Abstract Mutation-derived neoantigens are important targets of T-cell mediated reactivity towards tumors. Their unique tumor-restriction poses an advantage compared to shared tumor antigens in that they are in principle both foreign and tumor specific, hence presumably less impacted by T-cell tolerance and for therapeutic applications less prone to mediate immune-related destruction of noncancerous tissue. Moreover, the mutational burden and predicted number of neoantigens correlate to favorable clinical outcome and benefit from immune checkpoint therapy. Neoantigen-reactive T-cells have been detected across a number of solid cancers, ranging from immunogenic tumors such as melanoma and non-small cell lung cancer to less immunogenic tumors such as breast cancer. Renal cell carcinomas (RCCs) are among medium-range mutational burden tumors and present with the highest pan-cancer number and proportion of frameshift mutations, a mutation type considered to be highly immunogenic. However, to our knowledge, yet no reports have described neoantigen-specific T-cells in this malignancy. In this study, the mutational landscape and HLA (human leukocyte antigen) profile of tumors from six renal cell carcinoma patients were analyzed by whole-exome sequencing (WXS) of DNA from tumor fragments (TFs), autologous tumor cell lines (TCLs) and tumor-infiltrating lymphocytes (TILs, germline reference). Hereafter the online MuPeXi tool was used to predict binding of mutated peptide sequences of 9-11mer length to the HLAs of each patient, using a rank score < 2 for selection of peptide binding, hereby creating patient-specific libraries of putative neo-peptides. TILs extracted from the patients tumors were screened for T-cell recognition of the peptide libraries by use of a novel high-throughput platform based on DNA barcode labeled peptide-MHC multimers, and responses were verified by conventional fluorochrome labeled MHC multimers. In four of six patients WXS was performed on both TF and TCL, in two of six patients only on TF. The average mutational burden of the six patients was 271 for TF (range 146–381, n=6) and 289 for TCL (range 182-404, n=4). Prediction of HLA-restricted peptides within the mutated sequences resulted in patient specific libraries of average 269 peptides for TF and TCL combined (range 126-443, n=6). Half of the peptides were predicted from both sources (52%, range 28-74%, n=4) compared to 20% (range 8-31%, n=4) predicted solely from TF and 29% (range 18-41%, n=4) predicted solely from TCL. The proportion of predicted peptides derived from frameshift mutations out of total mutations was 16% (range 7-24%, n=6). A total of 67 neoantigen-specific T-cell responses were detected across all patients by use of a novel high-throughput DNA barcode screening platform, with the number of detected responses ranging from 4-30 and spanning 3-5 HLA restrictions per patient. Of note, we detected a number of T-cell responses towards HLA-C restricted peptides, which have previously been poorly described. For several patients, the number of HLA-C restricted T-cell responses observed was substantially higher than for both HLA-A and -B, highlighting the importance of including this HLA type for neoepitope analyses. In the four patients in whom peptides were predicted from both TF and TCL, the distribution of responses was 37% on TF (range 14-60%, n=4), 36% on TCL (range 29-43%, n=4) and 27% (range 0-43%) on TF+TCL combined. The proportion of responses towards frameshift mutations was 17% (range 0-24%, n=6) of total responses. The identification of neoantigen-specific T-cells within tumors from RCC patients is an important step towards the use of neoantigens as therapeutic targets and predictors of response to immunotherapy in this cancer subtype. Moreover, our study points toward the importance of broad peptide prediction platforms covering multiple sources for WXS and mutational analyses covering both point and frameshift mutations. Citation Format: Sofie Ramskov, Ulla Kring Hansen, Anne-Mette Bjerregaard, Amalie Kai Bentzen, Marco Donia, Rikke Andersen, Zoltan Szallasi, Inge Marie Stentoft Svane, Aron Charles Eklund, Sine Reker Hadrup. Tumor infiltrating T-cells from renal cell carcinoma patients recognize neoantigens derived from point and frameshift mutations [abstract]. In: Proceedings of the Fourth CRI-CIMT-EATI-AACR International Cancer Immunotherapy Conference: Translating Science into Survival; Sept 30-Oct 3, 2018; New York, NY. Philadelphia (PA): AACR; Cancer Immunol Res 2019;7(2 Suppl):Abstract nr B092.
Read full abstract