Transfer of T cell surface molecules to dendritic cells upon CD4+ T cell priming involves two distinct mechanisms

Activation of CD4(+) T cells by APCs occurs by multiple Ag recognition events including the exchange of costimulatory signals and cytokines. Additionally, the T cells acquire APC-derived surface molecules. Herein, we describe for the first time the transfer of human and murine T cell surface receptors to APCs after Ag-specific interaction. This transfer occurs in two qualitatively different phases. The first group of molecules (e.g., CD2) derived from the T cell surface was transferred rapidly after 2 h of interaction, was strongly bound on the DC surface (acid wash-resistant), was strictly dependent on dendritic cell-T cell contact, and transferred independently of T cell activation. The second group, including the CD3/TCR complex, CD27, and OX40, was of intracellular origin, transferred later after 10-16 h in a cell-cell contact-independent fashion, was noncovalently bound, and was strictly dependent on Ag-specific T cell activation. Functionally, murine dendritic cells that received TCR molecules from OVA-specific CD4(+) T cells after Ag-specific interaction were less efficient in priming naive CD4(+) T cells of the same specificity without losing their ability for CD8(+) T cell stimulation, indicating that the transferred TCR molecules mask the Ag-bearing MHC II molecules, thereby reducing their accessibility to following Ag-specific CD4(+) T cells. While the first group of transferred T cell surface molecules might facilitate the detachment of the CD4(+) T cell from the dendritic cell during the early scanning phases, the second group could play an important immunomodulatory role in intraclonal competition of T cells for APC access, making the physical presence of CD4(+) T cells unnecessary.     

CD8-independent tumor cell recognition is a property of the T cell receptor and not the T cell

The CD8 coreceptor enhances T cell function by stabilizing the TCR/peptide/MHC complex and/or increasing T cell avidity via interactions with the intracellular kinases Lck and LAT. We previously reported a CD4(+) T cell (TIL 1383I), which recognizes the tumor-associated Ag tyrosinase in the context of HLA-A2. To determine whether CD8 independent tumor cell recognition is a property of the TCR, we used retroviral transduction to express the TIL 1383I TCR in the CD8(-) murine lymphoma, 58 alpha(-)/beta(-). Immunofluorescent staining of TCR-transduced cells with human TCR V beta subfamily-specific and mouse CD3-specific Abs confirmed surface expression of the transferred TCR and coexpression of mouse CD3. Transduced effector cells secreted significant amounts of IL-2 following Ag presentation by tyrosinase peptide-pulsed T2 cells as well as stimulation with HLA-A2(+) melanoma lines compared with T2 cells alone or HLA-A2(-) melanoma cells. Further analysis of TCR-transduced clones demonstrated a correlation between T cell avidity and cell surface expression of the TCR. Therefore, the TIL 1383I TCR has sufficient affinity to mediate recognition of the physiologic levels of Ag expressed by tumor cells in the absence of CD8 expression.     

Peptide-specific intercellular transfer of MHC class II to CD4+ T cells directly from the immunological synapse upon cellular dissociation

The transfer of membrane proteins from APC to T cells was initially described in the 1970s, and subsequent work has described two mechanisms of transfer: APC-derived exosomes and direct transfer of small packets, while cells remain conjugated. Using fibroblast APC expressing a GFP-tagged I-E(k) molecule with covalently attached antigenic peptide, we observed a third mechanism in live cell imaging: T cells spontaneously dissociating from APC often capture MHC:peptide complexes directly from the immunological synapse. Using two I-E(k)-restricted murine TCR transgenic T cells with different peptide specificity, we show in this study that the MHC transfer is peptide specific. Using blocking Abs, we found that MHC:peptide transfer in this system requires direct TCR-MHC:peptide interactions and is augmented by costimulation through CD28-CD80 interactions. Capture of the GFP-tagged MHC:peptide complexes correlates with an activated phenotype of the T cell, elevated CD69 with down-modulated TCR. The transferred MHC:peptide molecules transferred to the T cell are associated with molecules that imply continued TCR signaling; p56(lck), phosphotyrosine, and polarization of the actin cytoskeleton.     

