Proc. Natl. Acad. Sci. USA Vol. 85, pp. 2293-2297, April 1988 Immunology Induction and properties of syngeneic murine anti-immunoglobulin D

SEIJI HABA*t, G. JEANETTE THORBECKEt, AND ALFRED NISONOFF* *Department of Biology, Rosenstiel Research Center, Brandeis University, Waltham, MA 02254; and tDepartment of Pathology, New York University Medical Center, New York, NY 10016 Contributed by Alfred Nisonoff, December 21, 1987

ABSTRACT High titers of autoantibodies directed to ever, the greatly reduced numbers of remaining 4+, 5- B isotypic determinants of IgD were produced by inoculation of cells in such mice were shown to be actually hyperresponsive syngeneic monoclonal IgD, conjugated covalently to keyhole to antigens injected i.v. (6, 8). limpet hemocyanin, into adult or neonatal inbred mice. Anti- idiotypic antibodies were induced at the same time. The MATERIALS AND METHODS average affinity of the mouse antibodies (Ka 1079 M-1) is similar to that of rabbit anti-IgD and of syngeneic anti-IgE Mice. BALB/c mice were obtained from The Jackson induced by the same procedure. Results indicate that B cells of Laboratory. Mice were 5-8 weeks old at the start of each the mice are not tolerant to serum IgD and that tolerance is experiment. Neonatal mice were 2 days old at the first maintained at the level of T cells. Direct interaction of the antigen inoculation. syngeneic anti-IgD with cell-surface IgD was minimal, and Monoclonal IgD. Two IgDK preparations were used; both there was no convincing evidence that cell-surface IgD was were derived from BALB/c myeloma cell lines (TEPC-1033 down-regulated in the anti-IgD-producing mice. Further stud- and TEPC-1017) developed by F. Finkelman, M. Potter, and ies, preferably employing monoclonal anti-IgD, are required collaborators (11). The cell lines were grown as ascites cells to determine whether epitopes on cell-surface IgD can be in pristane-primed BALB/c mice. Purification of IgD was recognized by syngeneic anti-IgD. The ability to generate in done by modifying slightly the method of Finkelman et al. vivo high titers of anti-IgD should facilitate the production of (11). Our procedure includes two precipitations with ammo- such monoclonal antibodies. nium sulfate followed by chromatography on DEAE- cellulose in 0.03 M sodium phosphate buffer, pH 8.3. The Recently we reported the induction of high titers (to 1 mg/ IgD was obtained together with IgG in the fall-through ml) of syngeneic anti-IgE with isotypic specificity in adult fraction; IgG was removed by repeated passage through a A/J mice (1). The immunogen used was a copolymer of protein A-Sepharose column (Sigma) in a high-salt buffer keyhole limpet hemocyanin (KLH) with an A/J monoclonal (0.1 M sodium phosphate/i M sodium chloride, pH 8.3). antibody (mAb) of the IgE class. Results indicated that mice Finally, gel filtration was done on Sephacryl S-300 (Phar- are not tolerant at the level of B cells to their own IgE. The macia). Purification was monitored by double diffusion in fact that the copolymer was immunogenic, whereas mono- agarose gel, immunoelectrophoresis, and NaDodSO4/ meric IgE was not, indicated that tolerance is mediated by T PAGE. In a nonreducing gel the final product migrated as a cells and can be broken by the use of a foreign carrier single major band with an apparent Mr of 125,000 (cf. ref. molecule, KLH. These observations permitted the produc- 12); gel filtration indicated a Mr for both TEPC-1033 and tion of syngeneic anti-IgE mAbs (2), which we are using in TEPC-1017 somewhat higher than that of IgE, a result studies of regulation of IgE synthesis. consistent with the presence of a dimer of IgD (11). A The capacity to produce high titers of syngeneic anti-IgE portion of each purified IgD preparation was trace-labeled suggested that the same methods might be applied to the with 1251 by the chloramine-T method (13). Samples were induction of syngeneic anti-IgD because IgD and IgE both tested for degradation by gel filtration shortly before using exist at very low concentrations in