Production of syngeneic autoreactive monoclonal antibodies specific for isotypic determinants of IgE

Journal of Immunological Methods, 105 (1987) 193-199 Elsevier

193

JIM 04577

Production of syngeneic autoreactive monoclonal antibodies specific for isotypic determinants of IgE * Seiji H a b a and Alfred Nisonoff Rosenstiel Research Center, Department of Biology, Brandeis University, Waltham, MA 02254, U.S.A. (Received 13 July 1987, accepted 27 July 1987)

Methods are described for the production of syngeneic mouse anti-IgE monoclonal antibodies (mAb). Hybridomas were prepared by using spleen cells from mice immunized with a conjugate of keyhole limpet hemocyanin with monoclonal IgE. Serum titers varied from approximately 40 to 1000/~g/ml. The anti-IgE mAb were isolated by affinity chromatography on columns containing immobilized monoclonal IgE. The mAb are specific for isotypic determinants of IgE and do not react with other immunoglobulin isotypes. One of the mAb, which has a high affinity for IgE (K a = 4.7 X 108 M-l), should be useful for studies of regulation of IgE. The applicability of the mAb to quantitative assays for IgE was demonstrated. Key words: Autoantibody; Monoclonal anti-IgE; Syngeneic anti-IgE; Immune tolerance

Introduction

The spontaneous occurrence of auto-antibodies to IgE has been described in patients with hyperIgE syndrome, allergic respiratory disease, atopic dermatitis (Quinti et al., 1986) or allergic asthma (Inganas et al., 1981). The concentrations of autoanti-IgE were not reported but the highest appear to be < 1 t~g/ml. Marshall and Bell (1985) induced an anti-IgE response in BN or PVG. RT1 u rats by inoculation of rat IgE myeloma proteins (LOU/c WS1 strain) precipitated in alum. Assays for anti-IgE were Correspondence to: S. Haba, Rosenstiel Research Center, Department of Biology, Brandeis University, Waltham, MA 02254, U.S.A. * Supported by Grants AI-24272 and AI-22058 from the National Institutes of Health. Abbreviations: At, p-azobenzenearsonate; BSA, bovine serum albumin; CAFI, (BALB/cXA/J)F1; CFA, complete Freund's adjuvant; G aMFc, goat anti-mouse Fc; KLH, keyhole limpet hemocyanin; mAb, monoclonal antibody or antibodies; TNP, trinitrophenyl.

carried out by a solid-phase method using IgEcoated wells and the serum containing anti-IgE, with 125I-labeled sheep anti-rat IgG as the developing reagent. Titers of anti-IgE were not directly specified but the amounts of anti-IgE/ml of undiluted anti-IgE antiserum were sufficient to bind several hundred nanograms of the labeled sheep antibody. Inoculation of the IgE myeloma protein reduced the apparent IgE serum titers in PVGRT u rats and, as shown subsequently (Marshall et al., 1987) altered the proportions of mast cell types in recipient rats. We recently reported a method for generating high titers of auto-antibodies to IgE in A / J mice (Haba and Nisonoff, 1987). The method is based on the use, as immunogen, of a covalent conjugate of KLH with monoclonal A / J IgE. The conjugate is inoculated into adult A / J mice in complete Freund's adjuvant. Anti-IgE titers as high as 1 mg/ml were obtained by this method. The results suggest an absence of tolerance to IgE at the B cell level in adult mice. T cell tolerance is evidently broken by the use of the KLH conjugate.

0022-1759/87/$03.50 © 1987 Elsevier Science Pubfishers B.V. (Biomedical Division)

194

In this paper we describe methods for the production of monoclonal anti-IgE antibodies that are reactive with isotypic determinants of IgE. Because these antibodies are syngeneic with the IgE they should be useful in studies of regulation of IgE biosynthesis. One of these mAb (AEll) is of fairly high affinity (K a = 4.7 × 108 M -1 ) and may therefore be of particular value for such studies.

