Proc. Natl. Acad. Sci. USA Vol. 82, pp. 4217-4220, June 1985 induces chronic idiotype suppression (network regulation/regulatory /immunological memory) TAKESHI TOKUHISA AND KLAUS RAJEWSKY Institute for Genetics, University of Cologne, Weyertal 121, D-5000 Cologne 41, Federal Republic of Germany Communicated by Niels K. Jerne, February 11, 1985

ABSTRACT coupled to isologous gamma globulin . The following mouse hybridoma antibodies can substitute for anti-idiotypic in the induction of and their purification were used and have been described (6); neonatal chronic idiotype suppression. The immunologically B1-8 (IgMX1), P8-86.9 (IgGll), X-63 (IgGlK), Ac38 (IgGlK), suppressed animals fail to express an idiotypically defined and Ac146 (IgG2aK). The F(ab)'2 fragment of antibody Ac38 subset of anti-hapten antibodies in later life and harbor was prepared by peptic proteolysis and fractionation on regulatory T cells that prevent the expression of this same Sephacryl S-200 (Pharmacia) (7). The preparation of goat antibody subset by B cells. Suppression ceases after several anti-mouse IgG1 antibodies has been described (6). months, but memory of suppression persists and can be Neonatal Induction of Suppression and Immunization Pro- recalled by small doses ofhapten-carrier conjugate. These data tocol. One hundred micrograms of Ac38 or NP-conjugated show that an antigen present in a mouse at birth is able to mouse gamma globulin in 50 1.l phosphate-buffered saline generate a T-cell-dependent suppressive mechanism that con- (pH 7.2) was injected intraperitoneally as sterile solution into trols expression of antigen-specific antibodies through the newborn C57BL/6 mice within 24 hr after birth. For im- recognition of antibody idiotypes. The idiotypic network is thus munization, 100 ,ug of alum-precipitated NP-conjugated chick- involved in the control of tolerance and the available antibody en gamma globulin was injected intraperitoneally. The mice repertoire. were bled 12 days later, and levels of specific antibodies in the sera were measured by radioimmunoassay as described (6). The introduction of anti-idiotypic antibodies into an animal Idiotypically Defined Subsets of Al-Chain-Bearing Anti-NP commonly leads to the suppression of antibodies bearing the Antibodies and Their Measurement. Various idiotypically target idiotype. In idiotypically suppressed animals, regula- defined subsets of humoral anti-NP antibodies have been tory T cells have been observed that prevent idiotype identified on the basis of absorption studies (8). (Note in this expression by B cells in the absence of anti-idiotypic anti- context that we give an anti-idiotope antibody the same body (1). T-cell-mediated suppression ofthis type can be seen designation as the idiotope that it recognizes.) Subset 1 as a general mechanism by which the idiotypic network (2) antibodies are recognized by two anti-idiotope antibodies, controls immunological tolerance: the interaction of an an- Ac38 and Ac146 (Ac38+Ac146+). Antibodies of subset 2, on tigen receptor (e.g., an antibody on the surface) with the other hand, express idiotope Ac38 but not idiotope Ac146 a ligand (e.g., the anti-idiotypic antibody) results in the (Ac38+Ac146-). The concentration of antibodies of the two generation of regulatory T cells that suppress receptor subsets was determined as described (6). Briefly, individual expression through the recognition of the idiotype of the sera were incubated with Sepharose coupled with antibody receptor (3). In this picture, any macromolecule should be Ac38 or antibody X-63 for control. The absorbed sera were able to induce T-cell-mediated idiotypic suppression of the then applied to plastic plates coated with antibodies Ac38 or receptors specific for it. Indeed, idiotype-specific regulatory Ac146 and bound antibodies were determined with a radio- T cells have also been found in animals pretreated with labeled goat anti-IgGi antibody. antigen instead of anti-idiotypic antibody (1, 4). Here we Adoptive Transfer System. The system has been described show that chronic T-cell-controlled idiotypic suppression can (6). Briefly, T cells were purified from spleen cells over a be induced in newborn mice by antigen as well as by nylon wool column. For B-cell isolation, spleen cells were anti-idiotypic antibody. The animals eventually recover from treated with monoclonal anti-Thy-1.2 (New England Nu- suppression but suppression can be recalled by minute clear) and anti-Lyt-2.2 (gift of F. W. Fitch, Chicago) and amounts of antigen irrespective of whether it had been rabbit complement (Cedarlane Laboratories, Homby, ON, induced at birth by antigen or anti-idiotypic antibody. Canada). Purified B and T cells (10 x 106 each per host) were mixed and transferred into syngeneic irradiated (500 R) recipients. The recipients were immunized intraperitoneally MATERIALS AND METHODS with 100 ug of alum-precipitated NP-conjugated chicken Animals. C57BL/6 mice were bred in our animal facility. gamma globulin. . Chicken