- Email: [email protected]
The use of monoclonal antibodies in demonstrating the effect of antibody heterogeneity on immune complex size
CLINICAL
IMMUNOLOGY
AND
IMMUNOPATHOLOGY
32, 378-386 (1984)
The Use of Monoclonal Antibodies in Demonstrating of Antibody Heterogeneity on Immune Complex SUSUMU HOSOI, KEISUKE SHINOMIYA, Department of Pediatrics, Faculty of Medicine,
the Size1
AND HARUFCIMIKAWA
University of Kyoto, Sakyo-ku, Kyoto 606, Japan
The effect of antibody heterogeneity on immune complex size was investigated by using monoclonal antibodies. Five monoclonal antibodies which bind different antigenic determinants of human serum albumin were used to produce immune complexes. When one or two different monoclonal antibodies were used, under antigen-excess, antigenantibody equivalent, and antibody-excess conditions, hardly any immune complexes larger than mouse IgM were produced. On the other hand, in a representative experiment, when five different monoclonal antibodies were used, immune complexes larger than mouse IgM comprised 21.5% of the total antigen under antigen-antibody equivalent conditions, 18.2% under antibody-excess conditions, and 4.3% under antigen-excess conditions. These results indicate that antibody heterogeneity is one of the important factors determining immune complex size.
INTRODUCTION
Immune complexes are known to be associated with the pathogenesis of diseases such as rheumatoid arthritis, systemic lupus erythematosus, and other vasculitic syndromes (1, 2). The formation of immune complexes can be considered as a physiological response of the organism to eliminate foreign materials. These complexes are normally removed by the reticuloendothelial system. Defective clearance of immune complexes leads to prolonged circulation of the complexes, which are deposited in organs and tissues, causing tissue damage. The size of immune complexes is a critical factor in determining their clearance from the circulation. Multivalent immune complexes bind to Fe receptors of macrophages with increasing affinities in parallel with their increasing valency (3). Similarly dimeric, trimeric, and tetrameric complexes of IgG bind to Clq with affinities 9 x 106, 6 x 107, 7.5 x lo9 M-t, respectively (4). Large-sized immune complexes seem to be easily trapped by the reticuloendothelial system according to their multivalency. Therefore, it is important to elucidate the factors which determine the size of immune complexes. It has been speculated that the size of immune complexes depends on (a) the antigen to antibody ratio, (b) the affinity of binding of the antibody-combining sites to the antigenic determinants, and (c) the number of the antigenic determinants (5). In general, studies have focused on the antigen to antibody ratio and the affinities of antibodies (6, 7). Atsumi et al. reported the effect of antigenic valency on precipitin reactions of antihapten antibodies (8). They analyzed the reactions with rabbit antibenzylpen’ This work was supported by a Grand-in-Aid of Education, Japan. 0090-1229184 $1 SO CopyrIght All rights
D 1984 by Academic Press, Inc. of reproduction in any form reserved.
for Scientific Research 57480245 from the Ministry
378
EFFECT
OF ANTIBODY
HETEROGENEITY
ON IC SIZE
379
icilloyl (BPQ)2 antibodies of the IgG class and BP0 antigens having different numbers of antigenic determinants. However, as they described in another report, anti-BP0 antibody has heterogeneous combining sites and the attachment of BPQ groups to the carrier molecule is random (9). Therefore, even if antihapten antibody and a carrier molecule having a defined number of haptens are used, heterogeneity both of antibody and antigen cannot be controlled accurately. The methods of Kohler and Milstein for producing monoclonal antibodies against single antigenic determinants circumvented the heterogeneity of conventional antisera (10). Monoclonal antibodies also can be obtained in large quantities. They seem to be most suitable for elucidating experimentally the relationships between the size of immune complexes and the heterogeneity of antibodies. We investigated these relationships using five monoclonal antibodies against different antigenie determinants of human serum albumin. MATERIALS
AND METHODS
Preparation of monoclonal antibodies to human serum albumin. BALB/C male mice aged 8-10 weeks were immunized with 100 pg of human serum albumin (Chemical Dynamics Co.) with complete Freund’s adjuvant. After 3 weeks, a booster dose of 10 pg of human serum albumin precipitated with alum was given ip. Three days later, spleen cells were collected and were fused with P3-NSl/ lAg4-1 cells, essentially as described earlier (11). The production of antibodies was tested by passive hemagglutinations using sheep erythrocytes coated with human serum albumin by the chromic chloride methods (12). The hybrids which produced specific antibodies were cloned by limiting dilutions and were transferred to pristane-primed BALB/C mice ip. Ascitic fluid was collected after about 3 weeks. Monoclonal antibodies were purified from ascitic fluids with a protein A-Sepharose affinity column (13). lz51 iodination of monoclonal antibodies and human serum albumin. Monoclonal antibodies and human serum albumin were dissolved in phosphate-buffered saline (PBS), pH 7.4. After being centrifuged at 65,000 rpm (368,000g) for 1 hr, the upper half of the solution was kept and considered devoid of aggregates. Each solution was adjusted to a concentration of 1 mg/ml protein in PBS. The 1251 iodination was performed using Iodogen obtained from Pierce Chemical Company (14). Each solution (200 ~1) was mixed with 500 @i 1251in a tube coated with 5 fig of Iodogen and incubated 10 min at room temperature. 1251-Labeled protein was separated from free 1251using a Sephadex G-25 column. Reciprocal inhibition tests among monoclonal antibodies to human serum albumin. A polystyrene 96-well multiwell plate (Immulon 2 Plate, Dynatic Co.) was coated with human serum albumin by incubating with a solution of human serum albumin (I mg/ml) in 0.1 M carbonate-bicarbonate buffer, pH 9.0, at 4°C overnight. The plate was washed three times with PBS containing 0.05% Tween 20. Into each well 50 ~1 of a solution of 12?-labeled monoclonal antibody (0.1 yJLg/ml in PBS containing 0.1% ovalbumin) was added simultaneously with 50 JJJ of a solution of one of five different monoclonal antibodies. After overnight incubation 2 Abbreviations
used: BPO, benzylpenicilioyl;
PBS, phosphate-buffered
saline, pH 7.4.