Materials and methods

Mice A / J and (BALB/c × A / J ) F 1 (CAFl) mice were obtained from the Jackson Laboratory (Bar Harbor, ME). Monoclonal IgE antibodies Each IgE mAb was of the kappa type. SE20.2 (Haba et al., 1985) and SE21.1 (Haba and Nisonoff, 1987) are of A / J origin and are Arspecific; TIB-142 (American Type Culture Collection, Rockville, MD, donated by M. Wabl (Rudolph et al., 1981)) is from a BALB/c mouse and has specificity for the trinitrophenyl (TNP) hapten. Other mAb used (all kappa and all Arspecific) are SM1.5 (IgM; Robbins et al., 1986); SA131 (Ig,A; Haba and Nisonoff, 1987); R16.7 (IgG1), R22.4 (IgG2a), R9.3 (IgG2b), R13.4 (IgG3) (Lamoyi et al., 1980). Each mAb was affinity purified, using a specific hapten for elution, as described elsewhere (Haba and Nisonoff, 1985). Immunization of mice and preparation of syngeneic anti-IgE monoclonal antibodies Five 6-week-old A / J mice were immunized with 200/~g quantities of a conjugate of IgE (SE21.1) with KLH, prepared by using glutaraldehyde as described elsewhere (Haba and Nisonoff, 1987); the immunogen was emulsified in CFA. Mice were boosted with 200 btg of the antigen in CFA 14 days and 35 days after the initial inoculation and three of the mice were inoculated again 30 days later. To prepare anti-IgE hybridomas, spleens were removed 7 days after the last inoculation. Hybridomas were prepared by a modification (Gefter et al., 1977) of the method of K~Shler and Milstein (1976) using the non-secreting Sp2/0 cell

line (Shulman et al., 1978) for the fusion. Supernatants were assayed for anti-IgE activity by the solid-phase method described below. Selected hybridomas were recloned and grown as ascites in pristane-pretreated CAF1 mice.

Purification of anti-IgE mAb Ammonium sulfate was added to ascitic fluid to a final concentration of 45% of saturation. The precipitate was dissolved, treated again with the same concentration of ammonium sulfate, and the protein further purified by ion-exchange chromatography on DE52 (Whatman, Kent, England). Finally, the anti-IgE mAb were affinity purified by using a Sepharose 4B conjugate of the IgEx mAb, SE20.2 (not the immunogen). Elution was carried out with 3 M NaSCN. The isotype of each anti-IgE mAb was determined by Ouchterlony analysis. Solid-phase assay for anti-IgE antibody Polyvinylchloride microtiter plates were coated with mAb SE20.2 (1 /xg in 100 /~l/well), and saturated with 2.5% horse serum. After exposing to 50 #1 of the test sample/well, for 6 h at room temperature, the plates were washed and exposed overnight at room temperature to 100 ng in 100 /zl/well of 125I-labeled affinity-purified goat antimouse Fc (GaMFc) that had been adsorbed with mouse IgE. This reagent did not react with mouse IgE but reacted with each of the subclasses of mouse IgG. As a standard we used pooled polyclonal A / J anti-SE21.1 antiserum whose antibody concentration had been determined by measuring its maximum IgE-binding capacity/unit volume of serum by the liquid-phase assay. (Note that SE20.2, used for coating wells, was an IgE mAb other than the immunogen.) Liquid-phase assay for anti-IgE and affinity measurements These assays were carried out by using a fixed amount of anti-IgE (mAb or antiserum) and varying amounts of 125I-labeled SE20.2, with 5 /zl of normal A / J serum as carrier. The normal A / J serum was pre-adsorbed by passage through a rabbit anti-IgE-Sepharose column. Complexes were precipitated with rabbit anti-mouse Fc which had been preadsorbed with mouse IgE conjugated

195

to Sepharose. After standing overnight, the radioactivity of the precipitate was measured. Each assay mixture contained 22NaC1. This permitted estimation of the amount of supernatant remaining in the precipitate so that a washing procedure was not required. Data were processed with a Hewlett-Packard microcomputer. The concentration of anti-IgE in an antiserum was estimated by extrapolation of the curve of 1/b vs. 1 / f to l / f = O, where b and f are the bound and free concentrations of labeled IgE, respectively. Affinity constants were determined by Scatchard analysis (see Fig. 2).

