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.

A peptide encoded by the human MAGE3 gene and presented by HLA-B44 induces cytolytic T lymphocytes that recognize tumor cells expressing MAGE3.

The human MAGE3 gene is expressed in a significant proportion of tumors of various histological types, but is silent in normal adult tissues other than testis and placenta. Antigens encoded by MAGE3 may therefore be useful targets for specific antitumor immunization. Two antigenic peptides encoded by the MAGE3 gene have been reported previously. One is presented to cytolytic T lymphocytes (CTL) by HLA-A1, the other by HLA-A2 molecules. Here we show that MAGE3 also codes for a peptide that is presented to CTL by HLA-B44. MAGE3 peptides containing the HLA-B44 peptide binding motif were synthesized. Peptide MEVDPIGHLY, which showed the strongest binding to HLA-B44, was used to stimulate blood T lymphocytes from normal HLA-B44 donors. CTL clones were obtained that recognized not only HLA-B44 cells sensitized with the peptide, but also HLA-B44 tumor cell lines expressing MAGE3. The proportion of metastatic melanomas expressing the MAGE3/HLA-B44 antigen should amount to approximately 17% in the Caucasian population, since 24% of individuals carry the HLA-B44 allele and 76% of these tumors express MAGE3.

A nonapeptide encoded by human gene MAGE-1 is recognized on HLA-A1 by cytolytic T lymphocytes directed against tumor antigen MZ2-E.

We have reported the identification of human gene MAGE-1, which directs the expression of an antigen recognized on a melanoma by autologous cytolytic T lymphocytes (CTL). We show here that CTL directed against this antigen, which was named MZ2-E, recognize a nonapeptide encoded by the third exon of gene MAGE-1. The CTL also recognize this peptide when it is presented by mouse cells transfected with an HLA-A1 gene, confirming the association of antigen MZ2-E with the HLA-A1 molecule. Other members of the MAGE gene family do not code for the same peptide, suggesting that only MAGE-1 produces the antigen recognized by the anti-MZ2-E CTL. Our results open the possibility of immunizing HLA-A1 patients whose tumor expresses MAGE-1 either with the antigenic peptide or with autologous antigen-presenting cells pulsed with the peptide.

Combinatorial peptide library-based identification of peptide ligands for tumor-reactive cytolytic T lymphocytes of unknown specificity.

A novel approach for the identification of tumor antigen-derived sequences recognized by CD8(+) cytolytic T lymphocytes (CTL) consists in using synthetic combinatorial peptide libraries. Here we have screened a library composed of 3.1 x 10(11) nonapeptides arranged in a positional scanning format, in a cytotoxicity assay, to search the antigen recognized by melanoma-reactive CTL of unknown specificity. The results of this analysis enabled the identification of several optimal peptide ligands, as most of the individual nonapeptides deduced from the primary screening were efficiently recognized by the CTL. The results of the library screening were also analyzed with a mathematical approach based on a model of independent and additive contribution of individual amino acids to antigen recognition. This biometrical data analysis enabled the retrieval, in public databases, of the native antigenic peptide SSX-2(41-49), whose sequence is highly homologous to the ones deduced from the library screening, among the ones with the highest stimulatory score. These results underline the high predictive value of positional scanning synthetic combinatorial peptide library analysis and encourage its use for the identification of CTL ligands.

In vivo expression of natural killer cell inhibitory receptors by human melanoma-specific cytolytic T lymphocytes.

Natural killer (NK) receptor signaling can lead to reduced cytotoxicity by NK cells and cytolytic T lymphocytes (CTLs) in vitro. Whether T cells are inhibited in vivo remains unknown, since peptide antigen-specific CD8(+) T cells have so far not been found to express NK receptors in vivo. Here we demonstrate that melanoma patients may bear tumor-specific CTLs expressing NK receptors. The lysis of melanoma cells by patient-derived CTLs was inhibited by the NK receptor CD94/NKG2A. Thus, tumor-specific CTL activity may be decreased through NK receptor triggering in vivo.

CD28-negative cytolytic effector T cells frequently express NK receptors and are present at variable proportions in circulating lymphocytes from healthy donors and melanoma patients.

