Abstract 2562: Affinity-enhanced T-cell receptor (TCR) for adoptive T-cell therapy targeting MAGE-A4

Melanoma-associated antigens-A (MAGE-A) are members of an extensive family of cancer/testis (CT) antigens. Although emerging studies highlight the potential role of MAGE-A family members in promoting cancer growth, the biological function of MAGE-A4 remains poorly understood. MAGE-A4 is an attractive target because it has a high frequency of expression in multiple solid tumors according to The Cancer Genome Atlas (https://cancergenome.nih.gov/). In an attempt to further characterize its expression in solid tumors, immunohistochemistry analysis was performed in 534 resected non-small cell lung carcinoma (NSCLC) cases, stage I to IV with clinicopathological information including overall survival and recurrence. MAGE-A4 expression was observed in 24% of all NSCLC cases, with higher frequency observed in squamous cell carcinoma (SCC) (51%) versus adenocarcinoma (8%). MAGE-A4 expression was observed in 23% of the primary tumors and 48% of the recurrent tumor samples. Log-rank (Mantel-Cox) analysis showed that histology diagnosis, pathological stage, and MAGE-A4 expression are independent prognostic factors in NSCLC. However, no MAGE-A4 prognostic value was observed when analyzing the SCC subtype. An HLA-A*02-restricted MAGE-A4 TCR specific for MAGE-A4230-239 peptide GVYDGREHTV was identified from a pool of parental TCRs and engineered to optimize specificity and activity (MAGE-A4c1032). TCR fine specificity was assessed initially by mapping the response of the MAGE-A4c1032T-cells to panels of synthetic peptides with combinatorial substitutions at each amino acid with every other possible amino acid (X-Scan), to identify potentially cross-reactive peptides in the human and common pathogen proteomes. Secondly, MAGE-A4c1032T-cells were screened against a wide panel of normal primary cells from multiple organ systems, induced pluripotent stem cell-derived cells (iCells), and autologous whole blood to test for off-target reactivity, and against a panel of EBV-derived B-lymphoblastic cell lines expressing a wide range of HLA molecules to assess the risk of alloreactivity. Alloreactivity was observed in antigen negative cells expressing HLA-A*02:05, and therefore this allele is exclusionary. MAGE-A4c1032T-cell potency was assessed by a variety of in vitro assays, including proliferation, IFN-γ release and cytotoxicity in response to antigen-positive tumor lines in 2D and 3D culture, and cytokine release in response to freshly prepared antigen-positive primary tumor ma