Treatment of multiple sclerosis with T-cell receptor peptides: results of a double-blind pilot trial.

A T-cell receptor (TCR) peptide vaccine from the V beta 5.2 sequence expressed in multiple sclerosis (MS) plaques and on myelin basic protein (MBP)-specific T cells boosted peptide-reactive T cells in patients with progressive MS. Vaccine responders had a reduced MBP response and remained clinically stable without side effects during one year of therapy, whereas nonresponders had an increased MBP response and progressed clinically. Peptide-specific T helper 2 cells directly inhibited MBP-specific T helper 1 cells in vitro through the release of interleukin-10, implicating a bystander suppression mechanism that holds promise for treatment of MS and other autoimmune diseases.

Human TCR as antigen: homologies and potentially cross-reactive HLA-DR2-restricted epitopes within the AV and BV CDR2 loops.

The major function of the T-cell receptor is to confer antigen specificity to T cells. However, nascent TCR proteins that are not assembled into functional heterodimers may be processed and displayed with self MHC molecules on the T-cell surface, and are thought to be the genesis of autoregulatory T cells that can limit inflammatory responses through T-T network interactions. In previous work, we and others have exploited this natural regulatory system using TCR peptides to amplify regulatory T cells that potentially can treat human autoimmune diseases such as multiple sclerosis (MS) and arthritis. The development of this approach is limited by the diversity of human TCR V gene sequences, and by lack of knowledge of exactly which regions of the V gene proteins are immunogenic in association with various MHC alleles. To identify similar amino acid sequences within and among human V gene families that might have immunologic cross reactivity, we aligned 74 known AV and 109 known BV protein sequences into homologous groups using the ClustalX program. Moreover, with a focus on CDR2 peptides that have previously been used to induce regulatory T cells in clinical trials, we established homologous peptide groups, and then identified the optimal amino acid motifs for binding to two alleles, HLA-DRB1*1501 and DRB5*0101, that have been associated with susceptibility to MS. From this analysis, > 75% of AV and BV CDR2 sequences were predicted to bind with at least moderate avidity to each of the DR2 alleles, thus enhancing the likelihood that they could be antigenic. Further ordering of putative TCR contact residues revealed a different set of homology groupings, including many intrafamily sequence matches and some interfamily matches that might allow immunological cross reactivity. Particularly striking were DRB1*1501-restricted IH-S and IY-S motifs shared by BV11, BV12, and BV13 and BV3, BV12, BV13, and BV17 family members, respectively, and DRB5*0101-restricted RL-H and RL-Y motifs shared by BV11, BV12, and BV13 and BV13 and BV17 family members, respectively. This analysis may be useful in designing an array of clinically useful homologous peptides with optimal MHC binding properties and highly cross-reactive TCR binding motifs.

Combining site specificity of monoclonal antibodies to the organophosphate hapten soman.

The combining site specificities of eight monoclonal antibodies raised against the organophosphorous-containing hapten Soman are compared to monoclonal antibodies specific for a naturally occurring organophosphorous compound, phosphocholine (PC). Although these haptens share some structural and spatial features, differences in their chemical structures, most notably the presence or absence of a positive charge, appear to prevent significant cross-reactivity between antibodies specific for each. The murine memory response to PC-KLH has been shown previously to be characterized by the presence of two major groups of antibodies differentiated on the basis of their specificity for free PC and for the nitrophenyl derivative of PC, nitrophenylphosphocholine (NPPC). Interestingly, two groups of hybridoma antibodies were detected in the immune response to Soma-KLH which possess differential specificity for Soman and for a nitrophenyl derivative of Soman.

Rat RT1.B-transfected fibroblast lines process and present myelin antigens and activate T cells to induce experimental autoimmune encephalomyelitis.

The genes encoding the Lewis rat RT1.B molecule (MHC Class II I-A equivalent) were transfected and expressed in mouse DAP.3 fibroblast cells together with the gene encoding the mouse ICAM-1 molecule. Both molecules were stably expressed on the cell surface of DAP.3 cells under longterm culture conditions. The RT1.B/mICAM-1 transfectants presented antigen in a specific manner to a RT1. B-restricted rat T cell hybridoma specific for the 69-89 peptide of myelin basic protein (BP). In addition, the transfectants were able to present antigen to a BP69-89-specific rat T cell line. Presentation to a RT1.D (MHC Class II I-E equivalent)-restricted BP87-99-specific T cell line was minimal. Production of the Th1 cytokine IFN-gamma by BP69-89-specific T cells when stimulated by RT1.B/mICAM-1 transfectants correlated very well with proliferation to specific antigen. Moreover, RT1.B-transfected DAP.3 cells sufficiently stimulated BP69-89-specific T cells such that they were able to transfer experimental autoimmune encephalomyelitis (EAE) to Lewis rat recipients. Thus, the RT1.B molecule is functionally expressed on the surface of transfected Dap.3 fibroblasts and is capable of MHC Class II-restricted, antigen-specific presentation to rat T cells.