normal mouse serum. In the material in RIAs. Small amounts of breakdown products fact, existence of circulating autoantibodies to human IgD in were detected, but the major product was of high molecular rheumatic diseases has been reported (3). However, because weight. IgD is expressed on the cell surface of >90% of B cells, it Conjugation of IgD with KLH. Conjugates were prepared seemed possible that self-tolerance to IgD might be more (14) by mixing 10-mg quantities of purified IgD and KLH in firmly established than self-tolerance to IgE. We here de- a final volume of 2 ml in 0.1 M sodium phosphate buffer, pH scribe the induction of high titers of syngeneic anti-IgD and 6.0. Glutaraldehyde was added in an amount equal to 20 mol report on the affinities and specificity of such antibodies. per mol of IgD. After approximately one-half hour of incu- The results are relevant to questions of self-tolerance and bation at room temperature, the mixture became cloudy; the autoimmunity and may permit the eventual production of reaction was then quenched with 1 M L-lysine, and the syngeneic anti-IgD mAbs for use in studies of regulation of mixture was then dialyzed against neutral phosphate-buf- the immune system. fered saline. The significance of surface IgD in regulation has been Immunization of Mice. Neonatal or adult BALB/c mice demonstrated in many investigations (4-10). For example, it were immunized i.p. two or three times with IgD (TEPC- has been shown that treatment of mice from birth with 1017 or TEPC-1033) or with a conjugate of KLH with IgD; allogeneic anti-IgD suppresses cell-surface IgD expression complete Freund's adjuvant (CFA) was used in each case. and down-regulates the production of antibodies of all Schedules of immunization and amounts inoculated are spec- isotypes to antigens administered s.c. or i.p. (4, 5, 8). How- Abbreviations: CFA, complete Freund's adjuvant; FITC, fluores- The publication costs of this article were defrayed in part by page charge cein isothiocyanate; KLH, keyhole limpet hemocyanin; mAb, payment. This article must therefore be hereby marked "advertisement" monoclonal antibody; Fc, Fc fragment of IgG. in accordance with 18 U.S.C. §1734 solely to indicate this fact. tTo whom reprint requests should be addressed.

Downloaded by guest on October 4, 2021 2293 2294 Immunology: Haba et al. Proc. Natl. Acad. Sci. USA 85 (1988)

ified in table footnotes. Equal volumes ofantigen solution and to varying dilutions of serum. Development was done with CFA were used to prepare emulsions. '25I-labeled affinity-purified rabbit anti-IgD (40 ng in 0.1 ml). Antisera. Rabbit anti-mouse IgD antiserum was prepared As a standard we used purified IgD (TEPC-1033). against unconjugated TEPC-1017. Rabbit and goat anti-mouse Analysis of Cell Surface Immunoglobulin Phenotypes. These Fc (of IgG), used for double-diffusion analyses and for immu- analyses were done with a fluorescence-activated cell sorter noelectrophoresis, were prepared in our laboratory. For use (Orthocytofluorograph Spectrum II; Ortho Diagnostic Sys- in the solid-phase assay for anti-IgD, the goat anti-mouse Fc tems, Westwood, MA). Spleen cells or axillary and brachial was affinity-purified and trace-labeled with 125I by the chlo- lymph node cells were washed and suspended in phosphate- ramine-T method; the antibody reacted with mouse IgG but buffered saline/1% bovine serum albumin/0.1% sodium not with IgM. azide. The percentage of cells expressing surface IgM or IgD Assays for Anti-IgD. Two types of assay were used. In a was determined by indirect staining of 1 x 106 cells in 50 .ul solid-phase assay (15), wells of polyvinylchloride microtiter of medium using optimal concentrations of goat anti-mouse plates were coated with a purified IgD mAb (not the immu- IgM or goat anti-mouse IgD (provided by F. D. Finkelman, nogen). The coating solution contained 1 ;Lg of IgD in 0.1 ml. Uniformed Services University of the Health Sciences, Be- Horse serum (2.5%) was used to saturate surfaces. Coated thesda, MD), followed by fluorescein