Peripheral tolerance in T cell receptor-transgenic mice: Evidence for T cell anergy

T cell tolerance can be induced in adult mice by injection of soluble antigenic peptide. The underlying mechanism has been difficult to establish in normal mice due to the low precursor frequency of T cells specific for any given antigen. Therefore, we examined peripheral tolerance in mice transgenic for a T cell receptor specific for a cytochrome c peptide bound to I-E^k. Antigen-specific hyporesponsiveness could be induced in the transgenic mice. We followed the transgene-bearing T cells with a clonotypic monoclonal antibody and found similar numbers of clonotypic T cells in tolerized and control mice. To prevent de novo differentiation of T cells we analyzed thymectomized mice in which antigen-specific hyporesponsiveness was induced. Our analysis of thymectomized transgenic mice showed that antigen-specific T cell hyporesponsiveness following injection of peptide intravenously is not caused by gross elimination of T cells. These data provide evidence for the role of anergy in peripheral tolerance.     

Peripheral clonal deletion of antiviral memory CD8+ T cells

Antiviral cytotoxic memory CD8⁺ T cells adoptively transferred to mice which are persistently infected with lymphocytic choriomeningitis virus WE or DOCILE initially proliferated extensively; they either caused the death of the recipient or, alternatively, disappeared within a few days. Apparently, the complete and coordinated induction and stimulation by widely distributed viral antigen caused these memory T cells to die before virus had been eliminated from the host. Thus memory T cells are as susceptible to peripheral exhaustion/deletion as unprimed T cells. These results indicate possible limitations of exclusively CD8⁺ T cell-mediated adoptive immunotherapy against viral infections or tumors.     

Transient gain of effector function by CD8+ T cells undergoing peripheral tolerance to high-dose self-antigen

Induction of peripheral T cell tolerance is mediated by bone marrow-derived dendritic cells that cross-present self-antigen to self-reactive T cells. The current model for peripheral CD8(+) T cell tolerance is that TCR engagement by self-antigen in the absence of costimulation results in abortive activation without development of effector function. Here we demonstrate in vivo that high-dose self-antigen ("signal 1") can compensate for lack of costimulation ("signal 2"), leading to full activation of and development of effector function by self-reactive T cells. In the setting of low-dose self-antigen, acquisition of effector function by self-reactive T cells is dependent on costimulation via CD40 ligation in vivo. However, gain of effector function in either setting does not prevent eventual tolerance of self-reactive CD8(+) T cells. These results suggest that the mechanisms for peripheral CD8(+) T cell tolerance are more complex than the proposed "signal 1 in the absence of signal 2" hypothesis. Further exploration of these mechanisms will have direct impact on the design of effective immunotherapy for autoimmune diseases, chronic infections and cancers.     

Induction of CD4+ T cell depletion in mice doubly transgenic for HIV gp120 and human CD4

It has been suggested that loss of uninfected T cells in HIV infection occurs because of lymphocyte activation resulting in cell death by apoptosis. To address the question of whether cross-linking of CD4/HIV gp120 complexes by antibodies were sufficient to induce T cell depletion in vivo, we developed an animal model of continuous interaction between human CD4 (hCD4), gp120 and anti-gp120 antibodies in the absence of other viral factors. Double-transgenic mice have been generated in which T cells express on their membrane hCD4 and secrete HIV gp120. Although these mice have hCD4/gp120 complexes present on the surface of T cells, they do not show gross immunological abnormalities, and they are able to produce anti-gp120 antibodies following immunization with denaturated gp120. However, double-transgenic mice with antibodies to gp120, when immunized with tetanus toxoid, mount an IgG response that is significantly lower than that of double-transgenic mice without antibodies to gp120. Furthermore, the presence of anti-gp120 antibodies leads to CD4⁺ T cell depletion and immunodeficiency in the absence of HIV infection. Thus, the antibody response to gp120 can lead to CD4⁺ T cell attrition in vivo.     

