Negative allogeneic effects in vitro. I. Allogeneic T cells markedly suppress the secondary antibody-forming cell response.

Murine T cells can mediate a potent negative allogeneic effect on the capacity of primed cells to develop the secondary antibody-forming cell response to hapten carrier in vitro. This effect is detected when T cells confront responding cells differing at the major H-2 locus. The allosuppression is relatively sensitive to mitomycin treatment and to irradiation. The T cells responsible for the inhibition of antibody formation appear to express the Ly2 but not the Ly1 alloantigen. The secondary response of spleen cells in culture is quite insensitive to positive allogeneic effects. The usefulness of this model in elucidating the mechanism of allosuppression and the relevance of such effects to studies involving genetic restriction on cell interactions is discussed.     

Alloantigen-induced T helper activity. I. Minimal genetic differences necessary to induce a positive allogeneic effect.

Addition of histoincompatible lymphocytes can influence the course of ongoing immune responses. Such allogeneic effects may either augment or diminish immune responses. We describe here the minimal genetic differences necessary to generate positive allogeneic effects (allohelp) in a humoral immune response. The antibody response to sheep erythrocytes of T cell-depleted mouse spleen cells was reconstituted by addition of syngeneic or allogeneic nylon wool column-passaged spleen T cells. T cells were pretreated with mitomycin C before culture to prevent development of allo-suppression and cytotoxic lymphocytes. Positive allogeneic effects were operationally defined as superior helper effects (to generate greater antibody forming cell responses) with T cells allogeneic rather than syngeneic to the responding B cells. Thus, addition of allogeneic T cells resulted in many more antibody forming cells than did equal numbers of syngeneic T cells, and fewer allogeneic than syngeneic T cells were necessary to generate comparable responses. With congenic, recombinant, and mutant mouse lines, genetic differences in the H-2 complex and those associated with Mls were each sufficient to provide positive allogeneic effects. With intra-H-2 recombinants, differences at either I or D were sufficient. A disparity at H-2K alone, as provided by the H-2 mutant B6.C-H-2ba against the parental line C57BL/6By, also induced helper effects. The significance of these results is discussed.     

Mechanism of allosuppression: evidence for direct suppression of responding B cells.

The mechanism of allosuppression has been investigated. As previously described, negative allogeneic effects result when T cells recognize MHC-encoded alloantigens on responding lymphocytes, preventing the generation of a secondary anti-PFC response in vitro. Several experiments suggested that this suppression was not due to the generation of cytotoxic effectors. First, effective suppression occurred only when T cells, either unprimed or alloantigen activated, were added during the first 24 hr of the responding culture. Even previously generated cytotoxic effectors were relatively ineffective when added late in the response. Furthermore, no detectable cytotoxic effectors were found in suppressed cultures. The major target of suppression was the responding B cell. Only B cells carrying alloantigens (thus recognized by the T cells) were suppressed; bystander B cells were little affected. Thus allosuppression appears to involve the recognition by T cells of alloantigens on responding B cells and direct suppression of some early event in the development of these B cells into PFC. The responses of primed B cells were found to be preferentially sensitive to the suppressive effects of allo- T cells, whereas the response of unprimed B cells was influenced preferentially by the helper effects of alloantigen-activated T cells. It is possible that the state of differentiation of the B cell may determine the outcome of the interaction with regulatory T cells.     

Analysis of the negative allogeneic effect using B cells and helper T cell lines as an indicator system: B cells identified as target cells for suppression

The target cells for H-2b T lymphocytes mediating a negative allogeneic effect in vitro were analyzed by using carrier-specific helper T cell lines of H-2b, H-2d, or F1 origin and hapten-primed T-depleted spleen cells also expressing one or both of these haplotypes. The helper T cell lines were shown to be carrier specific and H-2b or H-2d restricted. Most importantly, the lines derived from H-2b homozygous mice were devoid of alloreactivity against H-2d and vice versa. Titration of naive H-2b T lymphocytes to the indicator cultures resulted in suppression of the secondary anti-DNP response of the indicator cells whenever the B cells expressed H-2d antigens. The lack of suppression observed in mixtures in which only the helper T cell lines expressed H-2d antigens was not reversed by the increased addition of naive H-2bxd cells, indicating that an insufficient amount of H-2d antigens present on the low number of helper T cells used did not account for this finding. Moreover, the polyclonal plaque-forming cell responses of F1 spleen cells to LPS were also suppressed by naive parental T cells. From these findings it is concluded that the suppressor T cells directly recognize and inhibit allogeneic B cells without the involvement of helper T cells. In addition, it was shown that the suppression of secondary anti-hapten responses by naive allogeneic T cells is blocked by monoclonal anti-Lyt-2 antibody added at the onset of culture. Addition late in culture had no effect, pointing to a functional role of the Lyt-2-bearing structure at an early stage of the suppressive events resulting in the negative allogeneic effect.     

