Exogenous IL2 partially reverts CML non-reactivity acquired during prophylactic immunosuppression with cyclosporin A of human allograft recipients

Proliferative and cytolytic lymphocyte responses and the influence of exogenous interleukin 2 (IL2) on cell-mediated lympholysis (CML) reactivity were evaluated in 12 allograft recipients. Responses were induced by mitogenic lectins or by donor and third-party cells. Patients were tested immediately before transplantation (Tx) and one and three months after grafting. Prophylactic immunosuppression consisted of Cyclosporin A (CyA) and low-dose prednisone (P). Analysis of post transplant cells revealed a reduced overall proliferative T cell responsiveness induced by both alloantigens and mitogenic lectins. No evidence for donor-specific reduction of MLC responses was seen. Overall CML reactivity of post-Tx lymphocytes was also impaired. This was accompanied by donor-specific CML non-reactivity in six of seven patients with quiescent grafts. In these patients, the cytolytic potential against donor cells could be restored when maximal T cell help via exogenous IL2 was provided.     

Requirements for Induction of Specific Suppressor T Cells and Detection of their H-2 Antigen-binding Receptors by Fractionation on Target Cell Monolayers

The MC-resistant specific suppressor T cells that inhibit DNA synthesis and CTL generation in MLC were induced in vivo by gamma-irradiated allogeneic lymphoid cells in a large dose. MLRs were inhibited only slightly when triggered by third-party cells, even neighbouring with the corresponding stimulators. Unlike the irradiated cells, intact allogeneic lymphoid cells induced a mixture of macrophage-like and T-cell suppressors with a pronounced non-specific component of the action. Syngeneic cells induced low active non-specific suppressors of macrophage type only. The suppression was not due to a cytotoxic effect, since specific T suppressors differed from CTL by conditions of induction and high sensitivity to gamma-irradiation and from CTL precursors by high sensitivity to CY and HC. The specific T-suppressors could be selectively removed by adherence to a macrophage monolayer of the corresponding donor. The subsequent elution of adherent lymphocytes with pronase resulted in enrichment of specific T suppressors by a factor of 30 and 2.6, as judged by reduction in the number of lymphocytes required for 50% inhibition of DNA synthesis and 33% inhibition of CTL generation, respectively. The high specificity of this enrichment is shown by using both syngeneic monolayers for fractionation and third-party stimulators in MLR for testing and by disappearance of slight non-specific suppression caused by non-fractionated suppression. Complete inactivation of the eluted suppressors with anti-Thy-1.2 and anti-T antibodies, their resistance to anti-Mls antibodies, carrageenin, and carbonyl iron, together with the data of autoradiographic study and total DNA synthesis in the population indicate that the eluted highly specific suppressors represent mainly DNA-synthesizing large and medium T lymphocytes.     

Reduced lysis by CD8+ cytotoxic T cells in mixed lymphocyte reactions induced via CD4+ T cells exposed to chemically modified antigen presenting cells.

The resistance by T lymphocytes to activation by antigen (anergy) is well documented for CD4+ T-helper (Th) cells, although less is known about CD8+ cytotoxic T lymphocytes (CTL). One widely used method of inducing anergy of CD4+Th is presentation of antigen by ECDI (1-ethyl-3-(3-dimethylamino-propyl)carbodiimide)-fixed antigen-presenting cells (APCs). We report here that in murine mixed lymphocyte reactions (MLRs), a marked reduction in detected cytotoxicity (which is mediated predominantly by CD8+ CTL) occurs on day 7 if the bulk cultures are restimulated 2 days previously with ECDI-fixed allogeneic splenocytes. No differences were seen between untreated cultures on days 5 and 7, or on day 7 of cultures to which were added unfixed allogeneic splenocytes, fixed or unfixed syngeneic splenocytes, or 'third-party' allogeneic splenocytes, 2 days previously. The effect is not mediated directly on CD8+ cells, since MLRs depleted of CD4+ cells immediately prior to exposure to fixed allogeneic splenocytes fail to show reduced lysis. On the other hand, reduced lysis did occur if CD4+ cells, purified from the MLRs on day 4, were exposed to ECDI-fixed allogeneic splenocytes and then returned to MLRs previously depleted of CD4+ cells. Moreover the effect is overcome using exogenous interleukin-2 (IL-2). We propose that CD4+ cells, restimulated by a regimen shown previously to induce their anergy, can cause a reduction in CD(8+)-mediated cytotoxicity in MLRs.     

