Central glucocorticoid receptor-mediated effects of the antidepressant, citalopram, in humans: A study using EEG and cognitive testing

Summary

Our previous work in cellular and animal models has shown that antidepressants activate glucocorticoid receptor (GR) translocation, induce GR down-regulation, and decrease GR-mediated effects in the presence of GR agonists. However, whether these effects can be extrapolated to the human brain is still unclear. In this study, the effects of four days of treatment with the antidepressant, citalopram (20 mg/day), or placebo, were assessed in a double-blind, placebo-controlled, cross-over study. Central GR-mediated effects were examined by the effects of a single dose of cortisol (30 mg, orally) on two measures known to be sensitive to glucocorticoid administration: EEG alpha power and working memory function. Twenty healthy male subjects aged between 18 and 33 years participated to the study. The results suggest that GR activation by antidepressants, and the subsequent decrease in GR-mediated effects in the presence of GR agonists, indeed occurs in the human brain. Specifically, pre-treatment with citalopram decreased the well-known ability of cortisol to increase EEG alpha power and to impair working memory: cortisol-induced increase in EEG alpha power was (anteriorly) +15 to +20% (p = 0.01) after placebo and +5 to +8% (p > 0.5) after citalopram; and cortisol-induced increase in working memory errors was (at level 12, on average) 2.50 vs. 4.55 (p < 0.05) after placebo and 4.10 vs. 3.35 (p > 0.05) after citalopram. No effects were detected on alerting. These results are consistent with the notion that citalopram treatment activates GR translocation and inhibits the functional consequences of the subsequent cortisol administration. Our study further emphasizes the importance of the GR as a target for antidepressant action in humans.

Introduction

Twenty year after the first demonstration that antidepressants directly regulate glucocorticoid receptor (GR) expression and function (Pepin et al., 1989), the relevance of this pharmacological action for the human brain is yet to be established (Pariante and Lightman, 2008). Together with the mineralocorticoid receptor (MR), the GR regulates the activity of the hypothalamic–pituitary–adrenal (HPA) axis, especially under conditions where glucocorticoid levels are high, such as during stress and depression. Most evidence in animals and humans supports the notion that antidepressant facilitate GR-mediated signaling, especially in tissues relevant for hypothalamic–pituitary–adrenal (HPA) axis function. For example, many studies in rodents have demonstrated that chronic (that is, weeks) antidepressant treatment increases GR expression in the hippocampus, and reduce basal and stress-induced corticosterone levels (reviewed in Carvalho and Pariante, 2008, Pariante, 2009, Anacker et al., 2011a). In humans, antidepressant treatment has been shown to increase GR-mediated negative feedback on the HPA axis and to decrease cortisol levels in both depressed patients (before or concomitantly to the therapeutic response) (Zobel et al., 2001) and healthy controls (Pariante et al., 2004). Indeed lack of normalization of HPA axis activity by antidepressants in depression has been shown as predicting incomplete recovery and early relapse (Ising et al., 2007). Moreover, polymorphisms in the GR or associated proteins predict risk of depression and treatment response (Binder, 2009, Claes, 2009, Otte et al., 2009). These data fit perfectly within the theoretical construct of major depression as a disorder characterized by impairment of GR function (Holsboer, 2000, Raison and Miller, 2003). However, whether these effects on GR function by antidepressants can be extrapolated from HPA axis tissues to other human brain areas is unclear.

Studies conducted in experimental models have helped clarify the molecular mechanisms underlying the effects of antidepressants on the GR, without the confounding effects of depression and of the associated GR and HPA axis abnormalities. We and others have extensively shown that, in cellular models, antidepressants induce GR activation and translocation from the cytoplasm to the nucleus (Okugawa et al., 1999, Budziszewska et al., 2000, Yau et al., 2001, Heiske et al., 2003, Carvalho et al., 2008, Carvalho et al., 2010, Otczyk et al., 2008). Miller et al. (2002) first proposed that activation of protein kinase A (PKA) and subsequent GR phosphorylation is crucial for the (ligand-independent) GR activation by antidepressants. Indeed, in our recent study in a human neuronal stem cells system, we have described that antidepressants directly increase neurogenesis by activating the GR, an effect that require GR phosphorylation by PKA (Anacker et al., 2011b).