+

S...

Electrophoresis in agarose gel This was carried out by using the 'Panagel' electrophoresis procedure (Princeton Separations, Freehold, N J). Electrophoresis was carried out for 50 min at 200 V and proteins were stained with Amido black.

Results

Monoclonal anti-IgE Anti-IgE hybridomas were prepared with A / J spleens as described above. Each of five donor mice was used in a separate fusion. The donors had anti-IgE serum titers of 43, 50, 760, 1000 and 1070 # g / m l , respectively. The three highest titers were in mice that received a total of four, rather than three inoculations of KLH-conjugated IgE. From a total of 14 positive wells eight were selected for recloning and three of these were used in this report; one was from a mouse having a high titer of anti-IgE and two from mice of low titer (43 and 50 #g/ml). The hybridomas were finally grown as an ascites in CAF 1 mice, and monoclonal antibodies were purified as described above. The subclasses of the anti-IgE hybridoma products (all kappa) were as follows: five IgGl, four IgG2a, five IgG2b. The agarose gels shown in Fig. 1 provide evidence for purity of the isolated products. After affinity purification each mAb formed a single band.

Properties of anti-IgE mAb These are summarized in Table I. One mAb

i

1

2

3

4

5

6

7

8

Fig. 1. Electrophoretic patterns obtained in agarose gels (Veronal buffer, pH 8.6). The arrow indicates the position of application of the sample. Samples numbered 1 and 8 are undiluted normal A / J serum; 2, 4, 6: ascitic fluids containing hybridomas A E l l , AE2, and AE1, respectively. 3, 5, 7: affinity purified AEll, AE2, AE1 (15/tl of a 3 mg/ml solution).

was IgG1K and two were IgG2ax. The affinity of the IgGlx mAb was about two orders of magnitude greater than those of the other two. The Scatchard plots upon which the derived affinities are based are shown in Fig. 2. Each experiment was carried out in duplicate, and both points are shown in the figure. Straight lines were fitted to the data by the method of least squares, and K a values were determined from the calculated slopes of the lines. The valences (maximum binding of

TABLE I PROPERTIES OF ANTI-IgE MONOCLONAL ANTIBODIES mAb

Isotype

Total isolated/vol. ascitic fluid

K a × 10- 7 (M - 1)

AEll AE2 AE1

IgGIK IgG2aK IgG2a•

5 mg/5 ml 6 mg/7 nil 5 mg/9 ml

47 0.43 0.32

196

TABLE II TESTS OF SPECIFICITY OF MONOCLONAL ANTI-IgE ANTIBODIES I0 x m

Wells coated with a

o

W t~

~o°8 o x

5

~1 ~

z

ti0

~0 2.0

Fig. 2. Scatchard plots of binding of 125I-labeled SE20.2 (IgEk) to affinity-purified anti-IgE mAb. The concentration of mAb was held constant at 200 ng/ml (curve A) or 5/~g/ml (curves B and C). (A) mAb AEll; (B) mAb AE2; (C) mAb AE1. r, moles IgE bound/mole anti-IgE; c, free concentration of IgE.

IgE per molecule of anti-IgE), derived from the intercepts on the abscissa, vary from 1.58 to 1.85. Differences f r o m the expected value of 2.0 could be due to partial denaturation, c o n t a m i n a t i o n not observed by electrophoresis, or errors in the assumed extinction coefficients of IgE or anti-IgE (E280 = 1.4 for a 1 m g / m l solution). The assumed molecular weights were 150000 for I g G and 190000 for IgE.