Gene transfer of human TCR in primary murine T cells is improved by pseudo-typing with amphotropic and ecotropic envelopes

BACKGROUND: T cell receptor (TCR) gene therapy represents an attractive anti-cancer treatment but requires further optimization of its efficacy and safety in clinically relevant models, such as those using a tumor antigen and TCR of human origin. Currently, however, there is no consensus as to what protocol is most optimal for retroviral human TCR gene transfer into primary murine T cells, most notably with respect to virus pseudo-type. METHODS: Primary murine T cells were transduced, expanded and subsequently tested for transgene expression, proliferation and antigen-specific function. To this end, murine leukemia virus (MLV) retroviruses were produced upon transfection of various packaging cells with genes encoding either green fluorescent protein (GFP) or TCRalphabeta specific for human melanoma antigen gp100(280-288) and the helper elements GAG/POL and ENV. Next to viral pseudotyping, the following parameters were studied: T cell densities; T cell activation; the amounts of IL-2 and the source of serum used to supplement medium. RESULTS: The pseudo-type of virus produced by packaging cells critically determines T cell transduction efficiencies. In fact, MLV-A and MLV-E pseudo-typed viruses derived from a co-culture of Phoenix-A and 293T cells resulted in T cell transduction efficiencies that were two-fold higher than those based on retroviruses expressing either VSV-G, GALV, MLV-A or MLV-E envelopes. In addition, T cell densities during transduction were inversely related to transduction efficiencies. Further optimization resulted in transduction efficiencies of over 90% for GFP, and 68% for both a murine and a human (i.e. murinized) TCR. Importantly, TCR-transduced T cells proliferate (i.e. showing a log increase in cell number in a few days) and show antigen-specific function. CONCLUSIONS: We set up a quick and versatile method to genetically modify primary murine T cells based on transient production of TCR-positive retroviruses, and show that retroviral gene transfer of a human TCR into primary murine T cells is critically improved by viral pseudo-typing with both MLV-A and MLV-E envelopes.     

Transduction of SIV-specific TCR genes into rhesus macaque CD8+ T cells conveys the ability to suppress SIV replication

Background

The SIV/rhesus macaque model for HIV/AIDS is a powerful system for examining the contribution of T cells in the control of AIDS viruses. To better our understanding of CD8⁺ T-cell control of SIV replication in CD4⁺ T cells, we asked whether TCRs isolated from rhesus macaque CD8⁺ T-cell clones that exhibited varying abilities to suppress SIV replication could convey their suppressive properties to CD8⁺ T cells obtained from an uninfected/unvaccinated animal.

Principal Findings

We transferred SIV-specific TCR genes isolated from rhesus macaque CD8⁺ T-cell clones with varying abilities to suppress SIV replication in vitro into CD8⁺ T cells obtained from an uninfected animal by retroviral transduction. After sorting and expansion, transduced CD8⁺ T-cell lines were obtained that specifically bound their cognate SIV tetramer. These cell lines displayed appropriate effector function and specificity, expressing intracellular IFNγ upon peptide stimulation. Importantly, the SIV suppression properties of the transduced cell lines mirrored those of the original TCR donor clones: cell lines expressing TCRs transferred from highly suppressive clones effectively reduced wild-type SIV replication, while expression of a non-suppressing TCR failed to reduce the spread of virus. However, all TCRs were able to suppress the replication of an SIV mutant that did not downregulate MHC-I, recapitulating the properties of their donor clones.

Conclusions

Our results show that antigen-specific SIV suppression can be transferred between allogenic T cells simply by TCR gene transfer. This advance provides a platform for examining the contributions of TCRs versus the intrinsic effector characteristics of T-cell clones in virus suppression. Additionally, this approach can be applied to develop non-human primate models to evaluate adoptive T-cell transfer therapy for AIDS and other diseases.     

Dual T cell receptor expressing CD8+ T cells with tumor- and self-specificity can inhibit tumor growth without causing severe autoimmunity

The engineering of Ag-specific T cells by expression of TCR genes is a convenient method for adoptive T cell immunotherapy. A potential problem is the TCR gene transfer into self-reactive T cells that survived tolerance mechanisms. We have developed an experimental system with T cells that express two TCRs with defined Ag-specificities, one recognizing a tumor-specific Ag (LCMV-gp(33)), the other recognizing a self-Ag in the pancreas (OVA). By using tumor cells expressing high and low amounts of Ag and mice expressing high and low levels of self-Ag in the pancreas (RIP-OVA-Hi and RIP-OVA-Lo), we show that 1) tumor rejection requires high amount of tumor Ag, 2) severe autoimmunity requires high amount of self-Ag, and 3) if Ag expression on tumor cells is sufficient and low in the pancreas, successful adoptive T cell therapy can be obtained in the absence of severe autoimmunity. These results are shown with T cells from dual TCR transgenic mice or T cells that were redirected by TCR gene transfer. Our data demonstrate that the approach of adoptively transferring TCR redirected T cells can be effective without severe side effects, even when high numbers of T cells with self-reactivity were transferred.     