In humans, NK receptors are expressed by natural killer cells and some T cells, the latter of which are preferentially alphabetaTCR+ CD8+ cytolytic T lymphocytes (CTL). In this study we analyzed the expression of nine NK receptors (p58.1, p58.2, p70, p140, ILT2, NKRP1A, ZIN176, CD94 and CD94/NKG2A) in PBL from both healthy donors and melanoma patients. The percentages of NK receptor-positive T cells (NKT cells) varied strongly, and this variation was more important between individual patients than between individual healthy donors. In all the individuals, the NKT cells were preferentially CD28-, and a significant correlation was found between the percentage of CD28- T cells and the percentage of NK receptor+ T cells. Based on these data and the known activated phenotype of CD28- T cells, we propose that the CD28- CD8+ T cell pool represents or contains the currently active CTL population, and that the frequent expression of NK receptors reflects regulatory mechanisms modulating the extent of CTL effector function. Preliminary results indicate that some tumor antigen-specific T cells may indeed be CD28- and express NK receptors in vivo.

In vivo localization of a bispecific antibody which targets human T lymphocytes to lyse human colon cancer cells.

A bispecific MAb was derived from the fusion of a hybridoma producing MAb CD3 with a hybridoma producing MAb L-DI (which is directed against a 41-kDa glycoprotein expressed in most gastro-intestinal and pancreatic carcinomas). Bispecific antibody molecules were isolated from parental antibody molecules by the use of hydroxylapatite-HPLC and shown to target human cytolytic T lymphocytes, irrespective of their original specificity, to specifically lyse human colon carcinoma cells. Localization studies in vivo using nude mice bearing human colon carcinoma xenografts showed significant accumulation of the HPLC-purified 125I-labelled bispecific antibodies into the tumor compared to 131I-labelled control CD3 antibody.

Flow cytofluorometric analysis of the binding of Vicia villosa lectin to T lymphoblasts: lack of correlation with cytolytic function.

The relationship between the binding of Vicia villosa (VV) lectin and the expression of cytolytic function in T lymphoblasts has been investigated using flow cytofluorometric techniques. Spleen cells activated in vitro in 5-day mixed leukocyte cultures (MLC) were incubated sequentially with VV, rabbit anti-V antiserum, and fluoresceinated sheep anti-rabbit IgG. When these stained MLC cells were passed on a flow cytometer gated to exclude nonviable cells and small lymphocytes, a single heterogeneous peak of fluorescence was seen, as compared to control MLC cells that had not been incubated with VV. Fluorescence of lymphoblasts was dependent upon lectin dose and was eliminated when staining was performed in the presence of N-acetyl-D-galactosamine, the appropriate competitive sugar for VV. T cell blast populations activated against H-2, Mls, or parasite antigens all had comparable levels of fluorescence after staining with VV, although the cytolytic activity of these cells varied widely. Furthermore, when MLC lymphoblasts binding large or small amounts of VV were sorted on the basis of their relative fluorescence intensity and tested for cytolytic function, no appreciable difference in activity between the 2 populations was observed. These results are inconsistent with the hypothesis that VV binds selectively to cytolytic T lymphocytes.

Direct analysis of peptide binding to cell-associated MHC class I molecules.

Exogenously added synthetic peptides can mimic endogenously produced antigenic peptides recognized on target cells by MHC class I-restricted cytolytic T lymphocytes. While it is assumed that exogenous peptides associate with class I molecules on the target cell surface, direct binding of peptides to cell-associated class I molecules has been difficult to demonstrate. Using a newly developed binding assay based on photoaffinity labeling, we have investigated the interaction of two antigenic peptides, known to be recognized in the context of H-2Kd or H-2Db, respectively, with 20 distinct class I alleles on living cells. None of the class I alleles tested, with the exception of H-2Kd or H-2Db, bound either of the peptides, thus demonstrating the exquisite specificity of peptide binding to class I molecules. Moreover, peptide binding to cell-associated H-2Kd was drastically reduced when metabolic energy, de novo protein synthesis or protein egress from the endoplasmic reticulum was inhibited. It is thus likely that exogenously added peptides do not associate with the bulk of class I molecules expressed at the cell surface, but rather bind to short-lived molecules devoid of endogenous peptides.

Human CD8(+) T cells engineered with supraphysiological TCRs are functionally inhibited via PD-1 and SHP-1 phosphatase

