Skip Internal Navigation
The goal of the laboratory is to understand the neuroendocrine mechanisms underlying the stress response, with emphasis on the regulation of the hypothalamic pituitary adrenal (HPA) axis. Not only during early development but also during adult life, the ability of the organism to adapt to acute and chronic stress situations is determined by genetic constitution and life experiences. The organism's degree of adaptability may lead to long-term consequences for the responsiveness of the HPA axis, with altered expression of hypothalamic corticotrophin releasing hormone (CRH) and circulating levels of glucocorticoids—hormones implicated in the pathogenesis of several psychiatric and metabolic disorders. Our laboratory studies the mechanisms of positive and negative regulation of expression of the hypothalamic hormones CRH and vasopressin (VP) and the proteins mediating adrenal steroidogenesis under various physiological situations as well as the impact of stress on HPA axis regulation. The influence of life experiences, especially during early development, on neuroendocrine regulation and the expression of genes involved in the stress response are important aspects of our research program. Elucidation of the mechanisms regulating the production of stress hormones and the effects of life experiences on the stress response is critical for understanding the mechanisms leading to HPA axis dysregulation and for developing diagnostic, preventive, and therapeutic tools for stress-related disorders.
Mechanisms for cyclic AMP–dependent regulation of CRH transcription
Studies in our laboratory have demonstrated that cAMP/phospho–CREB signaling is essential but not sufficient to activate CRH transcription. This finding led to the discovery that transcriptional activation of the CRH gene requires the interaction of CREB with the co-activator transducer of regulated CREB activity (TORC). Under resting conditions, the co-activator is found in a phosphorylated, inactive state in the cytoplasm. Its activation and nuclear translocation require protein kinase A (PKA)–mediated inhibition of protein kinases mediating TORC phosphorylation. The three TORC isoforms are encoded by different genes: TORC1, TORC2, and TORC3. Experiments examining the trafficking of TORC from the cytoplasm to the nucleus of CRH–producing neurons, as well as the effects of over-expression and knock-down of TORC on CRH transcription demonstrated that TORC is essential for CRH transcription. Paralleling stress-induced activation of CRH transcription, TORC shifts to the nucleus and binds to the CRH promoter as part of a complex with CREB. Of the three TORC isoforms, TORC2 appears to be the most important, though complete blockade of CRH transcription requires knock-down of both TORC2 and TORC3. We plan to examine the importance of TORC1 and TORC3 during physiological regulation of CRH transcription in vivo.
While cyclic AMP/PKA–dependent pathways are essential for activation of TORC and therefore transcriptional activation of the CRH gene, the main direct regulators of the CRH neuro—norepinephrine and glutamate—do not increase cyclic AMP production. We have previously demonstrated that stress induces type 1 CRH receptors in CRH neurons of the paraventricular nucleus (PVN), providing a mechanism for autocrine cyclic AMP–dependent regulation. In addition, the neuropeptide pituitary adenylate cyclase-activating polypeptide (PACAP) is released in the hypothalamus during stress and has been implicated in central control of the HPA axis. Studies during the past year conducted in collaboration with Lee Eiden used PACAP–knockout mice to show that CRH mRNA fails to increase in response to restraint stress in the absence of PACAP. The results suggest that PACAP is required for CRH transcription. Furthermore, PACAP can directly stimulate CRH hnRNA in primary cultures of hypothalamic neurons, indicating a direct effect. We are currently investigating the physiological role of PACAP in mediating cyclic AMP signaling in the CRH neuron along with the mechanisms involved.
