Depression and sleep: what has the treatment research revealed and could the HPA axis be a potential mechanism?

Abstract

Research indicates that insomnia improvement plays a critical role in depression symptom improvement. In line with the National Institute of Mental Health Experimental Therapeutics approach recent research focuses on identifying specific mechanisms; the present manuscript aims to review recent research on one potential mechanism, dysfunction in the HPA axis which is a shared biological substrate of both depression and insomnia. Over the past five years, research demonstrated a relationship between sleep disturbance and cortisol reactivity and recovery following a stressor. Meanwhile, research on the relationship between depression and HPA axis functioning is less consistent and is dependent on measurement of HPA axis. Experimental research that aims to determine a causal pathway between sleep, depression and HPA axis functioning is needed.

Depression and sleep have a complex and likely bidirectional relationship. Rates of insomnia among depressed patients are estimated to be 67–84% [1]. Likewise, rates of depression among insomnia patients are equally high [2]. Moreover, patient presentations that include depression comorbid with insomnia tend to include more severe depression symptoms, greater functional impairment and less responsiveness to both psychosocial and psychopharmacological interventions [3–8]. Moreover, insomnia is among the most common residual symptoms following even successful treatment for depression [9, 10]. Insomnia is also a significant predictor of both new [11] and recurrent depressive episodes [12].

The natural conclusion is to address both conditions with evidence based psychotherapy and psychopharmacology in a variety of combinations. Cognitive Behavioral Therapy for Insomnia (CBT-I) is the first line recommendation for treatment of insomnia [13]. Several recent reviews have outlined the efficacy and effectiveness of CBT-I for use among samples with comorbid depression and insomnia[14^••,15^••] As reviewed in Asarnow and Manber [14^••] some studies take a sequential approach, where providers treat one disorder at a time, and others take a concomitant approach where both disorders are treated simultaneously. While, randomized controlled studies clearly indicate that CBT-I effectively alleviates insomnia symptoms in adults and adolescents with comorbid depression [16,17^••], the results regarding depression treatment outcomes are more mixed. Results from studies that assessed the sequential approach (treating either insomnia before depression treatment or treating residual insomnia symptoms following depression treatment) showed promising results with regards to depression treatment outcome; however, more adequately powered randomized control trials with long-term follow-up are needed to evaluate how sequential strategies impact the course of depression over time [18,19^•]. There are only a handful of studies assessing whether a concomitant treatment of depression and insomnia results in improved depression outcomes and the results are mixed [16,17^••,20^••,21^••]. As noted in Cunningham and Shapiro [15^••], it is possible that the effects of CBT-I and antidepressant medications may not be additive. A recent finding that response to insomnia treatment mediated eventual remission from depression using the concomitant approach suggests that a focused effort to improve insomnia has the potential to enhance depression treatment outcomes [17^••]. This finding in combination with the sequential treatment literature demonstrates that, even while randomized control trials that augment treatment for depression with CBT-I may not always find depression symptom reduction to be significantly reduced by the addition of CBT-I, insomnia improvement still appears to play a critical role in depression symptom improvement.

In line with the National Institute of Mental Health Experimental Therapeutics approach new research is and has focused on identifying specific mechanisms that may be key components to understanding this complex bidirectional relationship between depression and disturbed sleep (i.e. insomnia). There are many possible biological mechanisms/processes that cut across both depression and disturbed sleep. The present manuscript aims to review one potentially important mechanism; dysfunction in the interaction between the hypothalamus, pituitary gland, and adrenal glands, referred to as the HPA axis. The HPA axis is a shared biological substrate of both depression and disturbed sleep. The research reviewed in this manuscript highlights some of the progress made over the past five years that elucidates this complex relationship between the HPA axis, depression and sleep.

HPA axis in response to a stressor

The HPA axis represents one of the body’s mechanisms for responding to acute stress. In response to a physiological or psychological stressor, the HPA axis is activated and a hormonal cascade is initiated starting with the secretion of corticotropin-releasing hormone (CRH) from the hypothalamus, which then stimulates the anterior pituitary gland to release adrenocorticotropic hormone (ACTH). ACTH then stimulates release of cortisol from the adrenal gland, resulting in a variety of physiological responses from the sympathetic nervous system (e.g. the release of adrenaline, increased heart rate, and increased blood pressure). Once the stressor has resolved, the response is terminated through a negative feedback loop, in which cortisol levels fall and suppress further release of ACTH and CRH. This capacity to adapt to a changing environment or stressful challenge is sometimes referred to as ‘allostasis’ and is critical to survival. While highly adaptive in the short-term, chronic exposure to stress can lead to wear and tear of the body characterized by dysregulated physiological responses to stress [22]. Cortisol, the end product of the HPA axis, serves a variety of crucial roles in promoting allostasis, including mediating and suppressing healthy stress responses [23].

In addition to cortisol’s important role in the stress response system, cortisol from the adrenal gland is also released in a classic circadian pattern. Cortisol production is normally at its highest 30–45 min after waking and declines steadily during the day, reaching its lowest point at bedtime [24]. The body’s central pacemaker, the suprachiasmatic nucleus (SCN), is responsible for the overall co-ordination of the HPA axis and synchronizing the time of day and neuroendocrine output, thereby regulating the diurnal pattern of cortisol [25]. HPA-axis regulation can thereby be measured in several distinct ways using cortisol measurements; 1) naturalistically by assessing diurnal cortisol release or 2) the cortisol awakening response, and 3) in a laboratory paradigm where cortisol reactivity is measured before and in response to an experimentally induced stressor [26].

