A unidirectional carrier effect

CELLULAR

IMMUNOLOGY

92,226-234 (1985)

A Unidirectional YOICHI

KOHNO,’

HAJIME

Carrier Effect

KAWAMURA,

AND JAY A. BERZOFSKY

Metabolism Branch, National Cancer Institute, National Institutes of Health, Bethesda, Maryland 20205 Received October 4, 1984; accepted December 6, I984 A surprising unidirectional carrier effect has been observed in the antibody response to myoglobin-ferritin conjugate. This conjugate servesas a hapten-carrier complex for myoglobinspecific T cells to help ferritin-specific B cells make anti-ferritin antibodies, but it does not function for ferritin-specific T cells to help myoglobin-specific B cells to make anti-myoglobin. Therefore, myoglobin-ferritin does not bypass the Ir gene defect of low responders to myoglobin. In contrast, myoglobin-fowl -y-globulin does induce anti-myoglobin antibodies in low responder mice and thus bypasses the Ir gene defect. Both complexes are covalently coupled. Since the myoglobin-ferritin conjugate serves for myoglobin-specific T cells to help myoglobin-specific B cells, the myoglobin in the conjugate is not altered in a way that would prevent recognition by myoglobin-specific B cells. Similarly, the conjugate serves for ferritinspecific helper T cells to help ferritin-specific B cells, so it can be recognized functionally by fenitin-specific T helper cells. Explanations such as unidirectional induction of or sensitivity to bystander help, or T-cell suppression, have been excluded. While the explanation for this unexpected observation is not yet certain, several possibilities are discussed to explain this novel phenomenon, which is believed to be the first example of such a unidirectional carrier effect between tW0 proteins. Q 1985 Academic Prw Inc. INTRODUCTION

Ever since the discovery of T cell-B cell collaboration in the generation of an antibody response (l-6), it has been widely believed that antigen can serve as a “bridge” to link an antigen (or carrier)-specific helper T cell to an antigen (or hapten)-specific B cell. In the caseof hapten-carrier complexes, with multiple copies of the hapten on one carrier molecule, one cannot ascertain which copy of the hapten is bound by the B cell which is activated, and where this hapten is located in relation to the site on the carrier bound by the T cell. However, in the case of monomeric natural protein antigens, each epitope (whether serving as hapten or carrier) is unique and occurs in a unique spatial relation to every other epitope. Previous studies in other Zr gene-controlled systemshad shown that low responders could be immunized by using the antigen attached to an immunogenic carrier (7, 8). In the course of attempts to induce low responder B cells immune to sperm whale myoglobin by immunizing with myoglobulin attached to ferritin as a carrier, we made an unexpected observation. Ferritin, although immunogenic in these mice, failed to act as a carrier for myoglobin. In contrast, fowl y-globulin when coupled ’ Present address: Department of Pediatrics, School of Medicine, Chiba University, Chiba, Japan.

0008-8749185$3.00 CopyrigJN8 1985 by Academic PITS, Inc. All rights of repmdwtion in any form revved.

UNIDIRECTIONAL

CARRIER EFFECT

227

to myoglobin was an effective carrier. Exploring this observation, we found that myoglobin could act as a carrier for a ferritin-specific response even though the fenitin was not effective as a carrier for myoglobin. We have excluded several obvious trivial explanations for this unidirectional carrier effect. We are not aware that this phenomenon, or the related observation of a carrier which works for only some haptens and not others, has ever been reported previously. MATERIALS

