A new generation of Melan-A/MART-1 peptides that fulfill both increased immunogenicity and high resistance to biodegradation: implication for molecular anti-melanoma immunotherapy.

Intense efforts of research are made for developing antitumor vaccines that stimulate T cell-mediated immunity. Tumor cells specifically express at their surfaces antigenic peptides presented by MHC class I and recognized by CTL. Tumor antigenic peptides hold promise for the development of novel cancer immunotherapies. However, peptide-based vaccines face two major limitations: the weak immunogenicity of tumor Ags and their low metabolic stability in biological fluids. These two hurdles, for which separate solutions exist, must, however, be solved simultaneously for developing improved vaccines. Unfortunately, attempts made to combine increased immunogenicity and stability of tumor Ags have failed until now. Here we report the successful design of synthetic derivatives of the human tumor Ag Melan-A/MART-1 that combine for the first time both higher immunogenicity and high peptidase resistance. A series of 36 nonnatural peptide derivatives was rationally designed on the basis of knowledge of the mechanism of degradation of Melan-A peptides in human serum and synthesized. Eight of them were efficiently protected against proteolysis and retained the antigenic properties of the parental peptide. Three of the eight analogs were twice as potent as the parental peptide in stimulating in vitro Melan-specific CTL responses in PBMC from normal donors. We isolated these CTL by tetramer-guided cell sorting and expanded them in vitro. The resulting CTL efficiently lysed tumor cells expressing Melan-A Ag. These Melan-A/MART-1 Ag derivatives should be considered as a new generation of potential immunogens in the development of molecular anti-melanoma vaccines.

The activatory receptor 2B4 is expressed in vivo by human CD8+ effector alpha beta T cells.

The membrane receptor 2B4 is a CD2 family member that is involved in lymphocyte activation. A fraction of human CD8+ alphabeta T cells up-regulate 2B4 in vivo, and here we demonstrate that this correlates with the acquisition of effector cell properties such as granzyme B and perforin expression, rapid IFN-gamma production, and down-regulation of the lymph node homing chemokine receptor CCR7. In PBLs from healthy donors, cytomegalovirus-specific effector T cells were 2B4 positive, whereas naive melanoma Ag (Melan-A/melanoma Ag recognized by T cells-1)-specific T cells were 2B4 negative. In melanoma patients, Melan-A-specific T cells up-regulated 2B4 in parallel with in vivo differentiation. This occurred in PBLs after vaccination with Melan-A peptides and in tumor-infiltrated lymph nodes, likely through disease-associated activation of Melan-A-specific T cells. Thus, 2B4 expression correlates with CD8+ T cell differentiation in vivo.

The alternatively spliced domain TnFnIII A1A2 of the extracellular matrix protein tenascin-C suppresses activation-induced T lymphocyte proliferation and cytokine production.

Several lines of evidences have suggested that T cell activation could be impaired in the tumor environment, a condition referred to as tumor-induced immunosuppression. We have previously shown that tenascin-C, an extracellular matrix protein highly expressed in the tumor stroma, inhibits T lymphocyte activation in vitro, raising the possibility that this molecule might contribute to tumor-induced immunosuppression in vivo. However, the region of the protein mediating this effect has remained elusive. Here we report the identification of the minimal region of tenascin-C that can inhibit T cell activation. Recombinant fragments corresponding to defined regions of the molecule were tested for their ability to inhibit in vitro activation of human peripheral blood T cells induced by anti-CD3 mAbs in combination with fibronectin or IL-2. A recombinant protein encompassing the alternatively spliced fibronectin type III domains of tenascin-C (TnFnIII A-D) vigorously inhibited both early and late lymphocyte activation events including activation-induced TCR/CD8 down-modulation, cytokine production, and DNA synthesis. In agreement with this, full length recombinant tenascin-C containing the alternatively spliced region suppressed T cell activation, whereas tenascin-C lacking this region did not. Using a series of smaller fragments and deletion mutants issued from this region, we have identified the TnFnIII A1A2 domain as the minimal region suppressing T cell activation. Single TnFnIII A1 or A2 domains were no longer inhibitory, while maximal inhibition required the presence of the TnFnIII A3 domain. Altogether, these data demonstrate that the TnFnIII A1A2 domain mediate the ability of tenascin-C to inhibit in vitro T cell activation and provide insights into the immunosuppressive activity of tenascin-C in vivo.

