Examination of resistance to nivolumab monotherapy through single-cell analysis of tumors from patients enrolled in the HCRN GU16-260 study of nivolumab monotherapy.

Abstract

698 Background: Nivolumab (nivo) monotherapy demonstrated anti-tumor activity in treatment-naïve renal cell carcinoma (RCC) in Part A of HCRN GU16-260 across all IMDC groups and in multiple histologies. Patient tumor samples were collected to characterize the tumor-immune microenvironmental (TME) determinants of effective anti-tumor immunity with nivo. Methods: Eligible patients (with clear cell or non-clear cell histology) underwent tumor biopsy prior to and/or at resistance to nivo monotherapy. Fresh tissue fragments were cryopreserved locally and centrally processed to extract viable single-cells from the RCC TME. Single-cell RNA-sequencing (scRNA-seq) and T cell receptor (TCR)-sequencing (scTCR-seq) was performed on all tumor samples. Gene expression signatures discovered through scRNA-seq were used to interrogate previously published bulk transcriptomic data from the CheckMate-009/010/025 trials of nivo in the treatment-refractory setting. Results: In total, 72,730 viable single-cells (56,900 immune and 15,830 tumor or stromal cells) were sequenced from 17 patients (8 with baseline only, 7 with progression only, and 2 with paired baseline and progression samples) across 7 trial sites. Trajectory inference analysis of tumor-infiltrating T cells revealed a bifurcating trajectory, starting with naïve T cells and ending either in PMCH+ terminally exhausted CD8+ T cells or SLAMF7+ CD8+ T cells. Notably, the SLAMF7+ T cell population expressed high levels of cytotoxic genes (including GZMA, GZMB, GNLY) and markers of tissue residency ( ZNF683/HOBIT and ITGAE/CD103), and had a restricted TCR diversity (normalized Shannon index = 0.57). Among patients with at least 100 sequenced T cells (n = 14), a higher percentage of SLAMF7+ CD8+ T cells (relative to total T cells) was associated with primary resistance (progressive disease [PD] as best response) to nivo (mean percentage in PD [n = 4] patients 32.7%; stable disease [n = 4] patients, 9.1%; complete or partial response [CR/PR; n = 6] patients, 2.2%; p = 0.019 for CR/PR vs PD). A signature derived using genes expressed in the T cell trajectory branch containing the SLAMF7+ CD8+ T cell population was used to interrogate bulk RNA-seq data from 172 pre-treatment tumors from the nivo arms of the CheckMate-009/010/025 trials. The signature score was enriched in patients with PD compared to CR/PR as best response (p = 0.032). Analysis of bulk whole exome sequencing and RNA-seq from patients enrolled in the HCRN GU16-260 is pending. Conclusions: Single-cell transcriptomic analysis uncovered a SLAMF7+ CD8 + T cell population with markers of cytotoxicity and tissue residency that was associated with resistance to nivo monotherapy in RCC. Further, the study highlights that scRNA-seq is a viable scientific strategy for deep correlative analysis in multicenter clinical trials.