In the late pregnant rat stimulated oxytocin secretion is inhibited, preserving the expanded neurohypophysial oxytocin stores for birth and minimising the risk of preterm labour. In virgin, but not pregnant rats, systemic administration of the cytokine interleukin-1β (IL-1β), which mimics infection, increases oxytocin secretion and the firing rate of supraoptic nucleus (SON) oxytocin neurons (1). Blood and brain levels of the neuroactive steroid allopregnanolone (AP; progesterone metabolite) increase during pregnancy and AP acts on GABA_A receptors on oxytocin neurons, to enhance inhibitory transmission (2). Here we tested whether AP is the pregnancy-related factor responsible for restraining oxytocin responses to IL-1β in pregnant rats, and whether its actions are central. Jugular vein cannulae were implanted under halothane anaesthesia 5 days prior to the experiments. Oxytocin secretion following IL-1β (500ng/kg i.v.) was measured (by radioimmunoassay) in plasma from: a) virgin and pregnant (day 21) Sprague Dawley rats treated with vehicle (oil) or finasteride (FIN, 25mg/kg s.c.; 5α-reductase inhibitor to block AP production) 20h and 2h before IL-1β. b) Pregnant rats given FIN and AP/oil; and virgins given AP (3mg & 1mg/kg s.c.) or oil pre-treatment (20h and 2h before IL-1β). To investigate central SON oxytocin neuron responses, perfuse-fixed brain sections from pregnant rats treated with oil/FIN + vehicle/IL-1β, and virgin rats treated with oil/AP + vehicle/IL-1β, were processed for Fos (indicator of neuronal activation) and oxytocin immunoreactivity. Results: IL-1β significantly increased oxytocin secretion (3.4-fold; p<0.001, two-way repeated measures ANOVA; n=7) and Fos expression in SON oxytocin neurons (4.4-fold; p<0.001, two-way ANOVA; n=8) in virgin, but not pregnant rats (1.2-fold for secretion and Fos; n.s., n=7). FIN had no further effect on oxytocin secretion in virgins, but significantly restored an oxytocin response to IL-1β in the pregnant rats (3.3-fold; p<0.001, two-way repeated measures ANOVA; n=6), while AP significantly reduced the oxytocin response to IL-1β in virgins by 50% (p<0.001, two-way repeated measures ANOVA; n=6) and reversed the effect of FIN in pregnant rats. Consistent with the secretion data, FIN significantly increased IL-1β-induced Fos expression in identified SON oxytocin neurones of pregnant rats (2.4-fold; p<0.005, two-way ANOVA; n=6), while AP significantly reduced the number of SON oxytocin neurones activated by IL-1β by 48% (p<0.05, two-way ANOVA; n=6). Thus, AP acts centrally to restrain oxytocin secretory responses to IL-1β in pregnancy. This mechanism will serve to reduce the likelihood of preterm labour and prevent depletion of posterior pituitary oxytocin stores, required for parturition. Acknowledgements: This work was supported by the BBSRC.Reference 1 : Brunton PJ et al (2006). Eur J Neurosci 23, 1241-1247.Reference 2 : Brussaard AB & Herbison AE (2000). Trends Neurosci 23, 190-195.