Insights into autotaxin- and lysophosphatidate-mediated signaling in the pancreatic ductal adenocarcinoma tumor microenvironment: a survey of pathway gene expression.

Abstract

Lysophosphatidate (LPA)-mediated signaling is a vital component of physiological wound healing, but the pathway is subverted to mediate chronic inflammatory signaling in many pathologies, including cancers. LPA, as an extracellular signaling molecule, is produced by the enzyme autotaxin (ATX, gene name ENPP2) and signals through six LPA receptors (LPARs). Its signaling is terminated by turnover via the ecto-activity of three lipid phosphate phosphatases (LPPs, gene names PLPP1-3). Many pharmacological developments against the LPA-signaling axis are underway, primarily against ATX. An ATX inhibitor against pancreatic ductal adenocarcinoma (PDAC), a very aggressive disease with limited systemic therapeutic options, is currently in clinical trials, and represents the first in-class drug against LPA signaling in cancers. In the present study, we surveyed the expression of ATX, LPARs, and LPPs in human PDACs and their clinical outcomes in two large independent cohorts, the Cancer Genome Atlas (TCGA) and GSE21501. Correlation among gene expressions, biological function and the cell composition of the tumor microenvironment were analysed using gene set enrichment analysis and cell cyber-sorting with xCell. ENPP2, LPAR1, LPAR4, LPAR5, LPAR6, PLPP1, and PLPP2 were significantly elevated in PDACs compared to normal pancreatic tissue, whereas LPAR2, LPAR3, and PLPP3 where downregulated (all P≤0.003). Only ENPP2 demonstrated survival differences, with overall survival favoring ENPP2-high patients (hazard ration 0.5-0.9). ENPP2 was also the only gene with enriched gene patterns for inflammatory and tissue repair gene sets. Epithelial (cancer) cells had increased LPAR2, LPAR5 and PLPP2 expression, and decreased ENPP2, LPAR1, PLPP1, and PLPP3 gene expression (all P<0.02). Tumor fibroblasts had increased ENPP2, LPAR2, LPAR4, PLPP1, and PLPP3 expression and decreased LPAR2, LPAR5, and PLPP2 expression in both cohorts (all P≤0.01). Immune cell populations were not well correlated to gene expression in PDACs, but across both cohorts, cytolytic scores were increased in high-expressing ENPP2, LPAR1, LPAR6, PLPP1, and PLPP3 tumors (P<0.01). Overall, in PDACs, ENPP2 may switch from an anti-to-pro tumor promoting gene with disease progression. LPAR2 and PLPP2 inhibition are also predicted to have potential therapeutic utility. Future multi-omics investigations are necessarily to validate which LPA signaling components are high-value candidates for pharmacological manipulation in PDAC treatment.