The effect of TCR V beta 8 peptide protection and therapy on T cell populations isolated from the spinal cords of Lewis rats with experimental autoimmune encephalomyelitis.

Vaccination or treatment of Lewis rats with TCR V beta 8 peptides can prevent or reverse the clinical signs of experimental autoimmune encephalomyelitis (EAE) which is mediated predominantly by V beta 8.2+ CD4+/CD45R lo T cells. However, rats protected or treated with V beta 8 peptides still developed histological lesions in the spinal cord (SC), even though they remained clinically well. We sought to discern phenotypic changes characteristic of these SC infiltrating lymphocytes. In particular, we focused on whether the immunoregulatory mechanism induced by TCR peptides caused a reduction of V beta 8.2+ T cells, or induced changes in CD45R lo or hi/CD4+ subpopulations that have been associated respectively with EAE induction or recovery. In the V beta 8 peptide vaccinated rats there was a dramatic decrease in the number of V beta 8.2+ T cells isolated from the SC early in disease. During the recovery phase, however, the number of V beta 8.2+ SC T cells was similar in protected and control groups; in contrast, there was striking reduction in the number and size of CD45R hi/CD4+ T cells in the protected animals. In rats treated with V beta 8.2 peptide, no changes were observed in the number of SC V beta 8.2+ T cells or expression of V beta 8.2 message, but similar to vaccinated rats, there was a marked decrease in the number of CD45R hi/CD4+ T cells. These data suggest that vaccination with TCR peptides prevented the initial influx of encephalitogenic V beta 8.2+ T cells into the central nervous system (CNS), whereas treatment appeared to inactivate V beta 8.2+ T cells already present in the CNS. In both cases, TCR peptide-induced inhibition of the encephalitogenic T cells apparently preempted the need for CD45R hi/CD4+ T cells that may normally be necessary to resolve the disease.

Multiple sclerosis: the frequency of allelic forms of tumor necrosis factor and lymphotoxin-alpha.

The cytokines LTa and TNF have been implicated as major mediators of tissue injury in multiple sclerosis (MS). In this study we have assessed the frequency of specific polymorphisms for these genes in MS (n = 53) and controls (n = 81) using a highly sensitive, two stage nested polymerase chain reaction (PCR), with the second stage using mutation-specific primers. Genomic DNA was extracted from blood cells and the results confirmed by direct dideoxy chain termination sequencing. The frequency of the -308 G to A mutation in the TNF promoter region in normal controls was 15% and in MS was 24%. For LTa gene the exon 3 polymorphism allele A was detected in 36% of controls and 34% of the MS patients. In MS, the combined genotype TNF G + A and LTa C + C was present 6 times more frequently (12%) than in controls (2%), and patients with this genotype showed the highest EDSS scores. We found the TNF and LTa polymorphisms to occur independently from the HLA class II DR2 allele distribution in MS. Whilst the G - A polymorphism in TNF gene promoter has been studied previously in MS, with conflicting results, this is the first study that has addressed the exon 3 polymorphism in LTa in MS. The results indicate that this polymorphism is not linked with the higher genetic predisposition for MS, but that combined TNF G + A and LTa C + C genotype might contribute to development of the disease.

IL-7 enhances Ag-specific human T cell response by increasing expression of IL-2R alpha and gamma chains.

Interleukin-7 has demonstrated potent enhancing effects on the growth and differentiation of several immature cell types, including thymocytes, and on survival of resting and antigen activated T cells. In this study, we evaluated the effects of IL-7 on post-thymic antigen-specific T cells from human blood. IL-7 was found to enhance proliferation responses and IFN-gamma secretion of myelin or recall Ag-specific Th1 cells through the selective up-regulation of the IL-2Ralpha and gamma but not beta chains in both an Ag-dependent and Ag-independent manner, but did not affect monocytes, B cells, or NK cells. These functions of IL-7 enhanced the detection of Th1 but not Th2 cell frequency by >2.5 fold, and promoted selection of Ag-specific Th1 cells by the limiting dilution method. Moreover, IL-7 pretreatment conferred increased resistance of CD4+ T cells to CD8+ cell lysis. These studies demonstrate that IL-7 promotes the growth and survival of circulating Ag-specific human Th1 cells through a mechanism that probably involves the gammac common receptor for IL-2 family members that includes IL-7.