380
HOSOI,
SHINOMIYA,
AND
MIKAWA
at 4°C the plate was washed three times with PBS containing 0.05% Tween 20. 1251-labeled antibodies which bound to human serum albumin coated to the plate were detached with a 10% acetic acid solution and were transferred to a tube for counting. Determination of binding affinities of monoclonal antibodies to human serum albumin. A 100~pl aliquot of each monoclonal antibody (20 p&ml) was mixed with 100 (~1 of increasing concentrations of 1251-labeled human serum albumin (20-70 pg/ml). Each mixture was incubated at 37°C for 2 hr, and further at 4°C overnight, To each tube a 200-1.1.1aliquot of a suspension of protein A-Sepharose CL 4B (Pharmacia) was added, which was preincubated with enough affinitypurified rabbit anti-mouse IgG antibodies to precipitate the mouse IgG. The mixture was rotated for 2 hr and then washed three times with PBS containing 0.05% Tween 20. The amount of human serum albumin bound by each monoclonal antibody was calculated by counting the radioactivities, and affinity constants were obtained by means of a Scatchard plot. Fractionation of immune complexes according to size. A 209.~1 aliquot of 1251labeled human serum albumin solution (5, 50, or 500 kg/ml of PBS) was mixed with 20 p,l each of 1, 2, 3, 4, or 5 kinds of monoclonal anti-human serum albumin antibodies at a concentration of 50 kg/ml. The total volume was adjusted to 120 ~1 with PBS. Each mixture was incubated for 2 hr at 37°C and then overnight at 4°C. After incubation, the mixture was resuspended and 100 ~1 of each was analyzed by 5-20% sucrose gradient ultracentrifugation (Model 7013-72, Hatachi Koki Co.). After ultracentrifugation at 50,000 rpm (270,OOOg) for 10 hr at 4°C each sample was separated into 20 parts by a fractionator (Model DGF-U, Hitachi Koki Co.) and transferred to tubes for gamma counting. Human serum albumin, mouse IgG, and mouse IgM were analyzed in the same ways and were used as standards for size. RESULTS
Preparation
of Monoctonal
Antibodies
to Human
Serum Albumin
A total of 15 hybrids which produced monoclonai anti-human serum albumin antibodies were obtained from three fusions between spleen cells, each of two mice immunized with human serum albumin and the mouse myeloma cell line NSI. As shown in Table 1, 13 hybrids produced IgG,; HSAl-5 produced IgG,, and HSA3-1 produced IgA. Isotypes of the monoclonal antibodies were determined by immunoprecipitations using class- or subclass-specific antisera (Miles). Their spent culture supernatants were examined for hemagglutination titers using sheep erythrocytes coated with human serum albumin. Characterization of Monoclonai Antibodies to Human Serum Albumin Six monoclonal antibodies were selected randomly. They were conjugated with iz51 using Iodogen. Their specific activities were between 1.2 x lo9 and 2.0 x lo9 cpm/mg. In order to determine whether their combining sites were different from each other, reciprocal inhibition tests were performed. Two representative results are presented in Figs. la and b. HSAl-3 and HSAl-6 completely mutually
EFFECT
OF ANTIBODY
HETEROGENEITY
38%
ON IC SIZE
TABLE 1 MONOCLONALANTI-HUMAN SERUM ALBUMIN ANTIBODIES Monoclonal
antibody
HSAl-1 HSAl-2 HSAI-3 HSAl-4 HSAl-5 HSAI-6 HSAI-7 HSAl-8 HSAl-9 HSAZ-I HSA2-2 HSA2-3 HSA2-4 HSA3-I
Isotype IgG, w, I@, I@, k% kG, kG, I@, W, Id-4 W, 16, Id-4 I&
Hemagglutination
titef
2’2 21” 2’2 2’1
2”
29 2’2 29 2’2 2’2 2’2 2’0 212 212
’ Hemagglutination titers were examined in spent culture supernatants using sheep erythrocytes coated with human serum albumin.