Lack of tumor recognition by hTERT peptide 540-548-specific CD8(+) T cells from melanoma patients reveals inefficient antigen processing.

Telomerase is a ribonucleoprotein complex responsible for the maintenance of the length of the telomeres during cell division, which is active in germ-line cells as well as in the vast majority of tumors but not in most normal tissues. The wide expression of the human telomerase catalytic subunit (hTERT) in tumors makes it an interesting candidate vaccine for cancer. hTERT-derived peptide 540-548 (hTERT(540)) has been recently shown to be recognized in an HLA-A*0201-restricted fashion by T cell lines derived from peptide-stimulated peripheral blood mononuclear cells (PBMC) from healthy donors. As a first step to the inclusion of this peptide in immunotherapy clinical trials, it is crucial to assess hTERT(540)-specific T cell reactivity in cancer patients as well as the ability of hTERT-specific CD8(+) T lymphocytes to recognize and lyse hTERT-expressing target cells. Here, we have analyzed the CD8(+) T cell response to peptide hTERT(540) in HLA-A*0201 melanoma patients by using fluorescent HLA-A*0201/hTERT(540) peptide tetramers. HLA-A*0201/hTERT(540) tetramer(+) CD8(+) T cells were readily detected in peptide-stimulated PBMC from a significant proportion of patients and could be isolated by tetramer-guided cell sorting. hTERT(540)-specific CD8(+) T cells were able to specifically recognize HLA-A*0201 cells either pulsed with peptide or transiently transfected with a minigene encoding the minimal epitope. In contrast, they failed to recognize hTERT-expressing HLA-A*0201(+) target cells. Furthermore, in vitro proteasome digestion studies revealed inadequate hTERT processing. Altogether, these results raise questions on the use of hTERT(540) peptide for cancer immunotherapy.

Subcellular localization of the melanoma-associated protein Melan-AMART-1 influences the processing of its HLA-A2-restricted epitope.

The peptide derived from the melanoma-associated protein Melan-A (Melan-A(26-35)/HLA-A2) is an attractive candidate for tumor immunotherapy but little is known about the intracellular processing of this antigen. Here we show that Melan-A is a single-pass membrane protein with an NH(2) terminus exposed to the lumen of the exocytic compartment. In transfected melanoma cells, Melan-A accumulates in the Golgi region. Inversion of the membrane topology leads to the retention of Melan-A in the endoplasmic reticulum. Most strikingly, melanoma cells expressing this form of Melan-A are more effectively recognized by specific CTL than those expressing either Melan-A in its native membrane orientation or Melan-A artificially localized in the cytosol. Our data are compatible with the notion that proteins retained in the endoplasmic reticulum are more efficiently degraded and produce more antigenic peptides.

Discrepancy between ELISPOT IFN-gamma secretion and binding of A2/peptide multimers to TCR reveals interclonal dissociation of CTL effector function from TCR-peptide/MHC complexes half-life.

Activation of CD8(+) cytolytic T lymphocytes (CTLs) by antigen is triggered by the interaction of clonotypic alphabeta T cell receptors (TCRs) with antigenic peptides bound to MHC class I molecules (pMHC complexes). Fluorescent multimeric pMHC complexes have been shown to specifically stain antigen-specific CTLs by directly binding the TCR. In tumor-infiltrating lymphocytes from a melanoma patient we found a high frequency of tyrosinase(368-376) peptide-specific cells as detected by IFN-gamma ELISPOT, without detectable staining with the corresponding A2/peptide multimers. Surprisingly, these T cells were able to lyse tyrosinase(368-376) peptide-pulsed target cells as efficiently as other specific T cells that were stained by multimers. Analysis of the staining patterns under different conditions of incubation time and temperature revealed that these results were explained by major differences in TCR-multimeric ligand interaction kinetics among the clones. Whereas no direct quantitative correlation between antigenic peptide concentration required for CTL effector functions and equilibrium multimer binding was observed interclonally, the latter was profoundly affected by the kinetics of TCR-ligand interaction. More importantly, our data indicate that similar levels of T cell activation can be achieved by independent CD8(+) T cell clonotypes displaying different TCR/pMHC complex dissociation rates.

Heterogeneous T-cell response to MAGE-A10(254-262): high avidity-specific cytolytic T lymphocytes show superior antitumor activity.

