Abstract

1. Introduction

The scientific understanding of ‘stress’ and its ramifications for the organism have continually evolved. Based on Claude Bernard's theory of the internal milieu, Walter Cannon first used the concept of homeostasis to explain the ‘fight-or-flight’ response of an organism presented with a threat (Cannon, 1932). In a biological sense, Hans Selye coined the term ‘stress’ as the non-specific response of the body to any homeostatic demand (Selye, 1936). While it is still generally accepted that the physiological role of the stress response is to coordinate autonomic, neuroendocrine, and immune responses to potential homeostatic threats, an emerging concept in stress neurobiology suggests that the primary role of stress responding is to mobilize energy to promote context-specific survival and not necessarily sustain homeostatic systems at levels maintained prior to a challenge (Dallman et al., 2006; Nederhof and Schmidt, 2013). Given this framework, responses to both acute and chronic stress are considered adaptive, up to a point, and prepare the organism for current and future demands. Thus, for the purpose of the current review, stress will be defined as a stimulus that mobilizes energetic systems to respond to an ongoing or anticipated challenge.

Responding to stress involves the concerted activity of multiple, interacting central stress-regulatory systems to mobilize energy for the organism. Activation of the stress response occurs either as a consequence of, or in anticipation of, a challenge (Myers et al., 2012b). Anticipatory responses require the organism to reference prior experiences to predict the need for energy mobilization, primarily mediated by multi-synaptic forebrain projections to the medial parvocellular paraventricular nucleus (PVN) of the hypothalamus. Systemic challenges are largely reflexive responses to physiological disruption generated by direct projections from the hindbrain to the PVN, though there is considerable overlap and integration at various nodes throughout the brain (Herman et al., 2012, 2003; Ulrich-Lai and Herman, 2009). Thus, the neuroendocrine response to stress is a highly-regulated, temporal process, involving the integration of sensory information from multiple modalities to rapidly activate, as well as inhibit the secretion of glucocorticoids.

The neuroendocrine stress cascade, comprising the hypothalamic-pituitary-adrenocortical (HPA) axis, begins with the release of adrencorticotropic hormone (ACTH) secretagogues from neurosecretory neurons in the medial parvocellular PVN, which project to hypophysial portal vessels in the external zone of the median eminance (Bruhn et al., 1984). Secretagogues travel via the portal veins to the anterior pituitary, where they can access corticotropes (De Wied et al., 1957; Gibbs and Vale, 1982; McCann and Fruit, 1957; Saffran and Schally, 1956). The pioneering work of Wylie Vale and colleagues provided initial identification of corticotropin releasing factor (CRF) as the primary driver of pituitary ACTH release (Bale and Chen, 2012; C. Rivier et al., 1983; Rivier et al., 1982; J. Rivier et al., 1983; Spiess et al., 1981; Swanson et al., 1983; Vale et al., 1981). Subsequent studies, also by Vale and colleagues, revealed the existence of several co-secretagogues that synergize with CRF, including arginine vasopressin (AVP) (Rivier and Vale, 1983a, 1983b; Sawchenko et al., 1984; Vale et al., 1983, 1981). By way of the systemic circulation, ACTH acts at the level of the adrenal cortex to induce the release of glucocorticoids, cortisol in some species (e.g., humans, non-human primates) and corticosterone in others (e.g., rats, mice) (Dallman and Jones, 1973; Dallman et al., 1987). At the adrenal, cortisol/corticosterone is released in pulses, the timing of which dictates the overall magnitude of both baseline activity and stress responses (Lightman et al., 2008; Young et al., 2004). Pulsatile patterns of glucocorticoid release are dictated by ultradian rhythms (for review see de Kloet and Sarabdjitsingh, 2008; Sarabdjitsingh et al., 2012a). This rhythmicity of glucocorticoid release is essential for maintaining cellular responsiveness and promotes wide-ranging glucocorticoid actions from gene transcription to behavior (Conway-Campbell et al., 2010; Sarabdjitsingh et al., 2010b). Glucocorticoids then travel throughout the body, exerting a multitude of effects in the periphery including glycogen breakdown and gluconeogenesis (Coderre et al., 1991; Exton, 1979; Munck et al., 1984; for detailed reviews of brain circuits regulating glucose homeostasis see Schwartz et al., 2000; Seeley and Woods, 2003; Woods et al., 1998). Glucocorticoids cross the blood-brain-barrier, and primarily bind to mineralocorticoid (MR) and glucocorticoid receptors (GR) in neurons and/or glia. In the brain, MR has high-binding affinity for corticosteroids and, consequently, is activated at basal levels (de Kloet and Sarabdjitsingh, 2008; de Kloet et al., 1998). Thus, MR is thought to sense resting levels of glucocorticoids and promote key functions associated with low glucocorticoid levels, including circadian drive of the HPA axis and mnemonic function (de Kloet et al., 2005). Conversely, GR has a lower binding affinity for glucocorticoids and is largely unoccupied at basal levels. Thus, GR is thought to be particularly important in signaling mediated by stress levels of glucocorticoids (Boyle et al., 2005; de Kloet and Reul, 1987; de Kloet et al., 2005; Reul and de Kloet, 1985). GR is abundantly expressed throughout the brain, including the primary stress-regulatory sites discussed in this review: medial prefrontal cortex (mPFC), hippocampus, amygdala, bed nucleus of the stria terminalis (BST), hypothalamus, and hindbrain (Fuxe et al., 1987; Meaney et al., 1985; Reul and de Kloet, 1986). MR has a more restricted distribution that overlaps with that of GR in several key regions, including the mPFC, hippocampus, and amygdala (de Kloet and Reul, 1987; Reul and de Kloet, 1985).

