Search

Pathogen recognition by T cells is dependent on their exquisite specificity for self-major histocompatibility complex (MHC) molecules presenting a bound peptide. Although this specificity results from positive and negative selection of developing T cells in the thymus, the relative contribution of these two processes remains controversial. To address the relation between the selecting peptide-MHC complex and the specificity of mature T cells, we generated transgenic mice that express a single peptide-MHC class I complex. We demonstrate that positive selection of CD8 T cells in these mice results in an MHC-specific repertoire. Although selection on a single complex is peptide promiscuous, mature T cells are highly peptide specific. Thus, positive selection imparts MHC and peptide specificity on the peripheral CD8 T cell repertoire.

Immunological synapse formation between T cells and target cells can affect the functional outcome of TCR ligation by a given MHC-peptide complex. Although synapse formation is usually induced by TCR signaling, it is not clear whether other factors can affect the efficiency of synapse formation. Here, we tested whether cytokines could influence synapse formation between murine CTLs and target cells. We found that IL-12 enhanced synapse formation, whereas TGFbeta decreased synapse formation. The enhanced synapse formation induced by IL-12 appeared to be functional, given that IL-12-treated cells could respond to weak peptides, including self-peptides, to which the T cells were normally unresponsive. These responses correlated with expression of functionally higher avidity LFA-1 on IL-12-treated CTLs. These findings have implications for the function of IL-12 in T cell-mediated autoimmunity.

The ongoing discovery of disease-associated epitopes detected by CD8 T cells greatly facilitates peptide-based vaccine approaches and the construction of multimeric soluble recombinant proteins (e.g. tetramers) for isolation and enumeration of antigen-specific CD8 T cells. Related to these outcomes of epitope discovery is the recent demonstration that MHC class I/peptide complexes can be expressed as single chain trimers (SCTs) with peptide, beta(2)m and heavy chain connected by linkers to form a single polypeptide chain. Studies using clinically relevant mouse models of human disease have shown that SCTs expressed by DNA vaccination are potent stimulators of cytotoxic T lymphocytes. Their vaccine efficacy has been attributed to the fact that SCTs contain a preprocessed and preloaded peptide that is stably displayed on the cell surface. Although SCTs of HLA class I/peptide complexes have been previously reported, they have not been characterized for biochemical stability or susceptibility to exogenous peptide binding. Here we demonstrate that human SCTs remain almost exclusively intact when expressed in cells and can incorporate a disulfide trap that dramatically excludes the binding of exogenous peptides. The mechanistic and practical applications of these findings for vaccine development and T cell isolation/enumeration are discussed.

Infection with West Nile virus (WNV) causes fatal encephalitis in immunocompromised animals. Previous studies in mice have established that T cell protection is required for clearance of WNV infection from tissues and preventing viral persistence. The current study assessed whether specific WNV peptide epitopes could elicit a cytotoxic T lymphocyte (CTL) response capable of protecting against virus infection. Hidden Markov model analysis was used to identify WNV-encoded peptides that bound the MHC class I proteins K(b) or D(b). Of the 35 peptides predicted to bind MHC class I molecules, one immunodominant CTL recognition peptide was identified in each of the envelope and non-structural protein 4B genes. Addition of these but not control peptides to CD8(+) T cells from WNV-infected mice induced IFN-gamma production. CTL clones that were generated ex vivo lysed peptide-pulsed or WNV-infected target cells in an antigen-specific manner. Finally, adoptive transfer of a mixture of envelope- and non-structural protein 4B-specific CTL to recipient mice protected against lethal WNV challenge. Based on this, we conclude that CTL responses against immundominant WNV epitopes confer protective immunity and thus should be targets for inclusion in new vaccines.