Primary in vitro cytotoxic response of F1 T lymphocytes against parental antigens.

Spleen cells from adult female (AKR/J x BALB/(c)F1 mice can respond to mitomycin C-treated spleen from AKR/J mice and can generate effector CTL in a 5-day primary in vitro culture. The response is comparable in magnitude to the response to allogeneic H-2K or H-2D antigens. The response is T cell mediated and is directed to antigen(s) present only on the parental cells. The target cell must be homozygous at H-2Kk to be lysed and H-2Dk antigens do not serve as a target in this response. Spleen cells from (B10.BR x B10.D2) hybrids that have been stimulated with AKR/J lyse B10.Br as well as AKR/J target cells. Similar H-2k/d hybrid F1 anti-H-2k parent responses are seen in certain other strain combinations. A number of possible interpretations of these responses are discussed.     

H-2K/H-2D and Mls and I region-associated antigens stimulate helper factor(s) involved in the generation of cytotoxic T lymphocytes

This study examines the antigen that stimulate production or release of a soluble helper factor(s) involved in development of cytotoxic T lymphocytes (CTL). Antigens associated with the Mls locus, I and K/D regions of the MHC were all capable of stimulating responder cells in MLC to produce helper factor. These supernatant fluids were all capable of providing "help" for the generation of cytotoxic T lymphocytes in MLC in which spleen cells are stimulated by allogeneic heat-treated thymocytes or splenocytes. Previous reports from our laboratory as well as others have shown that heat-treated cells do not stimulate a cytotoxic response. Heat-treatment of Mls, I, and H-2K/H-2D region incompatible stimulatory cells in MLC eliminated their ability to induce responder cells to produce helper factor, suggesting this is the mechanism whereby heat-treatment reduces the ability of cells to stimulate cell-mediated lympholysis (CML). The inability of supernatant fluids, from MLCs in which heat-treated cells were the stimulators, to assist in the generation of cytotoxic T cells did not appear to be the result of any suppressive factor induced by such treatment. Further, the antigens that stimulate pre-killer cells appear functionally distinct from those heat labile antigens (Mls, I, H-2K/H-2D associated) that stimulate helper factor production since heat-treated allogeneic cells served as stimulators of cytotoxicity provided helper activity was added to the MLC.     

T cell recognition in the mixed lymphocyte response. II. Ia-positive splenic adherent cells are required for non-I region-induced stimulation

Ia-positive splenic adherent cells (SAC) have been shown to be the predominant stimulators of a mixed lymphocyte response (MLR) to whole H-2 differences, in which most of the proliferative response is directed against I region-encoded determinants. The present studies were undertaken to examine the ability of several purified lymphoid subpopulations to activate T cells in response to the non-H-2-linked MIs products or to products of the K or D regions of H-2. The results demonstrated that adherent cell-depleted populations of T and B cells were nonstimulatory, whereas SAC were potent stimulators for responses involving each of these genetic differences. Treatment of these SAC with anti-Ia and C abrogated their MLR-stimulating ability. In contrast, whereas treatment of SAC with anti-Ia and C abrogated their ability to stimulate an MLR directed against K or D region-encoded determinants, this treatment had no effect on their ability to generate a cytotoxic T lymphocyte response against these same determinants. These findings suggest that in addition to presenting allogeneic I region-encoded determinants, Ia-positive SAC also play a unique role in the presentation of non-I region-encoded alloantigens to proliferating T cells.     