Analysis of post-transplant immune status in recipients of liver/bone marrow allografts

The aims of this study were to assess the effect of donor bone marrow infusion on the reactivity of recipient peripheral blood lymphocytes (PBL) to mitogen and to donor and third-party cells after primary liver allotransplantation and to identify any correlation between altered immunoreactivity and HLA mismatches, occurrence of rejection, and immunosuppression. The immunoreactivity of recipient PBL toward frozen donor splenocytes was evaluated in mixed lymphocyte culture (MLC) (n = 29) and cell-mediated lympholysis (CML) (n = 27) assays in time intervals ranging from 0.7 to 27 months after transplant. Overall, the mean anti-donor MLC stimulation index (SI) fell from 25.6 ± 5.2 preoperatively to 4.8 ± 1.7 after transplantation (p < 0.002), with 14 out of 29 (48.3%) patients developing donor-specific MLC hyporeactivity. HLA class II mismatches were significantly associated with recipient post-transplant immune profile (p < 0.05): MLC donor specific hyporesponsiveness was observed in 70%, versus 37% of patients who shared a class II antigen, versus those that did not. Of the control group, 61.1% developed donor-specific nonreactivity versus 27.2% in the donor bone marrow cells (DBMC) group (p = 0.02). Donor-specific CML hyporeactivity was observed after transplantation, independent of DBMC infusion, with mean percentage values of pre- and post-transplant donor-specific lysis of 22.4% ± 4.1% versus 3.1% ± 1.6%, p = 0.0004, respectively. Our results suggest that DBMC infusion favors development of nonspecific MLC hyporesponsiveness to donor and third-party alloantigen, with maintenance of reactivity to mitogen and no additional effect on T-cell cytotoxicity.     

Regulation of human cytotoxic lymphocyte responses. I. Non-cytolytic suppression mediated by alloantigen-activated cells.

Alloantigen sensitized human lymphocytes obtained from a 2-3 day mixed lymphocyte culture (MLC) suppressed the in vitro generation of alloreactive cytotoxic lymphocytes (CTL). The inhibition of CTL responses was demonstrated in MLC after both 1 day and 14 days' allosensitization. The suppressor cells were nylon wool non-adherent, lacked Fc receptors and adhered to histamine columns. The MLC-activated suppressor cell population had an associated very low and transient cytotoxic response directed against the allogeneic sensitizing cell. Several procedures were used to dissociate this activity from suppressor cell function: (1) donors were preselected which showed minimal cross-killing between allogeneic stimulating cells, (2) suppressor cultures were added to the test cultures prior to the development of maximal CTL activity, (3) suppressor cultures were irradiated preventing cell proliferation associated with differentiating CTL. Lastly, by increasing stimulating antigen concentration suppressor cell activity was increased rather than being competitively diminished, which would be predicted if suppression was occurring through a cytolytic or inactivation of stimulating antigen. It was therefore concluded that alloantigen stimulation in MLC activates a non-cytolytic regulatory cell population which is capable of inhibiting CTL responses to third-party allogeneic lymphocytes.     

Role of blood transfusions on the induction of antibodies against recognition sites on T lymphocytes in renal transplant patients

We have tested sera from 23 renal allograft recipients to study the effects of blood transfusions on the induction of antibodies directed against recognition sites on T lymphocytes. The results demonstrate that antibodies capable of inhibiting responses in MLC could be induced by blood transfusion. This inhibition in MLC is observed by treatment of responder lymphocytes with serum plus rabbit complement and is mediated by IgG antibodies. Also, the inhibitory effect is specific for certain responder cells and is not mediated by antibodies against common surface antigens of either the responder or the stimulator lymphocytes. The antibodies inhibiting proliferative responses in MLC against antigens present on the kidney donor were demonstrable in renal transplant recipients with functional allografts, but not in patients who had rejected the graft. The data suggest that antibodies directed against recognition sites on T lymphocytes could be induced by blood transfusions and these antibodies may be associated with prolonged graft survival.     