A breast cancer-associated antigen, mammaglobin-A, is specifically expressed in 80% of primary breast tumors. The definition of immune responses against this highly expressed breast cancer-specific antigen should be of great value in the development of new therapeutic strategies for breast cancer. Thus, the purpose of this study was to identify HLA-A2-restricted mammaglobin-A-derived epitopes recognized by CD8+ cytotoxic T lymphocytes (CTL). We identified seven mammaglobin-A-derived candidate epitopes that bind the HLA-A2 molecule (Mam-A2.1-7) by means of a HLA class I-peptide binding computer algorithm from the Bioinformatics & Molecular Analysis Section of the National Institutes of Health. Subsequently, we determined that CD8+ CTLs from breast cancer patients reacted to the Mam-A2.1 (83-92, LIYDSSLCDL), Mam-A2.2 (2-10, KLLMVLMLA), Mam-A2.3 (4-12, LMVLMLAAL), Mam-A2.4 (66-74, FLNQTDETL), and Mam-A2.7 (32-40, TINPQVSKT) epitopes using an IFN-gamma ELISPOT assay. Interestingly, healthy individuals also showed high reactivity to the Mam-A2.2 epitope. Two CD8+ CTL lines generated in vitro against TAP-deficient T2 cells loaded with the candidate epitopes showed significant cytotoxic activity against the Mam-A2.1-4 epitopes. These CD8+CTL lines recognized a HLA-A2+breast cancer cell line expressing the Mam-A2.1 epitope. In addition, DNA vaccination of HLA-A2+/human CD8+ double-transgenic mice with a DNA construct encoding the Mam-A2.1 epitope and the HLA-A2 molecule induced a significant expansion of epitope-specific CD8+ CTLs that recognize the same HLA- A2+/Mam-A2.1+ breast cancer cell line. In conclusion, these results demonstrate the immunotherapeutic potential of mammaglobin-A for the treatment and prevention of breast cancer.

The self-restricted T cell repertoire exhibits a high frequency of alloreactivity. Because these alloreactive T cells are derived from the pool of cells selected on several different self MHC alleles, it is unknown how development of the alloantigenic repertoire is influenced by homology between a self MHC allele and an alloantigen. To address this, we used the 2C transgenic TCR that is selected by K(b), is alloreactive for L(d), and cross-reacts with L(q). L(q) is highly homologous to L(d) and binds several of the same peptide ligands, including p2Ca, the peptide recognized by 2C. We find that L(d)/p2Ca is a high avidity agonist ligand, whereas L(q)/p2Ca is a low avidity agonist ligand for 2C T cells. When mice transgenic for the 2C TCR are bred to L(q)-expressing mice, 2C(+) T cells develop; however, they express lower levels of either the 2C TCR or CD8 and require a higher L(d)/p2Ca ligand density to be activated than 2C(+) T cells selected by K(b). Furthermore, the 2C T cells selected in the presence of L(q) fail to detect L(q)/p2Ca complexes even at high ligand density. Thus, despite possessing the identical TCR, there is a functional avidity difference between 2C(+) T cells selected in the presence of L(q) vs K(b). These data provide evidence that homology between the selecting ligand and an alloantigen can influence the avidity of the T cell repertoire for the alloantigen, and suggest that thymic selection can fine tune T cell avidity independent of intrinsic TCR affinity.

The molecular basis for the difference in the strength of T cell responses to self vs alloantigens is unknown, but may reflect how T cells are selected in the thymus. Because T cells with a high affinity for foreign as opposed to self MHC molecules are able to mature, it has been proposed that alloreactive T cells may be more strongly dependent upon interaction with MHC residues than are self-restricted T cells. This study was undertaken to rigorously address this hypothesis. Whereas other studies have compared self vs alloantigen recognition of different MHC alleles by a single T cell clone, we have compared self vs alloantigen recognition of a single MHC allele, H-2Ld, by a large panel of self-restricted and alloreactive T cell clones. Target cells expressing Ld molecules mutated at several different potential TCR contact residues were analyzed to determine which residues are important for recognition by self-restricted vs alloreactive T cells. We unequivocally demonstrate that self-restricted and alloreactive T cells do not differ, but rather are comparably dependent on interaction with MHC residues. Importantly, both self-restricted and alloreactive T cells are dependent upon the same MHC residues as primary contacts and, in addition, share a common recognition pattern of Ld. Furthermore, our analysis enables us to provide a model for allotype-specific T cell recognition of Ld vs Kb class I molecules.