The CD3-zeta chimeric antigen receptor overcomes TCR Hypo-responsiveness of human terminal late-stage T cells

Adoptive therapy of malignant diseases with tumor-specific cytotoxic T cells showed remarkable efficacy in recent trials. Repetitive T cell receptor (TCR) engagement of target antigen, however, inevitably ends up in hypo-responsive cells with terminally differentiated KLRG-1(+) CD57(+) CD7(-) phenotype limiting their therapeutic efficacy. We here revealed that hypo-responsiveness of CMV-specific late-stage CD8(+) T cells is due to reduced TCR synapse formation compared to younger cells. Membrane anchoring of TCR components contributes to T cell hypo-responsiveness since dislocation of galectin-3 from the synapse by swainsonine restored both TCR synapse formation and T cell response. Transgenic expression of a CD3-zeta signaling chimeric antigen receptor (CAR) recovered hypo-responsive T cells to full effector functions indicating that the defect is restricted to TCR membrane components while synapse formation of the transgenic CAR was not blocked. CAR engineered late-stage T cells released cytokines and mediated redirected cytotoxicity as efficiently as younger effector T cells. Our data provide a rationale for TCR independent, CAR mediated activation in the adoptive cell therapy to avoid hypo-responsiveness of late-stage T cells upon repetitive antigen encounter.     

Challenges in T cell receptor gene therapy

The function of T lymphocytes as orchestrators and effectors of the adaptive immune response is directed by the specificity of their T cell receptors (TCRs). By transferring into T cells the genes encoding antigen-specific receptors, the functional activity of large populations of T cells can be redirected against defined targets including virally infected or cancer cells. The potential of therapeutic T cells to traffic to sites of disease, to expand and to persist after a single treatment remains a major advantage over the currently available immunotherapies that use monoclonal antibodies. Here we review recent progress in the field of TCR gene therapy, outlining challenges to its successful implementation and the strategies being used to overcome them. We detail strategies used in the optimization of affinity and surface expression of the introduced TCR, the choice of T cell subpopulations for gene transfer, and the promotion of persistence of gene-modified T cells in vivo. We review the safety concerns surrounding the use of gene-modified T cells in patients, discussing emerging solutions to these problems, and describe the increasingly positive results from the use of gene-modified T cells in recent clinical trials of adoptive cellular immunotherapy. The increasing sophistication of measures to ensure the safety of engineered T cells is accompanied by an increasing number of clinical trials: these will be essential to guide the effective translation of cellular immunotherapy from the laboratory to the bedside.     

A novel cell type responsible for marrow graft rejection in mice. T cells with NK phenotype cause acute rejection of marrow grafts

Acute rejection of allogeneic and semiallogeneic marrow grafts has long been considered to be a function of the natural immune system because it shares many features with NK activity in mice. With the use of a recently developed in vivo adoptive transfer assay in which spleen cells are transferred from mice able to reject a particular marrow graft into mice that fail to do so, we show that the cells responsible for induction of marrow graft rejection indeed display the phenotype of NK cells: they lack the T cell Ag CD4 and CD8 but express the NK Ag NK1 and ASGM1. The rejection induced by adoptively transferred cells is exquisitely specific--a feature that points to a specific recognition process by the transferred cells. To elucidate what the recognition structure on these cells may be we found that they express CD3 and most likely the beta-chain of the TCR. Highly purified responder cells with the NK1+, CD3+, CD4-, CD8- phenotype, when transferred into nonresponder recipients, cause specific marrow graft rejection. We conclude that the acute rejection of bone marrow grafts is caused by a cell that expresses NK phenotype but is of T cell lineage. This may suggest the specificity of acute marrow graft rejection is caused by a specific recognition process that involves TCR.     

