Comprehensive multi-site profiling of the malignant pleural mesothelioma micro-environment identifies candidate molecular determinants of histopathologic type

Abstract

Pleural mesothelioma (PM) comprises sarcomatoid, epithelioid and biphasic histologic subtypes. Bulk PM RNA-sequencing identifies a histology-associated molecular gradient with features of epithelial-mesenchymal (EM) transition but cannot parse malignant, stromal, and immune tumor components. The mechanisms driving PM malignant cell phenotype and associated histology is not well-characterized. Here, we use single-cell RNA-sequencing (scRNA-seq) paired with exome, bulk RNA-sequencing, and histologic analysis of adjacent samples to characterize malignant cell EM state, parse the tumor microenvironment (TME), and identify candidate drivers of PM cell fate. We observe EM variation in malignant cells analogous to bulk samples. We characterize epithelioid and sarcomatoid malignant cell programs and identify a new uncommitted malignant cell EM phenotype enriched in biphasic histology samples. Using inferred CNVs we observe that single individual PM clones consist of cells exhibiting all three EM cell states. We find that distinct non-malignant microenvironments associated with tumors consisting of mostly cells in each state, and identify WNT inhibition, GAS6-AXL, and HBEGF-EGFR signaling as pathways associated with distinct EM cell states. These findings provide deeper insight into the molecular drivers of PM malignant cells and identify non-malignant cell signals as potential EMT and growth drivers in PM.