Abstract
BACKGROUND: Despite the success of immune checkpoint inhibitors (ICI) for the treatment of renal cell carcinoma (RCC), many patients do not receive durable clinical benefit. Therefore, an understanding of resistance to ICIs is critical for the treatment of this disease. Using scRNA-seq, we previously found increased tissue-resident ZNF683 + (Hobit) SLAMF7+ CD8+ exhausted T cells (T-exh-SLAMF7) in human RCC resistant to PD-1 blockade in the HCRN GU16-260 trial. A T-exh-SLAMF7 gene expression signature (GES) was associated with worse clinical outcomes with PD-1 blockade in multiple validation cohorts. Here, through bulk RNA-seq of RCC biospecimens from 90 patients enrolled in this trial, we identified higher tertiary lymphoid structures (TLS) in patients responsive to PD-1 blockade and investigated the interplay between TLS and T-exh-SLAMF7 in shaping therapeutic responses. METHODS: Bulk RNA sequencing was performed on tumor samples from 90 RCC patients enrolled in the HCRN GU16-260 clinical trial. GES scores were calculated for each sample by z-scoring all genes and then computing the average expression of the genes comprising each signature of interest. To validate TLS presence at the protein level, 19 FFPE RCC tumor samples were analyzed using a 7-plex multiplex immunofluorescence panel targeting DAPI, CD20, CD3, CD21, CD4, PD-1, and FOXP3. TLS were manually quantified based on the colocalization of B and T cell markers in organized structures. Patients were stratified by high versus low TLS and T-exh-SLAMF7 GES scores (≥ or < median) for downstream analyses of clinical response and progression-free survival. RESULTS: Patients with complete/partial response had a higher (≥ median) TLS GES score compared to patients with progressive disease (P = .0004). Similarly, high TLS signature scores were associated with improved progression-free survival (PFS; HR = 2.08, 95% CI: 1.28–3.4, P = .0032), indicating significantly higher risk of progression in patients with low TLS scores. We confirmed that tumors with a high TLS GES score had a higher number of TLS detected by multiplex immunofluorescence (P = .028). Finally, we analyzed the interplay between TLS and tissue-resident exhausted CD8+ T cells. We divided patients into four categories based on median split TLS and T-exh-SLAMF7 GES score values: TLS high SLAMF7 low (n = 28), TLS high SLAMF7 high (n = 15), TLS low SLAMF7 high (n = 28), and TLS low SLAMF7 low (n = 15). Patients with both high TLS and low T-exh-SLAMF7 GES scores had substantially improved PFS compared to all other patients (HR = 0.45, 95% CI: 0.26-0.79, P = .0052), with 60.7% PFS at 12 months compared to 25.9% in the remaining groups. CONCLUSIONS: These findings support a paradigm where both high TLS and low T-exh-SLAMF7 cells are required for optimal response to PD-1 blockade in RCC. Ongoing studies will use spatial transcriptomics and functional assays to evaluate the interaction between TLS and T-exh-SLAMF7 cells and to define the roles of SLAMF7 and Hobit in regulating CD8+ T cell effector functions within the RCC tumor microenvironment. Note: Encore Presentation; recently published in Cancer Discovery (PMID: 39992403)