MAGE-encoded antigens, which are expressed by tumors of many histological types but not in normal tissues, are suitable candidates for vaccine-based immunotherapy of cancers. Thus far, however, T-cell responses to MAGE antigens have been detected only occasionally in cancer patients. In contrast, by using HLA/peptide fluorescent tetramers, we have observed recently that CD8(+) T cells specific for peptide MAGE-A10(254-262) can be detected frequently in peptide-stimulated peripheral blood mononuclear cells from HLA-A2-expressing melanoma patients and healthy donors. On the basis of these results, antitumoral vaccination trials using peptide MAGE-A10(254-262) have been implemented recently. In the present study, we have characterized MAGE-A10(254-262)-specific CD8(+) T cells in polyclonal cultures and at the clonal level. The results indicate that the repertoire of MAGE-A10(254-262)-specific CD8(+) T cells is diverse both in terms of clonal composition, efficiency of peptide recognition, and tumor-specific lytic activity. Importantly, only CD8(+) T cells able to recognize the antigenic peptide with high efficiency are able to lyse MAGE-A10-expressing tumor cells. Under defined experimental conditions, the tetramer staining intensity exhibited by MAGE-A10(254-262)-specific CD8(+) T cells correlates with efficiency of peptide recognition so that "high" and "low" avidity cells can be separated by FACS. Altogether, the data reported here provide evidence for functional diversity of MAGE-A10(254-262)-specific T cells and will be instrumental for the monitoring of peptide MAGE-A10(254-262)-based clinical trials.

Ex vivo analysis of tumor antigen specific CD8+ T cell responses using MHC/peptide tetramers in cancer patients.

The development of soluble tetrameric MHC/peptide complexes has opened the possibility to directly identify and monitor antigen-specific CD8+ T cells in different clinical situations. This represents a technological breakthrough for the field of cell-mediated immunity. For example, the direct identification and enumeration of tumor-specific CD8+ T cells at the tumor site and in blood has recently provided compelling evidence that strong anti-tumoral responses naturally occur in some cancer patients. Moreover, the use of tetramers plays an essential role in the design of vaccination protocols aimed at inducing a strong and protective CD8+ T cell-mediated anti-tumoral response in cancer patients. The monitoring of antigen-specific T cell responses elicited by various peptide-based vaccines tested in phase I clinical trials clearly indicates that tumor-specific CD8+ T cells can be activated effectively at least in some cancer patients. Thus, multiparameter monitoring of antigen-specific T cell responses that combines ex vivo tetramer staining with various phenotyping and functional assays provides a novel approach to assess the functional potential of tumor-specific T lymphocytes and may also facilitate the optimization of vaccination protocols.

Combinatorial peptide libraries as an alternative approach to the identification of ligands for tumor-reactive cytolytic T lymphocytes.

The recent identification of molecularly defined human tumor antigens recognized by autologous CTLs has opened new opportunities for the development of antigen-specific cancer vaccines. Despite extensive work, however, the number of CTL-defined tumor antigens that are suitable targets for generic vaccination of cancer patients is still limited, mostly because of the painstaking and lengthy nature of the procedures currently used for their identification. A novel approach is based on the combined use of combinatorial peptide libraries in positional scanning format (positional scanning synthetic combinatorial peptide libraries, PS-SCLs) and tumor-reactive CTL clones. To validate this approach, we herein analyzed in detail the recognition of PS-SCLs by Melan-A-specific CTL clones. Our results indicate that, at least for some clones, most of the amino acids composing the native antigenic peptide can be identified through the use of PS-SCLs. Interestingly, this analysis also allowed the identification of peptide analogues with increased antigenic activity as well as agonist peptides containing multiple amino-acid substitutions. In addition, biometrical analysis of the data generated by PS-SCL screening allowed the identification of the native ligand in a public database. Overall, these data demonstrate the successful use of PS-SCLs for the identification and optimization of tumor-associated CTL epitopes.

Ex vivo IFN-gamma secretion by circulating CD8 T lymphocytes: implications of a novel approach for T cell monitoring in infectious and malignant diseases.