Identification of an HLA-DQ6-derived peptide recognized by mouse MHC class I H-2Db-restricted CD8+ T cells in HLA-DQ6 transgenic mice

Summary CD8⁺ T cells from C57BL/6(B6) mice show cytotoxicity to B cell blasts prepared from syngeneic transgenic mice expressing HLA-DQ6 molecules in a mouse MHC class I H-2D^b restricted manner. Although these results suggest that CD8⁺ T cells recognize peptides derived from DQ6 molecule bound to H-2D^b on target cells, no direct evidence so far has been obtained. To clarify this, we synthesized 23 peptides corresponding to DQ6α orβ chain and carrying the motifs of D^b-binding peptides, and examined their capacity to induce cytotoxicity in the CD8⁺ T cell line. We show here that DQA1-2, one of these peptides, induced cytotoxicity of the CD8⁺ T cells when this peptide was pulsed to H-2D^b expressing target cells, as efficiently as HLA-DQ6 expressing target cells did. Thus, our results suggest that DQA1-2 can be naturally processed from DQ6 molecules and recognized by the CD8⁺ T cells in the context of H-2D^b molecules. These results suggest that allogeneic HLA class II molecules are involved in the rejection not only as the ligand for T cell receptor of alloreactive CD4⁺ T cells but also as self-peptides bound to HLA class I molecules recognized by CD8⁺ T cells.     

CD4+8− help prevents rapid deletion of CD8+ cells after a transient response to antigen

We have followed the fate of mature CD8⁺ T cells with a male-specific transgenic T cell receptor after antigenic stimulation with hemopoietic cells in the absence or presence of help. Our data show that mature CD8⁺ T cells can be deleted after a 3-week period of transient activation and that help, e.g. in the form of interleukin-2, can considerably delay the deletion. These experiments have implications for the design of protocols aiming at the establishment of specific immunological tolerance in T cells.     

Tolerance is overcome in beef insulin-transgenic mice by activation of low-affinity autoreactive T cells

To gain insight into the factors controlling the maintenance or loss of T cell self tolerance we produced beef insulin (BI)-transgenic BALB/c mice. Transgenic mice express BI under control of the human insulin promoter and secrete physiological amounts of beef insulin. Although these mice are tolerant to BI, as evidenced by the lack of insulin-specific IgG antibody production following intraperitoneal immunization, tolerance is not complete. Footpad immunization results in a weak antigen-specific T cell proliferative response, indicating the presence of self- reactive BI-specific T cells in the periphery. These T cells are functional in vivo, providing support for IgG1, IgG2a, and IgG2b BI-specific antibody production, but require higher concentrations of antigen than nontransgenic T cells (both in vivo and following recall responses in vitro) to become activated. In vitro, BI-specific T cell proliferation in BI-transgenic mice can be largely restored by addition of interleukin-2, indicating that a significant component of T cell tolerance is mediated by anergy. To characterize the autoreactive T cells that become activated when tolerance is broken, BI-specific T cell hybridomas were generated from transgenic mice and compared to a panel of hybridomas previously derived from nontransgenic BALB/c mice. The majority of BI-transgenic hybridomas recognized the immunodominant A1–14 beef insulin peptide but with lower avidity than BALB/c hybridomas. Consistent with this, none of the dominant T cell receptor rearrangements found in the BALB/c BI-specific T cell receptor repertoire were found in the transgenic hybridomas. These results indicate that, despite evidence for clonal inactivation of many BI-specific T cells in BI-transgenic mice, loss of tolerance results from activation of low-affinity antigen-specific T cells that appear to have escaped this process.     