Purpose/Objective: Protective CD8+ T cell responses rely on TCRdependent recognition of immunogenic peptides presented by MHC I. Cytolytic T lymphocytes directed against self/tumor antigens express TCRs of lower affinity/avidity than pathogen-derived T lymphocytes and elicit less protective immune responses due to mechanisms of central and peripheral tolerance. Anti-tumor T cell reactivity can be improved by increasing the TCR-pMHC affinity within physiological limits, while intriguingly further increase in the supraphysiological range (KD < 1 lM) leads to drastic functional declines. We aim at identifying the molecular mechanisms underlying the loss of T cell responsiveness associated with supraphysiological TCRpMHC affinities in order to improve effectiveness of TCR-engineered T cells used in adoptive cell transfer (ACT) cancer immunotherapy. Materials and methods: Using a panel of human CD8+ T cells engineered with TCRs of incremental affinity for the HLA-A2-resticted tumor cancer testis antigen NY-ESO-1, we performed comparative gene expression microarray and TCR-mediated signaling analysis together with membrane receptors level analysis. Results: As compared to cells expressing TCR affinities generating optimal function (KD from 5to 1 lM), those with supraphysiological affinity (KD from 1 lM to 15 nM) had an overall reduced expression of genes implied in signaling, cell activation and proliferation, and showed impaired proximal and distal TCR signaling capacity. This correlated with a decline in surface expression of CD8b, CD28 and activatory TNFR superfamily members. Importantly, expression of inhibitory receptor PD-1 and SHP-1 phosphatase was upregulated in a TCR affinity-dependent manner. Consequently, PD-L1 and SHP-1 blockade restored the function of T cells with high TCRs affinity. Moreover, SHP-1 inhibition also augmented functional efficacy of T cells with TCRs of optimal affinity. Conclusions: Our findings indicate that TCR affinity-associated regulatory mechanisms control T cells responsiveness at various levels to limit potential auto-reactive cytotoxic effects. They also support the development of ACT therapies combined with blockade of inhibitory molecules such as SHP-1 to enhance effectiveness of T cell immunotherapy.

Efficient in vivo induction of CTL by cell-associated covalent H-2Kd-peptide complexes.

A novel procedure is presented describing the induction of antigen-specific cytolytic T lymphocytes (CTL) in vivo, that uses as immunogen syngeneic Concanavalin A stimulated spleen cells expressing H-2Kd (Kd) molecules photocrosslinked with a photoreactive peptide derivative. The Kd restricted Plasmodium berghei circumsporozoite (PbCS) peptide 253-260 (YIPSAEKI) was conjugated with photoreactive iodo-4-azidosalicylic acid (IASA) at the NH2-terminus and with 4-azidobenzoic acid (ABA) at the TCR contact residue Lys259 to make IASA-YIPSAEK(ABA)I. Selective photoactivation of the IASA group allowed specific photoaffinity labeling of cell-associated Kd molecules. Optimal peptide derivative binding to Kd molecules of concanavalin A stimulated spleen cells was obtained upon 4-6 h incubation at 26 degrees C in the presence of human beta 2 microglobulin. Photocrosslinking prevented the rapid dissociation of cell-associated Kd-peptide derivative complexes at 37 degrees C. The photoaffinity labeled cells were injected i.p. into syngeneic recipients. After 10 days, the peritoneal exudate lymphocytes were harvested and in vitro stimulated with peptide derivative pulsed P815 mastocytoma cells. The resulting bulk cultures displayed high cytolytic activity that was specific for IASA-YIPSAEK(ABA)I and YIPSAEK(ABA)I. In contrast, peritoneal exudate lymphocytes from mice inoculated with concanavalin A blasts that were pulsed, but not photocrosslinked, with IASA-YIPSAEK(ABA)I expressed only marginal levels of IASA-YIPSAEK(ABA)I-specific cytolytic activity. This immunization strategy, using neither adjuvants nor potentially hazardous transfected/transformed cells, is safe and should be universally applicable.

Identification of new MHC class I-restricted human tumor antigens for immunotherapy

Summary Cancer is a leading cause of morbidity and mortality in Western countries (as an example, colorectal cancer accounts for about 300'000 new cases and 200'000 deaths each year in Europe and in the USA). Despite that many patients with cancer have complete macroscopic clearance of their disease after resection, radiotherapy and/or chemotherapy, many of these patients develop fatal recurrence. Vaccination with immunogenic peptide tumor antigens has shown encouraging progresses in the last decade; immunotherapy might therefore constitute a fourth therapeutic option in the future. We dissect here and critically evaluate the numerous steps of reverse immunology, a forecast procedure to identify antigenic peptides from the sequence of a gene of interest. Bioinformatic algorithms were applied to mine sequence databases for tumor-specific transcripts. A quality assessment of publicly available sequence databanks allowed defining strengths and weaknesses of bioinformatics-based prediction of colon cancer-specific alternative splicing: new splice variants could be identified, however cancer-restricted expression could not be significantly predicted. Other sources of target transcripts were quantitatively investigated by polymerase chain reactions, as cancer-testis genes or reported overexpressed transcripts. Based on the relative expression of a defined set of housekeeping genes in colon cancer tissues, we characterized a precise procedure for accurate normalization and determined a threshold for the definition of significant overexpression of genes in cancers versus normal tissues. Further steps of reverse immunology were applied on a splice variant of the Melan¬A gene. Since it is known that the C-termini of antigenic peptides are directly produced by the proteasome, longer precursor and overlapping peptides encoded by the target sequence were synthesized chemically and digested in vitro with purified proteasome. The resulting fragments were identified by mass spectroscopy to detect cleavage sites. Using this information and based on the available anchor motifs for defined HLA class I molecules, putative antigenic peptides could be predicted. Their relative affinity for HLA molecules was confirmed experimentally with functional competitive binding assays and they were used to search patients' peripheral blood lymphocytes for the presence of specific cytolytic T lymphocytes (CTL). CTL clones specific for a splice variant of Melan-A could be isolated; although they recognized peptide-pulsed cells, they failed to lyse melanoma cells in functional assays of antigen recognition. In the conclusion, we discuss advantages and bottlenecks of reverse immunology and compare the technical aspects of this approach with the more classical procedure of direct immunology, a technique introduced by Boon and colleagues more than 10 years ago to successfully clone tumor antigens.