Modulation of Thy-1 alloantibody responses: donor cell-associated H-2 inhibition and augmentation without recipient Ir gene control

In this study we have found that in vivo PFC responses to Thy-1 are strongly modulated by H-2 gene products in at least 2 ways. First, a profound inhibition of primary PFC responses occurs when foreign H-2 antigens are expressed on Thy-1 incompatible donor cells. This interference effect does not reflect a requirement for H-2-restricted antigen presentation by donor cells, since it is also seen using semi-syngeneic antigenic cells that share a full H-2 haplotype with the recipient. Interference acts more profoundly on slower primary responses than on more rapid secondary responses and requires associative recognition of the H-2 and Thy-1 antigens. By contrast, strong augmentation of the anti-Thy-1 response was obtained when foreign H-2 antigens were expressed in the recipient, as shown by a poor response of an H-2k/k recipient to Thy-1.1 on AKR cells (H-2k) compared with high responses of H-2k/b recipients. A gene controlling this phenotype was mapped to the H-2IA or H-2K regions. However, subsequent experiments revealed that donor recognition of recipient H-2 antigens was required for these high responses; thus, an Ir-gene effect was excluded and an 'intimate form' of an allogeneic effect was postulated. Thus, the immune response to Thy-1 is regulated by at least 3 factors acting at the level of the donor cell, including non-H-2 helper alloantigens, H-2 interference, and H-2-associated allogeneic effects.     

Specificities of killing by T lymphocytes generated against syngeneic SV40 transformants: studies employing recombinants within the H-2 complex.

T lymphocyte effectors to syngeneic SV40-transformed cells, generated by secondary in vitro sensitization of immune spleen cells, lyse SV40 transformed targets that are syngeneic at the H-2 locus. In this study we have employed recombinants within the H-2 region to examine in detail this H-2 specificity. H-2b effectors were found to lyse SV40-transformed targets from recombinants bearing either H-2Kb or H-2Db.H-2k effectors recognized only SV40-transformed H-2Kk, and not H-2Dk target cells. By using the same protocol for sensitization, no effector cells could be detected in H-2d mice. Effectors generated in H-2 recombinant mice showed that the response capacity resides with K and D. For example, HTG, which is H-2d except at the D locus (H-2Db), produced effector cells specific for SV40-transformed H-2Db targets. Thus, the secondary in vitro response to SV40 transformants was found to depend only on the K and D alleles and not to be modified by the I region to any measurable extent.     

The allogeneic effect: M-locus differences substitute for differences in the H-2 major histocompatibility complex.

The primary immune response against sheep red blood cells in T cell-deficient spleen cell cultures from nude mice was tested in the absence and presence of allogeneic spleen cells. The allogeneic spleen cells differed either in regard to the major histocompatibility complex (H-2) or only with respect to the M-locus. Surprisingly the M-locus different spleen cells were almost as efficient in enhancing the anti-sheep red blood cell response in nude cultures as were the cells differing on the complete H-2 complex. Evidence is presented that AKR anti-theta serum sensitive T cells are responsible for the M-locus-dependent effect edscribed. This effect is shown to be mediated by a factor released from actived T cells stimulated in M-locus different mixed lymphocyte cultures. Since almost identical parameters have been observed in both the M-locus-dependent situation as in the "classical" allogeneic situation we concluded that an allogeneic effect can be induced by T cells responding to a complete set of the major histocompatibility complex (H-2) as well as to lymphocyte-activating determinants (M-locus) alone.     

Properties of H-2L locus products in allogeneic and H-2 restricted, trinitrophenyl-specific cytotoxic responses.

The role of the recently defined L antigen (a second D region product) in allogeneic and TNP-specific syngeneic primary CML responses has been investigated. The lysis by anti-L specific cytotoxic effector cells was not inhibited when the target cells were pretreated with an antiserum directed against K and D, whereas an antiserum against L completely abrogated this response. Therefore, H-2L products are recognized on the target cell independently of H-2K and H-2D locus products. Both A.SW cells as well as B10 cells were found to respond to Ld alloantigens, in addition to Dd alloantigens when stimulated by cells differing only in the D region. The results of cold target blocking and antiserum inhibition experiments failed to detect cytotoxic cells with specificity of L antigens in association with TNP, under conditions in which TNP-specific effectors to K and D antigens were demonstrable. These findings suggest that there is a more limited involvement of H-2L locus products than the H-2K or H-2D locus products in the induction and specificity of these responses.     

