Marker genes of decidualization: activation of the decidual prolactin gene

Decidualization of human endometrial stromal (ES) cells in vitro is induced by cAMP analogues and ligands that elevate cellular cAMP levels in a manner resembling the gonadotrophins, prostaglandin E2 and relaxin (RLX). This differentiation process is marked by the onset of decidual prolactin (PRL) production in the late luteal phase of the cycle. Using transfection assays and a primary ES cell culture system, we have demonstrated that decidual PRL gene transcription is driven by an alternative upstream promoter (dPRL), approximately 6 kb upstream of the pituitary transcription start site. In primary cell cultures, RLX not only acutely but also permanently elevated cellular cAMP levels and induced PRL secretion after 6 days. Northern and Western blot analyses revealed all regulatory subunit isoforms (RIα, RIβ, RIIα, RIIβ) and catalytic subunits Cα and Cβ of protein kinase A (PKA) in ES cells. Transcript levels of PKA subunit isoforms are not altered during decidualization, but in decidualized ES cells exposed to elevated cellular cAMP levels by stimulation with RLX for >6 days, RIα protein levels were significantly reduced, whereas levels of all other forms remained unchanged. Reducing the availability of R subunits changed the R:C subunit ratio in favour of C and increased kinase activity. In transient transfections of undifferentiated ES cells, the dPRL promoter was activated by 8-Br-cAMP and by C subunit (Cβ) of PKA. This induction, and the differentiation-dependent activity of the dPRL promoter in transfected decidualized cells, was effectively abolished by the co-expression of protein kinase inhibitor (PKI). A fragment of 332 bp of 5'-flanking region of the dPRL transcription start site was sufficient to mediate full inducibility by cAMP. cAMP activation of the dPRL promoter in ES cells was biphasic as an initial weak induction within 12 hours was followed by a subsequent, much more intense induction after 12 hours. The secondary induction was not seen with a control construct driven by a consensus cAMP response element (CRE) linked to a minimal promoter. The early response of the dPRL promoter depended upon a non-palindromic CRE at position-12 and mutation of this sequence led to omission of the early, but not of the delayed, induction. The major activation of the dPRL promoter depended upon a different region between position -332 and -270 since its deletion significantly reduced inducibility by cAMP. Its action was probably indirect as its kinetics