isothiocyanate (FITC)- wells were exposed to varying dilutions of antisera for 6 hr, labeled rabbit anti-goat immunoglobulin (Miles). The percent- and wells were developed with 50 ng of 125I-labeled affinity- age of cells expressing cell-surface Thyl.2 antigen was deter- purified goat anti-mouse Fc fragment. As a standard we used mined by direct staining with FITC-anti-Thyl.2 (ICN pooled anti-IgD serum, the titer of which was determined by Immunobiologicals, Lisle, IL). Labeled cells were analyzed a liquid-phase assay. in the cell sorter at 1 x 106 cells per ml. To determine whether Assays in the liquid phase were done by mixing varying mouse anti-IgD interacted with surface IgD, cells were incu- amounts of 1251I-labeled purified IgD (TEPC-1017) with a bated first with affinity-purified syngeneic mouse anti-IgD fixed dilution ofantiserum. Five microliters of normal mouse (100 ug/ml), or with dilutions of mouse serum initially con- serum was added as carrier. Complexes were precipitated taining 70 ,ug of anti-IgD per ml, and then stained with with rabbit antiserum directed against the Fc fragment of FITC-labeled goat anti-mouse IgG (American Qualex, Lami- mouse IgG. In the absence of anti-IgD a precipitate was rada, CA). In some experiments cells were exposed to mouse obtained, but it contained <1% of the added radioactivity. anti-IgD for 10 min before addition of goat anti-IgD or F(ab')2 Measurement of IgD-Binding Capacities and Average Ka fragments of rabbit anti-IgD to determine whether mouse Values of Anti-IgD (TEPC-1033) Antibodies. The IgD-binding anti-IgD would interfere with binding of the heterologous capacity of an antiserum was determined by the liquid-phase anti-IgD. Cells were stained with appropriate FITC-labeled assay using increasing amounts of labeled IgD (to 10 ,ug) and reagents, specific for goat or rabbit immunoglobulin. estimating the degree of binding at saturation from a plot of 1/b vs. 1/f, where b andf are bound and free concentrations RESULTS of IgD. For the extinction coefficient of IgD at 280 nm we Induction of Syngeneic Anti-IgD Antibodies in BALB/c used the value 14 for a 1% (wt/vol) solution. M, of TEPC- Mice. Table 1 shows that significant titers of anti-IgD anti- 1017, which is a dimer (11), was estimated as 250,000. bodies (to =220 ,ug/ml) can be induced in adult BALB/c The liquid-phase assay using 125I-labeled IgD (TEPC-1017) mice by two inoculations in CFA of a conjugate of KLH with was also used to estimate values for average K.. Fractions of IgD (TEPC-1033). Sera were assayed on wells coated with a IgD bound and free were determined for a series of concen- different myeloma IgD (TEPC-1017). Significant, but consid- trations of total IgD, with antiserum amount held constant. erably lower, titers of anti-IgD were also elicited with The value for average K. was obtained from a Scatchard plot unconjugated IgD as the immunogen. Anti-IgD antibodies on the assumption of a Sips distribution (16, 17) of binding could also be elicited when the first inoculation was given to affinities. Binding curves were all nonlinear, indicating het- BALB/c mice at 2 days of age (Table 2); in this case erogeneity of anti-IgD antibodies with respect to their bind- unconjugated TEPC-1033 was about as effective as the ing affinities. KLH-TEPC-1033 polymer. Concentrations of anti-IgD elic- Affinity-Purification of Rabbit and Mouse Anti-IgD. Rabbit ited by the polymer were considerably lower, however, in anti-TEPC-1017 was affinity-purified from antiserum on a neonatal than in adult mice (Tables 1 and 2). In contrast to Sepharose conjugate of TEPC-1033. BALB/c anti-TEPC- TEPC-1033, TEPC-1017 was a very poor immunogen in 1033 (IgG fraction) was affinity-purified on a Sepharose- neonatal mice, although both IgD preparations induced TEPC-1017 conjugate. NaSCN (3 M) was used for elution. anti-IgD in adult recipients when administered as the copoly- Assay for IgD in Serum. Wells of microtiter plates were mer with KLH. Neonatal mice given unconjugated TEPC- coated with affinity-purified rabbit