To elucidate the functional heterogeneity of Ag-specific T lymphocyte populations, we combined labeling of lymphocytes with MHC/peptide tetramers and a cell surface affinity matrix for IFN-gamma. Magnetic cell sorting of IFN-gamma-positive lymphocytes allowed the selective enrichment and identification of live Ag-specific cytokine-secreting cells by flow cytometry. Naive, memory, and effector Ag-specific populations were evaluated in healthy HLA-A2 individuals. Significant fractions of influenza- and CMV-specific cells secreted IFN-gamma upon challenge with cognate peptide, consistent with an effector/memory status. The sensitivity of the approach allowed the detection of significant numbers of CMV-specific IFN-gamma-secreting cells ex vivo (i.e., without Ag stimulation). This was not apparent when using previously described assays, namely, ELISPOT or intracellular IFN-gamma staining (cytospot). CD8+ T cells specific for the melamoma-associated Ag Melan-A/MART-1 did not produce IFN-gamma upon challenge with cognate peptide, reminiscent with their naive functional state in healthy individuals. In contrast, CD45RA(low) Melan-A/MART-1 tumor-specific cells from three of three melanoma patients presented levels of activity similar to those found for influenza- or CMV virus-specific lymphocytes, compatible with a functional differentiation into competent effector/memory T lymphocytes in vivo. Notably, a sizable fraction of Melan-A/MART-1-specific cells from a patient secreted IFN-gamma ex vivo following peptide-based vaccination. Thus, the high sensitivity of the assay provides a valuable tool to monitor effector T cell responses in different clinical situations.

CD8+ T-cell response to NY-ESO-1: relative antigenicity and in vitro immunogenicity of natural and analogue sequences.

We have shown previously that HLA-A*0201 melanoma patients can frequently develop a CTL response to the cancer testis antigen NY-ESO-1. In the present study, we have analyzed in detail the relative antigenicity and in vitro immunogenicity of natural and modified NY-ESO-1 peptide sequences. The results of this analysis revealed that, although suboptimal for binding to the HLA-A*0201 molecule, peptide NY-ESO-1 157-165 is, among natural sequences, very efficiently recognized by specific CTL clones derived from three melanoma patients. In contrast, peptides NY-ESO-1 157-167 and NY-ESO-1 155-163, which bind very strongly to HLA-A*0201, are recognized less efficiently. In agreement with previous data, substitution of peptide NY-ESO-1 157-165 COOH-terminal C with various other amino acids resulted in a significantly increased binding to HLA-A*0201 molecules as well as in an increased CTL recognition, although variable at the clonal level. Among natural peptides, NY-ESO-1 157-165 and NY-ESO-1 157-167 exhibited good in vitro immunogenicity, whereas peptide NY-ESO-1 155-163 was poorly immunogenic. The fine specificity of interaction between peptide NY-ESO-1 C165A, HLA-A*0201, and T-cell receptor was analyzed at the molecular level using a series of variant peptides containing single alanine substitutions. The findings reported here have significant implications for the formulation of NY-ESO-1-based vaccines as well as for the monitoring of either natural or vaccine-induced NY-ESO-1-specific CTL responses in cancer patients.

Expansion and functional maturation of human tumor antigen-specific CD8+ T cells after vaccination with antigenic peptide.

Peptide-based vaccines are currently being tested for their ability to induce or augment tumor antigen (Ag)-specific CD8+ T-cell responses in cancer patients. Here we report that the frequency of circulating CD8+ T cells directed against the Melan-A/MART-1 Ag increased >20-fold in an HLA-A2 melanoma patient immunized repeatedly with the corresponding antigenic peptide, as assessed by staining with HLA-A2/peptide tetramers. Multiparameter flow cytometric analysis demonstrated that the increase in total Melan-A-specific cell number was accompanied by a marked increase in the proportion of the cells that expressed an activated/memory surface phenotype. As assessed by ELISPOT assays and intracellular staining, the absolute number of Melan-A-specific cells able to secrete IFN-gamma increased >50-fold upon vaccination. When tested directly after cell sorting on the basis of tetramer staining, Melan-A-specific cells were weakly cytolytic but became highly active after in vitro restimulation. Altogether, these results indicate that large numbers of functionally active tumor Ag-specific CD8+ T cells can be obtained and maintained at high levels after in vivo activation by repeated peptide-based vaccination.

Melanoma patients respond to a cytotoxic T lymphocyte-defined self-peptide with diverse and nonoverlapping T-cell receptor repertoires.