AND METHODS

Animals. BlO.D2/nSn, BlO.BR/SgSn, and B6D2Fr mice were purchased from Jackson Laboratories, Bar Harbor, Maine, and were 8 to 20 weeks of age at the time of immunization. Antigens.Sperm whale myoglobin (Mb)* was obtained from the Accurate Chemical and Scientific Company (Hicksville, N.Y.). The major chromatographic component, IV, purified as described previously (9) by the method of Hapner et al. (lo), was used throughout these studies. Fowl y-globulin (chicken) (FyG) from United States Biochemical Corp. (Cleveland, Ohio) was used. Horse fenitin (Fer) was a six-times crystallized preparation from Miles Laboratories (Pentex 96-027-2, Lot 16). Rabbit antisera to horse Fer were obtained from Miles Laboratories (Elkhart, Ind.; Code 65- 122, Lot R75). The immunoglobulin fraction was coupled to cyanogen-bromideactivated Sepharose4B (Pharmacia) according to the manufacturer’s instructions. Mb-coupled F’TG or Fer. Preparation of Mb-coupled FiG (Mb-FyG) or Fer was performed by a modification of the method described by Schroer et al. (11) and the preparation of Mb-FyG was described in detail in a previous paper (12). Mbcoupled Fer (Mb-Fer) was prepared similarly as follows: m-maleimidobenzoyl-Nhydroxysuccinimide ester (MBS; Pierce Chemical Co., Rockford, Ill.; 1.2 mg) in 0.06 ml of dimethyl formamide (DMF; Pierce Chemical Co.) was reacted with 0.5 pmol of Mb in 0.4 ml of 0.1 M sodium phosphate buffer (NaPi), pH 7.1, for 1 hr at 20°C and the product dialyzed against 0.05 M NaPi, pH 7.1, for 1 hr and 30 min. Fer (26 nmol) was reacted with 0.31 mg of methyl-4-mercaptobutyrimidate hydrochloride (MBI; Pierce Chemical Co.) in 0.57 ml of 0.05 it4 NaPi, pH 7.1, with 0.1 mM Drdithiothreitol (DTT) (Sigma Chemical Co., St. Louis, MO.) for 1 hr at 20°C. The MBI-Fer solution was dialyzed against 0.05 it4 NaPi, pH 7.1, with 0.01 mM DTT for 1 hr and 30 min under nitrogen. After centrifugation of both, the MBI-Fer and MBS-Mb were mixed and allowed to react for 3 hr at 20°C under nitrogen with occasional stirring. The Mb-Fer solution was kept in 0.05 M NaPi, pH 7.2, with 0.4 mM DTT, overnight in order to block residual active sites of maleimide on the MBS and was purified by gel filtration on Sephadex G-75 (Pharmacia Fine Chemicals, Piscataway, N.J.). Although Fer (M, N 700,000) and FyG (M, N 150,000) differ in size, the molar substitution of Mb per 100,000 MW of carrier was comparable for both complexes, namely 1.24 for the Mb-Fer complex and 1.0 for the Mb-FyG complex. These concentrations were determined from the absorbance at 280 and 410 nm at pH 7.2 using extinction coefficients for Fer at 1 ’ Abbreviations used: BlO, C57BL/lOSn; CFA, complete Freund’s adjuvant; DMF, dimethyl foxmamide; DTT, DLdithiothreito~ Fer, fenitin; FyG, fowl y-globulin; GAT, a random terpolymer of 60% Glu, 30% Ala, and 10% Tyr; Mb, sperm whale myoglobin; MBI, methyl4mercaptobu~midate hydrochloride; MBS, m-maleimidobenzoyl-N-hydroxysuccinimide este.r,Pi, inorganic phosphate; PBS, phosphate buffered saline; RaMB, rabbit anti-mouse brain associatedantigen.