anti-IgD (TEPC-1017; 1 1017 did exhibit low-to-significant titers (to 7 ,g/ml) at a ,ug/ml), saturated with 2.5% horse serum, and then exposed later bleeding (day 73; data not shown). Table 1. Production of anti-IgD antibodies against syngeneic IgD in adult BALB/c mice Anti-IgD, ,uglml Dose, t.g per Day 28 Day 42 Group* Immunogen inoculation Range Mean ± SEM Range Mean ± SEM 1 KLH-IgDt 200 23-56 38 ± 4 119-218 138 ± 21 2 IgDt 100 6-22 10 ± 2 10-54 21 ± 5 3 KLH 100 All <0.1 All <1.5 4 Saline + CFA (0.2 ml) All <0.1 All <0.1 Mice were inoculated i.p. with 0.1 ml of the antigen emulsified with 0.1 ml of CFA on days 0 and 14 and bled on the day specified. Serum anti-IgD was assayed in wells coated with TEPC-1017. (The IgD used for immunization was TEPC-1033.) *Each group contained eight mice. tKLH and IgD were copolymerized with glutaraldehyde. tUnconjugated. Downloaded by guest on October 4, 2021 Immunology: Haba et al. Proc. Natl. Acad. Sci. USA 85 (1988) 2295 Table 2. Production of anti-IgD antibodies against syngeneic IgD in neonatal BALB/c mice Anti-IgD, ,mg/ml Mice, Dose, u±g per Day 28 Day 42 Day 52 Group no. Immunogen inoculation Range Mean ± SEM Range Mean ± SEM Range Mean ± SEM 1 4 KLH-IgD* (TEPC-1033) 50, 100, 100 7-25 14 ± 4 12-33 19 ± 5 14-25 20 ± 3 2 4 KLH-IgD* (TEPC-1017) 50, 100, 100 3-18 10 ± 3 8-16 11 ± 2 11-24 16 + 3 3 5 IgD (TEPC-1033) 25, 50, 50 1-8 3 ± 1 1-6 3 ± 1 7-39 16 ± 1 4 5 IgD (TEPC-1017) 25, 50, 50 All <0.1 All <1.2 All <1.2 5 5 KLH 25, 50, 50 All <0.1 All <0.1 All <0.1 6 5 Saline + CFA (50, 100, 100 1LI) All <0.1 All <0.1 All <0.1 See footnotes, Table 1. For these experiments three inoculations were given on days 0, 14, and 42; mice were 2 days old at the first inoculation. Total volume of antigen emulsion administered was 50, 100, and 100 jJ for the three inoculations. Antibodies prepared against TEPC-1033 were assayed on wells coated with TEPC-1017 and vice versa. *Polymerized with glutaraldehyde. Specificity of the anti-IgD antibodies is shown by Tables 3 wells were developed with affinity-purified '25I-labeled rab- and 4; labeled anti-Fc of IgG was used as the developing bit anti-mouse IgM or goat anti-mouse Fc fragment. The reagent in these assays. Binding to IgD-coated wells was relative binding efficiencies of the labeled anti-IgM and inhibited to a large extent by 10 1Lg of added IgD but not by anti-Fc were determined by using known amounts of IgM or 10 tug of monoclonal mouse immunoglobulins of other IgG mAbs specific for the phenylarsonate hapten on trays isotypes (Table 3). The failure to observe complete inhibition coated with a bovine serum albumin-phenylarsonate conju- by soluble IgD suggests that the effective affinity (avidity) of gate. Four different pools of syngeneic anti-IgD antisera were binding of anti-IgD is much greater for IgD coated onto a used. (The pools are those listed in Table 5.) The calculated surface than for IgD in solution-probably because of biva- ratios of IgM to IgG antibodies were 1:10, 1:23, 1:32, and 1:58 lent attachment of antibodies to surface-bound IgD. for the four pools. We were unable to detect significant Further evidence for specificity was obtained with 125I- concentrations of either IgM or IgG anti-IgD antibodies in labeled affinity-purified anti-IgD, prepared from a pool of normal adult BALB/c sera used as controls. sera from adult BALB/c mice immunized with the KLH- Affinities of Syngeneic (BALB/c) and Rabbit Anti-Mouse TEPC-1033 conjugate (Table 4). Significant binding to wells IgD. Data on average binding affinities are shown in Table 5. coated with IgD was seen, but not to wells coated with For adult mice and for mice that received their initial immunoglobulins of other isotypes. Data in this table also inoculation at 2 days of age, average Ka values range from provide evidence for specificity of the affinity-purified rabbit 0.6-3.5 x 107 M -1. There are no major differences among anti-IgD. The results obtained with rabbit anti-mouse Fab adult and neonatal recipients. Affinity-purified rabbit anti- (controls) demonstrate that the wells were effectively coated IgD (pooled serum from two rabbits) had an average Ka of with immunoglobulin. 