Abstract 1689: Intranodal self-correlation analysis reveals differences in gene-to-gene interactions between lymph nodes draining cold vs. hot triple-negative breast tumors

Abstract

Abstract Hot tumors (i.e., tumors with more infiltrating lymphocytes) are generally associated with better prognosis and response to immune checkpoint blockade therapies in TNBC. Higher tumor mutation burden (TMB) has been associated with hot tumors and considered a potential reason why hot tumors may have more tumor infiltrating lymphocytes (TILs) compared to cold tumors. However, TMB does not fully explain immune infiltration. We hypothesized that tumor draining lymph nodes (TDLNs) play an important role in lymphocyte infiltration into tumors. To study the functional features of LNs draining cold vs. hot tumors, we characterized the expression of 730 immune functional genes of 15 tumor-free TDLNs from paired cold (n = 7) and hot (n = 8) tumors based on low (<10%) or high (>60%) TIL percentages defined by pathologists in H&E stained slides. By standard differential gene expression (DGE) analysis, there were similar transcriptomic profiles in TDLNs between cold and hot cohorts. Since DGE analysis only provides comparison of average gene expression, it cannot compare gene-to-gene interactions. Therefore, to further investigate differences in intranodal gene-to-gene interactions, we implemented self-correlation analysis (i.e., generating clustered gene-to-gene correlations) to both cohorts. Results showed that TDLNs generally present weaker intranodal regulations (i.e., less significantly correlated gene pairs and smaller organized clusters) in the cold cohort. By further comparing specific gene-to-gene correlations, the GATA3-CXCR1 correlation in the cold cohort were found to be negative (rCold = -0.56), while positive (rHot = 0.90) in the hot cohort. Similar opposite correlations were also found in TBX21-CXCR1 pair (rCold = 0.85, rHot = -0.88). Since CXCR1 would be downregulated during the maturation of dendritic cells (DCs) and T cell differentiation, these results suggest that matured dendritic cells within TDLNs from cold tumors may preferably prime naïve CD4+ T cells to T helper 2 (Th2) cells. In contrast, TDLNs from hot tumors have an opposite preference to T helper 1 (Th1) cells. In addition, a positive CD4-STAT6 correlation (r = 0.88, p-value = 0.0084) was also observed, which further indicated a preference to Th2 cells in TDLNs from cold tumors. In summary, by applying intranodal self-correlation analysis to TDLNs from cold and hot tumors, opposite preferences of CD4+ naïve T cell differentiation in TDLNs are suggested. The weaker regulation and preference of Th2 cells in TDLNs from cold tumors may hinder lymphocyte infiltration into the tumor. Citation Format: Weihua Guo, Minhui Lim, Jiayi Tan, Lei Wang, Ting-fang He, Shawn Solomon, Colt A. Egelston, Diana L. Simons, Daniel Schmolze, James Waisman, Peter P. Lee. Intranodal self-correlation analysis reveals differences in gene-to-gene interactions between lymph nodes draining cold vs. hot triple-negative breast tumors [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 1689.