Activation of specific suppressor cells with heat-treated allogeneic tumor cells

The ability of heat-treated allogeneic cells to induce suppressor cells was examined. The tumor cell lines EL-4 (H-2^b) and P815-X2 (H-2^d), were heated to 56 °C for 10 min and injected intravenously into mice of the DBA/2J (H-2^d) and C57BL/6J (H-2^b) strains, respectively. After 4 days, the splenocytes of the treated mice were mixed with normal spleen cells and cultured for 5 days with allogeneic tumor cells. The cytotoxic T-cell response was reduced in cultures of these cell mixtures. An allogeneic difference was required to induce suppression because the syngeneic combination did not induce suppressor cell activity. Furthermore, the induction of cytotoxic T cells to the C118 cell line (H-2^k) was not suppressed by this procedure, which suggests that the suppression was haplotype specific. These suppressor cells were sensitive to anti-Thy 1.2 and complement, cortisone, and cyclophosphamide, but insensitive to irradiation. These are characteristics similar to suppressor cells activated by intact cells. Heat treatment abrogated the tumor cell's ability to induce a proliferative and a primary, but not a secondary, cytotoxic T-cell response. The heat-treated cells also lost their ability to function as cold target inhibitor cells, but retained the same quantity of serologically detected antigens as the intact cells. These results suggest that the serologically detected antigens are responsible for the activation of the suppressor cells of the cytotoxic T-cell response.     

Frequency of allogeneic helper T cells responding to whole H-2 differences and to an H-2K difference alone.

Limiting dilution analysis was used to determine the frequency of splenic T cells that are stimulated by alloantigen to give help in a primary antibody response to SRBC. Several haplotype combinations were tested. A semilogarithmic plot of the fraction of nonresponding culture as a function of the number of T cells added to excess B cells gave a straight line intercepting with the origin. Thus a single cell-type was limiting, which was required to help B cells respond to SRBC. The frequency of syngeneic precursors of T helper cells specific for SRBC ranged from 1/10,000 to 1/55,000 with a mean of about 1/20,000. Allohelpers generated by whole H-2 differences gave precursor frequencies that ranged from 1/1000 to 1/7000 with a mean of about 1/2500. Thus allohelpers to whole H-2 differences were approximately 8-fold more frequent than SRBC-specific helpers. When the stimulation was limited to the H-2K difference between the mutant B6.C-H-2ba and wild-type B6, frequencies of from 1/2600 to 1/7900 allohelpers were found with a mean of about 1/5000, approximately half the frequency of allohelpers to whole H-2 differences. Thus some, but probably not all, of the magnitude of allogeneic halp can be attributed to the high frequency of helper T cells that respond to a given alloantigen.     

Cell-sized, supported artificial membranes (pseudocytes): response of precursor cytotoxic T lymphocytes to class I MHC proteins

Novel cell-sized, supported artificial membranes bearing class I antigens have been prepared by a simple dialysis procedure and then used to study the requirements for antigen recognition by precursor cytotoxic T lymphocytes (CTL). The membranes were made by mixing lipid, H-2 antigen, and C18 alkylated 5 microns silica beads in deoxycholate, and dialyzing to remove the detergent. The H-2 antigen-bearing, cell-sized beads, termed pseudocytes (artificial cells), were able to simulate generation of secondary CTL responses with the same specificity as alloantigen-bearing spleen cells. Comparative analyses demonstrated that the size of an antigen-bearing structure, and thus its potential for multivalent interaction, was a critical determinant of effectiveness of antigen recognition, and showed that H-2 antigen was recognized as effectively on cell-sized beads as on allogeneic spleen cells. Generation of a response to antigen on the cell-sized beads was completely dependent on addition of lymphokines to the cultures. Thus, unlike liposomes, H-2 antigen on beads was not available to accessory cells for stimulation of Ia-dependent production of lymphokines by T helper cells. These results, as well as direct observations by microscopy, strongly indicate that antigen is recognized on the surface of the beads. Despite effective stimulation of secondary CTL responses, antigen on beads was completely inactive in stimulating a primary CTL response by naive spleen cells. The results of mixing experiments by using beads and alloantigen-bearing cells or plasma membrane vesicles indicate that the lack of a primary response may result from a requirement for a soluble factor(s) that is not needed for generation of secondary responses. The unique advantages of cell-size supported membranes for studying antigen recognition by T cells are discussed. The beads can be handled and used like antigen-bearing cells in functional assays, while possessing well-defined, readily varied, and easily quantitated composition.     