Maturation of the antibody response to a protein-coupled form of the organophosphorus toxin soman.

We have analysed the serum antibody response of BALB/c mice to the organophosphorus toxin soman coupled to the protein carrier keyhole limpet haemocyanin (So-KLH) and compared the specificity of the serum antibodies to that of hybridomas described previously. The relative inhibitory capacities of various soman analogues for serum antibodies correlated with those for the monoclonal antibodies. Our results also demonstrate that immune memory to this organophosphorus hapten is stable for greater than 1 year. Interestingly, maturation of the serum antibody response is accompanied by fine specificity changes resulting in increased binding to soman-protein conjugates but not in significant changes in binding to free hapten analogues of soman. This finding suggests that contributions made by the protein carrier or bridge structure, including those made by amino acid side chains involved in the linkage, may play a significant role in the maturation process of antibodies recognizing protein-coupled organophosphorus haptens such as So-KLH. Structurally related but charge-dissimilar organophosphate haptens such as nitrophenylphosphocholine were poorly recognized, even when conjugated to protein with the same diazophenyl linkage used to conjugate soman. This is consistent with maintenance of high specificity in the memory immune response to soman-coupled protein.

Molecular analysis and fine specificity of antibodies against an organophosphorus hapten.

We have identified four fine specificity groups reactive with the organophosphorus hapten Soman among 46 hybridomas generated in specific response to immunization with Soman-keyhole limpet hemocyanin (KLH). The different fine specificity groups do not appear to correlate with the use of particular V genes. Molecular analysis of VH genes demonstrates predominant use of VH J558 family members in hybridomas of all fine specificity groups although several different VH genes within this family as well as others are able to contribute. Diversity of VH gene usage was also apparent in primary IgM-producing hybridomas. In contrast, there appears to be restricted L chain usage; a large number (18/46) used the V kappa 1 family. Interestingly, the V kappa 1 family also plays an important role in the memory response to phosphocholine (PC)-KLH, a related organophosphate hapten which shares several structural features with Soman, particularly when coupled to protein carriers. The V kappa 1 C gene appears to predominate in the PC-KLH response. Restriction analysis of DNA from the V kappa 1-positive Soman-KLH-specific hybridomas suggests that a single V kappa 1 gene may be utilized by 17/18 but that this gene is different from V kappa 1 C and may be V kappa 1 A. We propose that members of the V kappa 1 family contribute favorably in generating combining sites that recognize all or part of the structural features shared by the two haptenic structures Soman and PC when they are coupled to protein. This most likely involves recognition of the phenyl linker moiety as the dominant feature. It appears that the L chain rather than the H chain may play a more significant role in forming the phenyl-Soman-specific combining site and perhaps the combining sites for phenyl or ring structures in general.

Some characteristics of thymus suppression of antibody production in vitro in Xenopus laevis, the South African clawed toad.

Organ cultures were used to test whether thymus suppression of antibody production by spleen cells of Xenopus laevis is antigen inducible, antigen specific, or genetically restricted when a thymus-dependent (TD) antigen, horse erythrocytes (HRBC), is injected in vivo. The capacity of mammalian thymus-independent (TI) antigens, e.g. TNP-Ficoll (TI-2) and TNP-LPS (TI-1) to induce thymus suppression of the anti-HRBC) response was also tested. Additionally, the susceptibility of TI spleen responses to thymus suppression was studied. Auto- and allogeneic thymus and spleen combinations were compared to spleen cultured alone with respect to hemagglutinin titers secreted into the culture medium. The results suggest that thymus suppression of a TD response in this species is antigen inducible. Moreover, depending on the antigen used either antigen specific or unspecific suppression will be generated. Both can function across allogeneic boundaries. Responses to TI antigens are not subject to thymus suppression.

Coculture of TCR peptide-specific T cells with basic protein-specific T cells inhibits proliferation, IL-3 mRNA, and transfer of experimental autoimmune encephalomyelitis.