inhibited the binding to human serum albumin. In other combinations the binding of each 12SI-labeled monoclonal antibody was not inhibited by other antibodies but was inhibited completely by the same antibody (Fig. lb). These reciprocal inhibition tests demonstrated that HSAl-1, HSAl-2, HSAI-5, HSAl-6, and HSAI-7 recognize different antigenic determinants of human serum albumin, an that HSAl-3 and HSAl-6 recognize the same or overlapping antigenic determinants. Analysis of binding affinities of these monoclonal antibodies revealed that HSAl-2 and HSAl-7 have relatively high affinity constants, but the differences between those of five monoclonal antibodies are not enormous (Table 2). Relationship between the Size of Immune Complexes nnd the Heterogeneiry of Antibodies No aggregate was produced during the procedure of 1251iodination and, immediately before being incubated with iz51-labeled human serum albumin, each solution of monoclonal antibody was ultracentrifuged to remove aggregates. This assures that the relationship between the size of immune complexes and the heterogeneity of antibodies reported in this paper excludes the effect of aggregates. HSAl-1, HSAl-2, HSAl-5, HSAl-6, and HSAl-7 were used to produce immune complexes as described under Materials and Methods. Human serum albumin, mouse IgG, and mouse IgM were used for standards. One representative example of the results of fractionations of immune complexes is shown in Fig. 2. Large-sized immune complexes which adhered to the bottom of the ultracentrifnge tubes were counted after removal of the contents, and the fractions were grouped into five parts (Fig. 2). Fart 1 contained mainly free human serum albumin, Part 2 mainly free antibodies, and Part 5 immune complexes which were as large as mouse IgM. Percentages of totals of each part were calculated and plotted. Figure 3 shows the representative results of experiments which used HSAl-I, and HSAl-2, in the mixture of two kinds of monoclonal antibodies:
382
HOSOI, SHINOMIYA,
10-d
10-3 concentration
10-2 10-l of inhibitors
AND MIKAWA
mg,m,
FIG. 1. Reciprocal inhibition tests. (a) Inhibition tests to ‘2SI-HSA1-6. ‘“51-HSAl-6 (0.1 Kg/ml) was incubated simultaneously with unlabeled monoclonal antibodies (O-O, HSAl-1; e---O, HSAI-2, A-& HSAl-3, A---A, HSAl-5; Cl---O, HSAl-6; m-m, HSAl-7) in 96-well polystyrene multiwell plates coated with human serum albumin. (b) Inhibition tests to 12’1-HSAl-5. *251-HSAl-5 (0.1 kg/ml) was incubated simultaneously with unlabeled monoclonal antibodies (O-O, HSAl-1; O---O, HSAI2; &-A, HSAI-3; O---O, HSAl-5; U-4, HSAl-6; A--A, HSAl-7) in 96-well polystyrene multiwell plates coated with human serum albumin.
HSAl-1, HSAl-2, and HSAI-5 with three; HSAl-1, HSAl-2, HSAl-5, and HSAI-6 with four; HSAl-1, HSAl-2, HSAl-5, HSAI-6, and HSAl-7 with five. Combinations of different antibodies resulted in similar ultracentrifuge patterns. Under both antibody-excess (antibody to antigen 8.7 to 1) and antigen-antibody equivalent conditions, almost identical patterns of size distribution of immune complexes were obtained. The immune complexes produced with one or two antibodies were as small as those produced under antigen-excess conditions. However, as the number of the monoclonal antibodies used to produce immune complexes increased, more immune complexes larger than mouse IgM were produced. When five different monoclonal antibodies were used to produce immune complexes, the relative radioactivities of 12SI-labeled human serum albumin in the TABLE ASSOCIATIONCONSTANTS
Monoclonal
2
OFMONOCLONALANTIBODIESAGAINSTHUMAN
antibody
HSA-1 HSAl-2 HSAl-5 HSAl-6 HSAI-7
SERUM
Association constant W- ‘) 3.2 2.3 5.6 8.2 1.2
x x x x x
lo6 lo7 IO6 106 107
ALBUMIN
EFFECT
OF ANTIBODY
Part I -+--
t1
Part II + I 5
HETEROGENEITY
Part III ---+-
10
Part I\ +-
15
383
ON IC SIZE
part\
+
part\1
20 Fr.
FIG. 2. The sucrose density (5-20%) ultracentrifugation pattern of 1251-labeled human serum albumin, HSAl-1 and HSAl-2 complexes under the antigen-antibody equivalent condition (27Q,OOOg, 10 hr). ‘251-labeled human serum albumin (50 ug/ml) was mixed with HSAl-1 and HSAl-2 (each 50 pgiml). After incubation, the mixture was resuspended and 100 ~1 was analyzed with sucrose density (5-20%) ultracentrifugation. Twenty fractions were grouped into five parts. Part 6 was a count of large-sized immune complexes which adhered to the ultracentrifuge tube. Markers for albumin, IgG, and IgM are shown by arrows.
immune complexes, larger than mouse IgM, were 21.5% under antigen-antibody equivalent conditions and 18.2% under antibody-excess conditions. Under antigen-excess conditions, distribution patterns of immune complexes were only slightly influenced by the number of antibodies. However, even under these antigen-excess conditions (antigen to antibody 11.5 to 1) the amount of large-sized immune complexes increased slightly as the number of antibodies increased. These results indicate that with a combination of five different monoclonal antihuman serum albumin antibodies, the size of immune complexes depends mainly on the number of monoclonal antibodies used. DISCUSSION
This paper reports the effect of antibody heterogeneity on immune complex size. We used five monoclonal antibodies which bind different antigenic determinants of human serum albumin to produce immune complexes. Association constants of these antibodies were between 3.2 x lo6 and 2.3 x IO7 M-". When one or two monoclonal antibodies were used to produce immune complexes, those larger than mouse IgM were hardly produced not only under antigen-excess conditions, but also under .antigen-antibody equivalence and antibody-excess conditions. Theoretically, large lattice formation can be made between specific antibody and antigens having two different antigenic determinants. It is possible that the reason our two monoclonal antibodies did not produce large complexes was either because of their low affinities or because the steric rela-
384
HOSOI,
SHINOMIYA,
AND
MIKAWA
part
b
c
I I
II
Ill
IV
v
VI
part
I
I’I
Ii1
IV
v
VI
part
FIG. 3. The size distributions of human serum albumin anti-human serum albumin antibody complexes (O-O, with one monoclonal antibody; O---O, with two different monoclonal antibodies; A---A, with three different monoclonal antibodies; A---A, with four different monoclonal antibodies; W-M, with five different monoclonal antibodies), (a) Under antigen-excess conditions 500 pgiml lz51labeled human serum albumin and 50 pg/ml each monoclonal antibody. (b) Under antigen-antibody equivalent conditions 50 kg/ml 1251-labeled human serum albumin and 50 pgiml of each monoclonal antibody. (c) Under antibody-excess conditions 5 pg/mI L251-labeled human serum albumin and 50 pgi ml of each monoclonal antibody.