HLA-A2+ melanoma patients develop naturally a strong CD8+ T cell response to a self-peptide derived from Melan-A. Here, we have used HLA-A2/peptide tetramers to isolate Melan-A-specific T cells from tumor-infiltrated lymph nodes of two HLA-A2+ melanoma patients and analyzed their TCR beta chain V segment and complementarity determining region 3 length and sequence. We found a broad diversity in Melan-A-specific immune T-cell receptor (TCR) repertoires in terms of both TCR beta chain variable gene segment usage and clonal composition. In addition, immune TCR repertoires selected in the patients were not overlapping. In contrast to previously characterized CD8+ T-cell responses to viral infections, this study provides evidence against usage of highly restricted TCR repertoire in the natural response to a self-differentiation tumor antigen.

Human CD8(+) T cells expressing HLA-DR and CD28 show telomerase activity and are distinct from cytolytic effector T cells.

Cycling lymphocytes may express the enzyme telomerase which is involved in maintenance of telomere length and cell proliferation potential. In CD8(+) T cells freshly isolated from peripheral blood, we found that in vivo cycling cells expressed HLA-DR. Furthermore, CD28-positive cells are known to have longer telomeres than CD28-negative T cells. Therefore we used HLA-DR- and CD28-specific antibodies to sort CD8(+) T cells and measure telomerase activity ex vivo. Relatively high levels of telomerase activity were found in HLA-DR/CD28 double-positive cells. In contrast, HLA-DR-negative and CD28-negative cells had almost no telomerase activity. In summary, HLA-DR expression correlates with proliferation, and CD28 expression with proliferative potential. We have previously identified that ex vivo cytolytic CD8(+) T cells are CD56 (NCAM) positive. Here we show that HLA-DR(+) cells were rarely CD56(+) and vice versa. This demonstrates that telomerase-expressing and cytolytic CD8(+) T cells can be separated on the basis of the cell surface markers HLA-DR and CD56. Thus, activated CD8(+) T cells specialize and exert distinct functions correlating with surface molecule expression.

Frequent cytolytic T-cell responses to peptide MAGE-A10(254-262) in melanoma.

MAGE genes encode tumor-specific shared antigens that are among the most interesting candidates for cancer vaccines. Despite extensive studies, however, CD8+ T-cell responses to MAGE-derived epitopes have been detected only occasionally in cancer patients, even after vaccination. In contrast with these findings, we report here that HLA-A2 melanoma patients respond frequently to the recently identified peptide MAGE-A10(254-262). Indeed, as assessed by staining with fluorescent HLA-A2/peptide MAGE-A10(254-262) tetramers, CD8+ T cells directed against this peptide were readily detectable in a large proportion of HLA-A2+ melanoma patients. These results provide new insight into the immunogenicity of MAGE antigens and underline the potential usefulness of MAGE-A10 peptide-based cancer vaccines.

Efficient simultaneous presentation of NY-ESO-1/LAGE-1 primary and nonprimary open reading frame-derived CTL epitopes in melanoma.

Recent studies have shown that CTL epitopes derived from tumor-associated Ags can be encoded by both primary and nonprimary open reading frames (ORF). In this study we have analyzed the HLA-A2-restricted CD8(+) T cell response to a recently identified CTL epitope derived from an alternative ORF product of gene LAGE-1 (named CAMEL), and the highly homologous gene NY-ESO-1 in melanoma patients. Using MHC/peptide tetramers we detected CAMEL(1-11)-specific CD8(+) T cells in peptide-stimulated PBMC as well as among tumor-infiltrated lymph node cells from several patients. Sorting and expansion of tetramer(+) CD8(+) T cells allowed the isolation of tetramer(bright) and tetramer(dull) populations that specifically recognized the peptide Ag with high and low avidity, respectively. Remarkably, only high avidity CAMEL-specific CTL were able to recognize Ag-expressing tumor cells. A large series of HLA-A2-positive melanoma cell lines was characterized for the expression of LAGE-1 and NY-ESO-1 mRNA and protein and tested for recognition by CAMEL-specific CTL as well as CTL that recognize a peptide (NY-ESO-1(157-165)) encoded by the primary ORF products of the LAGE-1 and NY-ESO-1 genes. This analysis revealed that tumor-associated CD8(+) T cell epitopes are simultaneously and efficiently generated from both primary and nonprimary ORF products of LAGE-1 and NY-ESO-1 genes and, importantly, that this occurs in the majority of melanoma tumors. These findings underscore the in vivo immunological relevance of CTL epitopes derived from nonprimary ORF products and support their use as candidate vaccines for inducing tumor specific cell-mediated immunity against cancer.