Human suppressor T cells induced in vitro with an autologous renal allograft-derived T cell line. I. Suppressor cell induction, function, and specificity

The functional characteristics of T suppressor (Ts) cells generated from the peripheral blood lymphocytes (PBL) of a kidney transplant recipient who had excellent graft function for 1 year were examined. Ts cells were induced by co-culture of PBL with an autologous alloreactive cytotoxic T lymphocyte (CTL) line (EE-1) previously grown from a routine renal allograft biopsy of this patient performed 10 days posttransplant. The EE-1 line included CD3+ T cells of CD8+ and CD4+ phenotypes with cytotoxic specificity for disparate class 1 (HLA-B8) and class II (HLA-DR1 and 3) antigens of the kidney donor (JC). The EE-1 induced Ts cell lines (designated TsEE) were found to significantly suppress (50%-95%) autologous fresh responder EE-PBL stimulation by donor EBV-transformed cells (JC-EBV) in mixed lymphocyte reaction (MLR) assay. TsEE cells were CD3+ (98%) and predominantly CD8+ (68-80%), showed no cytotoxic activity, and were suppressive only at the early phase of MLR stimulation. In three-party cell test MLR assays, TsEE-mediated suppression appeared restricted to responder cells sharing HLA-B7 with the suppressor line, and was not abrogated by the addition of exogenous interleukin-2 (IL-2). TsEE cells also showed restricted suppression of CTL generation but not mature CTL activity. The restricted suppressor activity of TsEE lines was dependent upon their induction and restimulation with the autologous EE-1 line.     

Antibodies against the transferrin receptor block the induction of cytotoxic T lymphocytes. A new method for antigen-specific negative selection in vitro

The induction of cytotoxic T lymphocytes (CTL) specific for alloantigen in a primary mixed lymphocyte culture (MLC) was used as a model system to study the potential usefulness of monoclonal antibodies directed against the murine transferrin receptor in eliminating antigen-activated T cells. Two monoclonal antibodies were used which are able to block the transferrin-mediated uptake of iron by growing cells. One of these antibodies is cytotoxic in the presence of complement (C). Treatment of MLC cells from control, uninhibited cultures with the cytotoxic anti-transferrin receptor antibody plus C completely knocked out all CTL effector activity showing that all of the effector cells expressed the receptor. However, surviving cells from such antibody-plus C-treated populations were not depleted of precursors since they were able to mount a strong CTL response to the original alloantigen upon restimulation. Blocking antibody against the transferrin receptor, in the absence of C, when present during the course of a primary MLC, completely inhibited the generation of effector CTL. Conditions were optimized such that antibody present during the course of the MLC could completely and specifically eliminate all antigen-activated T cells including precursors as assessed by secondary restimulation with the original, plus third-party antigens, in the absence of blocking antibody. This method for producing specifically unreactive T cell populations appears to be superior to previously described techniques.     

抗TCRαβ单抗联合供体骨髓细胞输注诱导小鼠皮肤移植耐受的研究

探讨抗TCRαβ单克隆抗体联合大剂量供体骨髓细胞输注方法对同种异基因小鼠皮肤移植耐受诱导的促进作用。第 0天 ,C5 7BL/ 6 (H 2 b,B6 )小鼠尾静脉注入 2× 10 8BALB/c (H 2 d,B/c )来源的骨髓细胞 ,同时腹腔注射抗TCRαβ单克隆抗体5 0 0 μg ;第 6天进行皮肤移植 ;在不同的时间对B6小鼠进行迟发型超敏反应 (DTH )、混合淋巴细胞反应 (MLR )等耐受状态进行检测 ,并进行MLR的IL 2逆转实验、过继转移实验和嵌合状态分析以初步探讨耐受形成机制。结果显示 ,经抗TCRαβ单克隆抗体联合大剂量供体骨髓细胞输注处理的B6小鼠的供体皮肤移植物平均存活 5 0 4d ,显著长于其他各组 (P <0 0 0 1)。相对于正常对照组小鼠 ,耐受B6小鼠在DTH和MLR中均表现出显著的低反应性 (P <0 0 0 1)。IL 2逆转实验结果表明克隆无能参与了移植耐受的形成。体内、外过继转移实验均显示耐受B6小鼠脾细胞中存在抑制细胞活性。嵌合体检查结果表明在耐受B6小鼠的胸腺和脾脏中均形成了混合嵌合体 ,在皮肤移植后第 15、 30和 70天时脾脏内供体来源嵌合体水平依次为 15 86 % ,10 5 7%和 1 77% ,胸腺中依次为 8 19% ,5 72 %和 1 87% ,嵌合体水平随时间而下降。这些表明 ,抗TCRαβ单抗联合大剂量供体骨髓细胞输注可以在异基因成年小?     