Specificity of anti-IgE mAb Two types of analyses were carried out to establish specificity of the anti-IgE m A b for IgE. In the first, wells of polyvinylchloride microtiter plates were coated with normal A / J I g G or with one of a variety of mAb, differing in isotype; wells were exposed to 1/~g of the coating protein in 100 /d of neutral buffer. After incubating overnight, the plates were washed and saturated with 2.5% horse serum. The washed plates were exposed to 20 ng in 100 /~1 of affinity purified a25I-labeled anti-IgE m A b for 6 h at r o o m temperature. After washing, the individual wells were counted in a Packard g a m m a counter. The total a m o u n t of

None NIgG IgE (SE20.2) IgE (SE21.1) IgE (TIB142) IgM (SM1.5) IgA (SA131) IgG (R16.7) IgG2a (R22.4) IgG2b (R9.3) IgG3 (R13.4)

Counts/min 125Ibound 125I-labeled antibody b AE11 c

AE2

AE1

GaMFc d

330 320 83 700 71000 61500 540 320 400 420 390 320

250 290 10500 8 200 7 750 200 200 300 280 280 280

400 420 13100 11350 11200 320 300 370 400 400 490

2000 119000 1500 1300 4 000 2 000 4000 87 000 42 500 45 000 15000

a Each well was exposed to 1 /~g/100 /tl of the specified protein. Except for the normal A / J IgG (NIgG) each protein was an m A b and each was of the kappa type. b 20 ng in 100 /tl for the mAb; 100 ng for G a M F c . The total amounts of radioactivity added ( c o u n t s / m i n ) were: 103 000 for A E l l ; 88000 for AE2; 155000 for AE1; 208000 for G a M F c . Assays were done in duplicate. A E l l , AE2 and AE1 are affinity purified anti-IgE m A b

(Table I). d Affinity purified goat anti-mouse Fc.

radioactivity used in each experiment is indicated in Table II, which also presents the results. It is evident that the labeled anti-IgE m A b adhered appreciably only to wells coated with IgE. As a control, affinity purified 125I-labeled goat antimouse Fc was used in place of the labeled anti-IgE. (The Fc was derived from pooled normal A / J IgG.) Binding to each subclass of I g G was observed, but there was little or no binding to wells coated with A / J IgE, I g A or I g M (Table II). A second set of assays for specificity of the anti-IgE m A b was carried out by measurements of inhibition of binding. Wells of microtiter plates were coated with the m A b SE20.2 (IgEx) as described in the materials and methods section. A n lzSI-labeled anti-IgE m A b was preincubated for 1 h with the material to be tested for inhibition. 100 /~1 of a mixture containing 20 ng of a25I-labeled anti-IgE and varying amounts of inhibitor was then added to an IgE-coated well and allowed to stand for 6 - 8 h at r o o m temperature. The results

197

AI 2b

M\O/

OxOI

~2o 3

~I00 o u

tx

cl

z 50 ~"o.,S E20.2 E21.1 TIBI42 %

~

o o3 H if) r,/

IO

I O0

effective, at the 50% inhibition level, as SE21.1 (the least effective of the IgE inhibitors). These results indicate that the isotypic determinants on different IgE mAb recognized by AE1 or AE2 are not completely identical in their fine structure. This could be due to an influence of carbohydrate variability or of V regions.

Use of anti-IgE mAb for assays of IgE

1,0OO fig INHIBITOR

I O,OOO

Fig. 3. Inhibition of binding of 10 ng ]2SI-labeled anti-IgE (mAb A E l l ) to wells coated with IgE mAb SE20.2 (see materials and methods section). Unlabeled inhibitors (IgE mAb SE20.2, SE21.1 or TIB142) were preincubated with labeled A E l l . The symbols in the upper right comer show the degree of inhibition by 30000 ng of unlabeled IgM (M), IgA (A), IgG1 (1), etc. The symbol x represents non-specific A / J IgG.

obtained using anti-IgE mAb AE11 are shown in Fig. 3. It is evident that three different unlabeled IgE mAb, including the immunogen (SE21.1), yielded very similar inhibition curves; 23-35 ng was required for 50% inhibition. Two of the mAb are of A / J origin and one (TIB142) is BALB/c. As indicated in Fig. 3, upper fight corner, unlabeled mAb of other isotypes failed to cause significant inhibition when 30000 ng was present in the mixture. All isotypes except IgD were tested as inhibitors. Similar results were obtained for the other two anti-IgE mAb (AE1 and AE2), except that considerably larger amounts of unlabeled IgE were needed to cause the same degree of inhibition (data not shown). This is consistent with the lower affinities of mAb AE1 and AE2 as compared to A E l l (Table I). All mAb expressing isotypes other than IgE (see Fig. 3) were essentially non-inhibitory; 30000 ng caused less than 5% inhibition in each case. The amounts of IgE mAb needed for 50% inhibition ranged from 180 to 2000 ng, and 30000 ng of IgE caused more than 92% inhibition in each case. In contrast to anti-lgE mAb A E l l (Fig. 3) quantitative differences in inhibitory capacity of the three IgE mAb were noted, with TIB142 (the best inhibitor) being 9 or 13 times as

