n 1.5 and 2 mol/1. The ascitic fluid used to prepared the immunoadsorbent had a calculated
A280 6
..,
I
NaCI 1M
5
AFP
10
1
20
Urea 4M
25
30
35
F r a c t i o n no.
Fig. 4. P u r i f i c a t i o n o f A F P b y a f f i n i t y c h r o m a t o g r a p h y o n a 1-ml c o l u m n containing about 2 mg m o n o c l o n a l a n t i - A F P c o u p l e d t o S e p h a r o s e 4B. S i x t y m l o f a m n i o t i c fluid was a p p l i e d t o t h e c o l u m n . T h e A F P in the first 30 ml was a d s o r b e d o n t h e c o l u m n a n d e l u t e d w i t h 4 mol/1 urea. A F P was m o n i t o r e d b y i m m u n o d i f f u s i o n .
344
binding capacity o f 3.6 mg AFP. After precipitation of the Ig fraction and coupling to Sepharose, 16% of the calculated capacity was recovered. Portions of 10 and 20 ttg of AFP were submitted to isoelectric focusing in polyacrylamide gel. One strong band isoelectric at pH 4.7 and 3 weaker bands were observed (Fig. 3A). This microheterogeneity is typical of human AFP. Polyacrylamide electrophoresis in the presence of SDS gave only one band (Fig. 3B). These tests show that the AFP preparation probably contained less than 1% of contaminating protein. DISCUSSION
The aim of the present study was to investigate the usefulness of monoclonal antibodies for assay and purification of clinically important proteins. AFP was chosen as a model because it was available in pure form in sufficient amounts. In addition, determination o f AFP is clinically important and difficult to standardize (Sizaret et al., 1979). A continuous supply of homogeneous antibody might help to resolve this problem. In this study the antibody produced b y one monoclonal cell line was characterized. With this antibody we could demonstrate some o f the advantages inherent in the monoclonal technique. For RIA work, however, the antibody did n o t prove superior to conventional antibodies, and this has also been the experience o f groups working with other antigens (Accolla et al., 1980; Miggiano et al., 1980). Because of the only moderate affinity of the antibody, the sensitivity obtained with it in RIA was a b o u t 10 times less than that of our conventional RIA for AFP. In affinity chromatography our monoclonal antibody demonstrated useful properties which m a y be typical of monoclonal antibodies. The antigen could be eluted from the affinity column with mild reagents, i.e., 1.5--2 mol]l urea. This was probably n o t only due to the moderate affinity of the antibody, b u t m a y in part be accounted for by the fact that monoclonal antibody binds monomeric antigen at one site only. High avidity resulting from simultaneous binding to t w o antibodies is thus avoided. With gradient elution the adsorbed AFP was eluted as a fairly sharp peak. This was expected, since monoclonal antibody had a constant affinity, not a spectrum from low to high. Absence o f multipoint binding might also be important in this respect. Rather than attempting to prepare an affinity column with maximal capacity, we maximized coupling o f the limited a m o u n t of antibody available. However, the very high antibody concentration that can be obtained in ascites, in combination with low concentrations of other proteins, makes it easy to prepare concentrated antibody and high capacity columns with low nonspecific adsorption. Monoclonal antibody columns have earlier been used to purify antigenically similar H L A antigens (Parham, 1979). This unique property together with those observed in the present study make monoclonal antibodies ideal for affinity chromatography.
345 REFERENCES Accolla, R.S., S. Correl and J.P. Mach, 1980, Proc. Natl. Acad. Sci. U.S.A. 77,563. KShler, G. and C. Milstein, 1975, Nature 256,495. Laemmli, U.K., 1970, Nature 227,680. Melchers, F., M. Potter and N.L. Warner (eds.), 1978, Lymphocyte Hybridomas. Curr. Top. Microbioi. Immunol. 81, 1. Miggiano, V., C. St~/hli, P. H~iring, J. Schmidt, M. Le Doin, T. Glatthaar and T. Staehling, 1980, Protides of the Biological Fluids, Belgium, May, 1980, Abstr. 24. Parham, P., 1979, J. Biol. Chem. 254, 8709. Porath, J., R. Ax6n and S. Ernback, 1967, Nature 215, 1491. Ruoslahti, E. and W.D. Terry, 1976, Nature 260, 804. Sizaret, P., N. Breslow et al., 1979, Clin. Chim. Acta 96, 59. Thorell, J.I. and B.G. Johansson, 1971, Biochim. Biophys. Acta 251,363.