gamma globulin was purified from whole serum obtained from Behringwerke (Marburg/Lahn, RESULTS F.R.G.), by ammonium sulfate precipitation. Mouse gamma globulin was prepared from normal C57BL/6 mouse serum The Experimental System. The primary anti-NP response of by chromatography on DE52 (Whatman, Maidstone, U.K.). C57BL/6 mice consists of a large family of structurally and Bovine serum albumin was purchased from Behringwerke. genetically related A1-chain-bearing antibodies (5). Within (4-Hydroxy-3-nitrophenyl)acetyl (NP)-conjugated chicken this family, subsets of antibodies can be defined on the basis and mouse gamma globulins and (4-hydroxy-3-iodo- of their idiotypic phenotype as analyzed by monoclonal 5-nitrophenyl)acetyl-conjugated bovine serum albumin were anti-idiotope antibodies (8). Subset 1 antibodies are charac- prepared as described (5). terized by co-expression of the idiotopes Ac38 and Ac146, recognized by the anti-idiotope antibodies Ac38 and Ac146. The publication costs of this article were defrayed in part by page charge Subset 2 antibodies carry the Ac38, but not the Ac146 payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact. Abbreviation: NP, (4-hydroxy-3-nitrophenyl)acetyl. 4217 Downloaded by guest on September 23, 2021 4218 Immunology: Tokuhisa and Rajewsky Proc. Natl. Acad. Sci. USA 82 (1985) idiotope, and occur in similar frequency as antibodies of (Fig. 2). In this situation only short-term suppression is subset 1 in primary anti-NP sera. The two subsets do not observed and subset 1 recovers from suppression earlier than differ from each other in their affinity to the NP hapten and most other anti-NP antibodies including subset 2. This result the anti-idiotope Ac38 (6). is in agreement with our previous data (6), which indicated Antigen and Anti-Idiotope Antibody Induce Chronic Sup- that subset 1 antibodies do not have a higher affinity for the pression of the Same Antibody Subset. We have previously hapten than antibodies of subset 2. shown that injection of the anti-idiotope antibody Ac38 into Antigen and Anti-Idiotope Antibody Induce Regulatory T newborn mice results in acute suppression of both subset 1 Cells Suppressing the Same Antibody Subset. We have previ- and subset 2 anti-NP antibodies. Acute suppression of the ously demonstrated that the spleens of mice treated at birth two subsets is followed by chronic suppression of subset 1 with anti-idiotope antibody harbor idiotype-specific regula- only (6). The results of experiments in which newborn tory T cells. These T cells do not allow the expression of C57BL/6 mice were given either 1 ,ug ofantibody Ac38 or 100 subset 1 anti-NP antibodies by normal B cells (6). The data ,tg of NP-conjugated mouse gamma globulin intraperitone- in Fig. 3 show that, similarly, T cells taken from 14-week-old ally and groups of the animals were immunized with NP- mice injected at birth with 100 ,ug of NP-conjugated mouse conjugated chicken gamma globulin at 6, 8, 9, 12, 14, 16, and gamma globulin selectively prevent the production of subset 20 weeks of age are shown in Fig. 1. The mice were bled 12 1 antibodies by B cells from normal animals when mixtures days after immunization and the sera were titrated for total, of T and B cells are transferred to syngeneic irradiated subset a, and subset b anti-NP antibodies. The results animals and the latter are immunized with NP-conjugated obtained in animals suppressed by anti-idiotope antibody are chicken gamma globulin. As in the case of suppression by taken from previous work (6). It is apparent from the data in anti-idiotope antibody, B cells from the suppressed mice are Fig. 1 that the recovery of subset 1 antibodies from suppres- themselves unable to produce subset 1 antibodies, when sion follows the same kinetics in the two groups of sup- combined with normal T cells and as in the previous study pressed mice. In both cases subset 2 antibodies recover much cell-mixing experiments fail to produce clear evidence for earlier. In the animals injected with NP-conjugated mouse active suppression by suppressor T cells. gamma globulin there is also short-term suppression of the Memory of Anti-Idiotope-Induced Suppression Can Be total anti-NP response. This is not found in anti-idiotope- Triggered by Antigen. At 36 weeks of age, mice injected at mediated suppression. birth with 100 ug anti-idiotope antibody or NP-conjugated Chronic suppression of subset 1 antibodies is not seen mouse gamma globulin have completely recovered from when adult mice are given 100 ,tg of NP-conjugated mouse suppression in that they produce normal levels of anti-NP gamma globulin intraperitoneally at various times before antibodies (including subsets 1 and 2) upon immunization immunization with NP-conjugated chicken gamma globulin with NP-conjugated chicken gamma globulin (Fig. 4). When such animals receive an intraperitoneal injection of 1 ,ug of NP-conjugated mouse gamma globulin 3 weeks before the immunization, the production of subset 1 antibodies is