Mechanisms of CML hyporesponsiveness in long-term renal allograft recipients

Long-term, HLA disparate, renal allograft recipients were assessed by a dose-dependent CML (cell-mediated lympholysis) assay for evidence of donor-specific CML hyporesponsiveness. The frequency of cytolytic T-cell (CTL) precursors capable of responding to donor alloantigen was also examined by a limiting-dilution assay and application of Poisson distribution statistics. Four recipients were found to have depressed CML responses against donor class I HLA antigens (mean 2-5% specific lysis), whereas significant responses against third party targets were noted (15-60% specific lysis). These data are consistent with an antigen-specific defect in the recipient CTL effector mechanism. To determine whether or not this may be due to a deletion of anti-donor alloreactive cells, a sensitive limiting-dilution assay was developed in conjunction with Poisson distribution statistics to obtain the frequency of both anti-donor and anti-third party CTLp present in recipient PBLs. A simultaneous reduction in the number of alloreactive clones and the presence of an active suppressor population were defined, suggesting that several mechanisms may operate concurrently in long-term renal allograft recipients with well-functioning grafts.     

Effect of in Vivo Exposure to Allogeneic Cells upon Subsequent in Vitro T Cell Responses and upon Allograft Rejection

The effect of intravenous injection of 10⁸ BALB/c spleen cells into C57B1/6J recipients was assayed by both mixed lymphocyte culture (MLC) of recipient lymphocytes, and by grafting donor (BALB/c) thyroids into recipient mice. It was observed that a single intravenous injection produced depression of proliferative and cytotoxic T cell responses in MLC of spleen, lymph node and peripheral blood lymphocytes of the recipients. This effect was specific for the sensitizing genotype (MLC of recipient und third-party C8A/H lymphocytes was unaffected), and persisted for several days after sensitization. The pattern of this diminished response suggested that the effect was due to a combination of recruitment of reactive lymphocytes from peripheral lymphoid populations, and the generation of alloantigen (H-2?)-specific suppressor T cells in the spleen. In contrast to these findings in vitro, a similar sensitization led only to accelerated rejection of thyroid allgrafts in vivo.     

Change in mixed lymphocyte culture reactivity following allosensitization between DLA-identical dog sibs

The mixed lymphocyte culture (MLC) is known to react to major histocompatibility complex disparities but is insensitive to minor histocompatibility antigens. To determine which modifications could amplify reactivity to these antigens, we have studied the effects of presensitization on normally nonresponsive MLC in dogs. Lymphocytes, primed in vitro to and restimulated with cells of a DLA-identical sib, did not react to these specific cells, in contrast to normal proliferation to other stimulators. However, lymphocytes obtained after in vivo sensitization against a DLA-identical donor showed increased reactivity to the donor in approximately half of the recipients. In comparison to primary MLC responses between DLA-mismatched individuals, the [³H] thymidine uptake was less important in this reaction, and peak incorporation was usually not accelerated. This response appeared to be specific in that responsiveness to third-party cells was not increased, and, alternatively, immunization against a DLA haplo-identical donor did not modify reactivity to DLA-identical cells. This increased reactivity after alloimmunization may reflect differences in minor histocompatibility antigens although the skin graft survival time was not significantly shorter in MLC responders than in nonresponder recipients. It seems that in vivo priming is required to expand the number of cells specific for these antigens and to enable detection of proliferation in MLC.     

Difference between antigen-binding receptor repertoires in effector cytotoxic T lymphocytes and their secondary precursors (memory cells) specific to H-2Kb.

The fine specificity of antigen-binding receptors was compared in pCTL-2 and secondary effector CTL (cytotoxic T lymphocytes) induced in vivo with the H-2Kb alloantigen in recombinant inbred mice. The lymphocyte preparations were enriched by elution from macrophage monolayers of various origins, including the donor (B6 strain), the H-2Kb mutant bm1, the H-2Kk allele B10.A(4R) and the recipient strain B10.D2(R101) as a control. Anti-Kb pCTL-2 eluted from third-party bm1 or B10.A(4R) monolayers gave rise to CTL progeny that lysed, equally well, both donor TC and those third-party TC from whose monolayer the pCTL-2 had been eluted, but which were unable to lyse irrelevant third-party TC. The lytic activities of secondary CTL whose precursors had been eluted from bm1 or B10.A(4R) monolayers were 6 and 12 times lower, respectively, than pCTL-2 eluted from the donor monolayer. Opposite results were shown for receptors of enriched secondary anti-Kb effector CTL. Irrespective of their elution source, whether donor, mutant or allele variant, the eluted effector CTL were able to lyse the donor TC to a similar degree and much more than the given third-party TC; moreover, they retained cross-reactivity in all cases. It is suggested that CTL receptors are homogeneous in specificity for a whole composite immunodominant epitope and differ from each other only in the affinity/lability of the combining site. In contrast, pCTL-2 can be separated into fractions: receptors of each fraction are strictly specific (with high affinity) to a particular portion of the same composite CTL epitope. It seems likely that the pCTL-2 receptor antigen-binding site is modified during pCTL-2 in vivo differentiation into effector CTL.     