We developed a novel approach to probe the molecular basis of TCR recognition of the MHC class I-peptide complex and to determine how constraints placed on peptide binding by the class I molecule influence T cell recognition. We synthesized peptide pairs derived from the N- and C-terminal regions of class I peptide ligands in which the TCR contacts and dominant binding residues were placed together or were separated. Complementary peptide pairs derived from two well-characterized Ld peptide ligands, tum- (QNHRALDL) and p2Ca (LSPFPFDL), were tested for the ability to sensitize targets for recognition by peptide-specific cytotoxic T lymphocytes (CTL). The tum-derived tetramer QNHR, containing both primary TCR contact residues (H17 and R18), is recognized only when used in combination with ALDL which contains the primary binding residues (A19, D21 and L22). This suggests that both peptides of the pair contribute to positioning of the TCR contacts. Remarkably, CTL clone P24 recognized target cells sensitized with a trimer (QNH) combined with a pentamer (RALDL), demonstrating that TCR recognition can occur when the TCR contacts are separated (placed on separate peptide subunits). For the p2Ca peptide LSPFPFDL, the C-terminal tetramer PFDL, which contains both the primary TCR contact residue (P) and the dominant binding residue (L), is sufficient for recognition. In addition, PFDL was able to bind effectively to Ld and to activate naive antigen-specific T cells. These data suggest that peptide subunits and complementary peptide pairs composed of trimeric, tetrameric or pentameric peptides can bind independently to the Ld molecule in the same register and orientation as they do when contained within the parent peptide.

We have examined the ability of beta 2-m- mice to produce CD4-8+ T cells by generating CD8+ CTLs to a defined ligand. We report here the first demonstration of peptide-specific, self-class I MHC-restricted CTLs from beta 2-m-deficient mice. We have used the KOD mouse, an H-2d beta 2-m- strain, to generate CTLs that recognize the class I MHC molecule Ld in association with one of two Ld-binding immunogenic peptides. Testing of these CTLs on a panel of Ld-binding peptides reveals a high degree of peptide specificity. Peptide-specific CTL bulk cultures from KOD mice differ from those generated in beta 2-m+ mice in that they possess altered affinities for their peptide ligands. In addition, we show that CTLs generated from beta 2-m- mice in the presence of beta 2-m+ stimulator cells and exogenous peptide are specific either for the exogenous peptide or for endogenous peptides that are present in association with Ld on the surface of beta 2-m+ cells, but are not present at detectable levels on beta 2-m- cells. These results demonstrate that positive selection of CD8+ CTLs can occur in vivo on the very low levels of class I MHC found in the KOD mouse. Furthermore, CTLs from the KOD mouse maintain a high degree of peptide specificity despite reduced levels of class I MHC.

The P911 variant of the P815 mastocytoma was shown by Lurquin et al. (Cell 58:293, 1989) to elicit rapid tumor rejection in a syngeneic host. This rejection was mediated by Ld-restricted cytotoxic T lymphocytes (CTL) for which targets could be sensitized by the synthetic peptide designated tum- (P91A-.12-24). In a previous study, T cell clones specific for Ld-tum- complexes displayed very restricted TCR usage and a characteristic TCR motif in the V alpha CDR3 region, predicted to interact with peptide. However, in contrast to the majority of Ld peptide ligands that are nonamers, the tum- peptide is a 13-mer and its sequence does not fit the Ld binding motif. Thus, to define shorter versions of the tum- 13-mer and residues involved in TCR recognition, nonamer derivatives were synthesized and compared in several different binding and functional assays. From these comparisons, the peptide TQNHRALDL was found to be the optimal nonamer. CTL recognition of Ala-substituted analogues of this peptide indicated that the His and Arg residues at positions 4 and 5 are important for TCR contact. We propose that these basic residues of the tum- peptide interact with the previously defined acidic residues in the CDR3 region of several TCR known to recognize Ld-tum- complexes.