Tolerance induction by clonal deletion of CD4+8+ thymocytes in vitro does not require dedicated antigen-presenting cells

The cellular requirements of T cell tolerance induction in the thymus by clonal deletion was investigated by using an in vitro assay: thymocytes from mice expressing a transgenic TcR specific for lymphocytic choriomeningitis virus (LCMV) and H-2D^b were co-cultured with various H-2^b cell types as antigen-presenting cells in the presence of the antigenic LCMV peptide. The results revealed that all cell lines examined including embryonic and transformed fibroblasts, melanoma cells, cortical thymic epithelial cells, lymphomas and neuronal cells induced an antigen dose-dependent deletion of CD4⁺8⁺ thymocytes. Similarly, highly enriched accessory cell populations from thymus and spleen (macrophages, dendritic and cortical epithelial cells, i.e. thymic nurse cells) could induce antigen-specific depletion of immature CD4⁺8⁺ thymocytes. Depending on the cell type examined micromolar to picomolar concentration of LCMV peptide were required to induce deletion. The effectiveness of deletion by the different cell types did not correlate with their major histocompatibility class I expression level; it was, however, influenced by the presence of ICAM-1 adhesion molecules.     

CD8+ regulatory T cells are responsible for GAD-IgG gene-transferred tolerance induction in NOD mice

Our previous studies demonstrated that lipopolysaccharide (LPS)-stimulated splenocytes, retrovirally transduced with a glutamate decarboxylate 65 (GAD) and immunoglobulin G (IgG) fusion construct, can protect non-obese diabetic (NOD) mice from diabetes by inducing GAD-specific tolerance, and also that there are increased numbers of CD4(+) regulatory T cells (Tregs) in GAD-IgG-treated NOD mice. However, little is known about the role of CD8(+) Tregs in GAD-IgG gene-transferred tolerance induction in NOD mice. Here, we found that GAD-IgG-transduced splenocytes induced an increase in the number of CD8(+) Foxp3(+) Tregs in vitro. Using a T-cell depletion assay, we found that, compared with undepleted groups, NOD recipients transfused with CD8(-) or CD8(-) CD25(-) GAD-IgG-transduced splenocytes showed a decrease in the percentage of CD8(+) Foxp3(+) T cells, a high incidence of diabetes, serious insulitis, GAD-specific hyperresponsiveness at both the cellular and humoral levels, and changes in cytokine expression. These results indicate that CD8(+) Tregs, which were induced in vitro by GAD-IgG-transduced splenocytes, were also responsible for GAD-IgG gene-transferred tolerance induction in NOD mice.     

Alloreactive cytotoxic T cells can develop and function in mice lacking both CD4 and CD8

Using the technique of homologous recombination in embryonic stem cells, a mouse strain without functional CD4 and CD8 genes has been generated. Surprisingly, these mice contain significant numbers of αβ T cells. Although mice deficient for CD8 only do not show any cytotoxic response when their T cells are stimulated with either alloantigen or viral antigen, the CD4^?8^? mice do generate alloreactive cytotoxic T cells. These cytotoxic T cells bear the αβ T cell receptor and recognize major histocompatibility complex class I antigens. In addition fully allogeneic skin transplants were rejected but skin transplants expressing only minor transplantation antigens were not. Virus-specific cytotoxic T lymphocytes were also not detected. It seems that alloreactive cytotoxic T cells can be induced and exert their effector function in vitro and in vivo in the absence of CD8, and that they can develop and mature in vivo without the CD8 molecule or the signals it might provide.     

Deletion of CD4+ T cells and thymocytes by apoptosis in mouse mammary tumor virus (C4)-infected Vβ2 transgenic mice

Mouse mammary tumor virus MMTV(C4) encodes a Vβ2-specific superantigen. In Vβ2 transgenic (TG2) mice more than 98 % of peripheral T cells express Vβ2. Infection of Tg2 mice with MMTV(C4) at birth through their mothers' milk or at 6–8 weeks of age by intravenous injection resulted in massive deletion of peripheral CD4⁺ T cells and suppressed thymopoiesis. The number of peripheral CD8⁺ T cells was not affected in neonatally infected mice. In older mice injected with MMTV(C4), splenic CD8⁺ T cells were significantly elevated. Suppressed thymopoiesis was observed in both neonatally infected and older mice injected with MMTV(C4). Thymocytes which expressed high level CD3 or Vβ2 were deleted. To determine if T cells or thymocytes were deleted through apoptosis, DNA fragmentation was examined by flow cytometry and diphenylamine (DPA) binding assay. Approximately 31 % of CD4⁺ T cells from MMTV(C4)-infected Tg2 mice as compared to 6% from normal Tg2 mice contained fragmented nuclear DNA by flow-cytometric analysis. The DPA binding assay showed significantly increased total soluble DNA in lymph node cells and thymocytes from MMTV(C4)-infected mice. The kinetics of T cell and thymocyte apoptosis correspond to their deletion, supporting apoptosis as the mechanism of T cell and thymocyte deletion. CD4⁺ T cell and thymocyte deletion by MMTV(C4) in Tg2 mice provides a sensitive system for the analysis of retrovirus superantigen-induced apoptosis.     