228

KOHNO,

KAWAMURA,

AND

BERZOFSKY

mg/ml of A2s0= 10.44, A4rO= 2.13 and for Mb at 1 mi!4 of A280= 26.8 and AdI0 = 140. For Mb-FyG, see reference (12). The concentrations (&ml) specified in the cultures are those of the Fer or FrG moiety. The covalent linkage of Mb to Fer was confirmed by the nearly complete binding of the Mb in the complex to an antiFer Sepharosecolumn (data not shown). Immunization schedule. Mice were immunized intraperitoneally (ip) with 150 pg of purified Mb or with 200 pg of Fer, FyG, Mb-F-yG, or Mb-Fer in PBS emulsified 1:l in complete Freund’s adjuvant (CFA) (H37Ra, Difco, Detroit, Mich.) in a total volume of 0.1 ml per animal. Three weeks after immunization, the mice were boosted ip with 0.1 pg of purified Mb, Fer, FyG, Mb-FyG, or Mb-Fer in PBS and they were sacrificed 1 to 2 weeks later. Preparation of T cells. T cells were purified by passage over nylon wool and irradiated with 250 R from a 13’Cssource to eliminate memory B cells, as described (12, 13). Preparation of B cells and accessory cells. B cells and accessorycells were spleen cells treated with rabbit anti-mouse brain associated antigen and complement as described ( 14). Cell cultures. The culture system, modified from that of Mishell and Dutton (15), was described in a previous paper (14). Briefly, 2.5 X lo6 spleen cells from immunized mice were cultured with a concentration of antigen found to be optimal, based on dose-response studies for each antigen (not shown), namely 1 or 0.1 Z.&ml of Mb, 0.01 or 0.001 &ml of Mb-FyG or FyG, or 0.001 pg or 0.0 1 pg/ml of Mb-Fer or Fer in 1.5 ml of medium as described (14) for 10 days at 37°C 6% CO2 on a rocking platform. On the 4th day, 1 ml of supernatant was exchanged for fresh medium. On the 10th day, culture supernatants were harvested in order to measure secreted antibodies. Radioimmunoassay for antibodies to Mb, FyG, or Fer. The assay of antibodies to Mb in the culture supernatants was described in detail elsewhere (14). Briefly, polyvinyl chloride 96-well microtiter plates were coated with 200 &ml of Mb and remaining sites on the plastic blocked with bovine serum albumin. After washing, the coated wells were incubated with 50 ~1 of test sample (culture supernatant fluid) at room temperature for 2 hr, washed, and then incubated with 50 ~1 of 3H-labeled affinity-purified goat anti-mouse Fab (a kind gift of Dr. Thomas Chused, NIH). After washing, the wells were cut out and the 3H bound was determined by scintillation counting. The amount of antibody to FyG or Fer was measured similarly by a solid-phase radioimmunoassay technique using the plates coated with FyG (200 &ml) or Fer (200 pg/ml) in PBS. RESULTS Fer as a carrier does not overcome H-d-linked low responsivenessto Mb. Many previous studies of MHC-linked Zr gene control showed that the Zr defect could be bypassed by attaching the antigen (which is not itself immunogenic in the Zr low responder) to an immunogenic carrier (7, 8, 16-18). Presumably, this maneuver allows carrier-specific helper T cells to substitute for helper cells specific for the antigen under Zr gene control, which are not activated or do not function with syngeneic B cells in the low responder (19). We wanted to take advantage of this technique to insure priming of B cells to Mb in the low responder (12, 19). The

UNIDIRECTIONAL

229

CARRIER EFFECT

first carrier we tried was Fer, which has a molecular weight of about 700,000 and bound about 1.2 Mb molecules (MW 17,800) per lo5 kDa (see Methods). Despite the fact that spleen cells from both high responder BlO.D2 mice and low responder BlO.BR mice immunized with Mb-Fer made similar levels of antibodies to Fer, and made at least as much as when immunized with Fer alone (right panel of Fig. l), the same BlO.BR low responder spleen cells from mice immunized with Mb-Fer failed to make antibodies to Mb just as they failed to respond to Mb alone (left panel of Fig. 1). Controls immunized with Mb alone or Fer alone demonstrated Zr gene control of the response to Mb but not to Fer. Thus, mice which were low responders to Mb had adequate Fer-specific helper activity, but those helper cells did not appear to help a response to Mb in the presence of Mb-Fer. In contrast, FyG as carrier did overcome the low responsivenessof either B 1O.BR or B 10 mice to Mb (Fig. 2 and reference (19)). Thus, both strains which were low responders to Mb had Mb-specific B cells which could be helped by appropriate carrier-specific helper T cells. In T cell-B cell mixing experiments using a strain which responds to both antigens, Mb acts as a carrier for Fer but Fer does not act as a carrier for Mb. We

