Discussion
In the present study, we hypothesized that the regulation of GRs and MRs in the PVN may influence depressive-like behaviors by modulating CRF in the PVN. After exogenous CORT exposure via drinking water for 21 days, plasma corticosterone rhythms were clearly altered in the present rat model (Supplementary Fig. S2), which resulted in the overproduction of CRF and reductions of GR and MR expression in the PVN. Intra-PVN administration of only the combination of the GR antagonist RU486 and MR antagonist spironolactone inhibited depressive- or anxiety-like behaviors and restored MR, GR, and CRF levels and the MR/GR ratio in the PVN to normal. However, intra-PVN administration of either RU486 or spironolactone alone did not inhibit depressive-like behaviors. Intra-PVN administration of spironolactone alone inhibit anxiety-like behaviors.
Glucocorticoid receptors are richly expressed in the medial parvocellular PVN and co-localize with CRF, thus placing the PVN in a prime position to control the output of neurons that activate the HPA axis (Liposits et al., 1987; Uht et al., 1988). Although the PVN is well known to be the major site of CRF synthesis in the brain (De Souza, 1995), with projections to many extrahypothalamic areas that are involved in various behaviors, its role in stress-related behaviors has not been studied extensively. Chronic variable stress reduces the expression of GRs in the medial parvocellular PVN. CRF expression is negatively regulated by GCs, and GRs may play a role in the observed upregulation of CRF during chronic stress (Herman et al., 1995). This negative correlation between GR and CRF levels (i.e., downregulation of GRs and upregulation of CRF) was also observed in the PVN in the present rat model. However, the intra-PVN infusion of RU486 returned GRs level in the PVN to normal, had no effect on MR, decreased the MR/GR ratio, and increased CRF levels, and depressive- and anxiety-like behaviors did not improve. These results showed that simply increasing the expression of GRs in the PVN without reversing CRF to normal levels could not improve either depressive- or anxiety-like behavior. Therefore, one unresolved issue is why the recovery of GRs did not return CRF levels to normal but rather further increased the abnormally high levels of CRF. One possibility is the recruitment of other mechanisms beyond the negative regulation of CRF by GRs. Lamberts et al. reported that RU486 activated the HPA axis, resulting in a resetting of this system at a higher level at which the diurnal rhythm and responsiveness to CRF stimulation were maintained, whereas the sensitivity to dexamethasone diminished, and these changes were caused by the induction of partial cortisol receptor resistance during RU486 therapy (Lamberts et al., 1991). Based on its mechanism of action and evidence that the restoration of normal hypothalamic CRF activity by RU486 appears to be the rate-limiting step in HPA axis recovery (Muglia et al., 2000), central GR blockade may deprive the hypothalamus from GR activity, thereby stimulating CRF-producing neurons in the hypothalamus and resulting in the further enhancement of CRF.
The HPA axis has long been recognized for its involvement in depression, with a focus on cortisol/CORT and less of an emphasis on aldosterone as a stress hormone. A recent study found that aldosterone acted selectively in mood-regulating brain areas, without competing with cortisol/CORT (Murck et al., 2014). Although receptors for aldosterone (i.e., MRs) are known to be expressed in the PVN, its role in the local regulation of HPA axis function has been less explored. The MR antagonist spironolactone inhibits the effects of mineralocorticoids by displacing them from MRs in the PVN. Wu et al. (2013) found that chronic subcutaneous CORT treatment triggered several depressive-like behaviors and downregulated MR expression in parallel in the hippocampus and hypothalamus. On the other hand, they also reported that intraperitoneal spironolactone administration for 7 days in animals that were subcutaneously treated with CORT reduced immobility time in the FST and improved performance in a novel object recognition test (Wu et al., 2013). In the present study, CORT-exposed rats via drinking water also exhibited lower MR expression in the prefrontal cortex, hippocampus, and amygdala (data not shown). Bitran et al. (1998) hypothesized that the anxiolytic effects of an MR antagonist may be more pronounced in animals whose hippocampal MRs are activated by CORT or the MR agonist aldosterone. In the present study, the intra-PVN infusion of spironolactone restored MR levels in the PVN to normal and improved anxiety-like behaviors in the NSFT and SIT, without altering CRF levels in the PVN, but it did not improve depressive-like behaviors in the FST or SPT. However, still needing to be confirmed is whether the anxiolytic effects were caused by the restoration of MRs in the PVN in the present rat model. Thus, further studies should seek to determine the pharmacological specificity of the effects of intra-PVN spironolactone administration to exclude possible actions of spironolactone at sites other than MRs.