EFFECT
OF
A~IBOD~
H~~EKO~E~EIT~
ON
I(? Sf~E
385
tionship between the two antigenic determinants of human serum albumin did not permit stable lattice formation. In contrast, when five different monoclonal antibo~ies were used, immune complexes larger than mouse IgM comprised 21.5% of the total amount of antigen under antigen-antibody equivalent conditio~s~ 18.2% under antibody-excess conditions, and even 4.3% under antigen-excess conditions. These results indicate that with combinations of five different monoclonal anti-human serum albumin antibodies, the sizes of immune complexes deemed manly on the number of monoclonal antibodies used. Because the binding of artisan and antibody is reversible, it cannot be excluded that some immune complexes were dispersed during ultracentrifugation for 10 hr. Therefore, the sizes of the immune complexes might have been underestimated or the amount of free antigen might have been overestimated. These observations were made on a molecule with a variety of antigenic determin~ts rather than recurrent ones. The presence of a recurrent antigen in a molecule theoretically would allow large lattice formation with even homogeneous antibody. Atassi was the first to report the completed antigenic structure for a native protein antigen (15). He determined five different antigenic determinants of sperm-whale myoglobin and aiso observed that, though a reactive antigenic region is always a reactive region, the contribution of a given antigenic reactive region to the total reaction of a protein antigen may vary with the individual animal immunized. In (NZB x NZW)F, mice, which are genetically predisposed to develop autoimmunity, anti-DNA autoantibodies are clonally heterogeneous, but the majority share a common idiotype (16). Isoelectric focusing of the serum antibodies to fragment A (MW 20,000) of diphtheria toxoid disclosed that in a majority of normal humans who were booster-immunized with diphtheria toxoid, Ig~-anti-fr~ment A spectrotypes were restricted between three to eight distinct bands (17). This evidence suggests that humoral immune responses to even macromolecular antigens do not always result in heterogeneous antibodies against different antigenic determinants of the antigens. incus et aE. reported that three of ten rabbits injected intravenously daily with bovine serum albumin for periods longer than 10 weeks developed chronic glomerulonephritis, and that a major part of the antibodies produced by rabbits with chronic nephritis lacked precipitating properties (18). Devey et al. found that in mice selectively bred to produce low-affinity antibody against protein antigens, the incidence of chronic immune complex disease was 61% compared to 21% in mice selectively bred to produce high-affinity antibody (7). However, they did not discuss the number of the antigenic determinants that were combined with the low- or high-affinity antibody. It is possible that in mice selectively bred to produce low-affinity antibody against protein antigens, the humoral immune respouse is directed to a small number of antigenic determinants, because mixing monoclonal antibodies yields an enhanced affinity for antigen (19). Therefore, the number of the antigenic determinants that are recognized by the organism appears to be one of the important factors involved in immune complex disease. Some studies examining the effect of the number of the antigenic determinants on immune complex size, using antihapten antibodies have been reported (8). However, even when antihapten antibody and a molecule having a restricted number of
386
haptens are used, (9). Monoclonal obtained in large different epitopes heterogeneity on
HOSOI,
SHINOMIYA,
AND
MIKAWA
heterogeneity both of antibody and antigen cannot be excluded antibody binds to a single antigenic determinant and can be quantities. Thus, we used five monoclonal antibodies against of human serum albumin to examine the effect of antibody immune complex size. REFERENCES
1. Cochrane, C. G., and Koffler, D., Zn “Advances in Immunology” (F. J. Dixon and H. G. Kunkel, Eds.), Vol. 16 p. 185, Academic Press, New York, 1973. 2. Christian, C. L., and Sergent, J. S., Amer. J. Med. 61, 385, 1976. 3. Segal, D. M., and Hurwitz, E., J. Zmmunol. 118, 1338, 1977. 4. Wright, J. K., Tschopp, J., Jaton, J. C., and Engel, J., Biochem. J. 187, 775, 1980. 5. Germuth, J. G., and Rodriguez, E., “Immunopathology of the Renal Glomerulus,” Little Brown, Boston, 1973. 6. Wilson, C. B., and Dixon, F. J., J. Exp. Med. 134, 7S, 1971. 7. Devey, M. E., Beasdale, K., Collins, M., and Steward, M. W., Znt. Arch. Allergy Appl. Zmmunol. 68, 47, 1982. 8. Atsumi, T., Nishida, K., Kinoshita, Y., and Horiuchi, Y., Immunochemistry 8, 271, 1981. 9. Atsumi, T., Nishida, K., Kinoshita, Y., Shibata, K., and Horiuchi, Y., J. Zmmunol. 99, 1286, 1967. 10. Kohler, G., and Milstein, C., Nature (London) 256, 495, 1975. il. Hosoi, S., Nakamura, T., Higashi, S., Yamamuro, T., Toyama, S., Shinomiya, K., and Mikawa, H., Cancer Res. 42, 654, 1982. 12. Gold, E. R., and Fudenberg, H. H., J. Zmmunol. 99, 859, 1967. 13. Ey, P. L., Prowse, S. J., and Jenkin, C. R., Immunochemistry 15, 429, 1978. 14. Markwell, M. A. K., and Fox, C. E, Biochemistry 17, 4807, 1978. 15. Atassi, M. Z., Immunochemistty 12, 423, 1975. 16. Marion, T. N., Lawton, A. R., III, Keamey, J. E, and Briles, D. E., J. Zmmunol. 128, 688, 1982. 17. Morrow, C. D., Dorey, F., and Stevens, R. H., J. Zmmunol. 130, 818, 1983. 18. Pincus, T., Haberkern, R., and Christian, C. L., J. Exp. Med. 127, 819, 1967. 19. Ehrlich, P. H., Moyle, W. R., Moustafa, Z. A., and Canfield, R. E., J. Zmmunol. 128, 2709, 1982. Received February 22, 1984; accepted with revisions May 14, 1984.