tested the functional carrier activity of the Mb-Fer complex using T and B cells from a strain, B6D2F1, which was a high responder to both antigens (Fig. 3). Mice were immunized with Mb alone or Fer alone and their splenic T and B cells prepared and cultured with either the Mb-Fer complex or a mixture of unconjugated Mb and Fer. Syngeneic mixes of T cells and B cells immune to the same or different antigens were cocultured at a ratio of 1:2 in the presence of Mb plus Fer or the Mb-Fer complex. Culture supernatants were tested for antibodies binding to Mb (Fig. 3A) or to Fer (Fig. 3B). Groups 1, 5, 10, and 14 were controls to show that IMMUNIZED _

WI ANTIBODIESTO

MYOGLOBIN

ANTIBODIES

0

AND CULTURED

WITH:

MYOGLOBIN-FERRITIN MYOGLOBIN

D2

BR

TO FERRITIN

FIG. 1. Immunization with Mb-Fer does not overcome the Ir gene control of the response to Mb. BlO.D2 and BlO.BR mice were immunized with Mb or Mb-Fer and cells from these strains were cultured with Mb (1 p&ml), Fer (0.001 &ml), or Mb-Fer (0.00 1 p&ml). These antigen concentrations have been found to be optimal in dose-response studies. The antibodies specific for Mb or Fer in the culture supematants were measured by solid-phase radiobinding immunoassay. The antibody concentrations are expressed as mean Acpm + SEM. D2, BlO.D2/nSn; BR, BlO.BR/SgSn; immunized and cultured with Mb-Fer, n ; Mb, 0; Fer, q .

230

KOHNO, KAWAMURA,

BR

AND BERZOFSKY

0

CULTURED

WITH

Mb,l~g/ml

m

CULTURED

WITH

Mb-FyG,

0.01 pglml

CULTURED

WITH

Mb-FyG,

0.001 pgl

MICE

IMMUNIZED

D2

BR

SPLEEN

CELLS

02 FROM

610

D2

WITH

Mb-FIG

El0

FIG. 2. Immunization with Mb-FyG does overcome the Zr gene control of the responseto Mb. BlO.D2, C57BL/lO, or BlO.BR mice were immunized with Mb-FyG and spleen cells from these mice were cultured with Mb (1 &ml, Cl), or Mb-FyG (0.01 &ml, m, or 0.001 &ml, 0). These antigen concentrations have been found to be optimal in dose-response studies. The antibodies specific for Mb or FyG in the culture supematants were measured by solid-phase radiobinding immunoassay. The antibody concentrations are expressed as mean Acpm + SEM. D2, BlO.DZ/nSn; BR, BIO.BR/SgSn. The difference between the BlO.D2 and BlO response to Mb-FyG in the two bars at the far right is not a consistent finding. Both responsesare quite high, and are usually comparable.