The nature of alloreactivity to MHC molecules has been enigmatic, primarily because of the observation that allogeneic responses are considerably stronger than syngeneic responses. To better determine the specificity potential of allogeneic responses, we have generated alloreactive CTL specific for exogenous, viral-derived peptide ligands. This approach allowed us to critically evaluate both the peptide- and MHC-specificity of these alloreactive T cells. Exploiting the accessibility of the H-2Ld class I molecule for exogenous peptide ligands, alloreactive CTL were generated that are specific for either murine cytomegalovirus (MCMV) or lymphocytic choriomeningitis virus (LCMV) peptides bound by Ld alloantigens. Peptide specificity was initially observed in bulk cultures of alloreactive CTL only when tested on peptide-sensitized T2.Ld target cells that have defective presentation of endogenous peptides. Subsequent cloning of bulk alloreactive CTL lines generated to MCMV yielded CTL clones that had exquisitely specific MCMV peptide recognition requirement. All of the MCMV/Ld alloreactive CTL clones were also exquisitely MHC-specific in that none of the CTL clones lysed targets expressing MCMV/Lq complexes, even though Lq differs from Ld by only six amino acid residues and Lq also binds the MCMV peptide. This observation clearly demonstrates that alloreactive CTL are capable of the same degree of specificity for target cell recognition as are syngeneic CTL in MHC-restricted responses.

The primary immune response of cytotoxic T lymphocytes in H-2d and H-2q mice to infection with lymphocytic choriomeningitis virus is directed mostly towards the common major T cell epitope of amino acids 112-132 on the viral nucleoprotein (NP). The molecules responsible for presentation of the T cell epitope NP112-132 are in both haplotypes the MHC class I L antigens (Ld, Lq). Truncations of the amino and carboxy termini of the NP 112-132 sequence revealed the nonapeptide RPQASGVYM (NP118-126) as a most effective peptide antigen, but even the tetrapeptide GVYM was recognized by CTL of both haplotypes in a class I antigen-restricted specificity. When tyrosine (Y) or methionine (M) were substituted with alanine, CTL recognition of the altered nonamer required 10(6) to 10(8) times higher peptide concentrations and in one case (Y----A on Ld) the peptide was not recognized at all. Up-modulation of the expression of Ld and Lq class I antigens as measured by flow cytometry correlated with the ability to present the peptide antigens. The only exception was peptide NP118-126 (M----A), which was recognized by T cells on L-Ld and L-Lq target cells but failed to up-regulate Ld and Lq antigens.

We have taken advantage of some unique properties of H-2Ld to investigate the determinant density requirements for cytotoxic T lymphocyte (CTL) priming versus effector function and to correlate the determinant density requirements with CD8 dependency. In a previous study (Lie, W.-R., N. B. Myers, J. Gorka, R. J. Rubocki, J. M. Connolly, and T. H. Hansen. 1990. Nature [Lond.]. 344:439), we demonstrated that culturing normal cells with peptides known to be restricted by H-2Ld led to a two- to fourfold increase in surface Ld expression. In the present study, we demonstrate the generation of Ld-restricted, peptide-specific in vitro primary CTL by culturing spleen cells with murine cytomegalovirus or tum- peptide at concentrations previously shown to result in maximum induction of Ld expression. Target cells can be sensitized for recognition by these CTL with lower dose of peptide than are required for the primary sensitization. This demonstrates differences in the determinant density requirements for priming versus effector function. The in vitro primary CTL generated with peptide can weakly lyse target cells that express the determinant endogenously, and CTL lines and clones capable of strong lysis of endogenous expressors are easily obtained. In both cases, target cells treated with exogenous peptide are lysed better than target cells expressing antigen endogenously. This suggested that there are differences in the determinant density of peptide-fed versus endogenous targets. This interpretation was substantiated when it was observed that the level of lysis of target cells expressing endogenous determinants correlated inversely with the amount of peptide required to sensitize targets for recognition by various tum- -specific CTL clones. Furthermore, simultaneous titration of both the peptide used to treat target cells and the antibody to CD8 revealed that the various CTL clones analyzed displayed widely disparate CD8 dependencies. In each case, the CD8 dependency correlated inversely with the determinant density requirement. Therefore, CD8 dependency of CTL is relative, but shows an absolute and quantitative correlation with their dependency on determinant density. These findings suggest that under physiologic conditions, where only low determinant densities are likely to be encountered, all CTL clones will show at least partial CD8 dependency.