IMMUNOLOGY
AND
IMMUNOPATHOLOGY
32, 378-386 (1984)
The Use of Monoclonal Antibodies in Demonstrating of Antibody Heterogeneity on Immune Complex SUSUMU HOSOI, KEISUKE SHINOMIYA, Department of Pediatrics, Faculty of Medicine,
the Size1
AND HARUFCIMIKAWA
University of Kyoto, Sakyo-ku, Kyoto 606, Japan
The effect of antibody heterogeneity on immune complex size was investigated by using monoclonal antibodies. Five monoclonal antibodies which bind different antigenic determinants of human serum albumin were used to produce immune complexes. When one or two different monoclonal antibodies were used, under antigen-excess, antigenantibody equivalent, and antibody-excess conditions, hardly any immune complexes larger than mouse IgM were produced. On the other hand, in a representative experiment, when five different monoclonal antibodies were used, immune complexes larger than mouse IgM comprised 21.5% of the total antigen under antigen-antibody equivalent conditions, 18.2% under antibody-excess conditions, and 4.3% under antigen-excess conditions. These results indicate that antibody heterogeneity is one of the important factors determining immune complex size.
INTRODUCTION
Immune complexes are known to be associated with the pathogenesis of diseases such as rheumatoid arthritis, systemic lupus erythematosus, and other vasculitic syndromes (1, 2). The formation of immune complexes can be considered as a physiological response of the organism to eliminate foreign materials. These complexes are normally removed by the reticuloendothelial system. Defective clearance of immune complexes leads to prolonged circulation of the complexes, which are deposited in organs and tissues, causing tissue damage. The size of immune complexes is a critical factor in determining their clearance from the circulation. Multivalent immune complexes bind to Fe receptors of macrophages with increasing affinities in parallel with their increasing valency (3). Similarly dimeric, trimeric, and tetrameric complexes of IgG bind to Clq with affinities 9 x 106, 6 x 107, 7.5 x lo9 M-t, respectively (4). Large-sized immune complexes seem to be easily trapped by the reticuloendothelial system according to their multivalency. Therefore, it is important to elucidate the factors which determine the size of immune complexes. It has been speculated that the size of immune complexes depends on (a) the antigen to antibody ratio, (b) the affinity of binding of the antibody-combining sites to the antigenic determinants, and (c) the number of the antigenic determinants (5). In general, studies have focused on the antigen to antibody ratio and the affinities of antibodies (6, 7). Atsumi et al. reported the effect of antigenic valency on precipitin reactions of antihapten antibodies (8). They analyzed the reactions with rabbit antibenzylpen’ This work was supported by a Grand-in-Aid of Education, Japan. 0090-1229184 $1 SO CopyrIght All rights
D 1984 by Academic Press, Inc. of reproduction in any form reserved.
for Scientific Research 57480245 from the Ministry
378
EFFECT
OF ANTIBODY
HETEROGENEITY
ON IC SIZE
379
icilloyl (BPQ)2 antibodies of the IgG class and BP0 antigens having different numbers of antigenic determinants. However, as they described in another report, anti-BP0 antibody has heterogeneous combining sites and the attachment of BPQ groups to the carrier molecule is random (9). Therefore, even if antihapten antibody and a carrier molecule having a defined number of haptens are used, heterogeneity both of antibody and antigen cannot be controlled accurately. The methods of Kohler and Milstein for producing monoclonal antibodies against single antigenic determinants circumvented the heterogeneity of conventional antisera (10). Monoclonal antibodies also can be obtained in large quantities. They seem to be most suitable for elucidating experimentally the relationships between the size of immune complexes and the heterogeneity of antibodies. We investigated these relationships using five monoclonal antibodies against different antigenie determinants of human serum albumin. MATERIALS
AND METHODS
Preparation of monoclonal antibodies to human serum albumin. BALB/C male mice aged 8-10 weeks were immunized with 100 pg of human serum albumin (Chemical Dynamics Co.) with complete Freund’s adjuvant. After 3 weeks, a booster dose of 10 pg of human serum albumin precipitated with alum was given ip. Three days later, spleen cells were collected and were fused with P3-NSl/ lAg4-1 cells, essentially as described earlier (11). The production of antibodies was tested by passive hemagglutinations using sheep erythrocytes coated with human serum albumin by the chromic chloride methods (12). The hybrids which produced specific antibodies were cloned by limiting dilutions and were transferred to pristane-primed BALB/C mice ip. Ascitic fluid was collected after about 3 weeks. Monoclonal antibodies were purified from ascitic fluids with a protein A-Sepharose affinity column (13). lz51 iodination of monoclonal antibodies and human serum albumin. Monoclonal antibodies and human serum albumin were dissolved in phosphate-buffered saline (PBS), pH 7.4. After being centrifuged at 65,000 rpm (368,000g) for 1 hr, the upper half of the solution was kept and considered devoid of aggregates. Each solution was adjusted to a concentration of 1 mg/ml protein in PBS. The 1251 iodination was performed using Iodogen obtained from Pierce Chemical Company (14). Each solution (200 ~1) was mixed with 500 @i 1251in a tube coated with 5 fig of Iodogen and incubated 10 min at room temperature. 1251-Labeled protein was separated from free 1251using a Sephadex G-25 column. Reciprocal inhibition tests among monoclonal antibodies to human serum albumin. A polystyrene 96-well multiwell plate (Immulon 2 Plate, Dynatic Co.) was coated with human serum albumin by incubating with a solution of human serum albumin (I mg/ml) in 0.1 M carbonate-bicarbonate buffer, pH 9.0, at 4°C overnight. The plate was washed three times with PBS containing 0.05% Tween 20. Into each well 50 ~1 of a solution of 12?-labeled monoclonal antibody (0.1 yJLg/ml in PBS containing 0.1% ovalbumin) was added simultaneously with 50 JJJ of a solution of one of five different monoclonal antibodies. After overnight incubation 2 Abbreviations
used: BPO, benzylpenicilioyl;
PBS, phosphate-buffered
saline, pH 7.4.