the B cells were adequately depleted of T-cell help. Similarly, groups 2, 6, 11, and 15 were controls to show that the T-cell preparations were depleted of B cells and could not make antibodies alone. The Mb-primed B cells were helped by Mb-primed T cells (groups 3 and 7), but failed to be helped by Fer-primed T cells when cultured with either the Mb-Fer complex (group 8) or the mixture of the two antigens (group 4). This result confirmed our interpretation of the experiment in Fig. 1 that Fer did not act as a carrier for Mb even in the high responder strains. The Mb-Fer complex stimulated significant anti-Mb production in the presence of Mb-specific helper T cells (group 7), although less than that stimulated by Mb (group 3), probably because of the lower concentrations of Mb in the complex. Therefore the Mb sites on the complex were still recognizable by Mb-specific B cells. Moreover, the helper-T-cell sites on Fer had not been destroyed or blocked by coupling Mb to it, as the Mb-Fer complex could serve in vitro to stimulate just as much anti-Fer antibody production by Fer-immune T and B cells (group 16) as could native Fer in vitro (group 12). Surprisingly, however, Mb-specific T cells in the presence of Mb-Fer complex helped Fer-specific B cells make almost as much anti-Fer (group 17) as did Fer-specific T cells (group 16). Thus, Mb (MW 17,800), which was intended to act as the “hapten,” in fact functions as a carrier for Fer. This help is not due to a nonspecific bystander effect, since the same combination of cells, when cultured with a mixture of Mb plus Fer instead of the complex, did not result in significantly greater antibody production than the purified B-cell control (compare group 13 with group 10, and group 4 with group 1). This result proves functionally that the Mb-Fer complex is covalently coupled in a way that it

UNIDIRECTIONAL

231

CARRIER EFFECT

Cells in Culture Group Number

C.&s I-

1

-

5

-

B Cells Immune

ANTI-MYOGLOBIN

Mb

0

1000

CONCENTRATION,

“g/ml

Cultured

With Mb + Fer

Cultured

With Mb-Fer

Conjugate

2ooo

ANTI-FERRITIN CONCENTRATION (Acpm in f7adkimmunoassayl

FIG. 3. Unidirectional carrier effect of Mb-Fer. B6D2Fi mice were immunized with Mb or Fer. T cells were prepared by collection of nylon-nonadherent cells and were irradiated with 250 R from 13’Cssource in order to eliminate memory B cells. B cells were prepared by pretreatment with RaMB, monoclonal anti-Thy I .2 and monoclonal anti-Lyt 1.2 and C. The indicated mixtures of T cells and B cells were cultured at the ratio of 1:2.5 (1 X 106 T cells and 2.5 X 106 B cells per well) either with Mb plus Fer (1 &ml and 0.0 I &ml; 0) or with Mb-Fer (0.01 &ml; n), respectively (optimal concentrations for each, based on dose-response studies). In groups 9 and 18, a mixture of the T cells from mice immuniz.ed with Mb and those from mice immunized with Fer were used, each at 1 X 106/well. The antibodies specific for Mb (A) or Fer (B) in the culture supernatants were measured by solid-phase radiobinding immunoassay. The antibody concentrations are expressed as means of triplicate cultures + SEM in ng/ ml for anti-Mb or in Acpm for anti-Fer for which no adequate standard was available for conversion to rig/ml.

can serve as a hapten-carrier complex. More importantly, it also demonstrates a surprising unidirectional carrier effect (i.e., Mb acts as a carrier for Fer, but Fer does not act as a carrier for Mb), a paradox for which we have no completely satisfying explanation. The unidirectionality is not due to induction of suppression by the Mb-Fer complex. First, the mice used in the experiment of Fig. 3 were all immunized with Mb or Fer alone, not the complex. Second, addition of Fer-immune T cells, which failed to help, to the mixture of Mb-immune T and B cells in the presence of the Mb-Fer complex did not suppress the anti-Mb response (compare groups 7 and 9). Thus, the failure of the Fer-immune T cells to help an anti-Mb response in the presence of the complex was not due to their contamination with suppressor cells that could suppress Mb-specific B cells either nonspecifically or via a bridge of Mb-