TCR peptides, namely V beta 8.2-39-59 or the minimal idiotope, V beta 8-44-54, can treat experimental autoimmune encephalomyelitis (EAE) in Lewis rats, presumably by activating naturally induced TCR peptide-specific T cells that arise in response to the focused appearance of V beta 8.2+ encephalitogenic T cells. The purpose of the present study was to evaluate the mechanisms by which TCR peptides inhibit EAE. We found that treatment of EAE with the V beta 8.2-39-59 peptide did not induce any evidence of DNA fragmentation (apoptosis) in spinal cord cells isolated from clinically well rats, implicating a regulatory rather than a deletional mechanism. TCR peptide-specific T cell lines failed to inhibit EAE induced by already activated BP-specific T cells when the two T cell specificities were co-injected. However, coculturing the encephalitogenic T cells in the presence of the regulatory T cells during the activation step before transfer almost completely inhibited the induction of EAE. Inhibition could be induced by direct contact between the two cell types or by soluble factors produced in a transwell system, but was greatly enhanced when soluble V beta 8.2-39-59 peptide was used to optimally activate the regulatory T cells. The inhibition was regulatory cell dose dependent, and was reflected in vitro by reduced proliferation response and mRNA production for IL-3, and to a lesser extent, IFN-gamma and IL-2. These results indicate that regulation induced by TCR peptides involves cell-cell interactions that lead to the production and release of soluble factors that locally inhibit the activation of encephalitogenic T cells expressing MHC-bound idiotopes of the target V beta-chain, and possibly "bystander" specificities expressing different V beta-chains.

Analysis of V beta 8.2 CDR3 sequences from spinal cord T cells of Lewis rats vaccinated or treated with TCR V beta 8.2-39-59 peptide.

Experimental autoimmune encephalomyelitis (EAE) in the Lewis rat can be induced with the administration of Gp-BP. This disease appears to be mediated at least in part by V beta 8.2+CD4+T cells, which specifically recognize the BP72-89 encephalitogenic peptide. Treatment or protection with V beta 8.2 CDR2 39-59 peptide can suppress or prevent clinical signs of EAE, presumably through the activation of regulatory T cells. Interestingly, V beta 8.2+ T cells continue to persist in the spinal cord of protected animals, although their appearance in the central nervous system (CNS) is delayed when compared with control animals with EAE. As part of our effort to elucidate the mechanism(s) of peptide protection and therapy, we sought to determine whether the V beta 8.2+ T cells in the spinal cord of protected or treated rats were truly representative of those found in rats with clinical EAE. Therefore, we examined the following CNS samples for the Asp96Ser97 motif, which has been identified previously in V beta 8.2+ BP-specific, encephalitogenic T cell clones: 1) rats protected with V beta 8.2-39-59 peptide, 2) rats treated with V beta 8.2-39-59 peptide, and 3) control rats with EAE. Our findings indicate that EAE-associated V beta 8.2+ sequences can still be found in both peptide-treated and peptide-protected rats. It appears that administration of V beta 8.2 CDR2 peptide does not prevent EAE-associated V beta 8.2+ T cells from infiltrating the CNS and that other mechanisms are at work to prevent the development of EAE.

Myelin basic protein-specific and TCR V beta 8.2-specific T-cell lines from TCR V beta 8.2 transgenic mice utilize the same V alpha and V beta genes: specificity associated with the V alpha CDR3-J alpha region.

Our analysis of TCR V gene usage in mice transgenic for the V beta 8.2 gene has demonstrated that T cells isolated from the spinal cord of these transgenic mice during active experimental autoimmune encephalomyelitis were significantly biased for V alpha 2 expression. This V alpha 2 bias was noted in T cells derived from the periphery as well but only after stimulation with specific antigen. Thus, spinal cord-derived pathogenic T cells had already undergone activation and expansion within the central nervous system environment of these mice. As part of an investigation of regulatory function in these V beta 8.2 transgenic mice, two T cell lines were selected. The first T cell line is encephalitogenic and specific for the dominant myelin basic protein peptide NAc1-11, while the second T cell line is specific for the V beta 8.2 protein. Surprisingly, polymerase chain reaction and sequence analysis of the TCR from both the T cell lines demonstrated that they utilize identical V beta, D beta, J beta, and V alpha gene segments. The only difference found was in their use of the J alpha gene segment, indicating that this region of the TCR molecule can play a key role in determining antigen specificity.

V beta CDR3 motifs associated with BP recognition are enriched in OX-40+ spinal cord T cells of Lewis rats with EAE.

