Elucidating the immunological effects of 5-azacytidine treatment in patients with myelodysplastic syndrome and identifying new conditional ligands and T-cell epitopes of relevance in melanoma.

This review is focused on research within three different areas of tumor immunology: discovery of new T-cell epitopes and a new immunological antigen (reported in Paper I and II), elucidation of the immunological effects of treatment with a hypomethylating drug (reported in Paper III) and discovery of new conditional ligands (reported in Paper IV). Many melanoma-associated T-cell epitopes have been described, but 45% of these are restricted to human leukocyte antigen (HLA)-A2, leaving the remaining 36 different HLA molecules with only a few described T-cell epitopes each. Therefore we wanted to expand the number of T-cell epitopes restricted to HLA-A1, -A3, -A11 and -B7, all HLA molecules frequently expressed in Caucasians in Western Europe and Northern America. In Paper I we focused on the proteins gp100, Mart1, MAGE-A3, NY-ESO-1, tyrosinase and TRP-2, all melanoma-associated antigens frequently recognized by T cells from HLA-A2 patients. On contrary, in Paper II we wanted to investigate the protein Nodal as a novel immunological target. We took advantage of a T-cell epitope mapping platform in which HLA ligands are predicted by computer-based algorithms, further tested in the laboratory by an ELISA-based method and used for flow cytometry-based detection of specific T-cell responses by use of combinatorial encoded major histocompatibility (MHC) class I multimers. This procedure resulted in 127 (Paper I) and 32 (Paper II) confirmed HLA ligands, respectively, which we used for screening of the T-cell recognition within peripheral blood mononuclear cell samples from melanoma patients. As spontaneous tumor-specific T-cell responses tend to be of very low frequency and probably below the detection threshold of the method, we incorporated a T-cell enrichment step prior to the detection of these responses. Our screening of 39 melanoma patients resulted in 26 (17 different) T-cell responses against the common melanoma-associated antigens and 10 (8 different) T-cell responses against Nodal. We were further able to show processing and presentation on the cell-surface in K562 and melanoma cells expressing relevant protein and HLA molecules of four of these peptide sequences from tyrosinase, gp100 (2 peptides) and Nodal, respectively. However, one of the gp100 peptides has previously been described as a T-cell epitope. In addition to identifying new melanoma-associated T-cell epitopes we could thus describe Nodal as a new immunological antigen found of relevance in melanoma patients. In Paper III we wanted to investigate if the hypomethylating drug 5-azactytidine (Vidaza, Celgene Inc.) modulates the immune system in patients with myeloproliferative diseases. It has previ-ously been shown that 5-azacytidine-mediated demethylation of gene promoter regions results in enhanced transcription and expression of tumor suppressor genes and cancer-testis antigens. Cancer-testis antigens have frequently been recognized by T-cells in many cancers, and we hypothesized that 5-azacytidine treat-ment in the clinic would increase their frequency with resulting enhanced anti-tumor reactivity. We investigated separately the effect on T cells and tumor cells, and found that tumor cells af-fected by the treatment were better recognized, resulting in higher numbers of activated T cells, than tumor cells not exposed to 5-azacytidine. No effects were observed on the T-cell population. A screen of the T-cell recognition of 43 cancer-testis antigens in blood from our patients revealed increased T-cell recognition upon start of therapy which, though, stabilized or declined at later time points. We further investigated the general immune effector and inhibitory cell populations and found only minor effects of drug exposure, suggesting that 5-azacytidine primarily affects the tumor cells. From these results we are currently initiating a phase I clinical trial of cancer-testis antigen-peptide vaccination in combination with 5-azacytidine therapy for patients with myeloproliferative diseases. In Paper IV we wanted to expand the library of conditional ligands for use with the UV light-mediated peptide-exchange method. This method enables high-throughput generation of MHC class I molecules with different peptide-specificities. These MHC monomers can be multimerized and used for detection of specific T cell populations by flow or mass cytometry. The HLA molecules are highly genetically variable and this necessitates unique design of conditional ligands for each HLA molecule. Thus, to screen for the T-cell recognition in a given setting within all patients or healthy donors present in a cohort, a broad library of conditional ligands is needed. We designed and evaluated conditional ligands for HLA-B*08:01, HLA-B*35:01 and HLA-B*44:02/03/05, all HLA-B molecules present in high frequency among Caucasians. In addition, we provided proof for the use of a conditional ligand first designed for HLA-B*15:02 in complex with HLA-B*15:01. We compared the staining patterns of HLA-B*15:01 and HLA-B*15:02 MHC multimers and found remarkable dissimilarities, although the two heavy chains in these MHC molecules only differ in a few amino acid positions.