Interplay between B cell and GABA metabolism (GABAm) and association with immune evasion in breast carcinoma (BC).

Abstract

1097 Background: GABAergic signaling has been reported to play a pivotal role in breast cancer (BC) tumorigenesis and metastasis, however, its role in immune modulation remains unclear. Recent in vitro and in vivo studies (Zhang et al., Nature, 2021) report the role of B cell-derived GABA metabolites in promoting anti-inflammatory macrophages (MM), thus limiting anti-tumor immunity. In this study, we aim to characterize the interplay between B cells and the GABAm pathway, as well as their associated immune infiltrates and cytokines. Methods: BC tumors (n = 9455) were analyzed by next generation sequencing (NextSeq, 592 Genes and WES, NovaSEQ) and whole transcriptome sequencing (WTS, NovaSeq) at Caris Life Sciences. Gene set variation analysis (GSVA) scores were used for GABAm pathway activity (GMPA). IFN score to test the likelihood of a tumor’s response to anti PD1 therapy and Immune cell fraction (quanTIseq) were assessed by mRNA analysis. Wilcoxon-Mann-Whitney test was applied (p without, q with multiple comparison correction). Correlation coefficients were calculated using spearman correlation. Results: GMPA demonstrated a statistically significant positive correlation with B cells fraction (r = 0.24, p < 0.0001). When stratified by classical molecular subtypes, the positive correlations were exclusive to HR+ and HER2+ BC, and absent in TNBC. GMPA was the most enriched in HR+ BC, followed by HER2+ and TNBC. BC tumors with high B cell infiltration were then grouped into GMPA-high (B+G+, cutoff > median for both) or GMPA-low (B+/G-), which likely represented tumors with B cell-derived high and low GMPA group, respectively. The GMPA-high group demonstrated significantly less fractions of MM1 (2.8 vs 3.7) and CD8+ T cells (0.8 vs 1.2) but greater MM2 (5.3 vs 4.9). mRNA levels of the MM2 marker IL10, a proposed marker of immune evasion, was significantly overexpressed in the B+/G+ group compared to the B+/G- group (fold change, FC = 1.39). mRNA levels of GAD1, a GABA-generating enzyme, were higher in B+/G+ than B+/G- (FC = 7.19). B+/G+ group had notably less IFN score than B+/G- group (-0.37 vs -0.27). When further stratified into molecular subtypes, concurrent more MM2 (5.4 vs 5.2) and less CD8+ T cell (0.74 vs 0.91) fractions were found in B+/G+ compared to B+/G- in HR+ tumors, but not in HER2+ or TNBC tumors. B+/G+ group also demonstrated a lower IFN score (-0.38 vs -0.32) in HR+ tumors. Additionally, IL10 and GAD1 were consistently overexpressed in B+/G+ regardless of subtype, reaching FC 7.9 in HR+ tumors. q < 0.0001 for all comparisons. Conclusions: Our study is the largest clinical dataset to demonstrate the association of interplay between B cell and GABAm with immunogenicity. Our results support the potential role of B cell-derived GABAm metabolites in immune modulation in BC in a subtype-specific manner. Targeting small metabolites to modulate immune evasion in BC warrants further investigation.