Cyclooxygenase2 (COX2) has been associated with cell growth, invasiveness, tumor progression and metastasis of colorectal carcinomas. However, the downstream prostaglandin (PG)-PG receptor pathway involved in these effects is poorly characterized.We studied the PG-pathway in gene expression databases and we found that PTGS2 (prostaglandin G/H synthase and cyclooxygenase) and PTGES (prostaglandin E synthase) are co-expressed in human colorectal tumors. Moreover, we detected that COX2 and microsomal Prostaglandin E 2 synthase 1 (mPGES1) proteins are both up-regulated in colorectal human tumor biopsies.Using colon carcinoma cell cultures we found that COX2 overexpression significantly increased mPGES1 mRNA and protein. This up-regulation was due to an increase in early growth response 1 (EGR1) levels and its transcriptional activity. EGR1 was induced by COX2-generated PGF 2b . A PGF 2b receptor antagonist, or EGR1 silencing, inhibited the mPGES1 induction by COX2 overexpression. Moreover, using immunodeficient mice, we also demonstrated that both COX2-and mPGES1-overexpressing carcinoma cells were more efficient forming tumors.Our results describe for the first time the molecular pathway correlating PTGS2 and PTGES in colon cancer progression. We demonstrated that in this pathway mPGES1 is induced by COX2 overexpression, via autocrine PGs release, likely PGF 2b , through an EGR1-dependent mechanism. This signaling provides a molecular explanation to PTGS2 and PTGES association and contribute to colon cancer advance, pointing out novel potential therapeutic targets in this oncological context.

The role of prostaglandin (PG) F2α has been scarcely studied in cancer. We have identified a new function for PGF2α in ovarian cancer, stimulating the production of Prostate Transmembrane Protein, Androgen Induced 1 (PMEPA1). We show that this induction increases cell plasticity and proliferation, enhancing tumor growth through PMEPA1. Thus, PMEPA1 overexpression in ovarian carcinoma cells, significantly increased cell proliferation rates, whereas PMEPA1 silencing decreased proliferation. In addition, PMEPA1 overexpression buffered TGFβ signaling, via reduction of SMAD-dependent signaling. PMEPA1 overexpressing cells acquired an epithelial morphology, associated with higher E-cadherin expression levels while β-catenin nuclear translocation was inhibited. Notwithstanding, high PMEPA1 levels also correlated with epithelial to mesenchymal transition markers, such as vimentin and ZEB1, allowing the cells to take advantage of both epithelial and mesenchymal characteristics, gaining in cell plasticity and adaptability. Interestingly, in mouse xenografts, PMEPA1 overexpressing ovarian cells had a clear survival and proliferative advantage, resulting in higher metastatic capacity, while PMEPA1 silencing had the opposite effect. Furthermore, high PMEPA1 expression in a cohort of advanced ovarian cancer patients was observed, correlating with E-cadherin expression. Most importantly, high PMEPA1 mRNA levels were associated with lower patient survival.