Spinal cord (SC) T cells were isolated at the onset of actively induced experimental autoimmune encephalomyelitis (EAE) and sorted for the presence of the OX-40 activation marker. Previously, we reported an enhanced bias in V beta 8.2 expression as well as enhanced proliferative responses to basic protein antigens among the OX-40+ SC T cells. Here we demonstrate that CDR3 motifs associated with EAE are present at a significantly higher frequency in V beta 8.2 sequences of OX-40+ SC T cells (16/17) compared with those of OX-40- SC T cells (5/17). Thus, the OX-40 antigen may be useful as a marker to isolate and characterize autoantigen-specific T cells from the site of inflammation in T-cell-mediated autoimmune diseases.

Definition of encephalitogenic and immunodominant epitopes of guinea pig myelin basic protein (Gp-BP) in Lewis rats tolerized neonatally with Gp-BP or Gp-BP peptides.

Two distinct epitopes of guinea pig basic protein (Gp-BP), residues 72-89 and 87-99, possess encephalitogenic activity in Lewis rats. The purpose of this study was to determine to what degree the 87-99 epitope functions in rats that have been injected with whole Gp-BP, and whether additional epitopes in Gp-BP are encephalitogenic. To address these questions, we induced neonatal tolerance to the dominant synthetic (S)72-89 peptide or to the combination of both S72-89 and S87-99 peptides, and evaluated resistance to experimental autoimmune encephalomyelitis (EAE) induced by Gp-BP, as well as T cell responses to peptides that encompassed most of the Gp-BP molecule. The results demonstrated that virtually all of the encephalitogenic activity of Gp-BP resides within the two described encephalitogenic epitopes. Moreover, deletion of responses to the dominant epitopes prompted T cell responses to other nonencephalitogenic epitopes of Gp-BP, a pattern of response observed previously in rats that had recovered from EAE and in those protected from EAE by vaccination with TCR peptides. These data may have relevance to human autoimmune diseases such as multiple sclerosis in that naturally or immunologically regulated responses to dominant epitopes that are likely to be encephalitogenic may be obscured by increased responses to relatively innocuous determinants of basic protein. Elevated responses to potentially pathogenic autoantigens will likely involve both types of determinants, thus, underscoring the importance of distinguishing encephalitogenic from nonencephalitogenic determinants.

Analysis of V beta 8-CDR3 sequences derived from central nervous system of Lewis rats with experimental autoimmune encephalomyelitis.

We have recently demonstrated that a strong bias for expression of V beta 8.2 is manifested early during the onset of experimental autoimmune encephalomyelitis (EAE) induced by guinea pig basic protein (Gp-BP) immunization of Lewis rats. More importantly, the V beta 8.2 bias was observed in T cells infiltrating the spinal cord (SC) and in cerebrospinal fluid (CSF), but was not present in T cells isolated from the periphery. Here, we report the V beta 8-CDR3 sequences found in unselected SC, CSF, and lymph node (LN) T cell populations at onset of Gp-BP-induced EAE. Striking similarities were observed among sequences derived from SC and CSF. Evidence for oligoclonal expansion of V beta 8.2 sequences associated with previously characterized encephalitogenic clones was observed in both SC and CSF, but not in LN. An AspSer CDR3 motif identified in encephalitogenic clones recognizing the dominant 72-89 epitope of Gp-BP was found in 9/22 SC cDNA clones, 11/24 CSF cDNA clones, and 1/16 LN cDNA clones. Interestingly, J beta 2.7 and J beta 1.3 were also highly represented in SC and CSF, but not in LN. Given that these sequences were derived from T cells present at the site of autoimmune attack and not selected by in vitro manipulation, the data offer compelling evidence that 1) selective recruitment and/or expansion of V beta 8.2+ T cells are occurring in the central nervous system; 2) these events are at least partially dependent on V beta residues which are likely to influence Ag binding; and 3) CSF-derived T cells provide a representative view of CNS events at the onset of EAE.

Suppression in Xenopus laevis: thymus inducer, spleen effector cells.