Targeting rare populations of murine antigen-specific T lymphocytes by retroviral transduction for potential application in gene therapy for autoimmune disease

CD4+ T cells are important mediators in the pathogenesis of autoimmunity and would therefore provide ideal candidates for lymphocyte-based gene therapy. However, the number of Ag-specific T cells in any single lesion of autoimmunity may be quite low. Successful gene transfer into autoantigen-specific CD4+ T cells would serve as an ideal vehicle for site-targeted gene therapy if it were possible to transduce preferentially the small number of autoantigen-specific T cells. In this study we have demonstrated that retroviral infection of CD4+ lymphocytes from either autoantigen-stimulated TCR transgenic mice, or Ag-activated immunized nontransgenic mice, with a retroviral vector (pGCIRES), resulted in the transduction of only the limited number of Ag-reactive CD4+ T cells. In contrast, polyclonal activation of the same cultures resulted in transduction of non-antigen-specific lymphocytes. Transduction of Ag-reactive CD4+ T cells with pGCIRES retrovirus encoding the regulatory genes IL-4 (IL4) and soluble TNF receptor (STNFR) resulted in stable integration and long-term expression of recombinant gene products. Moreover, expression of the pGCIRES marker protein, GFP, directly correlated with the expression of the upstream regulatory gene. Retroviral transduction of CD4+ T cells targeted specifically Ag-reactive cells and was cell cycle-dependent and evident only during the mitosis phase. These studies suggest that retroviral transduction of autoantigen-specific murine CD4+ T cells, using the pGCIRES retroviral vector, may provide a potential method to target and isolate the low frequency of autoantigen-specific murine CD4+ T cells, and provides a rational approach to gene therapy in animal models of autoimmunity.     

TCR transgenes and transgene cassettes for TCR gene therapy: status in 2008

The genetic introduction of T cell receptor genes into T cells has been developed over the past decade as a strategy to induce defined antigen-specific T cell immunity. With the potential value of TCR gene therapy well-established in murine models and the feasibility of infusion of TCR-modified autologous T cells shown in a first phase I trial, the next key step will be to transform TCR gene transfer from an experimental technique into a robust clinical strategy. In this review, we discuss the different properties of the TCR transgene and transgene cassette that can strongly affect both the efficacy and the safety of TCR gene transfer. This paper is a focussed research review based on a presentation given at the sixth annual meeting of the Association for Immunotherapy of Cancer (CIMT), held in Mainz, Germany,15–16 May 2008.     

Improving the ex vivo retroviral-mediated suicide-gene transfer process in T lymphocytes to preserve immune function

The retroviral-mediated transfer of a suicide gene into donor T cells has been proposed as a method to control alloreactivity after hematopoietic stem cell (HSC) transplantation. Gene-modified cells (GMC) may be infused into the patient either at the time of transplantation, together with a T-cell depleted HSC graft, or after transplantation, as a donor lymphocyte infusion. Administration of a so-called pro-drug activating the "suicide" mechanism only after occurrence of GvHD should selectively destroy the alloreactive GMC in vivo, eventually leading to GvHD abrogation. Although phase I-II clinical trials provided vital proof of the principle of GvHD control by suicide-gene therapy, this approach is still suboptimal. Indeed, current gene transfer strategies rely on gamma-retroviral vectors that require extensive T-cell activation and expansion for efficient transduction. Both in vitro and in vivo studies have shown that the activation, cell expansion, transduction and selection steps lead to TCR repertoire alterations and impairment of crucial T-cell functions, such as alloreactivity and anti-EBV reactivity. Thus, improvements of the suicide-gene transfer processes are required in order to preserve T-cell function. This could be achieved by using CD3/CD28 co-stimulation and immunomagnetic selection of transduced cells. In future clinical trials, lentiviral vectors may prove to be a better alternative to gamma-retroviral-mediated gene transfer, by reducing the need for prolonged ex vivo culture.     

Reconstitution of CD8+ T cells by retroviral transfer of the TCR alpha beta-chain genes isolated from a clonally expanded P815-infiltrating lymphocyte