The CD8 molecule on class I-reactive cytotoxic T lymphocytes (CTLs) is believed to function as a coreceptor along with the alpha beta T-cell receptor. Whereas the alpha beta T-cell receptor recognizes polymorphic residues in the alpha 1/alpha 2 domains of the class I molecule, the CD8 molecule is believed to recognize monomorphic class I residues. Our previous experiments suggested that residue 227 in the alpha 3 domain of major histocompatibility complex class I molecules contributes to the determinant recognized by CD8. By using a panel of site-directed mutants of H-2Dd, this observation has been extended herein. Our findings indicate that for recognition by CD8-dependent CTLs, residue 227 must be either glutamic acid or aspartic acid and cannot be either basic or uncharged. However, the recognition by CD8-independent CTLs is unaffected by any of the substitutions at position 227 of H-2Dd. Similarly, alterations of other charged residues at positions 222, 223, and 229 have an analogous effect to substitution at residue 227, whereas substitutions at residues 192 and 232 do not affect the reactivity of CD8-dependent or CD8-independent CTLs. In addition, mutant H-2Dd molecules that are not recognized by CD8-dependent CTLs are unable to stimulate a primary CTL response, yet they can stimulate a secondary CD8-independent H-2Dd-specific CTL response. These findings suggest that CD8 recognition is obligatory for the priming of class I-dependent CTL responses. Since endogenous class I molecules were expressed by all of the transfected cell lines, these findings provide direct genetic evidence that CD8 and the alpha beta T-cell receptor must interact with the same class I molecule.

The involvement of the different domains of the MHC class I molecule in CTL recognition was investigated. mAbs specific for the alpha 1/alpha 2 domains of H-2Ld interfered with both the primary and secondary generation and effector function of in vitro Ld-specific CTL. mAbs specific for the alpha 3 domain of H-2Ld interfered with the generation and function of primary in vitro Ld-specific CTL; however, there was no effect on the in vitro generation of secondary CTL and only partial inhibition of their function. In vivo treatment with graft-specific antibodies to both the alpha 3 domain and the alpha 1/alpha 2 domains together resulted in a dramatic enhancement of Ld- or Dd-disparate skin grafts, whereas the individual mAbs showed minimal effects. This suggested that the class I alpha 3 domain is recognized by alloreactive CTL. Several approaches were undertaken to examine whether recognition of the alpha 3 domain determinants is mediated by the Lyt-2 molecule. When mAbs specific for the alpha 3 domain of either H-2Ld or H-2Dd were used in vivo and in vitro, the resulting CTL population was not inhibited by antibody to the alpha 3 domain and was only partially inhibited by antibody to Lyt-2. We therefore observed a correlation between the effects of antibody to the class I alpha 3 domain of the target molecule and antibody to the Lyt-2 molecule on the CTL. To further test the relationship between CTL recognition of the alpha 3 domain and the involvement of Lyt-2, we used a cell expressing a mutation in the alpha 3 domain of the Dd molecule. The mutation resulted in a single amino acid substitution of glu to lys at residue 227 of the alpha 3 domain. Consistent with an earlier report, cells expressing the mutant Dd lys molecule were not lysed by CTL from a primary stimulation against the wild-type Dd glu molecule. However, this same cell line was killed by the Lyt-2-independent secondary Dd-specific CTL generated in the presence of antibody to the alpha 3 domain in vivo and in vitro. Furthermore, cells expressing the mutant Dd lys molecule failed to stimulate a primary response. In conclusion, several independent lines of evidence indicate that residues in the alpha 3 domain of the class I molecule are involved in recognition by the Lyt-2 molecule, and that Lyt-2-mediated recognition can be specifically blocked using mAb to determinants in the alpha 3 domain.