IL-6 enhances the generation of cytolytic T lymphocytes in the allogeneic mixed leucocyte reaction.

Cytolytic T lymphocytes (CTL) require soluble proteins termed lymphokines to develop lytic activity. In this report we have studied two of the lymphokines involved in the development of CTL during the allogeneic mixed leucocyte reaction (MLR). High doses of dendritic cells induced lytic activity from purified CD8+ cells in both the murine and human MLR. Under these conditions, IL-2 and IL-6 were endogenously produced and secreted. Antibodies to IL-2 or the IL-2 receptor blocked CTL formation; however, anti-IL-6 receptor antibodies only partially inhibited the response while anti-IL-6 antibodies were largely ineffective. When limiting numbers of antigen-presenting cells were used CTL failed to develop, and neither IL-2 nor IL-6 was secreted into the culture supernatant. Although the addition of IL-6 to such cultures was ineffective in generating CTL, the combination of IL-2 and IL-6 resulted in a 4-5-fold increase in lytic activity over that of IL-2 alone. We conclude that in the allogeneic MLR, IL-2 and IL-6 contribute to the generation of lytically active CD8+ cells, and the effect of IL-6 is evident when the dose of antigen-presenting cell is limited.     

Inhibition of the production of a soluble helper mediator by cyclosporin A results in the failure to generate alloreactive cytolytic cells in mixed-lymphocyte culture

The mechanism of action of cyclosporin A (CsA) in inhibiting the induction of alloreactive cytolytic T lymphocytes (CTL) in mixed-lymphocyte culture (MLC) was investigated. CsA at concentrations of 10(-3) to 10(-1) micrograms/ml completely prevented the generation of CTL. However, the addition of culture supernatants from mitogen-activated lymphocytes to MLC not only significantly reversed the suppressive effect of CsA but also fully restored the reactivity of lymphocytes already treated with CsA. By measuring the presence of a soluble helper mediator (SHF) in MLC supernatants, we found that CsA-treated lymphocytes produced no SHF, possibly interleukin 2 (IL-2). The effect of CsA on receptors for IL-2 was subsequently studied and it was found that the binding capacity of 125I-labeled IL-2 to lymphocytes was not altered by the presence of CsA. These findings suggest that the prevention of helper cells from producing SHF, rather than the inhibition of the response of effector cells to SHF, is a possible explanation for the immunosuppression mediated by CsA.     

Clonal restriction of the expansion of antigen-specific CD8+ memory T cells by transforming growth factor-{beta}

Recent evidence showed that transforming growth factor-beta (TGF-beta) regulates the global expansion of CD8+ T cells, which are CD44hi, a marker for memory cells. However, it is not clear whether this regulatory mechanism also applies to the antigen-specific CD8+ memory cells. By using a murine mixed lymphocyte culture (MLC) model, we examined the effect of TGF-beta on antigen-specific CD8+ memory cells [cytotoxic T lymphocyte (CTL)]. We found that the secondary CTL response in CD8+ memory cells from untreated MLC was not affected by TGF-beta but augmented by interleukin (IL)-2, whereas the CD8+ memory cells from TGF-beta-pretreated MLC (MLC-TGF-beta) failed to mount a significant, secondary CTL response, even when IL-2 was added. In exploring this dichotomy, in combination with flow cytometry analysis, we found that prolonged exposure to TGF-beta reduces the CTL activity in CD8+ memory cells. The increase by IL-2 and the reduction by TGF-beta of the CTL responses were clonal-specific. TGF-beta did not affect the CTL response to a third-party antigen or polyclonal T cell activation. Experiments performed with transgenic 2C cells gave similar results. Cell-cycle study performed with adoptive transfer of the cell tracker-labeled MLC cells revealed that the in vivo expansion of CD8+ memory cells from MLC-TGF-beta was restricted severely, and the restriction was clonal-specific, thus offering direct evidence to show that TGF-beta induces clonal restriction of CD8+ memory cell expansion.     