2.5 x 107 M-1 IgM and IgG Content of Syngeneic Anti-IgD. An estimation Content of Antiidiotypic Antibodies in Syngeneic Anti-IgD. of isotypes in the BALB/c anti-IgD (TEPC-1033) was made For these measurements we used BALB/c antiserum pooled by a solid-phase RIA using wells coated with IgD (TEPC- from adult mice immunized with KLH-TEPC-1033. Assays 1017). After exposure to varying dilutions of antiserum, were done in the liquid phase; complexes were precipitated Table 3. Specificity of serum antibodies prepared against with rabbit anti-mouse Fc fragment. As ligand we used syngeneic IgD increasing amounts of 125I-labeled TEPC-1033 (the immu- nogen) or TEPC-1017, together with a constant volume of Inhibition of binding, % Table 4. Specificity of syngeneic and heterologous Anti-IgD antiserum Pooled anti-IgD anti-IgD antibodies Inhibitor (10 /ug) 1 2 3 4 antiserum Labeled antibody, cpm of 1251 bound NIgG 3, 0 6, 0 1, 0 11, 9 0, 0 Well BALB/c anti-IgD Rabbit anti-IgD RaMFab* IgD (TEPC-1017) 87, 87 63, 60 79, 78 52, 45 69, 66 coating (20 ng) (20 ng) (100 ng) IgD (TEPC-1033) 87, 86 62, 59 82, 79 61, 42 56, 55 IgM (TEPC-183) 7, 0 0, 0 0, 0 4, 0 7, 0 None 1,160 510 2,070 IgG1 (MOPC-21) 5, 0 ND 0, 0 ND 2, 0 NIgG 1,650 900 62,550 IgG2a (LPC-1) 2, 0 10, 0 11, 0 0, 0 0, 0 IgD (TEPC-1017) 74,200 90,800 48,600 IgG2b (MOPC-195) 2, 0 2, 0 1, 0 6, 3 9, 7 IgD (TEPC-1033) 63,030 75,770 23,340 IgG3 (J-606) 0, 0 ND 0, 0 ND 7, 5 IgM (TEPC-183) 1,010 350 45,000 IgA (MOPC-460) 6, 0 8, 0 2, 0 7, 3 7, 5 IgM (MOPC-104E) 1,650 400 14,440 IgE (TIB-142) 0, 0 3, 2 4, 0 11, 8 4, 0 IgG1 (MOPC-21) 1,190 600 59,870 IgG2a (LPC-1) 1,620 700 52,360 Antisera were prepared in adult BALB/c mice against a copoly- IgG2b (MOPC-195) 1,600 700 59,730 mer of KLH and TEPC-1033 (IgD), except for the pool (last column) IgG3 (J-606) 1,820 450 39,540 that was prepared from sera of five other mice immunized with IgA (MOPC-460) 850 360 36,760 unconjugated TEPC-1033. RIAs were done on plates the wells of which were coated with IgD (TEPC-1017). The amount of antibody IgE (TIB-142) 780 400 36,360 present per well was 4-8 ng and was constant for a given antiserum. Wells were coated with the protein specified at a concentration of Inhibitors (10 ,ug of each) were purified BALB/c myeloma proteins 10 ,ug/ml and were then saturated with 2.5% horse serum. Coated or a hybridoma product (TIB-142), except for normal BALB/c IgG wells were exposed to 1251I-labeled, affinity-purified antibody for 6 hr (NIgG). Each experiment was done in duplicate, and both results at room temperature and then washed. Total cpm added to the wells are shown. The anti-IgD antibody titers of the four antisera and the were 186,000, 219,000, and 146,000 (columns 2, 3, and 4, respec- pool were 169, 88, 135, 103, and 7 ,ug/ml, respectively. ND, not tively). NIgG, normal BALB/c IgG. done. *Rabbit antibodies to mouse Fab fragments. Downloaded by guest on October 4, 2021 22% Immunology: Haba et al. Proc. Natl. Acad. Sci. USA 85 (1988) Table 5. Average Ka values for individual syngeneic anti-IgD rabbit antibody. Even at an extremely high level (50 ug) antisera or pools of antisera mouse anti-IgD caused only 70%o inhibition of binding of the Ka X 10-7 M-1 labeled rabbit anti-IgD, whereas 5 ,ug of unlabeled rabbit antibody caused 92% inhibition. Individual Pooled Concentrations of IgD in Normal and Immune Serum. Recipients Immunogen mice serum* Nonimmune sera of four adult BALB/c mice were found to Adult BALB/c mice KLH-TEPC-1033 0.6, 2.0, 2.1 2.5 have IgD concentrations ranging from 380 to 900 ng/ml with 2.4, 2.7, 3.5 a mean value of 560 ng/ml. IgD was not detectable (concen- Adult BALB/c mice TEPC-1033 0.8, 1.9, 1.9 0.7 tration <25 ng/ml) in the sera of four mice that had been Neonatal BALB/c KLH-TEPC-1033 0.6, 1.0, 2.0 2.6 immunized three times with KLH-TEPC-1033 in CFA. mice (2 days old) Test for Direct Binding to Cell Surfaces by Mouse Anti-IgD Neonatal BALB/c TEPC-1033 