380
HOSOI,
SHINOMIYA,
AND
MIKAWA
at 4°C the plate was washed three times with PBS containing 0.05% Tween 20. 1251-labeled antibodies which bound to human serum albumin coated to the plate were detached with a 10% acetic acid solution and were transferred to a tube for counting. Determination of binding affinities of monoclonal antibodies to human serum albumin. A 100~pl aliquot of each monoclonal antibody (20 p&ml) was mixed with 100 (~1 of increasing concentrations of 1251-labeled human serum albumin (20-70 pg/ml). Each mixture was incubated at 37°C for 2 hr, and further at 4°C overnight, To each tube a 200-1.1.1aliquot of a suspension of protein A-Sepharose CL 4B (Pharmacia) was added, which was preincubated with enough affinitypurified rabbit anti-mouse IgG antibodies to precipitate the mouse IgG. The mixture was rotated for 2 hr and then washed three times with PBS containing 0.05% Tween 20. The amount of human serum albumin bound by each monoclonal antibody was calculated by counting the radioactivities, and affinity constants were obtained by means of a Scatchard plot. Fractionation of immune complexes according to size. A 209.~1 aliquot of 1251labeled human serum albumin solution (5, 50, or 500 kg/ml of PBS) was mixed with 20 p,l each of 1, 2, 3, 4, or 5 kinds of monoclonal anti-human serum albumin antibodies at a concentration of 50 kg/ml. The total volume was adjusted to 120 ~1 with PBS. Each mixture was incubated for 2 hr at 37°C and then overnight at 4°C. After incubation, the mixture was resuspended and 100 ~1 of each was analyzed by 5-20% sucrose gradient ultracentrifugation (Model 7013-72, Hatachi Koki Co.). After ultracentrifugation at 50,000 rpm (270,OOOg) for 10 hr at 4°C each sample was separated into 20 parts by a fractionator (Model DGF-U, Hitachi Koki Co.) and transferred to tubes for gamma counting. Human serum albumin, mouse IgG, and mouse IgM were analyzed in the same ways and were used as standards for size. RESULTS
Preparation
of Monoctonal
Antibodies
to Human
Serum Albumin
A total of 15 hybrids which produced monoclonai anti-human serum albumin antibodies were obtained from three fusions between spleen cells, each of two mice immunized with human serum albumin and the mouse myeloma cell line NSI. As shown in Table 1, 13 hybrids produced IgG,; HSAl-5 produced IgG,, and HSA3-1 produced IgA. Isotypes of the monoclonal antibodies were determined by immunoprecipitations using class- or subclass-specific antisera (Miles). Their spent culture supernatants were examined for hemagglutination titers using sheep erythrocytes coated with human serum albumin. Characterization of Monoclonai Antibodies to Human Serum Albumin Six monoclonal antibodies were selected randomly. They were conjugated with iz51 using Iodogen. Their specific activities were between 1.2 x lo9 and 2.0 x lo9 cpm/mg. In order to determine whether their combining sites were different from each other, reciprocal inhibition tests were performed. Two representative results are presented in Figs. la and b. HSAl-3 and HSAl-6 completely mutually
EFFECT
OF ANTIBODY
HETEROGENEITY
38%
ON IC SIZE
TABLE 1 MONOCLONALANTI-HUMAN SERUM ALBUMIN ANTIBODIES Monoclonal
antibody
HSAl-1 HSAl-2 HSAI-3 HSAl-4 HSAl-5 HSAI-6 HSAI-7 HSAl-8 HSAl-9 HSAZ-I HSA2-2 HSA2-3 HSA2-4 HSA3-I
Isotype IgG, w, I@, I@, k% kG, kG, I@, W, Id-4 W, 16, Id-4 I&
Hemagglutination
titef
2’2 21” 2’2 2’1
2”
29 2’2 29 2’2 2’2 2’2 2’0 212 212
’ Hemagglutination titers were examined in spent culture supernatants using sheep erythrocytes coated with human serum albumin.
inhibited the binding to human serum albumin. In other combinations the binding of each 12SI-labeled monoclonal antibody was not inhibited by other antibodies but was inhibited completely by the same antibody (Fig. lb). These reciprocal inhibition tests demonstrated that HSAl-1, HSAl-2, HSAI-5, HSAl-6, and HSAI-7 recognize different antigenic determinants of human serum albumin, an that HSAl-3 and HSAl-6 recognize the same or overlapping antigenic determinants. Analysis of binding affinities of these monoclonal antibodies revealed that HSAl-2 and HSAl-7 have relatively high affinity constants, but the differences between those of five monoclonal antibodies are not enormous (Table 2). Relationship between the Size of Immune Complexes nnd the Heterogeneiry of Antibodies No aggregate was produced during the procedure of 1251iodination and, immediately before being incubated with iz51-labeled human serum albumin, each solution of monoclonal antibody was ultracentrifuged to remove aggregates. This assures that the relationship between the size of immune complexes and the heterogeneity of antibodies reported in this paper excludes the effect of aggregates. HSAl-1, HSAl-2, HSAl-5, HSAl-6, and HSAl-7 were used to produce immune complexes as described under Materials and Methods. Human serum albumin, mouse IgG, and mouse IgM were used for standards. One representative example of the results of fractionations of immune complexes is shown in Fig. 2. Large-sized immune complexes which adhered to the bottom of the ultracentrifnge tubes were counted after removal of the contents, and the fractions were grouped into five parts (Fig. 2). Fart 1 contained mainly free human serum albumin, Part 2 mainly free antibodies, and Part 5 immune complexes which were as large as mouse IgM. Percentages of totals of each part were calculated and plotted. Figure 3 shows the representative results of experiments which used HSAl-I, and HSAl-2, in the mixture of two kinds of monoclonal antibodies:
382
HOSOI, SHINOMIYA,
10-d
10-3 concentration
10-2 10-l of inhibitors
AND MIKAWA
mg,m,
FIG. 1. Reciprocal inhibition tests. (a) Inhibition tests to ‘2SI-HSA1-6. ‘“51-HSAl-6 (0.1 Kg/ml) was incubated simultaneously with unlabeled monoclonal antibodies (O-O, HSAl-1; e---O, HSAI-2, A-& HSAl-3, A---A, HSAl-5; Cl---O, HSAl-6; m-m, HSAl-7) in 96-well polystyrene multiwell plates coated with human serum albumin. (b) Inhibition tests to 12’1-HSAl-5. *251-HSAl-5 (0.1 kg/ml) was incubated simultaneously with unlabeled monoclonal antibodies (O-O, HSAl-1; O---O, HSAI2; &-A, HSAI-3; O---O, HSAl-5; U-4, HSAl-6; A--A, HSAl-7) in 96-well polystyrene multiwell plates coated with human serum albumin.