There is a vital need for glucocorticoid activity to be tightly regulated, requiring systems coordination from cellular to behavioral levels. In response to stress, glucocorticoid signaling promotes organismal adaptation to environmental conditions and helps to meet the resulting energetic demands. This adaptation requires the integration of multiple systems and engages key limbic-neuroendocrine circuits. Forebrain, hypothalamic, and hindbrain circuits are activated by glucocorticoids and participate in the coordination of physiological and behavioral output. However, when energetic drive does not appropriately match environmental demand, or the organism is chronically activating these systems, risk factors emerge for a variety of stress-related pathologies. The present article will review the actions of glucocorticoids in central stress-regulatory circuits, focusing on the rodent literature, in the context of the adaptive role of the stress response for the organism. The review will summarize the role of central glucocorticoid actions on synaptic, neuroendocrine, and behavioral regulation, highlighted by a discussion of the energetic integration of stress and the importance of context-specific regulation of glucocorticoids. The energy mobilizing effects of glucocorticoids require integration of cellular activity, circuit connectivity, and behavioral output to coordinate context-appropriate adaptation. We propose that glucocorticoid-mediated energetic drive generates an adaptive capacity in response to environmental demand; however, the cost of repeatedly or excessively driving adaptive systems may compromise performance under conditions of elevated environmental pressure. Thus, we will examine the integrative actions of glucocorticoids on the primary limbic sites mediating organismal stress responsiveness within the framework of context-specific adaptation.

2. Synaptic Actions of Glucocorticoids

The cellular actions of glucocorticoids are largely dependent on brain site and the relative expression of GR and MR (de Kloet, 2013a) (Table 1). Glucocorticoids also act in concert with monoaminergic and peptide neurotransmitters, particularly noradrenergic (Quirarte et al., 1997; Roozendaal et al., 2008, 2006a, 2006b, 2002) and CRF systems (Bale and Vale, 2004; Bale, 2005; Meng et al., 2011), which have been reviewed elsewhere (de Kloet, 2013b; Ferry and McGaugh, 2000; Heinrichs and Koob, 2004; Roozendaal, 2000). Further, the synaptic actions of glucocorticoids are critically affected by the recent stress history of the organism, a concept known as ‘metaplasticity’, which will be expanded on in the following section (for review see: Schmidt et al., 2013). Acute and chronic stress often yield different effects on cellular function to meet context-specific energetic demands placed on the organism. Thus, we will discuss the effects of glucocorticoids on cellular function in light of these considerations.