1.0, 1.2, 1.6 0.6 and Competition Experiments. Cell-staining data (Table 6) mice showed the following: (i) mouse anti-IgD, examined as Adult rabbit TEPC-1017 2.5t serum or as affinity-purified antibody, showed little binding to the surface of normal BALB/c spleen cells The immunogen was a copolymer of KLH with IgD (TEPC-1033) (column 3). (ii) or monomeric TEPC-1033 emulsified in CFA and administered i.p. High concentrations of mouse anti-IgD failed to inhibit signif- Adult mice received 200 ,ug of KLH-TEPC-1033 or 100 ,ug of icantly the binding of heterologous (goat or rabbit) anti-IgD to TEPC-1033 emulsified in CFA (three inoculations). Neonatal mice cell surfaces (columns 4 and 5). High percentages of cells received 50, 100, and 100 kug of the copolymer or half as much were stained by anti-I-Ad (positive control; column 6), using monomeric IgD per inoculation. Inoculations were given over a the same second antibody (FITC-labeled anti-IgG) that was period of 7 weeks. used for the experiments with syngeneic anti-IgD. *From five or six immunized mice and not including any individual Effect of Induction of Syngeneic Anti-IgD on Surface Im- mice of column 3. munoglobulin Expression. The percentages of IgM- or IgD- tAffinity-purified anti-IgD from a serum pool of two rabbits. bearing cells in anti-IgD-producing mice were only margin- antiserum. IgD-binding capacities of the antibodies in serum ally reduced in the spleen. The average percentages of were obtained by extrapolation of curves of 1/b vs. to spleen cells stained with goat anti-IgM or goat anti-IgD were 1/f 1/f 43% and 45%, respectively, for two mice immunized as = 0 (i.e., f = infinity), where b andf are the bound and free adults with KLH (controls) vs. 33% and 40% for three mice concentrations of labeled ligand. The binding capacity for immunized with KLH-IgD. The average values for three TEPC-1033 was found to be 330 A&g/ml; for TEPC-1017 it mice immunized with KLH-IgD from birth were somewhat was 210 ttg/ml. This suggests that 330 - 210, or 120 Ag of lower (28% and 27%, respectively). No significant differ- the antibodies per ml, is directed to idiotypic determinants of ences in staining for IgM or IgD, as compared with normal TEPC-1033. mice, were seen in lymph node suspensions from mice Comparison of Fine Specificities of Rabbit and Mouse Anti- immunized as adults or neonates. Percentages of lymph node IgD. To obtain information on this we question coated wells cells stained ranged from 18% to 25% for anti-IgM and from of polyvinylchloride microtiter plates with TEPC-1017 (10 15% to 28% for anti-IgD. Mean fluorescence values for Ag/ml). The anti-IgD antibodies used were rabbit anti-TEPC- lymph node cells stained with FITC-labeled goat anti-IgD 1017, affinity-purified on a column of TEPC-1033-Sepharose, were indistinguishable for anti-IgD-producing and control or BALB/c anti-TEPC-1033, affinity-purified on TEPC-1017- mice; that is, the levels of surface IgD were unchanged. Sepharose. 1251I-labeled anti-IgD antibodies (20 ng) were Staining with anti-mouse IgG did not show any increases mixed with varying amounts of unlabeled anti-IgD antibodies, over the background seen with KLH-immune mice: that is, and the degree of inhibition of binding of labeled anti-IgD was no detectable IgG anti-IgD bound to B cells despite the high determined. Unlabeled rabbit or mouse anti-IgD were about equivalent in their capacity to inhibit binding of the labeled concentration of circulating anti-IgD (data not shown). mouse antibody (Fig. 1). The unlabeled rabbit antibody was DISCUSSION much more effective, however, in displacing the labeled The data indicate that substantial titers (>200 ,ug/ml) of syngeneic anti-IgD can be induced in adult BALB/c mice by immunization with an IgD BALB/c myeloma protein conju- gated to KLH and emulsified in CFA. Unconjugated IgD 100 _ 1 0 was also effective, but lower concentrations of anti-IgD were produced in adult mice. Although some IgM anti-IgD anti- bodies were induced, 90-98% were IgG after hyperimmuni- 0 zation. Specificity