HSAl-1, HSAl-2, and HSAI-5 with three; HSAl-1, HSAl-2, HSAl-5, and HSAI-6 with four; HSAl-1, HSAl-2, HSAl-5, HSAI-6, and HSAl-7 with five. Combinations of different antibodies resulted in similar ultracentrifuge patterns. Under both antibody-excess (antibody to antigen 8.7 to 1) and antigen-antibody equivalent conditions, almost identical patterns of size distribution of immune complexes were obtained. The immune complexes produced with one or two antibodies were as small as those produced under antigen-excess conditions. However, as the number of the monoclonal antibodies used to produce immune complexes increased, more immune complexes larger than mouse IgM were produced. When five different monoclonal antibodies were used to produce immune complexes, the relative radioactivities of 12SI-labeled human serum albumin in the TABLE ASSOCIATIONCONSTANTS
Monoclonal
2
OFMONOCLONALANTIBODIESAGAINSTHUMAN
antibody
HSA-1 HSAl-2 HSAl-5 HSAl-6 HSAI-7
SERUM
Association constant W- ‘) 3.2 2.3 5.6 8.2 1.2
x x x x x
lo6 lo7 IO6 106 107
ALBUMIN
EFFECT
OF ANTIBODY
Part I -+--
t1
Part II + I 5
HETEROGENEITY
Part III ---+-
10
Part I\ +-
15
383
ON IC SIZE
part\
+
part\1
20 Fr.
FIG. 2. The sucrose density (5-20%) ultracentrifugation pattern of 1251-labeled human serum albumin, HSAl-1 and HSAl-2 complexes under the antigen-antibody equivalent condition (27Q,OOOg, 10 hr). ‘251-labeled human serum albumin (50 ug/ml) was mixed with HSAl-1 and HSAl-2 (each 50 pgiml). After incubation, the mixture was resuspended and 100 ~1 was analyzed with sucrose density (5-20%) ultracentrifugation. Twenty fractions were grouped into five parts. Part 6 was a count of large-sized immune complexes which adhered to the ultracentrifuge tube. Markers for albumin, IgG, and IgM are shown by arrows.
immune complexes, larger than mouse IgM, were 21.5% under antigen-antibody equivalent conditions and 18.2% under antibody-excess conditions. Under antigen-excess conditions, distribution patterns of immune complexes were only slightly influenced by the number of antibodies. However, even under these antigen-excess conditions (antigen to antibody 11.5 to 1) the amount of large-sized immune complexes increased slightly as the number of antibodies increased. These results indicate that with a combination of five different monoclonal antihuman serum albumin antibodies, the size of immune complexes depends mainly on the number of monoclonal antibodies used. DISCUSSION
This paper reports the effect of antibody heterogeneity on immune complex size. We used five monoclonal antibodies which bind different antigenic determinants of human serum albumin to produce immune complexes. Association constants of these antibodies were between 3.2 x lo6 and 2.3 x IO7 M-". When one or two monoclonal antibodies were used to produce immune complexes, those larger than mouse IgM were hardly produced not only under antigen-excess conditions, but also under .antigen-antibody equivalence and antibody-excess conditions. Theoretically, large lattice formation can be made between specific antibody and antigens having two different antigenic determinants. It is possible that the reason our two monoclonal antibodies did not produce large complexes was either because of their low affinities or because the steric rela-
384
HOSOI,
SHINOMIYA,
AND
MIKAWA
part
b
c
I I
II
Ill
IV
v
VI
part
I
I’I
Ii1
IV
v
VI
part
FIG. 3. The size distributions of human serum albumin anti-human serum albumin antibody complexes (O-O, with one monoclonal antibody; O---O, with two different monoclonal antibodies; A---A, with three different monoclonal antibodies; A---A, with four different monoclonal antibodies; W-M, with five different monoclonal antibodies), (a) Under antigen-excess conditions 500 pgiml lz51labeled human serum albumin and 50 pg/ml each monoclonal antibody. (b) Under antigen-antibody equivalent conditions 50 kg/ml 1251-labeled human serum albumin and 50 pgiml of each monoclonal antibody. (c) Under antibody-excess conditions 5 pg/mI L251-labeled human serum albumin and 50 pgi ml of each monoclonal antibody.