3. Connectivity and Integration of Glucocorticoid-Responsive Circuits

Stress integrative functions are regulated by forebrain circuits, primarily involving the above mentioned regions (mPFC, hippocampus, and amygdala) (Ulrich-Lai and Herman, 2009) (Table 2). Notably, these interconnected limbic forebrain sites do not send substantial projections to stress-effector neurons in the PVN. Thus, their influence on HPA axis output is communicated through intervening PVN-projecting neurons. Inhibitory GABAergic inputs to the PVN emanate from several structures in the basal forebrain and hypothalamus, including the BST, preoptic area (POA), and dorsomedial hypothalamus (DMH) (Boudaba et al., 1996; Radley et al., 2009; Roland and Sawchenko, 1993). In contrast, glutamatergic inputs to the PVN originate predominantly from hypothalamic nuclei, including the ventromedial hypothalamus (VMH), posterior hypothalamus (PH), as well as DMH (Ulrich-Lai et al., 2011). Neurons in these regions, including those that project to the PVN, show pronounced activation by stressful stimuli (Cullinan et al., 1996, 1995), consistent with a role in stress regulation. All hypothalamic PVN-projecting regions receive mixed GABAergic and glutamatergic input from other hypothalamic nuclei (Myers et al., 2013), which may be responsible for intrahypothalamic mechanisms governing the integration of forebrain limbic inputs and stress responsiveness based on metabolic demand.

Importantly, GR is expressed in numerous hypothalamic PVN-projecting neurons, including the POA, DMH, and arcuate nucleus (Fuxe et al., 1987). In the PVN, glucocorticoids signal by non-genomic mechanisms to rapidly inhibit activation of parvocellular neurons and consequent drive on the HPA axis (Evanson et al., 2010; Tasker and Herman, 2011). These ‘fast feedback’ glucocorticoid effects are non-genomic in nature, acting through an endocannabinoid-dependent mechanism to inhibit glutamatergic drive (Tasker and Herman, 2011). Fast inhibitory effects of glucocorticoids are attenuated in animals with PVN GR deletion, consistent with action via GR (Haam et al., 2010). However, the role of GR in other PVN-projecting hypothalamic cell groups remains to be determined.

4. Behavioral Effects of Glucocorticoids

Activation of glucocorticoid responsive pathways, particularly within the limbic system, can lead to pronounced changes in behavior that have long-term implications for organismal well-being (Arnett et al., 2011) (Table 3). Functional changes at the synaptic and circuit levels affect a variety of neurobehavioral systems. For instance, mice with deletion of GR in the forebrain (mPFC, hippocampus, and BLA) exhibit increased despair-like behavior in the forced swim test (FST) and tail-suspension test (TST), as well as anhedonia in the sucrose preference test (SPT) (Boyle et al., 2005). In addition, mice overexpressing MR in the forebrain exhibit decreased anxiety-like behavior in the open field test (OFT) and elevated-plus maze (EPM) (Rozeboom et al., 2007). These studies support a role for forebrain glucocorticoid signaling in inhibiting behavioral responses indicative of depression and anxiety. Although numerous psychiatric disorders are associated with dysregulation of stress systems, the precise role of glucocorticoid dyshomeostasis in pathology remains to be determined.