Gene transfer of TCR alphabeta-chains into T cells may be a promising strategy for providing valuable T lymphocytes in the treatment of tumors and other immune-mediated disorders. We report in this study the reconstitution of CD8(+) T cells by transfer of TCR alphabeta-chain genes derived from an infiltrating T cell into P815. Analysis of the clonal expansion and Vbeta subfamily usage of CD8(+) TIL in the tumor sites demonstrated that T cells using Vbeta10 efficiently infiltrated and expanded clonally. The TCR alpha- and beta-chain sequences derived from a tumor-infiltrating CD8(+)/Vbeta10(+) single T cell clone (P09-2C clone) were simultaneously determined by the RT-PCR/single-strand conformational polymorphism method and the single-cell PCR method. When P09-2C TCR alphabeta-chain genes were retrovirally introduced into CD8(+) T cells, the reconstituted T cells positively lysed the P815 tumor cells, but not the A20, EL4, or YAC-1 cells, in vitro. In addition, the CTL activity was blocked by the anti-H2L(d) mAb. Furthermore, T cells containing both TCR alpha- and beta-chains, but not TCR beta-chain alone, accumulated at the tumor-inoculated site when the reconstituted CD8(+) T cells were adoptively transferred to tumor-bearing nude mice. These findings suggest that it is possible to reconstitute functional tumor-specific CD8(+) T cells by transfer of TCR alphabeta-chain genes derived from TIL, and that such T cells might be useful as cytotoxic effector cells or as a vehicle for delivering therapeutic agents.     

SLAM-associated protein deficiency causes imbalanced early signal transduction and blocks downstream activation in T cells from X-linked lymphoproliferative disease patients

Deficiency of SAP (SLAM (signaling lymphocyte activation molecule)-associated protein) protein is associated with a severe immunodeficiency, the X-linked lymphoproliferative disease (XLP) characterized by an inappropriate immune reaction against Epstein-Barr virus infection often resulting in a fatal clinical course. Several studies demonstrated altered NK and T cell function in XLP patients; however, the mechanisms underlying XLP disease are still largely unknown. Here, we show that non-transformed T cell lines obtained from XLP patients were defective in several activation events such as IL-2 production, CD25 expression, and homotypic cell aggregation when cells were stimulated via T cell antigen receptor (TCR).CD3 but not when early TCR-dependent events were bypassed by stimulation with phorbol 12-myristate 13-acetate/ionomycin. Analysis of proximal T cell signaling revealed imbalanced TCR.CD3-induced signaling in SAP-deficient T cells. Although phospholipase C gamma 1 phosphorylation and calcium response were both enhanced in T cells from XLP patients, phosphorylation of VAV and downstream signal transduction events such as mitogen-activated protein kinase phosphorylation and IL-2 production were diminished. Importantly, reconstitution of SAP expression by retroviral-mediated gene transfer completely restored abnormal signaling events in T cell lines derived from XLP patients. In conclusion, SAP mutation or deletion in XLP patients causes profound defects in T cell activation, resulting in immune deficiency. Moreover, these data provide evidence that SAP functions as an essential integrator in early TCR signal transduction.     

Restoration of T cell receptor-mediated signal transduction by transfection of CD45 cDNA into a CD45-deficient variant of the Jurkat T cell line.

The TCR is a multimeric structure comprised of distinct Ag recognition and signal transduction components. Although none of the molecules that make up the TCR possess intrinsic protein tyrosine kinase (PTK) activity, stimulation of T cells via the TCR results in the rapid appearance of newly tyrosine phosphorylated proteins in cell lysates. Evidence suggests ligation of the TCR induces activation of a PTK that may be a member of the src family. One early consequence of this TCR-mediated PTK activation is the phosphorylation of the gamma 1 isoform of phospholipase C. This phosphorylation event is associated with increased enzymatic activity resulting in the hydrolysis of phosphatidylinositol 4,5 bisphosphate into two second messengers, inositol 1,4,5 trisphosphate and diacylglycerol. Recently, our laboratory and others have isolated mutant T cells that lack surface expression of CD45, the major surface tyrosine phosphatase expressed on lymphoid cells. Stimulation of the TCR on these cells fails to result in the expected activation events. We demonstrate that reconstitution of surface expression of the 180-kDa isoform of CD45 by gene transfer into a CD45-deficient mutant of the Jurkat T cell leukemic line restores the ability of the TCR to couple fully to its signal transduction machinery. These results support the role of CD45 tyrosine phosphatase activity in regulating the TCR-activated PTK.     