Of particular relevance to the present study, GR activation by antidepressants in cellular models is accompanied by a decrease in GR function in the presence of GR agonists such as dexamethasone, cortisol or corticosterone (Okugawa et al., 1999, Budziszewska et al., 2000, Yau et al., 2001, Heiske et al., 2003, Carvalho et al., 2008, Carvalho et al., 2010, Otczyk et al., 2008). Interestingly, antidepressant-induced GR activation in cells is associated with a rapid and persistent GR down-regulation (Pariante et al., 1997, Pariante et al., 2003a, Carvalho et al., 2008, Carvalho et al., 2010, Anacker et al., 2011b). Moreover, antidepressant treatment in rodents also induces GR down-regulation within the first few days of treatment (Reul et al., 1993, Yau et al., 2001), before the up-regulation occurs. Therefore, one possible explanation for the reduced GR function in the presence of GR agonists (after antidepressants) is that less GR is available for activation. However, it is also of note that we have recently shown, in the aforementioned study in stem cells, that antidepressants prevent the reduction in neurogenesis induced by dexamethasone via changes in GR phosphorylation and GR-dependent gene expression (Anacker et al., 2011b). Therefore, it is also possible that antidepressants induce functional changes in GR that then specifically counteract the effects of GR agonists.

In this study, we wanted to build on the evidence of the reduced GR function by antidepressants in the presence of GR agonists to test the hypothesis that GR activation by antidepressants occur in the human brain. Ultimately, such putative central effects may be more relevant to the therapeutic action of antidepressants and, more importantly, would be crucial in informing the debate on whether a successful antidepressant strategy should also involve targeting GR function. Therefore, we examined, in a double-blind, placebo-controlled, cross-over study in healthy volunteers, the impact of four days of treatment with the selective serotonin reuptake inhibitor (SSRI) antidepressant, citalopram (20 mg/day), or placebo, on the effects of a single oral dose of cortisol (hydrocortisone, 30 mg) on two measures known to be sensitive to glucocorticoid administration: increases in EEG alpha power (Tops et al., 2005) and impairment of working memory function (Lupien et al., 1999). Additionally, because of the widespread cognitive effects of even a single does of citalopram (Doerr et al., 2010, Oranje et al., 2011), we also examined alerting, orienting and executive control, to test for the regional and functional specificity of these hypothesized effects. The design of the study is based on the combination of our previous work in healthy volunteers, showing increased GR-mediated negative feedback after four days with citalopram (Pariante et al., 2004), and impairment of attention and spatial working memory by cortisol administration (Gallagher et al., 2009). Moreover, there is consistent evidence from cellular and animal studies that one-to-four days of treatment with antidepressants are effective in inducing GR activation and down-regulation: for example, amitriptyline induces GR down-regulation in the hippocampus of rats after three and seven days of treatment (Reul et al., 1993); in fibroblast and neuronal cell cultures, antidepressant-induced GR activation and down-regulation starts after as little as 12 h, and is still present after 4 days (Pariante et al., 2001, Pariante et al., 2003a, Anacker et al., 2011b); and, in human peripheral blood mononuclear cells (PMBCs), 24-h incubation with antidepressants leads to a decrease in GR function as shown by reduced inhibition of stimulated interleukin-6 production by dexamethasone (Carvalho et al., 2008, Carvalho et al., 2010). Therefore, based on the experimental evidence from cellular models described above, we hypothesized that the four days of citalopram would decrease the effects of the subsequent dose of cortisol on EEG and cognitive function, as a consequence of the antidepressant-induced GR activation in relevant brain areas.