Spatial organization of immune surveillance by germline mutation status in the Breast Cancer Family Registry.

Abstract

10566 Background: Germline pathogenic variants (gPVs) are associated with increased immune infiltration in resulting tumors; we previously reported increased CD4+ and CD8+ T-cells in gPV-associated tumors using single-cell spatial profiling. Here, we report the spatial features, including cell-cell co-localization, underlying these immune differences. Methods: We analyzed CosMx 6K single-molecule RNA transcriptomics in 222 samples from 144 women in the population-based Northern California Breast Cancer Family Registry. Across 1,083,345 cells and 6,583 genes, multicellular niches were identified by clustering cells based on spatial proximity and cell-type identity using a k-nearest neighbors approach, with the number of niches selected by elbow analysis. Niche proportions were compared between gPV-associated and wild-type tumors using linear models. Linear mixed-effects models were used to identify niche-specific enrichment of cell-type densities and pairwise cell–cell colocalization relative to all other niches. P-values were corrected using Benjamini–Hochberg. Results: Of 144 participants, 31 (22%) carried at least one gPV, including BRCA1 (n = 12), BRCA2 (n = 8), PALB2 (n = 3), FANCM (n = 3), ATM (n = 2), and CHEK2, RAD50, and TP53 (n = 1 each). We identified eight cell niches, two of which were enriched in gPV-associated tumors, including within ER/HER2 subtypes: niche 2 (10% of cells in gPV tumors vs 3.7% in non-gPV tumors; Wilcoxon FDR = 0.01) and niche 5 (17% vs 9.4%; FDR = 0.03). Niche 2 and niche 5 proportions were uncorrelated across tumors (r = −0.13, FDR = 0.14). Compared with other niches, niche 2 had higher proportions of immature B cells (FDR = 0.006), CD8⁺ T cells (FDR = 0.03), dendritic cells (FDR < 0.001), and macrophages (FDR < 0.001), with no specific cell-cell colocalization pattern. Niche 5 had higher proportions of CD4⁺ T cells (FDR < 0.001), myoepithelial cells (FDR < 0.001), and macrophage-derived stem cells (FDR < 0.001); niche 5 demonstrated increased colocalization across 95 cell-type pairs at FDR < 0.1, including cancer cells with macrophage-derived stem cells and pericytes with natural killer cells (FDR = 0.01 each). Comparing BRCA1 to BRCA2 gPV carriers, niche 2 was more prevalent in BRCA1 tumors (16% vs < 1% in BRCA2; p = 0.01); no significant difference was observed for niche 5. Conclusions: In this population-based cohort, gPV-associated breast cancers exhibited distinct spatial immune niches. Niche 2, especially enriched in BRCA1 gPV tumors, appears to reflect an inflamed immune environment, whereas niche 5 may represent a more spatially organized stromal-adaptive immune niche characterized by CD4⁺ T-cell enrichment and extensive non-random cell-cell interactions. These findings demonstrate that inherited cancer risk is associated with reproducible immune spatial architectures, with implications for immune-based prevention and interception strategies.