Detection and characterization of T cells specific for BDC2.5 T cell-stimulating peptides

Nonobese diabetic (NOD) mice expressing the BDC2.5 TCR transgene are useful for studying type 1 diabetes. Several peptides have been identified that are highly active in stimulating BDC2.5 T cells. Herein, we describe the use of I-Ag7 tetramers containing two such peptides, p79 and p17, to detect and characterize peptide-specific T cells. The tetramers could stain CD4(+) T cells in the islets and spleens of BDC2.5 transgenic mice. The percentage of CD4(+), tetramer(+) T cells increased in older mice, and it was generally higher in the islets than in the spleens. Our results also showed that tetAg7/p79 could stain a small population of CD4(+) T cells in both islets and spleens of NOD mice. The percentage of CD4(+), tetramer(+) T cells increased in cells that underwent further cell division after being activated by peptides. The avidity of TCRs on purified tetAg7/p79(+) T cells for tetAg7/p79 was slightly lower than that of BDC2.5 T cells. Although tetAg7/p79(+) T cells, like BDC2.5 T cells, secreted a large quantity of IFN-gamma, they were biased toward being IL-10-producing cells. Additionally, <3% of these cells expressed TCR Vbeta4. In vivo adoptive transfer experiments showed that NOD/scid recipient mice cotransferred with tetAg7/p79(+) T cells and NOD spleen cells, like mice transferred with NOD spleen cells only, developed diabetes. Therefore, we have generated Ag-specific tetramers that could detect a heterogeneous population of T cells, and a very small number of NOD mouse T cells may represent BDC2.5-like cells.     

Requirements for effective antitumor responses of TCR transduced T cells

Adoptive transfer of TCR gene-modified T cells has been proposed as an attractive approach to target tumors for which it is difficult or impossible to induce strong tumor-specific T cell responses by vaccination. Whereas the feasibility of generating tumor Ag-specific T cells by gene transfer has been demonstrated, the factors that determine the in vivo effectiveness of TCR-modified T cells are largely unknown. We have analyzed the value of a number of clinically feasible strategies to enhance the antitumor potential of TCR modified T cells. These experiments reveal three factors that contribute greatly to the in vivo potency of TCR-modified T cells. First, irradiation-induced host conditioning is superior to vaccine-induced activation of genetically modified T cells. Second, increasing TCR expression through genetic optimization of TCR sequences has a profound effect on in vivo antitumor activity. Third, a high precursor frequency of TCR modified T cells within the graft is essential. Tumors that ultimately progress in animals treated with this optimized regimen for TCR-based adoptive cell transfer invariably display a reduced expression of the target Ag. This suggests TCR gene therapy can achieve a sufficiently strong selective pressure to warrant the simultaneous targeting of multiple Ags. The strategies outlined in this study should be of value to enhance the antitumor activity of TCR-modified T cells in clinical trials.     

Display, engineering, and applications of antigen-specific T cell receptors

The use of T cell receptors (TCRs) as potential therapeutic agents provides an opportunity to target a greatly expanded array of antigens, compared to those now targeted with monoclonal antibodies. With the advent of new display technologies and TCR formats for in vitro engineering, it should be possible to generate high-affinity TCRs against virtually any peptide antigen that is shown to bind to a major histocompatibility complex (MHC) molecule (e.g. peptides derived from viral antigens or from self proteins that are associated with the transformed phenotype). What remains, however, are challenges associated with effective targeting of very low numbers of cell surface antigens (pepMHC), fewer than the case for conventional monoclonal antibody-based therapies. This hurdle might be overcome with the attachment of more effective payloads for soluble TCR approaches, or by using TCR gene transfer into T cells that can then be adoptively transferred into patients. There is considerable work to be done on the physiological aspects of either approach, including pharmacokinetic studies in the case of soluble TCRs, and T cell trafficking, persistence, and autoreactivity studies in the case of adoptively transferred T cells. As with the field of monoclonal antibodies, it will take time to explore these issues, but the potential benefits of TCR-based therapies make these challenges worth the effort.     

Induction of central deletional T cell tolerance by gene therapy

Transgenic mice expressing an alloreactive TCR specific for the MHC class I Ag K(b) were used to examine the mechanism by which genetic engineering of bone marrow induces T cell tolerance. Reconstitution of lethally irradiated mice with bone marrow infected with retroviruses carrying the MHC class I gene H-2K(b) resulted in lifelong expression of K(b) on bone marrow-derived cells. While CD8 T cells expressing the transgenic TCR developed in control mice reconstituted with mock-transduced bone marrow, CD8 T cells expressing the transgenic TCR failed to develop in mice reconstituted with H-2K(b) transduced bone marrow. Analysis of transgene-expressing CD8 T cells in the thymus and periphery of reconstituted mice revealed that CD8 T cells expressing the transgenic TCR underwent negative selection in the thymus of mice reconstituted with K(b) transduced bone marrow. Negative selection induced by gene therapy resulted in tolerance to K(b). Thus, genetic engineering of bone marrow can be used to alter T cell education in the thymus by inducing negative selection.