One-way positive cellular reactions between two HLA-A, B, C, D/DR genotypically identical brothers following active allogeneic immunization

The HLA-D/DR region in man encodes major determinants which stimulate T lymphocytes to proliferation. The genetic organization of this region is apparently complex and is at present largely unknown. One obstacle is the scarcity and quality of available typing reagents. In an attempt to obtain high quality anti-DR sera, a series of active immunizations was performed between highly selected, healthy unrelated donors and recipients.
One recipient (AR8) was immunized using cells incompatible for HLA-A2, B40 (w60), Cw3 and DIDRw6 and readily developed anti-A2 and B40 antibodies but no anti-C, DR, or other antibodies. When tested against his HLA genotypically fully identical brother using the cellular MLC, PLT, or CML techniques before immunization, results were mutually negative as expected. Following immunization, however, AR8 was able to mount MLC, PLT, and possibly CML responses against lymphocytes from the brother while the reverse combinations remained negative. When tested in the family the trait(s) thus identified seems to be maternally inherited.
These results suggest the existence of minor histocompatibility determinants encoded from regions not closely linked to HLA. The brother of AR8 and the immunizing donor thus seem to share one or more determinants not possessed by AR8.     

Tolerance Induction to H-2 Central Region Target Antigens: In Vivo/in Vitro Correlations

Tolerance was induced against cytotoxic target determinants coded for by genes of the I region. Neonatal recipients were immunized with high doses of cells from an I region incompatible donor. Nonreactivity in adult life did not reflect extensive donor cell chimerism, since the great majority of cells in spleens of animals rendered tolerant were of host phenotype. Although specific nonresponsiveness in CML could be induced by these protocols, the MLC proliferative response was in most cases still present alghough very much decreased. In only a very occasional animal was complete nonreactivity in MLC seen. The nonresponse in CML was paralleled by acceptance of thyroid allografts as measured by radioactive iodine incorporation and morphological studies.     

Role of CD4+ regulatory T cells in hyperbaric oxygen-mediated immune nonresponsiveness

We have previously shown that hyperbaric oxygen culture (HOC [95% O₂, 5% CO₂, 25 psi]) is an effective pretransplant tissue-modification technique that results in long-term allograft survival and the induction of systemic immune tolerance in a murine model. Here we address the immune modulatory effects of HOC-treatment of human immune responses using the in vitro mixed lymphocyte reaction (MLR). Pretreatment of allogeneic stimulator cells with HOC results in abrogation of cytotoxic T lymphocyte (CTL) activity, proliferative responses, and IFNγ production in a 7-day MLR. These responses can be restored either by the addition of IFNγ or IL-2 on day 0, or by blocking the activity of IL-4 and IL-10. The addition of IL-2 on day 4 does not restore allospecific CTL activity. The failure of HOC-treated cells to induce allospecific CTL is not due to the induction of anergy, demonstrated by the failure to restore responses after restimulation with allogeneic cells in the presence of IL-2. Removal of CD4⁺ cells prior to restimulation, results in restoration of CTL activity in MLR cultures restimulated with HOC-treated allogeneic cells. These results suggest that HOC-induced immune nonresponsiveness is mediated by the development of CD4⁺ regulatory cells in a Th2-type environment.     

Mutual tolerization of histoincompatible lymphocytes

T lymphocytes react strongly against foreign major histocompatibility complex encoded class I antigens by destroying incompatible tissue in vivo, and by generating cytotoxic T lymphocytes (CTL) in vitro. The absence of reactivity against self antigens may be due to clonal deletion of self-reactive T cells during their ontogeny in the thymus. The functional clonal deletion of mature T cells in the periphery was described recently: CTL recognizing antigen on other CTL are eliminated (Rammensee et al., Immunol. Today 1985. 6: 41). Hence, in a normal immune system only autoreactive cells would be eliminated. Here we show that injection of lymphocytes into class I-incompatible mice leads to abrogation of host anti-donor as well as donor anti-host reactivity, leaving a mixed population of host and donor T cells reactive against third-party antigens. The results demonstrate the existence of a peripheral failsafe mechanism for the elimination of autoaggressive CTL. Whether this failsafe mechanism is actually used under physiological conditions is a different question.