5. Clinical Relevance of Glucocorticoid Signaling

Recent technological advances have allowed for more in-depth analysis of glucocorticoid effects in humans. In healthy controls, glucocorticoids have profound time-dependent effects on memory and decision-making. For instance, the synthetic glucocorticoid hydrocortisone given 30 min prior to a memory encoding task impairs memory contextualization performance, while hydrocortisone given 210 min prior to the task has the opposite effect (van Ast et al., 2013). These findings suggest that the rapid effects of glucocorticoids promote memory of the most emotionally salient aspects of an experience, whereas the delayed effects of glucocorticoids may enhance cognitive function. Similarly, hydrocortisone given before sleep in the evening after the presentation of emotional imagery enhances memory consolidation of emotional information and suppresses amygdalar activation during retrieval (van Marle et al., 2013). Acute stress also shifts hippocampal-dependent learning toward habitual striatum-dependent learning, an effect that can be prevented by pretreatment with an oral MR antagonist (Schwabe et al., 2013). Additionally, this stress-induced shift in memory processing correlates with altered coupling of the amygdala with the hippocampus and dorsal striatum, marked by decoupling of amygdala-hippocampal connectivity and strengthening of amygdala-putamen connectivity. The effects of stress on alterations in amygdala connectivity with the hippocampus and dorsal striatum are also prevented by MR antagonist pretreatment (Schwabe et al., 2013). Thus, the importance of glucocorticoids for regulating cognitive function and the brain regions mediating glucocorticoid activity during these processes are beginning to be characterized in humans. While there has been a recent increase in the number of studies assessing the central actions of glucocorticoids in humans, there is still a vital need to establish the mechanistic effects of glucocorticoids in health as well as psychopathology.

Glucocorticoid signaling is essential for adaptation to stress and alterations in glucocorticoid activity may underlie transitions to pathology in humans. In fact, several polymorphisms of the glucocorticoid receptor gene have been reported in individuals with posttraumatic stress disorder (PTSD), as well as major depression. For instance, the single nucleotide polymorphism (SNP) Bcl-1 (C to G nucleotide change involving a restriction site in intron 2) is associated with glucocorticoid hypersensitivity and lower cortisol levels (Derijk et al., 2008; Hauer et al., 2011; Huizenga et al., 1998; Kumsta et al., 2008; van Rossum et al., 2006, 2003). Further, homozygous carriers of the Bcl-1 SNP(G allele) have more traumatic memories 6 months after intensive care unit care than heterozygous or non-carriers of the SNP (Hauer et al., 2011). Similarly SNPs in the steroid receptor chaperone FK506 binding protein 5 (FKBP5) increase glucocorticoid receptor sensitivity leading to lower basal cortisol levels and increased risk for PTSD following traumatic experience (Binder et al., 2008; Klengel et al., 2013; Wüst et al., 2004; Yehuda et al., 2009). Further, stress-level doses of hydrocortisone reduce PTSD symptomtology in patients following traumatic experience (Aerni et al., 2004; Delahanty et al., 2013; Schelling et al., 2003, 2001, 1999; Weis et al., 2006; Zohar et al., 2011). In healthy controls, stress enhances the consolidation of emotionally salient information (Andreano and Cahill, 2006; Buchanan and Lovallo, 2001), while also impairing memory retrieval (de Quervain et al., 2009, 2007). Therefore, hydrocortisone has been proposed to impair the retrieval of traumatic memories resulting in decreased PTSD symptoms following a traumatic experience (Hauer et al., 2013). Glucocorticoids also reduce local oxygen metabolism selectively in the amygdala and parahippocampal gyrus immediately following stress-level hydrocortisone exposure (Lovallo et al., 2010). Collectively, these studies suggest that stress-level doses of hydrocortisone may prove useful, particularly in patients with polymorphisms of the GR gene or genes affecting glucocorticoid sensitivity (Hauer et al., 2013; Schelling et al., 2013). In contrast to the low levels of cortisol generally associated with PTSD, patients with major depression typically display hypercortisolemia (Yehuda et al., 2004). Hypometabolism (decreased uptake of fluorine deoxyglucose) as well as hypermetabolism (increased regional cerebral blood flow) have been reported in the subgenual anterior cingulate cortex (Brodmann area 25) of patients with major depression (Drevets, 2000; Mayberg et al., 2005). Additionally, hydrocortisone decreases subgenual cingulate oxygen utilization evoked by sad stimuli, suggesting that this region is a key target for glucocorticoid-mediated effects on emotional processes (Sudheimer et al., 2013). Thus, the aggregate data suggest that glucocorticoids have pronounced effects on local metabolic activity within forebrain limbic circuits and alterations in glucocorticoid signaling within these areas may play a prominent role in stress-related neuropsychiatric disorders.