The three anti-IgE mAb were tested for their utility in quantitative assays for antibodies of the IgE class. Wells of microtiter plates were coated with TNP-BSA (100 /~g in 100 /~1) and saturated with 2.5% horse serum. Varying amounts of mAb TIB142 (IgEx, anti-TNP) were added to the wells. After standing for 6 h, the wells were washed and exposed overnight to 50 ng of 125I-labeled anti-IgE mAb in 100/~1 of buffer. The results are shown in Fig. 4. With each anti-IgE mAb a direct relationship is observed between the concentration of anti-TNP IgE used and the amount of labeled anti-IgE mAb bound to the well. The highest sensitivity was obtained with mAb A E l l which

IxlO 5

3

~3

o

E

•

I x l O4

g m

b lxiO l,a u

Ixl O'

0.I

IO ng A N T I - T N P TgE

IO

I00

Fig. 4. Use of 12SI-labeled, affinity purified anti-lgE mAb for quantitative assays of IgE antibody concentrations. Wells were coated with TNP-BSA (materials and methods section) and exposed to varying concentrations of anti-TNP IgE (TIB142). The developing reagent in each case is 50 ng 125I-labeled anti-lgE mAb. (A) AE11 (234000 counts/rain); (B) AE2 (210000 counts/min); (C) AE1 (348000 counts/min). The ordinate shows counts/min above background levels (wells exposed to the diluent only; the diluent is 0.5% BSA in saline-borate buffer). Background values were - 600- - 1000 counts/min.

198 also has the highest affinity for IgE (Table I). The lower limits of anti-TNP IgE detectable (cpm = two times background) varied from 220 pg (for m A b A E l l ) to 3.6 ng (for mAb AE2).

Discussion The data presented confirm the ability to generate high titers (up to 1 m g / m l ) of autoantibodies specific for isotypic determinants of IgE. The method makes use of a conjugate of K L H with IgE as the immunogen ( H a b a and Nisonoff, 1987). An IgE conjugate is the only preparation that we have observed so far to be strongly immunogenic in syngeneic adult mice. The ability to generate high concentrations of auto-anti-IgE is of theoretical interest from the standpoint of mechanisms of immunological tolerance. In the present paper, methods are described for the preparation and isolation of monoclonal antibodies with specificity for syngeneic IgE. Such m A b should be useful for studies of regulation of IgE in a syngeneic system. Data presented in this paper indicate that they can also be used successfully in immunoassays for IgE, although the property of being syngeneic may not be particularly advantageous for this purpose. The data presented also focus on the properties of three A / J anti-IgE mAb. The immunogen was the IgEK A / J mAb, SE21.1. One of the anti-IgE m A b is IgGIK and two are IgG2ax. The IgG1 protein ( A E l l ) may be especially useful because of its relatively high affinity ( K a = 4.7 × 108 M - 1 ) . This affinity is comparable to the highest average affinities we have observed in polyclonal anti-IgE prepared by this method of immunization ( H a b a and Nisonoff, 1987). Specificity of the anti-IgE m A b was demonstrated in two independent assays. One assay measured direct binding of labeled anti-IgE m A b to wells coated with m A b of various isotypes. The second was based on inhibition of binding of labeled anti-IgE to IgE by immunoglobulins of various isotypes. The ligand in the latter assay was an IgE preparation other than the immunogen; this avoids any contribution of idiotypic determinants. The data indicated that each m A b is specific for isotypic determinants of IgE and is non-reac-

tive with mAb of other isotypes. Since the tests were carried out with syngeneic IgE, the results suggested that the interactions do not involve allotypic determinants of IgE. However, the possibility existed that binding occurred to a region corresponding to the locus of an allotypic determinant. This apparently is not the case. We found that each of the A / J anti-IgE m A b was strongly reactive with IgE of the C 5 7 B L / 6 strain, which differs allotypically (Borges et al., 1981) from A / J IgE (data not shown). The methods described are now being tested with IgD as the immunogen. IgD has in common with IgE a very low concentration in normal mouse serum (Bargellesi et al., 1979; Finkelman et al., 1979).

