Glucocorticoids have powerful actions on metabolism and the immune system, especially when secretion is increased by stressors. In pregnancy glucocorticoid is important in fetal maturation, and in determining the onset of parturition via actions on the placenta. Protection of these fetal processes, and the mother’s metabolic and immune systems, from stress levels of maternal glucocorticoid may be advantageous for a successful pregnancy. One protective mechanism is placental 11β-hydroxysteroid dehydrogenase 2, which inactivates cortisol (human; corticosterone in rats); another is attenuation of maternal HPA axis responses to stress. Failure of these mechanisms in late pregnancy, as revealed by giving the synthetic glucocorticoid dexamethasone (which crosses the placenta), results in adverse fetal programming. Attenuated HPA axis stress responses in the last week of rat pregnancy (gestation is ca. 22 days) are seen with a mild emotional/physical stressor (e.g. forced swimming; [1]). Measurement by in situ hybridisation (ISH) of rapid gene expression changes in the hypothalamic parvocellular paraventricular nucleus (pPVN), reflecting activation of the corticotropin releasing factor (CRF)/ vasopressin (VP) neurones, shows they are less excited by stressors in late pregnancy. Since CRF and VP cause corticotropin (ACTH) release from the anterior pituitary, and hence glucocorticoid secretion from the adrenal cortex, reduced pPVN neurone activation explains reduced secretion of ACTH and glucocorticoid. Hence adaptations in pregnancy underlying reduced HPA axis responses are in the maternal brain. Physical stressors such as chemical signals of infection, lipopolysaccharide (LPS; endotoxin) and interleukin-1β (IL-1β, given to conscious rats via a jugular vein cannula), stimulate strong and prolonged HPA axis activation in virgins, but HPA responses are completely suppressed in late pregnant rats [2]. Similarly, neuropeptide orexigens that signal energy need activate the HPA axis in virgin but not in late pregnant rats [3]. To investigate causes of the profound suppression of HPA axis responses to IL-1β in late pregnancy, we tested the hypothesis that activation of an endogenous opioid mechanism in the brain prevents pPVN CRH/VP neurone responses to IL-1β, by a local action in the pPVN, and that this opioid mechanism is induced by actions of allopregnanolone (AP), a neuroactive steroid metabolite of progesterone. Progesterone, abundant in pregnancy, is metabolised by 5α-reductase (5α-R) and 3α-hydroxysteroid dehydrogenase to AP, in the brain and liver, resulting in high brain concentrations in late pregnancy [4]. If late pregnant rats are given the opioid antagonist naloxone (i.v.) a few minutes before IL-1β, IL-1β increases ACTH and corticosterone secretion and CRF mRNA expression in the pPVN, thus revealing an inhibitory endogenous opioid action [2]. IL-1β acts by stimulating noradrenergic neurones in the nucleus of the tractus solitarius (NTS) in the medulla oblongata, via prostaglandins generated by activation of endothelial IL type1 receptors on local blood vessels. ISH shows a lack of stimulation by IL-1β of cyclo-oxygenase-2 (COX-2) mRNA expression in the NTS in late pregnant rats, contrasting with stimulation of expression in virgin rats and indicating altered post-IL type1 receptor signalling in late pregnancy. Nonetheless, the number of NTS neurones expressing Fos protein (an indicator of recent neuronal activation) after i.v. IL-1β is similarly increased (compared with vehicle injection) in virgin and late pregnant rats [2]. Similar activation of Fos expression in central amygdala neurones in virgin and late pregnant rats is confirmation that IL-1β can signal into the brain in late pregnancy. In virgin rats activation of NTS neurones by i.v. IL-1β leads to noradrenaline release, measured via microdialysis, in the pPVN. This is not seen in late pregnant rats, but naloxone infused into the pPVN restores local release of noradrenaline after i.v. IL-1β [2]. These results show preterminal opioid inhibition on noradrenergic synapses in late pregnancy that blocks the stimulation, via NTS projections, of the HPA axis by i.v. IL-1β. The source of this opioid is likely to be the NTS neurones themselves, since pro-enkephalin (PENK) A mRNA expression in the NTS is increased at the end of pregnancy [2]. Overnight treatment of late pregnant rats with finasteride, a 5α-R antagonist, also restores HPA axis responses to i.v. IL-1β in late pregnant rats, and conversely allopregnanolone treatment reduces HPA axis responses to IL-1β in virgin rats, and induces PENK-A mRNA expression in the NTS selectively. Hence, increased allopregnanolone production in pregnancy, from luteal progesterone, suppresses maternal HPA axis responses to IL-1β through up-regulation of opioid gene expression in the maternal brainstem. Acknowledgements: Support: BBSRCReference 1 : Ma S et al. (2005) Endocrinology 146, 1626-1637.Reference 2 : Brunton PJ et al. (2005) J Neurosci 25, 5117-5126.Reference 3 : Brunton PJ et al. (2006) Endocrinology 147, 3737-3745.Reference 4 : Russell JA et al. (2006) Neuroscience 138, 879-889.