In ruminants, interferon produced by the trophectoderm (IFN-tau) is recognized as the embryonic signal responsible for maternal recognition of pregnancy. IFN-tau is believed to act by down-regulating estrogen receptors, thus preventing appearance of oxytocin receptors responsible for the release of prostaglandin F(2alpha) (PGF(2alpha)) by the endometrium. The present study was undertaken to determine in vitro the biological activities of different IFN-tau isoforms and document putative alternate luteotrophic mechanisms. Endometrial cells in primary cultures were treated with five different rIFN-tau isoforms: two ovine isoforms (ro-4 and ro-11) and three bovine isoforms (rb-1a, rb-2b and rb-3b). Their effect was quantified by measurement of PGE(2) and PGF(2alpha) production by ELISA and induction of cyclooxygenase (COX-2) by Western and Northern analysis and correlated with antiviral activity previously reported. The overall pattern of response to the IFNs tested suggests that low concentrations (<1 microg/ml) reduced the production of both PGs and higher concentrations (>1 microg/ml) stimulated preferentially PGE(2); however, exceptions were noted. Isoform rb-2b with high antiviral activity inhibited PG production in both cell types at all concentrations tested. IFNs rb-1a and ro-11 had similar antiviral activities, inhibiting PG at low concentrations and stimulating them at high concentrations. Isoform rb-3b stands out relative to the other IFNs tested because it induced a variable non-dose-dependent effect on PG production and low antiviral activity. An increase in COX-2 protein expression and messenger was correlated with increased PG production. The results showing two distinct responses to IFN-tau depending on its concentration and/or isoform and the absence of correlation with antiviral activity suggest that complex transduction mechanisms are involved.

In ruminants, the production of prostaglandins by the endometrium is critical for recognition of pregnancy. In the absence of an embryonic signal, luteolytic pulses of PGF 2␣ are released by the uterus. In contrast, the presence of a viable conceptus reduces the production of PGF 2␣ relative to PGE 2 and prevents luteolysis through the release of trophoblastic interferon (IFN-). Initially, it was thought that epithelial and stromal endometrial cells were specialized in the production of a single type of prostaglandin. However, purified cell populations of both types of cell can produce PGF 2␣ and PGE 2 ; therefore, selective production of PGF 2␣ and PGE 2 must be regulated within each type of cell. Two distinct prostaglandin synthases, cyclooxygenase 1 and cyclooxygenase 2, are involved in prostaglandin production and each may catalyse the production of a different prostaglandin. This possibility was investigated in cultured epithelial cells from bovine endometrium. Cells were treated with oxytocin or arachidonic acid, and expression of cyclooxygenase 1 and cyclooxygenase 2 proteins was monitored over time and correlated with prostaglandin accumulation. Cells were also treated with increasing doses of inhibitors of cyclooxygenase 1 or cyclooxygenase 2 (nonsteroidal anti-inflammatory drugs; NSAIDs) with or without arachidonic acid or oxytocin: flurbiprofen (0-50 mol l −1 ) was used as a non-selective inhibitor; valeryl salicylate (0-500 mol l −1 ) was used as a cyclooxygenase 1 inhibitor and NS-398 (0-1 mol l −1 ) was used as a cyclooxygenase 2 inhibitor. After stimulation with arachidonic acid or oxytocin, prostaglandin production and expression of cyclooxygenase 2 protein were increased. All inhibitors were able to block basal and stimulated prostaglandin production. These results indicate that in endometrium most, if not all, prostaglandin production is probably processed through cyclooxygenase 2.