100-

I 0!0_

C3 C~~~~~~ 10 0' .O5

0Il

Weeks of age FIG. 1. Recovery from suppression induced at birth with antigen <0.5 , or anti-idiotope antibody. C57BL/6 mice were injected at birth with 1 2 3 6 1 ,g of Ac38 or 100 ,g of NP-conjugated mouse gamma globulin and Weeks immunized with NP-conjugated chicken gamma globulin at the age of 6-20 weeks and bled 12 days after immunization. Subset b, antibod- FIG. 2. Recovery from suppression induced in adult mice with ies from mice treated with anti-idiotope (o-o) and NP-conjugated NP-conjugated mouse gamma globulin. Adult (12-week-old) mouse gamma globulin (o-o); subset a, antibodies from mice C57BL/6 mice were injected with 100 ,g of NP-conjugated mouse treated with anti-idiotope (-O) and NP-conjugated mouse gamma gamma globulin 1, 2, 3, or 6 weeks before immunization with globulin (u-u). The level of IgG1 anti-NP antibodies in NP- NP-conjugated chicken gamma globulin. Twelve days after im- conjugated mouse gamma globulin-injected mice is also shown munization, IgG1 anti-NP antibodies (o----o) and antibodies of (o---o). The experimental points represent geometric means of titers subsets a (o*-) and b (EJ-o) were determined. The experimental of 6-15 sera and are plotted as percentage of the corresponding titers points represent geometric means of titers of six sera plotted as in age-matched control mice. percentage of the mean titers in age-matched control mice. Downloaded by guest on September 23, 2021 Immunology: Tokuhisa and Rajewsky Proc. Natl. Acad. Sci. USA 82 (1985) 4219 selectively and strongly suppressed. This effect is not ob- 102o2 served in control mice that have not been injected at birth. Thus memory of neonatal idiotype suppression persists long beyond recovery from suppression and can be recalled by . 0 0 antigen irrespective of whether suppression had been in- 0 0 duced by antigen or anti-idiotype antibody. 0 0 _1u 0 DISCUSSION

0 The present data show that antigen can substitute for anti- 0 0 idiotypic antibody in the induction of neonatally induced 0 0 chronic idiotype suppression. Injection of the hapten NP 0 coupled to isologous gamma globulin at birth results in 0.,E _ 0 0 chronic suppression of a particular idiotypically defined 0 subset of anti-NP antibodies (subset 1) that has previously been shown to be the target of chronic suppression induced 0 0 0 0 by anti-idiotypic antibody. The requirement of 100 times * 0 0 more NP-conjugated mouse gamma globulin than anti-idi- 0 otope antibody to achieve equal suppression may be due to -2 different half lives of the two compounds in the body. The 1-2-o1_ restriction of chronic suppression to subset 1 antibodies cannot be explained by a particularly high affinity of this subset for the hapten (6). This is underlined by our finding 3x10~ that acute suppression induced by a large dose of NP- conjugated mouse gamma globulin in the adult affects subset 1 antibodies no more than antibodies of another subset (2) N N N S S NS N that is not the target of chronic suppression (Fig. 2). In Bcels N N S S NS addition, the anti-NP antibodies produced by animals chroni- Tcefls N S N S N NS NSNS N NS cally suppressed by either NP-conjugated mouse gamma FIG. 3. Cellular basis of neonatal suppression with NP- globulin or anti-idiotype have almost the same affinity for the conjugated mouse gamma globulin. B and T cells (1 x 107 cells of hapten as control antibodies (unpublished data). We con- each type) from 14-week-old C57BL/6 mice that had either been clude that chronic suppression of subset 1 antibodies induced given 100 ,g of NP-conjugated mouse gamma globulin at birth (S) or by NP-conjugated mouse gamma globulin or anti-idiotypic left untreated (N) were mixed in various combinations as indicated at birth is mediated by recognition ofthe idiotype of and transferred into irradiated syngeneic hosts together with alum- antibody precipitated NP-conjugated chicken gamma globulin. Subset 1 (Left) the suppressed antibody subset. We attribute this recognition and 2 (Right) anti-NP antibodies of the IgG1 class were determined to regulatory T cells that we have identified in both anti-idi- in the sera 12 days later. Each experimental point represents the otype- (6) and NP-conjugated mouse gamma globulin-sup- response of a single animal. Total anti-NP titers were indistinguish- pressed animals (Fig. 3). Whether the suppressive effect of able in all groups (data not shown). the regulatory T cells is based on active suppression or on the