The balance between MRs and GRs during chronic stress plays an important role in the etiology of depression (De Kloet et al., 1998). The pharmacological modulation of GRs and MRs also influences the endocrine stress response and depressive-like behaviors in rodents. The ways in which MRs and GRs interact to influence the endocrine system and related behaviors are still being clarified. In the present study, RU486 alone restored the downregulation of GR levels in the PVN to normal but did not influence the downregulation of MRs, thus making the MR/GR ratio significantly lower than normal and failing to suppress depressive- and anxiety-like behaviors (Fig. 3). Spironolactone treatment alone restored MR levels in the PVN to normal but did not influence the downregulation of GRs, thus making the MR/GR ratio significantly higher than normal and inhibiting anxiety-like behaviors in the NSFT and SIT (Fig. 4). When RU486 and spironolactone were administered together, MR and GR levels and the MR/GR ratio returned to normal, and depressive- and anxiety-like behaviors significantly improved (Fig. 5). The present findings suggest that the normalization of MR, GR, and CRF levels and the MR/GR ratio may be necessary conditions to inhibit depressive- and anxiety-like behaviors in the present rat model. Nonetheless, the restoration of MR levels in the PVN to normal may be a sufficient condition to improve anxiety-like behaviors.
CRF also acts within other regions of the brain where it directly contributes to stress-responsive behavior. Indeed, some depressed patients had high CSF levels of CRF, and these levels are normalized in patients who respond to antidepressant treatment, suggesting that these alterations of CRF levels contribute to the symptomology of depression (Pandey et al., 2019). In the central nervous system, patients had high levels of CRF in the PVN and in monoaminergic nuclei, including the locus coeruleus and raphe (Austin et al., 2003; Bissette et al., 2003). Elevations of plasma CRF levels in depressed individuals are ameliorated by successful electroconvulsive shock therapy (Nemeroff et al., 1991) and antidepressant treatment (Heuser et al., 1998). The causal relationship between depression and CRF hyperactivity and receptor expression is unclear. The present results suggest that a state of equilibrium between GRs and MRs in the PVN may improve the function of the HPA axis, which can improve depressive- and anxiety-like behaviors. However, the specific ways in which GRs, MRs, and CRF in the PVN are involved in the pathogenesis of depression and anxiety need to be elucidated.
The present study has some limitations. We only examined male rats to eliminate gender influences that are known to modulate experimentally induced emotion and HPA activity. The significance of sex differences in PVN function is virtually unexplored. Therefore, we do not know whether our results are generalizable to female rats. Furthermore, we did not detect the activity of HPA function. Therefore, unclear was the state of the HPA axis when antidepressant effects were induced by GR and MR blockade. One possibility is that the target sites of the antidepressant effects of RU486 and spironolactone may not be in the PVN.
Overall, the present study found that the combination of pharmacological GR and MR antagonism in the PVN prevented the pathogenesis of depressive- and anxiety-like behaviors in rats that were chronically treated with CORT, confirming that both MRs and GRs in the PVN play an important role in CRF function. Such a combined pharmacological approach may be a novel strategy to prevent hypercortisolism- or chronic stress-induced depression and its metabolic complications. Much work still needs to be done to achieve a more complete understanding of the role of GRs, MRs, and CRF in the PVN in depressive- and anxiety-like behaviors.