T cell repertoire and clonal deletion of Mtv superantigen-reactive T cells in mice lacking CD4 and CD8 molecules

CD4^?CD8^? double-negative T cells constitute a lymphocyte subpopulation within the thymus and peripheral lymphatic organs that express a unique T cell receptor (TCR) repertoire and do not undergo negative selection. To test whether these cells develop as a distinct lineage or due to altered selection in the absence of CD4 and CD8 expression, we analyzed the TCR repertoire in mice lacking both CD4 and CD8 accessory molecules after homologous recombination (CD4^0/0CD8^0/0). We show that mature T cells of CD4^0/0CD8^0/0 mice express an unbiased diverse TCR Vβ repertoire comparable to wild type mice. In addition, clonal deletion of mouse mammary tumor virus superantigen-reactive T cells did occur in CD4^0/0CD8^0/0 mice. These data show that the intrinsic lack of CD4 and CD8 expression has no effect on the mature TCR repertoire and that clonal deletion of superantigen-reactive cells is independent of CD4 and CD8 co-receptors.     

Human CD8{alpha} expression in NK cells but not cytotoxic T cells of transgenic mice

In our previous work, DNase hypersensitivity mapping was usedto identify an enhancer within the human CD8 (hCD8) gene whichallowed T cell-specific expression of a reporter construct intransiently transfected cell lines. To study the role of thisintronic enhancer in vivo, transgenic mice were made using humanCD8 genomic constructs. We found that while a 14 kb wild-typehuman CD8a (WThCD8) genomic construct did not lead to expressionin mature peripheral CD8⁺ T cells, this transgene was consistentlyexpressed in small populations of T cells and B cells, and ina subset of mouse NK cells. While murine CD8 is not normallyexpressed on resting NK cells, expression of the human CD8 transgeneon mouse NK cells is appropriate since CD8 is expressed on asubset of human NK cells. Deletion of the intronic enhancerresulted in a complete loss of transgene expression in mostlines and a loss of expression only in NK cells in one line.Our results indicate, firstly, that cis-acting sequences withinthe 14 kb genomic fragment are sufficient for NK cell-specificexpression. In addition, our results suggest that the enhancermay have dual roles in regulation of transgene expression. Itmay enhance general expression of the transgene and may alsobe required for NK cell-specific expression.     

Threshold tolerance in H-2Kb-specific TCR transgenic mice expressing mutant H-2Kb: conversion of helper-independent to helper-dependent CTL.

