The role of genetic variation in the glucocorticoid receptor (NR3C1) and mineralocorticoid receptor (NR3C2) in the association between cortisol response and cognition under acute stress

Highlights

• The effect of GR polymorphisms (including rs41423247) on HPA – axis reactivity is replicated.

• An interaction of MR polymorphisms and cortisol reactivity following acute stress on attentional performance is shown.

• Further individual factors contributing to the association between stress and cognition should be considered in the future.

Abstract

Although HPA – axis reactivity has repeatedly been related to cognitive functioning, ambiguity remains regarding the direction of the effect, i.e. whether it benefits or impairs functioning. Genetic factors that contribute to HPA – axis reactivity on the one hand and to cognitive functioning on the other could therefore help clarify the association between stress and cognition. We genotyped 10 single nucleotide polymorphisms (SNPs) on the NR3C1 gene (rs10482682, rs33389, rs10482633, rs10515522, rs2963156, rs4128428, rs9324918, rs41423247, rs6189, rs10052957) coding for the glucocorticoid receptor (GR) and 4 SNPs on the NR3C2 gene (rs6810951, rs4635799, rs11099695, rs2070950) coding for the mineralocorticoid receptor (MR) and required N = 126 healthy males to perform tasks assessing attention and reasoning before and after experiencing an acute laboratory stressor (the Socially Evaluated Cold Pressor Test, SECPT). Haplotype analyses revealed significant effects of NR3C1 (p = 0.011) and NR3C2 (p = 0.034) on cortisol stress response. NR3C2 also influenced attentional performance via an interaction with stress-induced cortisol response (p < 0.001). Neither NR3C1 haplotype nor NR3C2 haplotype was associated with reasoning abilities.

Results suggest that the association between stress induced cortisol reactivity and cognition strongly depends on genetic variation. The idea of an optimal arousal level depending on stress reactivity and genetic disposition is discussed.

Introduction

The Hypothalamic Pituitary Adrenal (HPA) axis is an important endocrine modulator of the stress reaction that helps us cope with everyday challenges. Consequently, a dysregulation of the HPA – axis or – more specifically – a disruption of the cortisol homeostasis leads to an impairment in dealing with stressors. Therefore, abnormalities in HPA activity have repeatedly been associated with affective psychopathologies, most prominently depression (De Rijk et al., 2008, Gold et al., 1995, Stetler and Miller, 2011). A large number of studies show altered cortisol levels in depressed patients following acute stress (Pariante and Lightman, 2008), as well as slower recovery from stress (Burke et al., 2005) and altered tonic cortisol levels, both after waking (Chida and Steptoe, 2009, Fries et al., 2009) and over the course of the day (Van den Bergh et al., 2008). Several explanations have been posited to account for the relationship between disrupted HPA axis functioning and depression. Some studies suggest an effect of adverse life events in childhood on cortisol levels in adulthood (McCrory et al., 2011). Genetic variance may also explain the relationship between depression and the HPA-axis. The NR3C1 gene is located on chromosome 5q31.3 and encodes for the glucocorticoid receptor (GR). This gene has been shown to influence cortisol levels in humans. Several polymorphisms located on this gene have also been associated with depression (van West et al., 2006, Schatzberg et al., 2014). Individual differences in some single nucleotide polymorphisms (SNPs) have also been found to influence the reactivity of the HPA axis and, in turn, cortisol levels following acute stress in healthy individuals (De Rijk, 2009, Taylor et al., 2014). Several haplotypes have been identified (see Li-Tempel et al., 2016), including some very prominent SNPs such as the BCL1 polymorphism (rs41423247), the TthIII (rs10052957), or the combined SNPs ER22/23EK (rs6189 and rs6190), although the latter show very little variance in Caucasians (the frequency of the minor allele is less than 1%) and, therefore, do not seem suitable candidates for the investigation of individual differences in a healthy population. The gene NR3C2, located on chromosome 4q31.1 and coding for the mineralocorticoid receptor (MR), has also been associated with depression and related traits such as hopelessness and rumination. In particular, a haplotype comprising the SNPs rs5522 (MR I180V; A → G resulting in an amino acid change from isoleucine to valine) and rs2070951 (MR-2; G → C) has been investigated in several studies (e.g. Klok et al., 2011, Vinkers et al., 2014, Vinkers et al., 2015). Furthermore, De Rijk et al. (2006) report an association between the NR3C2 MR I180V polymorphism (rs5522) and cortisol increase following the Trier Social Stress Test (TSST), in which carriers of the VAL/VAL variant displayed greater HPA-axis reactivity compared with VAL/MET carriers or participants homozygous for MET. In contrast, van Leeuwen et al. (2011) found that the MR I180V and the MR-2 (rs2070951) interact on in vitro MR expression rate and on salivary cortisol levels after undergoing the TSST: homozygous carriers of the MR-2C/MR I180 haplotype showed both higher expression rates and greater HPA-axis reactivity.

