Abstract LBA006: Identification of novel, tumor specific pHLA targets recognized by TILs from CPI responders in a high-throughput, high-diversity pHLA library screening platform (3T-TRACE)

Abstract Background The ideal cancer target antigen is selectively expressed on neoplastic cells, absent on normal tissues, and contributes to the survival of cancer cells. Recent clinical trial reports identified T cell receptors (TCRs) recognizing intracellular targets presented as peptides in the context of HLA (pHLA) molecules as the critical molecular mediators of deep and long-lasting anti-tumor responses in a few solid tumor patients with complete responses to checkpoint inhibitor (CPI) treatment. The pHLA targets recognized by these TCRs may thus represent the long sought-after, ideal targets for solid tumor targeting with high-potency therapeutic modalities. Here we describe an immune-response guided target identification platform (3T-TRACE) and its utility to identify novel, intracellular targets, and their corresponding T cell receptors (TCRs) for the treatment of solid tumors. We provide an overview of the frequency and expression levels of the most promising, novel cancer testis antigens (CTAs) and tumor associated antigens (TAAs) identified. Methods By employing single cell sequencing methods, we identified TCRs that are uniquely associated with CPI responses and generated recombinant TCRs for screening against yeast display libraries expressing synthetic peptides in the context of the matching HLA complex, reviewed in (Figure 1). We analyzed over 140,000 CD8/CD4+ TILs from over 90 patient tumor biopsies and/or peripheral blood samples, including colorectal, lung, breast, ovarian, melanoma, and renal cancer (Figure 2). High value TCRs were selected based on the following criteria: high TCR clonality, immunophenotype indicative of T cell activation, shared alpha and/or beta TCR chains across patient samples (Public TCRs) and TCRs with anti-tumor activities. Statistical- and machine learning algorithms were developed to predict human peptides from the library selection data. Predicted target peptides were tested in in vitro T cell co-cultures for TCR activation. Antigen processing assays or mass spec analysis was employed to validate immunological processing of the epitope to identify bona fide targets. Results Among the first 100 targets identified by 3T-TRACE, we found 15 novel cancer testis antigens (CTAs) and 42 tumor associated antigens (TAAs) that were not previously described as targets in CPI treated solid tumor patients (Figure 3). Conclusions By combining TCR repertoire profiling with a library-based, antigen identification approach we found a strong enrichment of CTA- and TAA targets recognized by TCRs on TILs from CPI responders (Figure 4). Our findings are consistent with previous reports, demonstrating long-lasting responses induced by epitope spreading and CD8 T cell responses towards self-antigens, including CTAs and TAAs. Citation Format: Hans-Peter Gerber, Marvin Gee, Leah Sibener. Identification of novel, tumor specific pHLA targets recognized by TILs from CPI responders in a high-throughput, high-diversity pHLA library screening platform (3T-TRACE) [abstract]. In: Proceedings of the AACR-NCI-EORTC Virtual International Conference on Molecular Targets and Cancer Therapeutics; 2021 Oct 7-10. Philadelphia (PA): AACR; Mol Cancer Ther 2021;20(12 Suppl):Abstract nr LBA006.