for IgD was established by two different assays: one involved direct binding and the other was a competition assay. Significant concentrations of antiidio- z 50- typic antibodies, specific for the immunogen, were also 0 present in immune sera. The average binding affinities, Ka, of the anti-IgD antibod- ies were of the order of 107 M 1; these affinities are only moderately high for antiprotein antibodies but are orders of magnitude greater than those generally reported for human rheumatoid factors (see ref. 19) and are similar to the 10 100 1,000 10,000 average affinity of four rabbit anti-IgD (Table 5). We should ng INHIBITOR indicate, however, that the rabbit preparation contained some antibodies of very high affinity that were not found in FIG. 1. Displacement of 20 ng of '25I-labeled mouse or rabbit anti-IgD from IgD-coated wells. The amount of unlabeled competi- the mouse anti-IgD (data not shown). tor is specified on the abscissa. Binding of labeled rabbit anti-IgD Average affinities of syngeneic anti-IgD are quite compa- was inhibited by unlabeled mouse anti-IgD (o) or unlabeled rabbit rable to those we recently reported for syngeneic anti-IgE, anti-IgD (o). Labeled mouse anti-IgD was inhibited by unlabeled whereas maximum serum concentrations of anti-IgD were mouse anti-IgD (A) or unlabeled rabbit anti-IgD (A). about one-fourth those obtained for anti-IgE (1). IgD and IgE Downloaded by guest on October 4, 2021 Immunology: Haba et A Proc. Natl. Acad. Sci. USA 85 (1988) 2297 Table 6. Failure of mouse anti-IgD to stain cells or prevent staining of normal adult BALB/c spleen cells by heterologous anti-IgD Cells stained, % Mouse anti-IgD* + Anti- Exp. NMS Mouse anti-IgD* NMS + heterol. anti-IgDt heterol. anti-IgDt I-Adt 1 9.2 6.2 34.6 31.3 49.0 2 1.8 8.9 29.6 (2 ,ug/ml) 27.1 (2 ;tg/ml) 34.1 30.6 (10 JLg/ml) 29.4 (10 ,ug/ml) The second (staining) reagent was FITC-goat anti-mouse Fc fragment (of IgG) for columns 2, 3, and 6 and FITC-anti-goat immunoglobulin (Exp. 1) or FITC-anti-rabbit immunoglobulin (Exp. 2) in columns 4 and 5. NMS, normal mouse serum; heterol., heterologous. *Mouse anti-IgD (70 ,ug/ml) was used as serum (Exp. 1), or as affinity-purified anti-IgD (100 j.g/ml, Exp. 2). tThe heterologous anti-IgD was afflinity-purified goat anti-IgD (20 Zg/ml, Exp. 1) or F(ab')2 fragments of affinity-purified rabbit anti-IgD at 2 or 10 ,ug/ml (Exp. 2). tIn this positive control, cells were incubated with anti-I-Ad of the IgG2a isotype (18) as primary antibody.

have the common property that both are present at very low IgD as well as of IgE in the mouse are evidently insufficient concentrations in normal adult mouse serum; average con- to induce B-cell tolerance. With mice immunized with IgD, centrations are -2 ,g/ml for the IgE content of adult syngeneic anti-IgD mAbs can potentially be produced in BALB/c serum and -0.6 ,ug/ml for IgD. quantity-such antibodies may be useful for investigating Fine Specificity of Syngeneic Anti-IgD. Competition experi- the apparent failure of syngeneic anti-IgD to interact with ments (Fig. 1) indicated that the specificity of the mouse cell-surface IgD. Also, if lack of cell staining by anti-IgD anti-IgD is somewhat less broad than that ofrabbit anti-mouse reflects adsorption of such anti-IgD by B cells, possibly IgD. Although both were about equally effective in displacing monoclonal syngeneic anti-IgD that interacts with mem- labeled mouse anti-IgD from IgD, unlabeled rabbit antibody brane IgD can be induced; such antibodies would be useful was much more effective than unlabeled mouse anti-IgD in for studies of immune regulation. displacing labeled rabbit anti-IgD. Only 70o of labeled rabbit The authors thank Christina D. Swenson for her expertise in the anti-IgD could be displaced by the mouse antibody. cell-staining analyses. This work was supported by National Insti- tutes of Health Grants AI-24272 and AI-22068 (to A.N.) and Although these results are not conclusive, our evidence AG-04980 (to G.J.T.). suggests that mouse anti-IgD does not react