In vivo imaging of T cell delivery to tumors after adoptive transfer therapy.

Adoptive transfer therapy of in vitro-expanded tumor-specific cytolytic T lymphocytes (CTLs) can mediate objective cancer regression in patients. Yet, technical limitations hamper precise monitoring of posttherapy T cell responses. Here we show in a mouse model that fused single photon emission computed tomography and x-ray computed tomography allows quantitative whole-body imaging of (111)In-oxine-labeled CTLs at tumor sites. Assessment of CTL localization is rapid, noninvasive, three-dimensional, and can be repeated for longitudinal analyses. We compared the effects of lymphodepletion before adoptive transfer on CTL recruitment and report that combined treatment increased intratumoral delivery of CTLs and improved antitumor efficacy. Because (111)In-oxine is a Food and Drug Administration-approved clinical agent, and human SPECT-CT systems are available, this approach should be clinically translatable, insofar as it may assess the efficacy of immunization procedures in individual patients and lead to development of more effective therapies.

Vaccination with a recombinant protein encoding the tumor-specific antigen NY-ESO-1 elicits an A2/157-165-specific CTL repertoire structurally distinct and of reduced tumor reactivity than that elicited by spontaneous immune responses to NY-ESO-1-expressing Tumors.

In a recent vaccination trial assessing the immunogenicity of an NY-ESO-1 (ESO) recombinant protein administered with Montanide and CpG, we have obtained evidence that this vaccine induces specific cytolytic T lymphocytes (CTL) in half of the patients. Most vaccine-induced CTLs were directed against epitopes located in the central part of the protein, between amino acids 81 and 110. This immunodominant region, however, is distinct from another ESO CTL region, 157-165, that is a frequent target of spontaneous CTL responses in A2+ patients bearing ESO tumors. In this study, we have investigated the CTL responses to ESO 157-165 in A2+ patients vaccinated with the recombinant protein. Our data indicate that after vaccination with the protein, CTL responses to ESO 157-165 are induced in some, but not all, A2+ patients. ESO 157-165-specific CTLs induced by vaccination with the ESO protein were functionally heterogeneous in terms of tumor recognition and often displayed decreased tumor reactivity as compared with ESO 157-165-specific CTLs isolated from patients with spontaneous immune responses to ESO. Remarkably, protein-induced CTLs used T-cell receptors similar to those previously isolated from patients vaccinated with synthetic ESO peptides (Vbeta4.1) and distinct from those used by highly tumor-reactive CTLs isolated from patients with spontaneous immune responses (Vbeta1.1, Vbeta8.1, and Vbeta13.1). Together, these results demonstrate that vaccination with the ESO protein elicits a repertoire of ESO 157-165-specific CTLs bearing T-cell receptors that are structurally distinct from those of CTLs found in spontaneous immune responses to the antigen and that are heterogeneous in terms of tumor reactivity, being often poorly tumor reactive.

Therapeutic cancer vaccines based on molecularly defined human tumor antigens.

The results of numerous phases I and II clinical trials testing the safety and immunogenicity of various cancer vaccine formulations based on cytolytic T lymphocytes (CTLs)-defined tumor antigens have been reported recently. Specific T cell responses can be detected in only a fraction of immunized patients. A smaller but significant fraction of these patients have objective tumor responses. Efficient therapeutic vaccination should aim at boosting naturally occurring anti-tumor responses and at sustaining a large contingent of tumor antigen-specific and fully functional effector T cells at tumor sites.