Melanoma patients were injected with various vaccines containing a MAGE-A3 peptide presented by HLA-DP4. Anti-MAGE-A3.DP4 T cells were not detectable in the blood before vaccination, but their frequencies after vaccination ranged from 2 Â 10 À6 to 2 Â 10 À3 among the CD4 + blood T lymphocytes of the patients. The CD4 + blood T lymphocytes that stained ex vivo with HLA-DP4 tetramers folded with the MAGE-A3 peptide were selected by flow cytometry and amplified under clonal conditions. About 5% of the CD4 + T-cell clones that recognized the MAGE-A3.DP4 antigen had a CD25 + phenotype in the resting state. These CD25 + clones had a high capacity to suppress the proliferation of another T-cell clone after peptide stimulation in vitro. Most of them had high FOXP3 expression in the resting state and an unmethylated FOXP3 intron 1. They produced active transforming growth factor-B but none of cytokines IFN-;, interleukin-2 (IL-2), IL-4, IL-5, and IL-10. About 20% of CD25 À clones had a significant but lower suppressive activity. Most of the CD25 À clonal populations contained cells that expressed FOXP3 in the resting state, but FOXP3 demethylation was not observed. We conclude that MAGE-A3.DP4 vaccination can produce CD4 + T cells that may exert regulatory T-cell function in vivo.

We report that melanoma cell lines expressing the interleukin-1 receptor exhibit 4-to 10-fold lower levels of mRNA of microphthalmia-associated transcription factor (MITF-M) when treated with interleukin-1b. This effect is NF-jB and JNKdependent. MITF-M regulates the expression of melanocyte differentiation genes such as MLANA, tyrosinase and gp100, which encode antigens recognized on melanoma cells by autologous cytolytic T lymphocytes. Accordingly, treating some melanoma cells with IL-1b reduced by 40-100% their ability to activate such antimelanoma cytolytic T lymphocytes. Finally, we observed large amounts of biologically active IL-1a or IL-1b secreted by two melanoma cell lines that did not express MITF-M, suggesting an autocrine MITF-M downregulation. We estimate that~13% of melanoma cell lines are MITF-M-negative and secrete IL-1 cytokines. These results indicate that the repression of melanocyte-differentiation genes by IL-1 produced by stromal cells or by tumor cells themselves may represent an additional mechanism of melanoma immune escape.

We have observed that human CTL clones lose their specific cytolytic activity and cytokine production under certain stimulation conditions, while retaining an antigen‐dependent growth pattern.These inactive CTL simultaneously lose their labeling by an HLA‐peptide tetramer, even though the amount of TCR‐CD3 at their surface is not reduced. The tetramer‐negative cells recover tetramer staining and cytolytic activity after stimulation with tumor cells in the presence of a supernatant of activated lymphocytes. Our results suggest the existence of a new type of functional defect of CTL. They also indicate that tetramers may fail to reveal some CTL bearing the relevant TCR, even when such functionally arrested CTL retain the potential to participate in immune responses because their defect is reversible.

"Cancer-germline" genes such as those of the MAGE family are expressed in many tumors and in male germline cells, but are silent in normal tissues. They encode shared tumor-specific antigens that have been used in therapeutic vaccination trials of cancer patients. It was previously demonstrated that MAGE-1 peptide KVLEYVIKV was presented by HLA-A 0201 molecules on the surface of a human breast carcinoma cell line, but no human specific CTL had been isolated so far. Here, we have used HLA-A2/MAGE-1 fluorescent multimers to isolate from blood cells three human CTL clones that recognized the MAGE-1 peptide. These clones killed efficiently HLA-A2 tumor cells expressing MAGE-1, whether or not they were treated with IFN-gamma, suggesting that the MAGE-1 antigen is processed efficiently by both the standard proteasome and the immunoproteasome. These results indicate that the MAGE-1.A2 peptide can be used for antitumoral vaccination.

"Cancer-germline" genes such as those of the MAGE family are expressed in many tumors and in male germline cells, but are silent in other normal tissues. They encode tumor specific antigens that are used in cancer immunotherapy trials. MAGE-4 antigens represent promising targets for cancer immunotherapy because gene MAGE-4 is expressed in more than 50% of carcinomas of the esophagus, lung, bladder, and head and neck. To identify new MAGE-4 antigenic peptides, we have folded HLA-A*2402 soluble molecules with candidate peptide NYKRCFPVI, which corresponds to amino acids 143 to151 of the MAGE-4 protein. A24/MAGE-4 multimers were used to isolate a cytolytic T cell clone that recognized the MAGE-4 peptide from the blood cells of a donor without cancer. This clone lysed specifically A24 carcinoma cells expressing MAGE-4. The antigenic peptide is processed more efficiently in tumor cells pre-treated with IFN-gamma. This MAGE-4 peptide could represent an interesting target for immunotherapy because it is presented by HLA-A24 molecules, which are widely expressed in different ethnic groups.

Supplementary Table 1 from The CD4⁺ T-Cell Response of Melanoma Patients to a MAGE-A3 Peptide Vaccine Involves Potential Regulatory T Cells

Supplementary Figure 5 from The CD4⁺ T-Cell Response of Melanoma Patients to a MAGE-A3 Peptide Vaccine Involves Potential Regulatory T Cells

Supplementary Figure Legends 1-6 from The CD4⁺ T-Cell Response of Melanoma Patients to a MAGE-A3 Peptide Vaccine Involves Potential Regulatory T Cells