Affective disorders are not only marked by emotional dysregulation but also by cognitive impairment (for a review see Marazziti et al., 2010). Similarly, the influence of stress on cognitive functioning is well documented (for a review see Lupien et al., 2007). However, the direction and the conditions of stress effects on cognition are poorly understood. Whereas some studies find enhancing effects of stress on cognitive functioning (e.g. Beste et al., 2013, Duncko et al., 2009), others find stress to be impairing with respect to cognitive functioning (e.g. Schoofs et al., 2009, Shansky and Lipps, 2013). Shields et al. (2016) review these inconsistencies in their recent meta-analysis. Overall, the authors come to the conclusion that acute stress tends to impair working memory and cognitive flexibility dependent on stress severity and task load whereas the meta-analytic results for cognitive inhibition seem to depend on type of inhibition task. In a previous meta-analysis, Shields et al. (2015) investigated the nongenomic and genomic effects of cortisol administration on cognitive functions and also found an overall impairment of working memory functions. In contrast to Shields et al. (2016), there were no effects of cortisol on cognitive flexibility and enhancing (nongenomic) effects of cortisol on inhibition abilities. Hence, the interplay between the induction of acute stress and subsequent cortisol effects should be considered.

We recently reported data showing that participants responding to acute stress with an increase in cortisol activity improve on tasks with low cognitive load, compared to people not responding to stress on neuroendocrine level (i.e. a cortisol increase following the acute stressor). However, there were no differences between cortisol responder groups in tasks requiring higher cognitive functioning (i.e. reasoning abilities) (Plieger et al., 2016).

The effects on cognitive functioning are significantly modulated by the binding of cortisol to the lower affinity GR and the higher affinity MR (De Kloet et al., 2005). GR and MR are widely distributed in the brain and occur in prefrontal and hippocampal areas. They are, therefore, very likely involved in the association between stress evoked cortisol expression and executive functioning and memory (De Kloet et al., 2009, Porcelli et al., 2008, Schwabe et al., 2010). Activity and density of GR and MR are highly heritable, so that individual differences in cognitive performance under acute stress should be observable.

Thus, there is mounting evidence that acute stress can increase as well as decrease cognitive functioning dependent on factors related to the complexity of the cognitive tasks, the amount of stress and most importantly, factors inherent in the individual.

Interestingly, little evidence considers the role of variation in the mineralocorticoid and the glucocorticoid receptor genes on the effects of stress on cognition. The NR3C2 gene in particular is, at least in human, insufficiently considered with respect to cognitive functioning, although there is evidence that the MR is involved in cognitive functions such as attention, working memory, memory, and emotional information processing and memory (e.g. De Kloet et al., 2016, Hamstra et al., 2015; Schwabe et al., 2013; Vogel et al., 2016). One exception is a study by Pan et al. (2011) who found evidence for an association between NR3C2 and IQ in a genome-wide association study (GWAS). Very recently, Keller et al. (2017) presented data dealing with the association of tonic (i.e. under basal/unstimulated conditions) cortisol levels and NR3C1 and NR3C2 polymorphisms on cognitive variables in depressed and healthy controls. The authors found an effect of NR3C1 polymorphisms on attentional and working memory tasks beyond the impact of tonic cortisol. NR3C2 polymorphisms were associated with long term memory encoding and retrieval. Given the expression of GR and MR in the prefrontal cortex, both are promising candidates for the investigation of stress effects on cognition (Spijker and Van Rossum, 2011, Vogel et al., 2016).

However, no study has thus far considered the effects of NR3C1/NR3C2 on cognition under acute stress. Therefore, the aim of the present study is to a) replicate previous findings reporting that variations on the NR3C1 gene could predict cortisol response to stress and b) to investigate the effects of polymorphisms on the NR3C1 and the NR3C2 genes on cognition under stress. We predict an interaction between genetic variation and endocrine stress response following an acute stressor on cognitive functioning. Such an outcome would delineate the path from genes to behavior via the endocrine system.