To evaluate the role of the structure of the class I molecule and associated peptide(s) in intrathymic selection and tolerance, mice expressing as a transgene (tg) a TCR specific for the H-2Kb alloantigen were crossed with mice expressing the mutant class I molecule H-2Kbm1 or H-2Kbm8. In H-2k/k TCR tg mice (in a situation of exclusive positive selection), peripheral tg TCR expressing (Ti+) CD8+ T cells showed high, suboptimal, and an absence of reactivity for H-2Kb, H-2Kbm1, and H-2Kbm8, respectively. In the peripheral lymphoid organs of TCR tg H-2k/k, H-2k/bm8, H-2k/bm1, and H-2k/b mice respectively, the tg TCR was expressed on T cells with decreasing intensity of surface CD8. Thymic subpopulations of TCR tg mice presented a pattern of negative selection with decreasing intensity from H-2k/b to H-2k/bm1 and H-2k/bm8. This suggests that a weak interaction between the TCR and H-2Kbm8 exists which partially results in negative, but not in positive, intrathymic selection. Results further indicate that expression of H-2Kbm8 does not induce tolerance to H-2Kb. In H-2k/bm1 mice, the peripheral Ti+ CD8lo cells express two distinct types of 'threshold' tolerance in vitro: (i) they generate cytotoxic T lymphocytes (CTL), in the presence of exogenous IL-2, which fail to respond to H-2Kbm1 but remain reactive to H-2Kb; and (ii) they do not make significant titers of IL-2 and do not significantly proliferate in response to H-2Kb, unlike the Ti+ CD8+ T cells from H-2k/k TCR tg mice which respond efficiently. These results show that tolerance is induced up to a level of non-reactivity within a given MHC environment: for the same TCR, CTL reactivity to H-2Kbm1 is totally lost, whereas CTL reactivity to H-2Kb is only slightly reduced. Additionally, proliferation and IL-2 production by Ti+ CD8+ cells in response to H-2Kb were strongly affected in H-2k/bm1 mice. Thus, in H-2k/k mice the Ti+ CD8+ cells behave as helper-independent, whereas in H-2k/bm1 mice CD8+ cells expressing the same TCR behave as helper-dependent CTL.     

Role of autoreactive CD8+ T cells in organ-specific autoimmune diseases: insight from transgenic mouse models

Summary: There is now convincing evidence that autoreactive CD8⁺ T cells can contribute to the pathogenesis of organ-specific autoimmune diseases. In the non-obese diabetic mouse, there is direct evidence that β-islet cell-specific CD8⁺ cytotoxic T cells have a pathogenic effect. In human diseases such as autoimmune diabetes and multiple sclerosis, indirect evidence also suggests a role for CD8⁺ T cells in tissue damage, although their antigen specificity is unknown. Transgenic mouse models as well as the use of knockout mite have been instrumental in the identification of the role of autoreactive CD8⁺ T cells. Spontaneous models of CD8+ T-cell-mediated autoimmunity generated through transgenesis should help delineate the effector mechanisms leading to tissue destruction. The study of autoreactive CD8⁺ T cells and the characterization of their antigenic specificity should help unravel the pathophysiology of organ-specific autoimmune diseases, help identify exacerbating foreign antigens, and allow the design of antigen-specific immunotherapy targeting the pathogenic autoreactive T cells.     

Activated CD4+ CD25+ T cells suppress antigen-specific CD4+ and CD8+ T cells but induce a suppressive phenotype only in CD4+ T cells

CD4(+) CD25(+) regulatory T cells are increasingly recognized as central players in the regulation of immune responses. In vitro studies have mostly employed allogeneic or polyclonal responses to monitor suppression. Little is known about the ability of CD4(+) CD25(+) regulatory T cells to suppress antigen-specific immune responses in humans. It has been previously shown that CD4(+) CD25(+) regulatory T cells anergize CD4(+) T cells and turn them into suppressor T cells. In the present study we demonstrate for the first time in humans that CD4(+) CD25(+) T cells are able to inhibit the proliferation and cytokine production of antigen specific CD4(+) and CD8(+) T cells. This suppression only occurs when CD4(+) CD25(+) T cells are preactivated. Furthermore, we could demonstrate that CD4(+) T-cell clones stop secreting interferon-gamma (IFN-gamma), start to produce interleukin-10 and transforming growth factor-beta after coculture with preactivated CD4(+) CD25(+) T cells and become suppressive themselves. Surprisingly preactivated CD4(+) CD25(+) T cells affect CD8(+) T cells differently, leading to reduced proliferation and reduced production of IFN-gamma. This effect is sustained and cannot be reverted by exogenous interleukin-2. Yet CD8(+) T cells, unlike CD4(+) T cells do not start to produce immunoregulatory cytokines and do not become suppressive after coculture with CD4(+) CD25(+) T cells.