6. Stress, Energetics, and Adaptation

Responses to stress occur at the level of the individual, requiring both appraisal systems (limbic forebrain) and effector sites (hypothalamus and hindbrain) in the brain, as well as peripheral physiological systems. Consequently, regulation of the HPA axis is a distributed process integrated by multiple sites with considerable overlap and redundancy to permit compensation. For instance, both the plPFC and the ventral hippocampus provide inhibition of the HPA axis via convergent projections onto PVN-projecting neurons in the aBST (Radley and Sawchenko, 2011). Neuroendocrine responses to stress are initiated by the brain and result in the secretion of glucocorticoids to promote energy mobilization for current or anticipated needs. In turn, glucocorticoids act as a signal that influences cellular function, gene transcription, circuit activity, and, ultimately, behavior. Thus, responses to both acute and chronic stress promote context-specific adaptation based on the real or perceived needs of the individual. Although responses to chronic stress may promote adaptation, there is a cost of adaptation. Consequently, the adaptive cost may be deleterious for the organism when stress effector systems are repeatedly activated to meet the energetic demands of severe or prolonged stress. The physiological and behavioral consequences of driving energetic systems may encompass both stress resilience, as well the transition to pathology. Thus, the stress response leads to a redistribution of energy resources to meet emergent or anticipated needs driven by a glucocorticoid signal that influences cellular function, neurocircuits, and behavior. This signal pushes physiological systems to adapt until the adaptive cost becomes greater than the adaptive capacity of the individual. In this framework, the adaptive capacity can be defined as the degree to which the organism can deploy energy mobilizing systems (e.g. HPA axis) to successfully meet current or anticipated needs.

The relationship between stress and energetics can be summarized by examining the relationship between systems performance, environmental demand, and energy input (Figure 1). The performance of the individual is largely dependent on energetic status, environmental demand (i.e. context), adaptive capacity, and adaptive cost. Glucocorticoids regulate the energetic status of the individual and are a crucial determinant of both adaptive capacity and cost. In turn, the adaptive capacity and cost directly influence the performance of the individual across varying ranges of environmental demand. The actions of glucocorticoids have been proposed to follow an inverted U-shape, such that high or low levels of glucocorticoids can impair systems performance (de Kloet et al., 1999; Herman, 2013). This is especially true for processes dependent on the hippocampus (Joëls, 2006). For example, a high dose of corticosterone or adrenalectomy increases behavioral reactivity to novel versus familiar objects, whereas a low dose of corticosterone normalizes this behavior (Oitzl et al., 1994). Glucocorticoid secretion must also be context appropriate. Hypersecreting glucocorticoids under conditions of low demand can be detrimental, but so can hyposecretion in the face of high demand. Therefore, the same concentration of glucocorticoids could be considered deleterious or beneficial depending on the energetic status of the individual. For instance, the same dose of exogenous corticosterone improves performance in a water maze task when animals are trained in warm water, but not cold water (Sandi et al., 1997). Thus, an optimal context-specific amount of glucocorticoid secretion is necessary for maintaining optimal performance (de Kloet et al., 1999; Herman, 2013; Joëls, 2006).

Open in a separate window

Figure 1

Stress, energetics, and adaptation. The relationship between stress and energetic can be explained visually by examining the relationship between systems performance and environmental demand. Under conditions of low environmental demand, high levels of systems performance can be achieved without an energetic cost to the individual. With increasing demand, maintaining systems performance requires energetic input to generate an adaptive capacity (indicated by green shaded area), which is opposed by the adaptive cost (yellow shaded area). Without the acquired limitations that generate the adaptive cost, organismal performance could hypothetically be maintained at high levels across a wide range of environmental demand (depicted by the yellow line). Conversely, the absence of energy input would lead to poor performance even under conditions of moderate environmental pressure (depicted by the green line). The actual response (depicted by the blue line) falls between these two theoretical extremes and is determined for each individual as a function of their adaptive capacity and adaptive cost. With optimal energy redistribution and appropriate appraisal of context, the organism is able to maintain systems performance across a wide range of demand. As the adaptive cost of the organism grows, greater energy input will be required to generate systems performance over a more narrow range of demand. Ultimately, the adaptive cost may be greater than the adaptive capacity, leading to a breaking point where systems performance is compromised and risk factors for pathology emerge.