Studies were carried out on suppressor function in the amphibian Xenopus laevis, the South African clawed toad. Suppression by the thymus of haemagglutinin (HA) production by spleen is antigen-dependent, partially specific and not MHC-restricted in this species (Ruben, Buenafe & Seivert, 1983). Three questions were considered in this study. Does the thymus effect suppression by stimulating peripheralized spleen effector cells, or do effector cells reside within the thymus? Do macrophages participate in the induction and/or expression of thymus-dependent suppressor function? Can thymus suppressor and helper functions be distinguished by using irradiation treatment? The capacity of immunized thymus to suppress HA when co-cultured with spleen fragments from immunized, cyclophosphamide (CyP)-injected animals was tested. Immunized thymus failed to suppress the high levels of HA production by spleen fragments from CyP-treated, immunized donors. Colloidal carbon injection resulted in blockade of macrophage function, and both the capacity of thymuses to suppress and of spleens to be suppressed in co-cultures. Finally, the effect of thymus exposure to gamma-irradiation in vitro was tested using autogeneic thymus/spleen combinations. This enabled the visualization of thymic helper function, which is MHC-restricted in Xenopus (Bernard et al., 1981). Four dosages of irradiation were tested after antigen challenge. The highest HA titres were produced by spleen co-cultures with thymuses which had received 1000 rads. We conclude that suppression of HA production in spleen is not the result of thymus suppressor effector cells, but that suppressor function is mediated by thymus inducer cells which stimulate suppressor effectors in spleen. Both the thymic inducers and effectors in the spleen are sensitive to CyP and macrophage blockade. Our studies further suggest that we are able to distinguish between the thymic functions of help and suppression in Xenopus by taking advantage of their differential sensitivities to irradiation. While it has been postulated, on other grounds, that suppression was one of the earliest thymic regulatory functions to have evolved (L.N. Ruben & R.H. Clothier, submitted), here we suggest the presence of sequential activities of more than one cellular subset, as early in evolution as the primitive anuran (tail-less) amphibia.

A TCR V alpha CDR3-specific motif associated with Lewis rat autoimmune encephalomyelitis and basic protein-specific T cell clones.

To investigate TCR V alpha gene expression in the Lewis rat model of experimental autoimmune encephalomyelitis, we obtained V alpha chain sequences from two V beta8.2+-encephalitogenic, BP72-89-specific T cell clones. Two different V alpha genes, a V alpha2 gene and a V alpha23 gene, are utilized, but both were found to contain an asparagine repeat (Asn3+) sequence present in the V alpha CDR3 region. This Asn3+ motif is also present in the previously reported sequence of a BP68-88-specific hybridoma, 510, which utilizes a different V alpha2 gene family member. In further experiments, spinal cord T cells were isolated at the onset of basic protein (BP)-induced disease and sorted for the OX-40 activation marker, which we have previously used to enrich for specifically activated T cells. Analysis of V alpha expression in the OX-40+ population revealed the biased use of three V alpha genes, V alpha1, V alpha2, and V alpha23. The Asn3+ motif was present in the V alpha CDR3 region of V alpha1, V alpha2, and V alpha23 cDNA derived from OX-40+ spinal cord T cells but found to be generally absent in the OX-40- spinal cord population. Since these Asn3+ motif-bearing V alpha chain sequences are nearly identical to those utilized by the BP-specific encephalitogenic clones described, it is likely that these V alpha sequences are derived from disease-associated T cells in the spinal cord. Thus, we demonstrate that the Asn3+ V alpha CDR3 motif is strongly associated with experimental autoimmune encephalomyelitis in the Lewis rat and propose that it plays a role in TCR recognition of a specific BP peptide/MHC complex.

In vitro thymus suppression of hemagglutinin production in Xenopus laevis: location, drug and temperature sensitivity.

Suppression of splenic hemagglutinin (HA) production by allothymuses in vitro was further studied in Xenopus laevis, the South African clawed toad. Since thymic capacity to suppress splenic HA secretion into the culture medium is retained in animals previously exposed to N-methyl-N-nitrosourea (NMU), which destroys the thymus cortex in this species, suppressor function must be located in the thymus medulla. Reciprocal thymus-spleen combinations showed that normal thymus can suppress immunised spleen fragments from NMU-treated animals. Since Xenopus exposed to NMU are also devoid of helper function, thymus suppression acts directly on the antibody forming system. Cyclophosphamide removed thymic suppression and enhanced spleen fragment antibody production, while Cyclosporin A enhanced thymic suppression and blocked HA production. Therefore, antigenic stimulation of thymic suppression and antibody production are inversely related in this species. Long-term storage of Xenopus in the cold increased lymphoid cellularity of the thymus medulla, but thymuses from cold-stored animals could not suppress normal spleen fragments and their spleens were not suppressed by normal thymuses. Thus, ectotherms may retain immunological capacity when subjected to prolonged cold by a loss of thymic suppression.