Rapid Central Corticosteroid Effects: Evidence for Membrane Glucocorticoid Receptors in the Brain1

Integrative and Comparative Biology, Sep 2005 by Tasker, Jeffrey G, Di, Shi, Malcher-Lopes, Renato

Indeed, we recently reported a rapid effect of glucocorticoids on synaptic glutamate currents recorded in putative PVN parvocellular neuroendocrine cells in hypothalamic slices (Di et al., 2003). In this study, dexamethasone and corticosterone suppressed glutamatergic synaptic currents within ~3 min in a dose-dependent fashion, with half-maximal effects occurring at stress corticosteroid levels (367 nM) (Fig. 1). Several lines of evidence suggested that this glucocorticoid effect on PVN parvocellular neurons was mediated by a membrane-associated receptor and a G protein-dependent mechanism (Fig. 2). The glucocorticoid effect was not blocked by antagonists of the intracellular type I and type II corticosteroid receptors. A dexamethasone-bovine serum albumin conjugate (10 �M) retained the inhibitory effect of dexamethasone on mEPSC frequency, and dexamethasone applied directly into the cytoplasm of parvocellular neurons via the patch pipette was without effect. The effect of dexamethasone on glutamate release was blocked by blocking protein kinase activity and, interestingly, by blocking G protein activity specifically in the postsynaptic parvocellular neurons by intracellular infusion of a G protein blocker via the patch pipette. The latter observation suggested that the corticosteroid effect was, in fact, mediated by activation of a receptor located postsynaptically on the parvocellular neurons and by the subsequent release of a retrograde messenger that acted on presynaptic glutamate terminals to suppress glutamate release.

Endocannabinoids have been shown recently to serve as retrograde messengers in the regulation of synaptic glutamate and GABA release (Auclair et al., 2000; Wilson and Nicoll, 2001), so we tested for a cannabinoid dependence of the rapid glucocorticoid suppression of glutamate release. We found that the glucocorticoid effect was completely blocked by antagonists of the CB1 cannabinoid receptor and that it was mimicked by a cannabinoid agonist (Fig. 3). These data suggested, therefore, that the rapid retrograde messenger activated by glucocorticoids was an endocannabinoid.

These effects of glucocorticoids on excitatory synaptic inputs mediated by endocannabinoid release in the PVN were found in CRH neurons, suggesting direct feedback inhibition of the HPA axis. However, they were also found in other parvocellular PVN neurons identified by single-cell RT-PCR, including thyrotropin releasing hormone (TRH)-, oxytocin- and vasopressin-expressing neurons (Di et al., 2003). Thus, glucocorticoid-induced suppression of glutamatergic synaptic inputs occurs in different parvocellular neuroendocrine cells of the PVN, which suggests a more generalized inhibitory feedback role of glucocorticoids in the regulation of neuroendocrine function.

These findings point to a rapid corticosteroid action mediated by the activation of a membrane-associated receptor and a G protein/protein kinase-dependent mechanism (Fig. 4A), and corroborate the increasing body of evidence for non-transcriptional corticosteroid effects mediated by putative membrane glucocorticoid receptors. Interestingly, our findings indicate that the activation of these receptors in the PVN leads to the suppression of glutamatergic synaptic inputs to PVN parvocellular neuroendocrine cells via a novel mechanism involving the retrograde release of an endocannabinoid (Fig. 4B). Indeed, we have preliminary evidence from liquid chromatography-mass spectrometry analyses in brain slices indicating that dexamethasone elicits a significant increase in the levels of the endocannabinoids anandamide and 2-arachidonoylglycerol in the rat PVN and supraoptic nucleus, but not in the cerebellum (Malcher-Lopes et al., 2004), which corroborates our electrophysiological findings and supports our model of glucocorticoid suppression of excitatory synaptic inputs to PVN parvocellular neurons via the retrograde release of endocannabinoids (Fig. 4). Additionally, we have preliminary confocal immunohistochemistry data showing colocalization of CB1 cannabinoid receptors with the vesicular glutamate transporter 2, a marker of glutamate synaptic boutons, in the PVN (Di and Tasker, 2003), which suggests CB1 expression in presynaptic glutamatergic synaptic terminals in the PVN and provides further support for our model. Although, at this point, this model seems the most parsimonious for explaining the observed rapid effects of glucocorticoids on synaptic glutamate inputs to PVN neuroendocrine cells, we cannot yet exclude other alternative models that might also account for these observations, including the possibility of a glial cell intermediate and neuronal-glial interactions to stimulate endocannabinoid release.