232

KOHNO,

KAWAMURA,

AND

BERZOFSKY

Fer complex. Also, in the case of the anti-Fer response, as expected, no suppression was detected when both types of T cells were mixed with Fer-specific B cells and the complex (group 18). DISCUSSION This report describes a paradoxical unidirectional carrier effect which we encountered unexpectedly in the course of trying to bypass the H-24inked Ir gene-controlled low responsiveness of BlO.BR and BlO mice to sperm whale Mb. Despite the fact that Fer was immunogenic in these mice and was covalently bound to the Mb, as confirmed by affinity chromatography on an anti-Fer column, it did not function as a carrier for Mb. In pursuing this observation in a strain which responded well in vivo to both antigens, we found that in T cell-B cell mixing experiments in vitro, the Mb-Fer complex was functionally active as a hapten-carrier complex, but unidirectionally so, only in the direction opposite from that intended. Thus, Mb (MW 17,800) acted as an effective carrier for Fer (MW approximately 700,000), but not vice versa. Fer-specific helper T cells could recognize sites on the Mb-Fer complex, since Mb-Fer in culture induced Fer-specific antibodies produced by Fer-immune T and B cells. We are not aware of any previous report of a situation in which an immunogenic protein carrier served to mediate help for some haptenic determinants attached to it but not others. We considered several possible mechanisms to explain our results. First, if the help we observed were noncognate, or bystander, help (that is, help due to activation of T cells by antigen to secrete lymphokines which induce antibody production nonspecifically, without antigen bridging), then the unidirectionality of help could result from either (1) a hypothetical greater sensitivity of Fer-specific B cells, compared to Mb-specific B cells, to respond to such bystander help, or (2) more effective production of bystander help by Mb-specific T cells than by Fer-specific T cells. These explanations were excluded by control parallel cultures with Mb plus Fer. Therefore, we believe we are dealing with a unidirectional carrier effect, not a unidirectional bystander effect. Second, the apparent unidirectional carrier effect could have been due to suppressor T cells among the Fer-immune T-cell population, even though they did not prevent the Fer-specific helper cells from helping an anti-Fer response. To test for this possibility, we mixed the same Fer-immune T-cell population with a culture of Mb-immune T cells and B cells in the presenceof Mb-Fer complex. The response was at least as great as that in the absence of the Fer-immune T cells. This experiment rules out bridging between a putative Fer-immune suppressor T cell and either the Mb-specific B cell or the Mb-specific helper T cell via the Mb-Fer complex, or induction of suppression by the complex in culture. Also, the equal response of Fer-specific B and T cells in the presence of the Mb-Fer complex or in the presence of a mixture of free Mb and free Fer rules out the possibility that the Mb-Fer complex induces suppressor cells which act on the Fer-specific helper cell. Thus, we could not account for the failure of the Fer-immune T cells to help Mb-immune B cells by the presence of suppressor cells in their midst. Third, the possibility that carrier determinants on the Fer molecule were damaged by coupling of Mb was excluded by the equal efficacy of the Mb-Fer complex and free Fer to induce anti-Fer antibodies in a culture of Fer-immune T and B cells.