Neonatal NP-MG at manipulation week 33 None

None +

Ac38

Ac38 + Z IIIffiv

NP-MG

RRRRRRARRRSF- NP-MG + -1-

100 101 102 103 IgG1, ig/ml

FIG. 4. Recall of idiotype-specific suppression by antigen. C57BL/6 mice injected at birth with 100 ,ug of Ac38 or NP-conjugated mouse gamma globulin (NP-MG) were given 1 /ig of NP-MG at 33 weeks of age or left untreated. At the age of 36 weeks the animals were immunized with alum-precipitated NP-conjugated chicken gamma globulin. Twelve days later the primary IgGl anti-NP response (_) and antibodies of subset 1 (M) and 2 (M) were measured in the sera. Results represent geometric means with standard deviation coefficients calculated from the titers of four to six sera. Downloaded by guest on September 23, 2021 4220 Immunology: Tokuhisa and Rajewsky Proc. Natl. Acad. Sci. USA 82 (1985) elimination ofidiotype specific helper T cells (or both) can be that antigen (and possibly others) in later life. Furthermore, decided from neither the previous nor the present cell-mixing the induction of neonatal idiotypic suppression, as that of experiments. The elucidation of this problem and of the fine other forms of suppression (12), generates memory that we specificity of T-cell-mediated idiotypic suppression will rely here report to persist in the animal for a good part of its life. on isolation of the regulatory cells, which has not yet been achieved. We are grateful to Drs. T. Takemori and T. Saito for helpful We think that chronic idiotype suppression is induced in discussions and suggestions and to Ms. Claudia Uthoff-Hachenberg the animal by idiotypic antibodies complexed with the for excellent technical help. This work was supported by the injected ligand-i.e., anti-idiotype or NP-conjugated mouse Deutsche Forschungsgemeinschaft through SFB 74. T.T. is a recipi- gamma globulin (3). As in the in vitro experiments of ent of a fellowship from the Alexander von Humboldt-Stiftung. Caulfield et al. (9), such complexes may induce regulatory T 1. Rajewsky, K. & Takemori, T. (1983) Annu. Rev. Immunol. 1, cells that recognize the idiotype of the complexed antibody 569-607. and maintain the suppressed state. There is no reason why 2. Jerne, N. K. (1974) Ann. d'Immunol. (Inst. Pasteur) 125C, this mechanism of chronic idiotype suppression should not 373-389. also operate at the level ofreceptors recognizing self-antigens 3. Rajewsky, K., Takemori, T. & Muller, C. E. (1984) in and thus participate in the control of self-tolerance (3). The Progress in Immunology, eds. Yamamura, Y. & Tada, T. physiological importance ofnetwork control of self-tolerance (Academic, Tokyo), Vol. 5, pp. 533-542. cannot be safely evaluated at present. In our experimental 4. Germain, R. N., Sy, S.-M., Rock, K., Dietz, M. H., Greene, system chronic suppression can be induced against only a M. I., Nisonoff, A., Weinberger, J. Z., Ju, S.-T., Dorf, M. E. small fraction of the total population of antibodies. A similar & Benacerraf, B. (1981) in Immunoglobulin Idiotypes: ICN- UCLA Symposia on Molecular and Cellular Biology, eds. restriction has been described in the case of chronic Janeway, C., Sercarz, E. E. & Wigzell, H. (Academic, New suppression (10), which is presumably based on a similar York), Vol. 20, pp. 709-723. mechanism. It is possible that these restrictions reflect the 5. Imanishi, T. & Makela, 0. (1973) Eur. J. Immunol. 3, 323-330. selection of a receptor repertoire in the regulatory cells that 6. Takemori, T. & Rajewsky, K. (1984) Eur. J. Immunol. 14, controls the expression of antibody (and possibly T-cell 656-667. receptor) variable-chain regions in a physiologically mean- 7. Nisonoff, A., Wissler, F. C., Lipman, L. N. & Woernly, D. L. ingful way, perhaps preventing auto-aggression (11). In the (1960) Arch. Biochem. Biophys. 89, 230-235. response of an animal to an artificial antigen such as NP such 8. Takemori, T. & Rajewsky, K. (1984) Immunol. Rev. 79, a control would be to affect 103-117. expected selectively certain 9. Caulfield, M. J., Luce, K. J., Proffitt, M. R. & Cerny, J. antibody subsets only, depending on their idiotypic specific- (1983) J. Exp. Med. 157, 1713-1725. ity. Be this as it may, the present data establish that an 10. Herzenberg, L. A. & Herzenberg, L. A. (1973) Contemp. Top. antigen present in the body at the time of birth is able to Immunobiol. 3, 41-75. induce a suppressive mechanism that presumably operates 11. Jerne, N. K. (1984) Immunol. Rev. 79, 5-24. through regulatory T cells, is based on idiotype recognition, 12. Loblay, R. H., St. Groth, B. F., Pritchard-Briscoe, H. & and modulates the repertoire of antibodies produced against Basten, A. (1983) J. Exp. Med. 157, 957-973. Downloaded by guest on September 23, 2021