Generation and characterization of malaria-specific human CD8(+) lymphocyte clones: effect of natural polymorphism on T cell recognition and endogenous cognate antigen presentationby liver cells.

CD8(+) cytolytic T lymphocytes (CTL) play a fundamental role in the clearance of malaria parasites from the liver in mouse models. In humans, however, only low levels of parasite-specific CD8(+) T lymphocytes have been observed in individuals living in endemic areas. In the present study, we identified high levels of circulating CD8(+) T lymphocytes specific for a previously described HLA-A2-restricted CTL epitope of the circumsporozoite (CS) protein of Plasmodium falciparum in an adult living in Burkina Faso, as evidenced by IFN-gamma ELISPOT assay and MHC-tetramer technology. After cloning by limiting dilution culture, T cell recognition of natural CS variants of P. falciparum was studied. The results demonstrate that naturally occurring variations drastically affect residues critical for T cell recognition as only two out of nine sequences analyzed were efficiently recognized by the CTL clones. These clones were also used to analyze T cell recognition of the endogenously presented cognate antigen. We observed efficient antigen recognition of both HLA-A*0201-transfected murine antigen presenting cells and liver cells from HLA-A*0201/K(b)-transgenic mice upon infection with recombinant vaccinia virus encoding the CS protein (WR-CS). More importantly, we demonstrate for the first time efficient recognition of WR-CS-infected human liver cells.

Ectopic human telomerase catalytic subunit expression maintains telomere length but is not sufficient for CD8+ T lymphocyte immortalization.

Like most somatic human cells, T lymphocytes have a limited replicative life span. This phenomenon, called senescence, presents a serious barrier to clinical applications that require large numbers of Ag-specific T cells such as adoptive transfer therapy. Ectopic expression of hTERT, the human catalytic subunit of the enzyme telomerase, permits fibroblasts and endothelial cells to avoid senescence and to become immortal. In an attempt to immortalize normal human CD8(+) T lymphocytes, we infected bulk cultures or clones of these cells with a retrovirus transducing an hTERT cDNA clone. More than 90% of transduced cells expressed the transgene, and the cell populations contained high levels of telomerase activity. Measuring the content of total telomere repeats in individual cells (by flowFISH) we found that ectopic hTERT expression reversed the gradual loss of telomeric DNA observed in control populations during long term culture. Telomere length in transduced cells reached the levels observed in freshly isolated normal CD8(+) lymphocytes. Nevertheless, all hTERT-transduced populations stopped to divide at the same time as nontransduced or vector-transduced control cells. When kept in IL-2 the arrested cells remained alive. Our results indicate that hTERT may be required but is not sufficient to immortalize human T lymphocytes.

Naturally occurring human lymphocyte antigen-A2 restricted CD8+ T-cell response to the cancer testis antigen NY-ESO-1 in melanoma patients.

Cancer testis (CT) antigens are particularly interesting candidates for cancer vaccines. However, T-cell reactivity to CT antigens has been detected only occasionally in cancer patients, even after vaccination. A new group of CT antigens has been recently identified using the SEREX technique based on immunoscreening of tumor cDNA expression libraries with autologous sera. We have used fluorescent HLA-A2/peptide tetramers containing an optimized antigenic peptide to directly identify HLA-A2-restricted CD8+ T cells specific for the SEREX-defined CT antigen NY-ESO-1 in melanoma patients. High frequencies of NY-ESO-1-specific CD8+ T cells were readily detected in peptide-stimulated peripheral blood mononuclear cells as well as in lymphocytes infiltrating melanoma lesions from patients with measurable antibody responses to NY-ESO-1. NY-ESO-1-specific CD8+ T cells were also detectable in peptide-stimulated peripheral blood mononuclear cells from some seronegative patients. Whereas the frequencies of NY-ESO-1-specific CD8+ T cells in circulating lymphocytes were usually below the limit of detection by tetramer staining, the presence of NY-ESO-1 CD8+ T cells displaying a memory phenotype was clearly detectable ex vivo in blood from a seropositive patient over an extended period of time. These results indicate that sustained CD8+ T-cell responses to CT antigens can naturally occur both locally and systemically in melanoma patients.