Energy regulatory systems must work in concert to meet the needs of the individual and use overlapping mechanisms to control and fine-tune energy reallocation (multiple feedback sites and pathways). As a result, the individual is able to maintain adaptive capacity across a broad range of environmental demand. For example, the behavioral and physiological changes accompanying responses to chronic stress are typically considered maladaptive. However, if these changes are interpreted based on the environmental context of the individual experiencing chronic stress, many responses could be considered adaptive for organismal survival. Chronic stress can be a time of high energetic need and glucocorticoids activate glycogen metabolism and glucose synthesis. Glucocorticoids also inhibit energy utilization by systems that are not necessary for immediate survival, including growth, immune, and reproductive processes (Munck et al., 1984). These steps in energy redistribution and reallocation are essential for increasing the adaptive capacity of the individual. In terms of behavior, the effects of chronic glucocorticoids on measures of anxiety, depression, and behavioral flexibility may be deleterious in some environmental contexts; however, in the face of chronic stress, diminished behavioral output could be considered beneficial. For instance, the neophobia and behavioral withdrawal assessed by many tests of anxiety- and depression-like behavior may serve to minimize risk exposure and energy expenditure in contexts of prolonged challenge. Further, a switch to habitual strategies under chronic stress may lead to greater efficiency in predictable tasks by allowing the individual to respond instinctually and bypass the appraisal process (Dias-Ferreira et al., 2009; Schwabe et al., 2013). Thus, limiting risky or energy-demanding behaviors and energy-intensive physiologic processes (e.g. growth, reproduction, immune responses, etc) represent major components of the organismal adaptive capacity. However, when these responses are inappropriate to environmental demand or are sustained of prolonged time periods, the individual generates an adaptive cost that can decrease overall performance.

Glucocorticoid responses promote survival during periods of environmental demand but, over longer time frames, glucocorticoid secretion can impair behavioral and physiological health. There is a certain cost of adaptation that is likely dependent on how the individual's genetic background and life history interact with their appraisal of environmental demand. Thus, inappropriate regulatory balance in stress systems may comprise a risk factor for pathology (increasing adaptive cost). Adaptive cost may be further increased by energy depletion due to overdrive of energetic systems, physiological energy ‘sinks’ (e.g. infection, inflammation, metabolic disease, etc), design faults (inefficient energy redistribution systems, genetic polymorphisms), or poor decision making (misinterpretation of energetic need). As mentioned previously, habitual strategies are beneficial under chronic stress when the task and environment remain the same. However, if the environment changes to a context that requires appraisal and flexible strategies, individuals that have adopted habitual strategies may be at a disadvantage. For example, a recent study found that chronically stressed humans are unable to appropriately appraise outcome values. Further, this behavior is associated with a shift from activation of associative areas to increased activation of sensorimotor areas, mPFC and caudate atrophy, and increased putamen volume (Soares et al., 2013). In this case, misappraisal and energy depletion likely increase the adaptive cost. Depression and PTSD are stress-related disorders characterized by cognitive, affective, and physiological responses that are inappropriate to environmental context. Thus, psychopathologies may emerge when adaptive costs outweigh adaptive capacity, leading to a ‘break point’ in the process of adaptation.

6. Conclusions

A substantial and evolving body of work indicates that glucocorticoids are vital for central stress integration. Accordingly, we propose that the actions of glucocorticoids on neural systems from synapses to circuits generate the energetic resources to promote behavioral stress adaptation. These actions also illustrate the importance of appropriate context appraisal for successful adaptive responses. Misappraisal and/or inadequate adaptive capacity increase the risk for stress-related disorders, including depression and PTSD. Future studies addressing the mechanisms by which glucocorticoid signaling is integrated across multiple brain regions and peripheral sites will improve our understanding of the etiology of stress-related illness. In order to meet this challenge we must consider the sensitivity of current models and endpoints for determining the transition from adaptive to deleterious stress responses, as well as ways to increase translation of these findings into improved clinical outcomes.

Acknowledgements

References