Abstract
BACKGROUND: Immune checkpoint blockade (ICB) therapies have revolutionized treatment for renal cell carcinoma (RCC). Despite this, only a subset of patients respond, and there remains an incomplete understanding of the cellular states that are associated with response. Defining these states may help elucidate potential mechanisms of response and assist in the development of next-generation immune therapies for RCC. METHODS: We utilized an existing tumor single-cell RNA sequencing (scRNA-seq) dataset from our lab of > 400 000 cells from 70 donors with RCC, with 49 of them having received ICB therapy (either anti-PD-1 alone or in different combinations). Each donor was also annotated as having benefit (complete or partial response) or non-benefit (progressive disease) after ICB therapy. RESULTS: In order to find new cell states that may impact response to ICB, we began by sub-clustering the CD4+ T cells, finding 10 clusters that existed. One cluster was strongly enriched in donors who experienced benefit after ICB therapy (Wilcoxon Test, P = .0042) and was marked by heat shock protein 70 genes (HSP70) such as HSPA6, HSPA1A, and HSPA1B. Previous work from our group has shown that many of these same HSP70 family genes are associated with response after ICB therapy in CD8+ T cells, so we hypothesized that high expression of these heat shock genes may be able to predict benefit from ICB therapy across diverse lymphocyte subsets (Kashima, ASCO, 2024). In order to test this, we made a module score of the upregulated genes from the CD4+ T cell heat shock cluster and applied it to all lymphocytes. We then calculated the correlation of the heat shock module score rank for each sub-cluster vs. the rank of how strongly each sub-cluster was associated with benefit from ICB. Strikingly, there was a nearly perfect correlation for CD4+ T cells (r = 0.89, P = .001), indicating that increased heat shock expression in CD4+ T cell clusters is associated with a better response to ICB. We repeated this test on NK and innate lymphoid cells (r = 0.77, P = .005), B cells (r = 0.67, P = .00013), and CD8+ T cells (r = 0.3, P = .325), finding similar trends each time albeit to varying extents. Next, we calculated the mean heat shock module score for each donor across all lymphocytes and calculated its difference between donors who had benefit vs. non-benefit. We found that there was a significant increase in heat shock score in donors who had benefit from ICB (Wilcoxon Test, P = .0031), and this difference was even more pronounced when only the top gene from the module (HSPA6) was used to calculate the mean difference (P = .001). To validate our findings, we turned to a pre-existing cohort of 17 donors from the HCRN GU16-260 trial where tumor scRNA-seq was available (Hugaboom et al., Cancer Discovery, 2025). After calculating the mean lymphocyte HSPA6 module score by donor, we discovered that donors who had a score greater than or equal to the median had greatly increased progression free survival (PFS) after ICB therapy (Kaplan-Meier, P = .0058; mean PFS of 14.3 months for HSPA6 high vs. 3.9 months for HSPA6 low). [Image: see text] CONCLUSIONS: We identified a heat shock signature in lymphocytes that is strongly associated with clinical benefit in RCC after ICB therapy, with the expression of the top gene (HSPA6) predicting PFS in a subset of donors from the HCRN GU16-260 clinical trial. Future studies will mechanistically test whether these genes are responsible for the therapeutic response themselves and uncover the stresses found in the tumor microenvironment that are necessary for their induction.