Acknowledgement The authors are grateful to Ms. Michele Dante for excellent technical assistance in the preparation of hybridomas.

References Bargellesi, A., Corte, G., Cosulich, E. and Ferrarini, M. (1979) Presence of serum IgD and IgD-containing plasma cells in the mouse. Eur. J. Immunol. 9, 490. Borges, M.S., Kumagai, Y., Okumura, K., Hirayama, N., Ovary, Z. and Tada, T. (1981) Allelic polymorphism of murine IgE controlled by the seventh immunoglobulin heavy chain allotype locus. Immunogenetics13, 499. Finkelman, F.D., Woods, V., Berning, A. and Scher, I. (1979) Demonstration of mouse serum IgD. J. Immunol. 123, 1253. Gefter, M.L., Margolies, D. and Scharff, M.D. (1977) A simple method for polyethylene glycol-promotedhybridization of mouse myeloma cells. Somatic Cell Mol. Genet. 3, 321. Haba, S. and Nisonoff, A. (1985) Quantitation of IgE antibodies by radioimmunoassayin the presence of high concentration of non-IgE antibodies of the same specificity. J. Immunol. Methods 85, 39. Haba, S. and Nisonoff, A. (1987) Induction of high titers of anti-lgE by immunization of inbred mice with syngeneic IgE. Proc. Natl. Acad. Sci. U.S.A. 84, 5009. Haba, S., Ovary, Z. and Nisonoff, A. (1985) Clearance of IgE from serum of normal and hybridoma-bearing mice. J. Immunol. 134, 3291. Inganas, M., Johansson, S.G.O. and Bennich, H. (1981) AntiIgE antibodies in human serum: occurrence and specificity. Int. Arch. Allergy Appl. Immunol. 65, 51.

199 K/Shler, G. and Milstein, C. (1976) Derivation of specific antibody-producing tissue culture and tumor lines by cell fusion. Eur. J. Immunol. 6, 511. Lamoyi, E., Estess, P., Capra, J.D. and Nisonoff, A. (1980) Heterogeneity of an intrastrain cross-reactive idiotype associated with anti-p-azophenylarsonate antibodies of A / J mice. J. Immunol. 124, 2834. Marshall, J.S. and Bell, E.B. (1985) Induction of an auto-antiIgE response in rats. I. Effects on serum IgE concentrations. Eur. J. Immunol. 15, 272. Marshall, J.S., Prout, S.J., Jaffery, G. and Bell, E.B. (1987) Induction of an auto-anti-IgE response in rats. II. Effects on mast cell populations. Eur. J. Immunol. 17, 445. Quinti, I., Brozek, C., Wood, N., Geha, R.S. and Leung,

D.Y.M. (1986) Circulating IgG antibodies to IgE in atopic syndromes. J. Allergy Clin. Immunol. 77, 586. Robbins, P.F., Rosen, E.M., Haba, S. and Nisonoff, A. (1986) Relationship of V H and V L genes encoding three idiotypic families of anti-p-azobenzenearsonate antibodies. Proc. Natl. Acad. Sci. U.S.A. 83, 1050. Rudolph, A.K., Burrows, R.D. and Wabl, M.R. (1981) Thirteen hybridomas secreting hapten-specific immunoglobulin E from mice with Ig a or Ig b heavy chain haplotype. Eur. J. Immunol. 11, 527. Shulman, M., Wilde, C.D. and K/Shler, G. (1978) A better cell line for making hybridomas secreting specific antibodies. Nature 276, 269.