UNIDIRECTIONAL

CARRIER EFFECT

233

For parallel reasons, the Mb in the complex must have been immunochemically intact as well. A fourth explanation that remains a possibility to explain this paradoxical unidirectional carrier effect is a steric restriction on the possible sites which can be bound concurrently by a helper T cell and a B cell to form a functional antigen bridge between them (20-25). The Fer-specific T cells bind the Mb-Fer complex at a site (or sites) which allows help for B-cell recognition of a Fer site (to produce Fer-specific antibodies), but not for B-cell recognition of any of the Mb sites (to produce Mb-specific antibodies). In contrast, the Mb molecules are attached in such a way as to favor the ability of T cells binding Mb to help B cells binding one or more of the Fer sites. Without knowing the exact structure of Fer and the sites thereon to which Mb is bound and to which antibodies are made, we cannot define more precisely the nature of the steric restrictions on this antigen bridging. Fong et al. (26) and Chen et al. (27) have studied the linear distance constraints on antigen bridging using a hapten, dinitrophenyl (DNP), separated from a small carrier, L-tyrosine-p-azobenzene-arsonate (RAT), by a semirigid poly-L-proline spacer of varying lengths. The maximum spacer length tested which was compatible with help was 22 proline residues (69 A), whereas a spacer of 31 proline residues (97 A) did not allow hapten-carrier help. On the other hand, a separation of less than 8 A was compatible with help (28). These results from a chemically better defined system give us a good idea of the constraints on linear distance between hapten and carrier. However, no equivalent studies have yet been possible to study constraints on haptens separated from carrier determinants around the surface of large globular protein antigens. These constraints may be different, especially for carriers the size of Fer (MW approx. 700,000), even if the hapten-carrier complex remains intact throughout the T-B helper interaction. Moreover, if the carrier undergoes proteolytic cleavage in the B cell during “processing” which might be required for T-cell recognition (29-3 l), then requirements for the hapten to be close to the appropriate carrier determinant may be even greater (20). For instance, processing by B cells, in contrast to macrophages, may be selective because the B cell takes up and processesantigen bound to surface immunoglobulin preferentially (32, 33), rather than free antigen; and the region of the carrier near the hapten sterically protected by the B cell’s immunoglobulin-combining site may be preferentially reexpressed on the surface in association with Ia (20). In this case, the constraints may result from selective processing rather than the geometry of a true antigen bridge, but the final outcome would be the same. This explanation would also be compatible with the parallel control of antibody and T-cell fine specificity by Ir genes for GAT (34), for staphylococcal nuclease (35, 36), and for Mb (14, 37), although such parallel control was not observed for insulin (38). It would also be compatible with the hypothesis of Benacerraf (39) to explain such observations, namely that if helper T cells must recognize on B cells the same complex of antigen and Ia involved in priming, then “the requirement for B-cell and T-cell determinants on the same fragment for adequate helper activity will result in selective activation of those B-cell clones that bound antigen or antigen fragments in a manner that permits specific interactions with their Ia molecules.” Why was such a unidirectional carrier effect not observed previously in more conventional hapten-carrier systems?One possibility is that so many copies of the hapten moiety are usually attached to the carrier, that even if some of these hapten moieties are in sterically unfavorable locations relative to the sites bound by carrier-

234

KOHNO, KAWAMURA,

AND BERZOFSKY

specific T cells, these would go undetected becauseother copies of the hapten would be available for activation of the same B cells. However, this remains a speculative explanation of these novel paradoxes of a unidirectional carrier effect and a carrier which does not work for certain haptens. ACKNOWLEDGMENTS We thank Dr. ha Green, Dr. Richard Hodes, Dr. David Sachs, Dr. Ronald Schwartz, Dr. Gene Shearer, and Dr. Alfred Singer for critical reading of the manuscript, and Dr. ha Berkower, Dr. Howard Streicher, Dr. Thomas Waldmann, and Ms. Gail Buckenmeyer for valuable discussion during the course of the work.