During the past year, we placed significant emphasis on the mechanisms regulating TORC activity. In basal conditions, CRH transcription is low because TORC remains in the cytoplasm, inactivated by phosphorylation through Ser/Thr protein kinases of the AMP–dependent protein kinases (AMPK) family, including salt-inducible kinase (SIK). To determine which kinase is responsible for TORC phosphorylation in CRH neurons, we measured AMPK, SIK1, and SIK2 mRNA in the PVN of rats by in situ hybridization. In basal conditions, we found low levels of the three kinases in the dorsomedial PVN, consistent with location in CRH neurons. One-hour restraint stress raised SIK1 mRNA levels while SIK2 and AMPK mRNA showed only minor increases. In 4B hypothalamic neurons, or primary cultures, SIK1 mRNA (but not SIK2 mRNA) was inducible by the cyclic AMP stimulator forskolin. Over-expression of either SIK1 or SIK2 in 4B cells reduced nuclear TORC2 levels and inhibited forskolin-stimulated CRH transcription. Conversely, the non-selective SIK inhibitor staurosporine raised nuclear TORC2 content and stimulated CRH transcription in 4Bcells and in primary neuronal cultures (heteronuclear RNA). Specific shRNA knock-down of endogenous SIK2 but not of SIK1 induced nuclear translocation of TORC2 and CRH transcription, suggesting that SIK2 mediates TORC inactivation under basal conditions while induction of SIK1 limits transcriptional activation. Our current research aims to test the hypothesis that, while SIK2 mediates sequestration of TORC in the cytoplasm under basal conditions, induction of SIK1 during stimulation of the CRH neuron mediates phosphorylation and nuclear export of TORC, thus contributing to the termination of the transcriptional response.
Mounting evidence suggests that CREB–dependent transcriptional activation of a number of genes requires the co-activator TORC. Given that CREB is highly involved in neuroendocrine and brain function, we conducted in situ hybridization studies to determine the topographic distribution of TORC1, TORC2, and TORC3 mRNAs in specific regions of the rat forebrain. The studies showed that TORC1 is the most abundant isoform in most forebrain structures, followed by TORC2 and TORC3. While TORC1 was widely distributed at high levels in the forebrain, TORC2 was found in discrete nuclei and TORC3 mostly in the ependyma and pia mater. In the paraventricular nucleus of the hypothalamus, TORC1 and TORC2 mRNAs were abundant in the parvicellular and magnocellular neuroendocrine compartments, whereas TORC3 expression was low. All three isoform mRNAs were found elsewhere in the hypothalamus, with the most prominent expression of TORC1 in the ventromedial nucleus, TORC2 in the dorsomedial and arcuate nuclei, TORC 1 and TORC2 in the supraoptic nucleus, and TORC2 in the suprachiasmatic nucleus. These differential distribution patterns are consistent with complex roles for all three TORC isoforms in diverse brain structures and provide a foundation for further studies on the mechanisms of CREB/TORC signaling in brain function.
TORC and adrenal steroidogenesis
The activity of the HPA axis is characterized by a circadian as well as by an ultradian pattern of glucocorticoid secretion with one secretory pulse per hour. Despite increasing evidence of the importance of pulsatility in regulating glucocorticoid-responsive gene transcription, little is known about the mechanism underlying the pulsatility of glucocorticoid synthesis and release. An important part of our research during the past year focused on the mechanisms determining pulsatile secretion at the adrenal level. Adequate levels of active steroidogenic acute regulatory protein (StAR) and side chain cleavage cytochrome P450 (P450scc), which are both rate-limiting proteins for steroidogenesis, are essential for normal glucocorticoid secretion. As for CRH, transcription of StAR and P450scc genes also involves CREB phosphorylation and the CREB co-activator TORC. In collaboration with Stafford Lightman, we investigated the relationship between TORC activation and ACTH-induced steroidogenesis in vivo, by examining the time-course of the effect of a low-dose ACTH injection, mimicking an ultradian pulse, on the transcriptional activity of StAR and P450scc genes and nuclear accumulation of TORC2 in rat adrenal cortex. ACTH produced rapid and transient increases in plasma corticosterone, with maximal responses between 5 and 15 minutes and declining to near basal values by 30 minutes. StAR and P450scc hnRNA levels rose 15 minutes following ACTH and fell towards basal by 30 minutes. Concomitant with a rise in nuclear phospho-CREB, ACTH injection induced nuclear accumulation of TORC2, with maximal levels at 5 min and returning to basal by 30 minutes. The decline in nuclear TORC2 was paralleled by increases in SIK1 hnRNA and mRNA 15 and 30 minutes after injection, respectively. The early rises in plasma corticosterone preceding StAR and P450scc gene transcription suggest that post-transcriptional and post-translational changes in StAR protein mediate the early steroidogenic responses. Furthermore, the direct temporal relationship between nuclear accumulation of TORC2 and the increase in transcription of steroidogenic proteins implicates TORC2 in the physiological regulation of steroidogenesis in the adrenal cortex. The delayed induction of SIK1 suggests a role for SIK1 in the declining phase of steroidogenesis.