effectively with IgD on mouse cell surfaces. Supporting data include the 1. Haba, S. & Nisonoff, A. (1987) Proc. Natl. Acad. Sci. USA 84, 5009-5013. following: (i) failure of excess mouse anti-IgD to prevent 2. Haba, S. & Nisonoff, A. (1987) J. Immunol. Methods 105, cell-surface staining by goat or rabbit anti-mouse IgD; (ii) 193-199. poor direct binding of mouse anti-IgD to cell surfaces; and 3. Pope, R. M., Keightley, R. & McDuffy, S. (1982) J. Immunol. (iiM) very weak inhibition by excess mouse anti-IgD of the 128, 1860-1863. binding of radiolabeled, affinity-purified rabbit anti-IgD to 4. Layton, J. E., Johnson, G. R., Scott, D. W. & Nossal, G. J. V. normal mouse spleen cells (data not shown). (iv) In addition, (1978) Eur. J. Immunol. 8, 325-330. that reacted with on B 5. Baine, Y., Chen, Y.-W., Jacobson, E. B., Pernis, B., Siskind, circulating anti-IgD IgD cells should G. W. & Thorbecke, G. J. (1982) Eur. J. Immunol. 12, 882-886. have increased the surface IgG level. However, no signifi- 6. Metcalf, E. J., Mond, J. J., Scher, I., LaVeck, M. A. & Finkel- cant surface IgG above the background of normal mice was man, F. D. (1982) Ann. N. Y. Acad. Sci. 399, 351-359. detected on B cells from anti-IgD-producing mice (data not 7. Xue, B. R., Coico, R. F., Wallace, D., Siskind, G. W., Pernis, shown). This, again, supports the interpretation that the B. & Thorbecke, G. J. (1984) J. Exp. Med. 159, 103-113. syngeneic anti-IgD reacts better with secreted than with 8. Xue, B., Hirano, T., Pernis, B., Ovary, Z. & Thorbecke, G. J. cell-surface IgD. (1984) Eur. J. Immunol. 14, 81-86. 9. These results are consistent with the possibility that our Coico, R. F., Xue, B., Wallace, D., Siskind, G. W. & Thor- becke, G. J. (1985) J. Exp. Med. 162, 1852-1861. rabbit anti-IgD reacts predominantly with epitopes on IgD 10. Wathen, P. M. & Manning, D. D. (1986) Eur. J. Immunol. 16, not recognized by mouse anti-IgD. The apparent absence of 969-974. syngeneic antibodies directed to cell-surface IgD could re- 11. Finkelman, F. D., Kessler, S. W., Mushinski, J. F. & Potter, flect tolerance of the mouse to those IgD epitopes exposed M. (1981) J. Immunol. 126, 680-687. on the surface of B cells or adsorption by B cells of such 12. Cheng, H. L., Blattner, F. R., Fitzmaurice, L., Mushinski, antibodies as they are produced. The possibility that a small J. F. & Tucker, P. W. (1982) Nature (London) 296, 410-415. percentage of high-affinity antibodies, which may be absent 13. Hunter, R. (1970) Proc. Soc. Exp. Biol. Med. 133, 989-992. 14. Daugharty, H., Hopper, J. E., MacDonald, A. B. & Nisonoff, in mouse anti-IgD, are responsible for staining cannot be A. (1969) J. Exp. Med. 130, 1047-1062. ruled out. 15. Klinman, N. R., Pickard, A. R., Sigal, N. H., Gearhart, P. J., That the number of circulating ,Lu, + B cells was almost Metcalf, E. S. & Pierce, S. K. (1976) Ann. Immunol. (Paris) normal, even in mice immunized at 2 days of age, suggests 127C, 489-502. that the circulating anti-IgD did not significantly down- 16. Sips, R. (1948) J. Chem. Phys. 16, 490-495. regulate B-cell maturation; this contrasts with previous 17. Nisonoff, A. & Pressman, D. (1957) J. Immunol. 80, 416-428. observations on mice injected with homologous allo-anti-IgD 18. Ozato, K., Mayer, N. & Sachs, D. H. (1980) J. Immunol. 124, from birth (5, 20, 21). 533-540. The results indicate that adult BALB/c are not 19. Dissanayake, S., Hay, F. C. & Roitt, I. M. (1977) Immunology mice 32, 309-318. tolerant at the B-cell level to autologous IgD, although they 20. Skelly, R. R., Baine, Y., Ahmed, A., Xue, B. & Thorbecke, may conceivably be tolerant to epitopes exposed on mem- G. J. (1983) J. Immunol. 130, 15-18. brane IgD. Tolerance at the level of T cells is suggested by 21. Jacobson, F. B., Baine, Y., Chen, Y.-W., Flotte, T., O'Neill, the greater effectiveness of KLH-IgD as an immunogen N. J., Pernis, B., Siskind, G. W., Thorbecke, G. J. & Tonda, P. compared with unconjugated IgD. Thus, concentrations of (1981) J. Exp. Med. 154, 318-332. Downloaded by guest on October 4, 2021