Spatial coupling of microbes and immune cells in solid malignancies [Cold_hot_tumors_mouse]

Solid tumors are composed of cancer cells and host immune cells that are distributed in a non-uniform pattern. Growing evidence shows that intratumoral microbes are associated with immune microenvironments in cancer. However, mechanisms that direct the recruitment of microbes to tumors remain poorly understood. Here, we show that intratumoral infiltration of immune cells and microbes are heterogeneous, and the distribution of microbes within tumors are orchestrated by the spatial heterogeneity of intratumoral lymphoid populations. Analysis of human solid tumors revealed that the spatial distribution of immune cells, particularly CD8+ T cells, is markedly heterogeneous. Compared to T cell-poor (“cold”) tumor nests, T cell-rich (“hot”) tumor nests displayed a significantly higher number of myeloid cells, B cells, and plasma cells. We performed laser capture microdissection (LCM) followed by RNA sequencing to identify unique gene signatures that define tumor epithelium and stroma of cold and hot tumor nests. Cold tumor nests expressed genes that promote tumor proliferation and fibrosis, whereas hot tumor stroma and epithelium showed upregulation of immune-related processes, including responses to bacteria, and receptors that mediate mucosal immune responses to microbes, respectively. Consistent with these findings, we detected elevated levels of microbes within hot tumor nests in human pancreatic and lung cancers as well as in mouse models of pancreatic cancer. Intratumoral T cell infiltration plays a causal role in spatial distribution of bacteria in tumor. Our data implicate intratumoral immune heterogeneity in defining microbial spatial distribution and highlight a potential role for crosstalks between microbes, cancer cells, and the host immune system in shaping constituents of the tumor microenvironment (TME). Overall design: RNA isolated from formalin-fixed, paraffin-embedded (FFPE) sections containing T cell-poor ("cold") and T cell-rich ("hot") tumors was analyzed using QuantSeq 3' mRNA sequencing.