REFERENCES I. Claman, H. N., Chaperon, E. A., and Triplet& R. F., Proc. Sot. Exp. Biol. Med. 122, 1167, 1966. 2. Miller, J. F. A. P., and Mitchell, G. F., Nature (London) 216, 659, 1969. 3. Davies, A. J. S., Leuchars, E., Wallis, V., Marchant, R., and Elliott, E. V., Transplantation 5, 222, 1967. 4. Mitchison, N. A., Eur. J. Immunol. 1, 18, 197 I. 5. W, M. C., Nature (London) 226, 1257, 1970. 6. Rajewsky, K., Schimnacher, V., Nase, S., and Jeme, N. K., J. Exp. Med. 129, 1131, 1969. 7. Green, I., Paul, W. E., and Benacermf, B., J. Exp. Med. 123, 859, 1966. 8. McDevitt, H. O., J. Zmmunol. 100,485, 1968. 9. Berzofsky, J. A., J. Immunol. 120,360, 1978. 10. Hapner, K. D., Bradshaw, R. A., HartzelI, C. R., and Gurd, F. R. N., J. Biol. Chem. 243,683, 1968. 11. Schroer, J. A., Inman, J. K., Thomas, J. W., and Rosenthal, A. S., J. Zmmunol. 123,670, 1979. 12. Kohno, Y., and Betzofsky, J. A., J. Exp. Med. 156,79 1, 1982. 13. Julius, M. H., Simpson, E., and Herzenberg, L. A., Eur. J Immunol. 3, 645, 1973. 14. Kohno, Y., and Betzofsky, J. A., J. Immunol. H&2458, 1982. 15. Mishell, R. I., and Dutton, R. W., J. Exp. Med. 126, 423, 1967. 16. Schwartz, M., Lifshitz, R., Givol, D., Mazes, E., and Haimovich, J., J. Zmmunol. 121, 421, 1978. 17. Benacerraf, B., and Germain, R. N., Immunol. Rev. 38,70, 1978. 18. Betzofsky, J. A., In “Biological Regulation and Development” (R. F. Goldberger, Ed.), Vol. II, pp. 467-594. Plenum Press,New York, 1980. 19. Kohno, Y., and Berzofsky, J. A., J. Exp. Med. 156, 1486, 1982. 20. Betzofsky, J. A., Survey Immunol. Res. 2, 223, 1983. 21. Cecka, J. M., Stratton, J. A., Miller, A., and Sercasz,E., Eur. J. Immunol. 6, 639, 1976. 22. Sercarz, E., Cecka, J. M., Kipp, D., and Miller, A., Ann. Immunol. (Paris) 128C, 599, 1977. 23. Campos-Neto, A., Levine, H., and Schlossman, S. F., J. Zmmunol. 121, 2235, 1978. 24. Campos-Neto, A., Levine, H., and Schlossman, S. F., Cell. Immunol. 69, 128, 1982. 25. Kawamura, H., Berkower, I., Glover, C., and Berzofsky, J., J. Cell. Biochem. Suppl. 8A, 211, 1984. 26. Fang, S., Nit&i, D. E., Cook, R. M., and Goodman, J. W., J. Exp. Med. 148, 817, 1978. 27. Chen, P., Nitecki, D. E., Lewis, G. K., and Goodman, J. W., J. Exp. Med. 152, 1670, 1980. 28. Alkan, S. S., Williams, E. B., Nitecki, D. E., and Goodman, J. W., J. Exp. Med. 135, 1228, 1972. 29. Ziegler, H. K., and Unanue, E. R., Proc. Natl. Acad. Sci. USA 79, 175, 1982. 30. Shimonkevitz, R., Kappler, J., Marrack, P., and Grey, H., J. Exp. Med. 158, 303, 1983. 31. Stmicher, H. Z., Berkower, I. J., Busch, M., Gurd, F. R. N., and Berzofsky, J. A., Proc. Nat/. Acad. Sci. USA 81, 683 I, 1984. 32. Kakiuchi, T., Chesnut, R. W., and Grey, H. M., J. Zmmunol. 131, 109, 1983. 33. Malynn, B. A., and Wortis, H. H., J. Immunol. 132,2253, 1984. 34. Bluestein, H. G., Green, I., Maurer, P. H., and Benacerraf, B., J. Exp. Med. 135,98, 1972. 35. Betzofsky, J. A., Schechter, A. N., Shearer, G. M., and Sachs,D. H., J. Exp. Med. 145, 123, 1977. 36. Schwartz, R. H., Berzofsky, J. A., Horton, C. L., Schechter, A. N., and Sachs, D. H., J. Zmmunol. 120, 1741, 1978.

37. Betzofsky, J. A., Richman, L. K., and Killion, D. J., Proc. Natl. Acad. Sci. USA 76,4046, 1979. 38. Barcinski, M. A., and Rosenthal, A. S., J. Exp. Med. 145,726, 1977. 39. Benacmaf, B., J. Immunol. 120, 1809, 1978.