Suppression of delayed-type hypersensitivity to third party "bystander" alloantigens by antigen-specific suppressor T cells

Delayed-type hypersensitivity (DTH) against alloantigens can be induced by sc immunization with allogeneic cells. The induction of DTH can be suppressed by iv preimmunization of the mice with similar allogeneic spleen cells, provided the cells are irradiated before injection. This suppression is mediated by T cells. The suppressor activity can be induced not only by H-2-and non-H-2-coded antigens, but also by H-2 subregion-coded antigens. Suppression induced by K, I, or D subregion-coded antigens is specific for that particular subregion as well as for its haplotype. I-J-coded alloantigens were found to not be necessary for the induction of antigen-specific suppressor T cells. After restimulation of suppressor T cells by the "specific" alloantigens, the DTH to simultaneously administered third-party alloantigens becomes suppressed as well. This nonspecific suppression of DTH to third party "bystander" alloantigens also occurs when the specific and the third-party antigens are presented on separate cells, provided that both cell types are administered together at the same site. The simultaneous presentation of both sets of alloantigens during the induction phase of DTH only is sufficient to prevent the normal development of DTH to the third-party antigens.     

Accessory cells in the in vitro generation of type C viruses specific T killer lymphocytes. I. Role of macrophages in primary anti-FMR reaction

The role of macrophages has been studied in in vitro cytolytic T lymphocytes- (CTL) mediated responses directed against the FMR cell-surface antigens induced by C type viruses. Macrophages, defined as Thy 1.2-negative, Ia-positive, adherent, phagocytic, radioresistant cells present in the spleen and the peritoneal cavity, are required to obtain primary in vitro anti-FMR responses. In most of our experiments they were not necessary in secondary reactions. In primary responses, macrophages and responder T cells must be compatible at least at the I region of the major histocompatibility complex. Anti-Ia antibodies inhibit the response. A rat soluble factor (Interleukin 2) can replace macrophages in primary anti-FMR CTL-mediated reactions. These results suggest that macrophages function during primary anti-FMR response by interacting with a helper T cell rather than with CTL precursors. In agreement with this hypothesis, it appears that the H-2 restriction of F1 hybrid anti-FMR CTL generated in vitro in primary reaction is not related to the H-2 specificities of the parental macrophages used and depends only on the H-2 antigens of the stimulating tumor cells.     

Suppression of antibody responses in allogeneic mice by products of lymphoid tissue. II. Lack of antigenic specificity and immunogenetic requirements of allogeneic suppressive factor (ASF).

Mice were irradiated, infused with thymocytes and immunized with a variety of antigens, i.e., sheep or horse red blood cells (SRBC or HRBC), diphtheria toxoid (DT) or bovine gamma-globulin (BGG). The spleen cells (T.Spleen cells) were harvested 5 days later and cellfree extracts were prepared. The extracts contained an allogeneic suppressive factor (ASF) that was capable of inhibiting IgM antibody responses of allogeneic or semi-allogeneic unirradiated mice. ASF had to be injected within 24 hr of immunization to be effective and a single injection delayed, rather than abolished, the antibody response at the cellular level. However, daily injections of ASF resulted in persistent suppression of antibody response. ASF activity was antigen nonspecific, i.e., the antigen used to stimulate ASF production did not have to be the same as the antigen used to test for ASF activity. C3H T.Spleen extracts were even immunosuppressive when prepared by exposure to C3BF1 alloantigens only; such extracts suppressed antibody responses of C3BF1 and DBA/2 mice. C3H ASF was removed from extracts after incubation with C3BF1 spleen cells but not after incubation with C3H spleen cells. C3BF1 spleen cells which had been preincubated with C3H ASF were unable to generate antibody-forming cells upon transfer to irradiated C3BF1 host mice. This suggests that the ASF molecule may be or include receptors for alloantigens. The immunogenetic requirements for ASF activity were evaluated by injecting extracts from C3H, C57BL, C3BF and BALB/c T.Spleen cells into C3H, CBA, C57BL, BALB/c, DBA/2, A or C3H.A recipient mice. All extracts tested had ASF activity. However, all allogeneic recipients were not suppressed by the extract material. The suppressive activity of ASF seemed to require two (or more) antigenic differences between donors and recipients of extract material, an H-2K or I antigen difference and a second antigen difference, possibility Ig-1. In the limited numbers of strain combinations tested, T.Spleen extracts suppressed IgM antibody response only if exposed to H-2 and Ig-1 antigens, e.g., BALB/c (H-2d, Ig-1a) ASF suppressed A (H-2a, Ig-1e) but not C3H.A (H-2a, Ig-1a) or DBA/2 (H-2d, Ig-1c). Separate ASF molecules may react with separate antigens on the cell surface, i.e., with H-2 and gammaG2a. Alternatively, one ASF molecule may react with two structurally associated antigens. If the latter is correct, it is conceivable that the beta2-microglobulin which is non-covalently linked to the major component of H-2 molecules expresses allotypic antigens coded for by Ig-1 and beta2-microglobulin is one of the antigens recognized by ASF.     