EFFECT
OF
A~IBOD~
H~~EKO~E~EIT~
ON
I(? Sf~E
385
tionship between the two antigenic determinants of human serum albumin did not permit stable lattice formation. In contrast, when five different monoclonal antibo~ies were used, immune complexes larger than mouse IgM comprised 21.5% of the total amount of antigen under antigen-antibody equivalent conditio~s~ 18.2% under antibody-excess conditions, and even 4.3% under antigen-excess conditions. These results indicate that with combinations of five different monoclonal anti-human serum albumin antibodies, the sizes of immune complexes deemed manly on the number of monoclonal antibodies used. Because the binding of artisan and antibody is reversible, it cannot be excluded that some immune complexes were dispersed during ultracentrifugation for 10 hr. Therefore, the sizes of the immune complexes might have been underestimated or the amount of free antigen might have been overestimated. These observations were made on a molecule with a variety of antigenic determin~ts rather than recurrent ones. The presence of a recurrent antigen in a molecule theoretically would allow large lattice formation with even homogeneous antibody. Atassi was the first to report the completed antigenic structure for a native protein antigen (15). He determined five different antigenic determinants of sperm-whale myoglobin and aiso observed that, though a reactive antigenic region is always a reactive region, the contribution of a given antigenic reactive region to the total reaction of a protein antigen may vary with the individual animal immunized. In (NZB x NZW)F, mice, which are genetically predisposed to develop autoimmunity, anti-DNA autoantibodies are clonally heterogeneous, but the majority share a common idiotype (16). Isoelectric focusing of the serum antibodies to fragment A (MW 20,000) of diphtheria toxoid disclosed that in a majority of normal humans who were booster-immunized with diphtheria toxoid, Ig~-anti-fr~ment A spectrotypes were restricted between three to eight distinct bands (17). This evidence suggests that humoral immune responses to even macromolecular antigens do not always result in heterogeneous antibodies against different antigenic determinants of the antigens. incus et aE. reported that three of ten rabbits injected intravenously daily with bovine serum albumin for periods longer than 10 weeks developed chronic glomerulonephritis, and that a major part of the antibodies produced by rabbits with chronic nephritis lacked precipitating properties (18). Devey et al. found that in mice selectively bred to produce low-affinity antibody against protein antigens, the incidence of chronic immune complex disease was 61% compared to 21% in mice selectively bred to produce high-affinity antibody (7). However, they did not discuss the number of the antigenic determinants that were combined with the low- or high-affinity antibody. It is possible that in mice selectively bred to produce low-affinity antibody against protein antigens, the humoral immune respouse is directed to a small number of antigenic determinants, because mixing monoclonal antibodies yields an enhanced affinity for antigen (19). Therefore, the number of the antigenic determinants that are recognized by the organism appears to be one of the important factors involved in immune complex disease. Some studies examining the effect of the number of the antigenic determinants on immune complex size, using antihapten antibodies have been reported (8). However, even when antihapten antibody and a molecule having a restricted number of
386
haptens are used, (9). Monoclonal obtained in large different epitopes heterogeneity on
HOSOI,
SHINOMIYA,
AND
MIKAWA
heterogeneity both of antibody and antigen cannot be excluded antibody binds to a single antigenic determinant and can be quantities. Thus, we used five monoclonal antibodies against of human serum albumin to examine the effect of antibody immune complex size. REFERENCES
1. Cochrane, C. G., and Koffler, D., Zn “Advances in Immunology” (F. J. Dixon and H. G. Kunkel, Eds.), Vol. 16 p. 185, Academic Press, New York, 1973. 2. Christian, C. L., and Sergent, J. S., Amer. J. Med. 61, 385, 1976. 3. Segal, D. M., and Hurwitz, E., J. Zmmunol. 118, 1338, 1977. 4. Wright, J. K., Tschopp, J., Jaton, J. C., and Engel, J., Biochem. J. 187, 775, 1980. 5. Germuth, J. G., and Rodriguez, E., “Immunopathology of the Renal Glomerulus,” Little Brown, Boston, 1973. 6. Wilson, C. B., and Dixon, F. J., J. Exp. Med. 134, 7S, 1971. 7. Devey, M. E., Beasdale, K., Collins, M., and Steward, M. W., Znt. Arch. Allergy Appl. Zmmunol. 68, 47, 1982. 8. Atsumi, T., Nishida, K., Kinoshita, Y., and Horiuchi, Y., Immunochemistry 8, 271, 1981. 9. Atsumi, T., Nishida, K., Kinoshita, Y., Shibata, K., and Horiuchi, Y., J. Zmmunol. 99, 1286, 1967. 10. Kohler, G., and Milstein, C., Nature (London) 256, 495, 1975. il. Hosoi, S., Nakamura, T., Higashi, S., Yamamuro, T., Toyama, S., Shinomiya, K., and Mikawa, H., Cancer Res. 42, 654, 1982. 12. Gold, E. R., and Fudenberg, H. H., J. Zmmunol. 99, 859, 1967. 13. Ey, P. L., Prowse, S. J., and Jenkin, C. R., Immunochemistry 15, 429, 1978. 14. Markwell, M. A. K., and Fox, C. E, Biochemistry 17, 4807, 1978. 15. Atassi, M. Z., Immunochemistty 12, 423, 1975. 16. Marion, T. N., Lawton, A. R., III, Keamey, J. E, and Briles, D. E., J. Zmmunol. 128, 688, 1982. 17. Morrow, C. D., Dorey, F., and Stevens, R. H., J. Zmmunol. 130, 818, 1983. 18. Pincus, T., Haberkern, R., and Christian, C. L., J. Exp. Med. 127, 819, 1967. 19. Ehrlich, P. H., Moyle, W. R., Moustafa, Z. A., and Canfield, R. E., J. Zmmunol. 128, 2709, 1982. Received February 22, 1984; accepted with revisions May 14, 1984.