Additional studies tested the hypothesis that pulsatile ACTH release is critical for optimal adrenocortical function. In these studies, rats under blockade of endogenous HPA activity by oral methylprednisolone received ACTH (4 ng/h), infused for 24 hours, either as a constant infusion or in 5-minute pulses at hourly intervals. Control methylprednisolone-treated rats had very low plasma corticosterone levels with undetectable pulses and exhibited steroidogenic acute regulatory protein (StAR) and cytochrome P450 side chain cleavage (P450scc) hnRNA levels that were approximately 50% of those in untreated animals. Pulsatile, but not constant, ACTH infusion restored pulsatile corticosterone secretion, which was accompanied by parallel rises in StAR and P450scc hnRNA levels during the rising phase of the corticosterone pulse, which then fell during the falling phase. The pulsatile pattern of StAR and P450scc was paralleled by pulsatile transcription of the melanocortin 2 receptor (MC2R) accessory protein MRAP. These findings show that pulsatile exposure of the adrenal cortex to ACTH is critical for pulsatile secretion of corticosterone and suggest that episodic transcription of the rate-limiting proteins is necessary to maintain the physiological pattern of glucocorticoid secretion.
Early-life stress and CRH transcription
During the past year, we placed considerable emphasis on the long-term consequences of early-life stress on the function of the HPA axis. It is well recognized that stress exposure during early development causes long-lasting alterations in behavior and HPA axis activity, including increased levels of CRH mRNA in the PVN. The aim of this study was to test the hypothesis that early-life stress causes epigenetic changes in the CRH promoter leading to increased CRH transcription. We evaluated ACTH and corticosterone as well as CRH primary transcript or hnRNA levels (as an index of CRH transcription) in groups of 8-week-old female and male rats that had been subjected to maternal deprivation (MD) between days 2 and 10 after birth, with or without 30- or 60-minute restraint stress. Groups of control and MD rats were also used for methylation analysis of the CRH promoter in the PVN and amygdala. Adrenal weight, basal levels of plasma corticosterone, and hypothalamic CRH hnRNA were elevated in MD females but not in males. However, plasma corticosterone and CRH hnRNA responses to acute restraint stress were elevated in MD animals of both sexes. DNA methylation analysis of the CRH promoter revealed, in both sexes, a lower percent of methylation specifically in two CpGs located immediately preceding (CpG1) and inside (CpG2) the cAMP–responsive element (CRE) at −230, both in the PVN and amygdala. This CRE has been shown to be an absolute requirement for activation of the CRH promoter. In contrast to the PVN, the percentage of methylation of CpG1 and CpG2 in the amygdala was identical in control rats and rats subjected to maternal deprivation. These findings demonstrate that HPA axis hypersensitivity caused by neonatal stress results in long-lasting enhanced CRH transcriptional activity in the PVN of both sexes.
In gel-shift assays, we examined the molecular consequences of CRE methylation on CREB binding, using CRH-promoter CRE DNA oligos, unmethylated or methylated either at CpG1(Met-C1) or CpG2 (Met-C2), or at both Met-C1 and Met-2. When we incubated labeled oligos with nuclear protein extracts from hypothalamic 4B cells and resolved them on an acrylamide gel, we observed a strong band corresponding to bound phosphorylated CREB (p-CREB) for both the unmethylated and Met-C1 oligos. Following incubation with a p-CREB antibody, we verified the identity of the shifted band by supershift. Methylation of the intra-CRE CpG Met-C2 significantly decreased pCREB binding (by 50%), and we observed a similar decrease in binding for the Met-1,2 oligo. The results demonstrate that the methylation state of the intra CRE CpG of the CRH promoter significantly affects transcription factor binding. Hypomethylation of the −230 CRE in the CRH promoter likely functions as a mechanism for the increased transcriptional responses to stress that are observed as a consequence of maternal deprivation in rats.