Induction of I region-restricted hapten-specific cytotoxic T lymphocytes.

Murine cytotoxic T lymphocytes (CTL) reactive to TNP-conjugated syngeneic target cells do lyse to a moderate but significant extent TNP-conjugated, I region compatible but H-2K or H-2D region incompatible target cells. Antibody inhibition experiments and "cold inhibition" experiments indicate that some CTL clones recognize TNP-conjugated targets in association with syngeneic I region determinants independently of H-2K or H-2D gene products. The data suggest that besides TNP-conjugated H-2K or H-2D gene products, in principle, also TNP-conjugated I region determinants do stimulate TNP-specific CTL precursor cells and act as target antigens of TNP-specific CTL.     

The allogeneic effect: the mechanism of allosuppression by Lyt-1, Ia- T cells

The kinetics of allohelp mediated by diffusable factors revealed that help by nonirradiated T cells (TOR) peaked at 48 to 72 hr, followed by a sharp decline if the T cells remained in the cultures. The temporal decrease in help after 72 hr was not mediated by suppressor lymphokines because mixtures of early (24 to 48 hr) and late (120-hr) allogeneic supernatants enhanced help synergistically. Lyt-1, Ia- T cells mediated the temporal decline in help and suppressed allogeneic B cell activation in co-cultures, and this "down-regulatory" activity (allosuppression) was radiosensitive. Help by irradiated T cells (T1000R) increased gradually until it plateaued between 96 and 120 hr. The helper activities of the allogeneic supernatants were directly proportional to their T cell growth factor (TCGF) activities. In addition, their kinetics were identical, and the removal of TCGF from 48-hr allogeneic supernatants by adsorption with TCGF-dependent HT-2 cells depleted both helper and TCGF activities. Help was restored to depleted 48-hr and 120-hr allogeneic supernatants by preparations of TCGF obtained from concanavalin A (Con A)-stimulated FS6-14.13 hybridoma cells that were adsorbed with lipopolysaccharide (LPS)-activated B cells or normal spleen cells (NS), but not with HT-2 cells. These results indicate that allohelp is dependent on TCGF. Moreover, help was dependent on at least one factor in addition to TCGF, because a high level of synergy occurred between TCGF and the "help-deficient" 120-hr allogeneic supernatant. In conclusion, the mechanism whereby Lyt-1, Ia- T cells regulated B cell activation with positive and negative allogeneic effects was through the production and subsequent exhaustion of TCGF, respectively. The production of TCGF and help was radioresistant, but exhaustion of TCGF and suppression was radiosensitive.     

Transplantation of cultured thymic fragments. VI. H-2 recombinant donors

Athymic (nude) mice were transplanted with cultured thymic fragments from syngeneic, allogeneic, and partially allogeneic (recombinant) mice. Lymphocyte proliferation and cytotoxicity in vitro were measured to assess immunologic reconstitution. Transplanted nude mice were immunocompetent whether donor and recipient were disparate for class I, class II, or both H-2 gene types. Furthermore, allotolerance for thymic H-2 class I antigens was achieved independently of class II antigen allotolerance. Class I antigen tolerance was not broken during lymphocyte responses to unrelated alloantigens, ruling out insufficient help as the tolerance mechanism. Splenocytes, isolated from nude mice transplanted with fully allogeneic or syngeneic thymic fragments and stimulated in vitro with trinitrophenyl-modified cells, displayed H-2-restricted, hapten-specific cytotoxicity. Cytotoxic cells from allotolerant mice were restricted to either host or thymic H-2 antigens, depending on the stimulating cell haplotype. Response levels for thymic and host trinitrophenyl-modified cells were comparable. We have shown that allogeneic thymic epithelium transplanted into adult nude mice can induce allotolerance to class I and II H-2 antigens equally, and permits T lymphocyte interaction with cells bearing thymic donor or host H-2 antigens. Our results are consistent with a model wherein T lymphocyte self-receptors retain their genomic repertoire but can be selectively mutated or expanded by appropriate H-2 antigen presentation by the thymus.