Sleep and the HPA axis

While, sleep problems have long been known to be associated with higher levels of stress measured behaviorally [27], more recent research has investigated the relationship between disturbed sleep and biological markers of stress. This body of research has identified a hyper-reactive cortisol in response to a laboratory stress exposure under conditions of sleep disturbance across distinct populations.

Among healthy adults, two experimental studies over the past five years have established a potential causal relationship between disturbed sleep and HPA axis dysregulation. Indeed, Minkel et al. [28^••] conducted a study in which 26 adults who were good sleepers completed one night of baseline sleep (with a 9 hour sleep opportunity) followed by randomization to either one night of total sleep deprivation or a normal-sleep control condition (with a 9 hour sleep opportunity). After the experimental sleep night, participants completed a laboratory stressor accompanied by multiple saliva samples before and after the stressor to be assayed for cortisol. Results indicated that sleep deprivation was associated with higher cortisol levels at baseline and an amplified cortisol response to the stressor relative to control participants. In the second experimental study, conducted by Massar et al. [29^••], the sleep habits of 59 adult male participants were monitored via actigraphy for one week followed by exposure to a laboratory stressor. Salivary cortisol was measured before and after the stress exposure. Low sleep efficiency the week before the stress exposure was associated with exaggerated cortisol reactivity in response to the stressor. Taken together, these recent experimental studies in adults suggest that both experimentally induced and naturalistically observed poor sleep adversely affects the body’s stress response system as measured by cortisol reactivity and recovery following exposure to a stressor.

While the research among adolescents on the relationship between sleep and the HPA axis has lagged behind the adult literature, in the previous five years there have been several significant contributions. Mrug et al. [30^••], studied the sleep and stress reactivity of 84 urban dwelling adolescents. The study found that more sleep problems and longer sleep duration measured using self-report predicted higher cortisol reactivity to a laboratory stressor and blunted recovery following the stressor. Interestingly, in this sample girls experienced greater cortisol elevation in response to stress than boys, as well as stronger effects of sleep problems on cortisol reactivity. In a naturalistic study conducted among a sample of 316 adolescents from the community, Chiang et al. [31^••] investigated the relationship between diurnal cortisol and sleep in the context of a real life stressor (family conflict). The researchers found that among adolescents with lower sleep efficiency and/or longer sleep onset latency, family stress was related to a blunted cortisol awakening response (CAR). Both the Mrug and Chiang studies suggest a relationship between naturally observed disturbed sleep (measured using self-report and objective sleep measures) and greater dysregulation of the HPA Axis both in response to an experimentally induced stressor and as a consequence of a naturally occurring stressor (family conflict) among adolescents. These studies are particularly important in the greater context of adolescence. A critical developmental period in brain development, the healthy adolescent brain is characterized by heightened emotional reactivity (HPA axis reactivity to a stressor) combined with underdeveloped behavioral control (prefrontal cortex) in comparison to adults, resulting from asynchronous development of the neural circuits involved in the response to stress. Thus, when experiencing sleep disturbances, adolescents who are already developmentally vulnerable to be emotionally reactive, may experience greater dysregulation in their responses to stress (heightened HPA axis responsivity to a stressor). This pattern is thought to be a pathway that may increase adolescents’ risk for depression.

Depression and the HPA axis

Similarly, cortisol hyper-reactivity and blunted recovery in response to a stress exposure has been documented among those with diagnosed depression [32]. Increased cortisol reactivity in stressful situations and blunted recovery following a stressor may be a biological substrate of depression, leading to heightened emotional reactivity and difficulties in stressful situations which manifests as poor coping response to stressors and may lead to increased depression symptoms over time [33,34^•]. However, chronic exposure to cortisol can lead to structural changes in the brain regions responsible for modulating the stress response (e.g. the hippocampus) and may contribute to the pathophysiology of depression [35–37]. Indeed, meta-analytic studies have indicated that major depressive disorder is associated with blunted cortisol reactivity to laboratory stressors in adults and youth [32,38].

In the past five years, research on the relationship between depression and dysregulation of the HPA axis has broadened to investigate sex differences, diurnal cortisol measurement and the cortisol awakening response. Powers et al. [39^••] examined sex differences in cortisol reactivity and responsivity to a stressor among 196 heterosexual dating couples. Depression and anxiety symptoms and diagnoses were assessed with questionnaires and diagnostic interviews, and cortisol reactivity and responsivity was measured in response to a discussion of an unresolved relationship conflict (a stressor). The researchers found that among women, depressive symptoms predicted attenuated cortisol levels, with a flatter response curve. In contrast, men’s depression symptoms and women’s anxiety symptoms predicted higher cortisol levels overall. These findings highlight the importance of examining sex differences in responses to interpersonal stressors for understanding HPA dysregulation in internalizing psychopathology as well as other psychiatric comorbidities.

One meta-analysis and one review of the relationship between CAR and depression were published within the past five years. In 2015, Dedovic and Ngiam [40^••] conducted a review of the key findings related to CAR and depression. While the review was not exhaustive, the authors observed that some studies observed heightened CAR and others blunted CAR in relation to depression. The authors concluded that while CAR is highly heritable, it is also greatly influenced by environmental factors and therefore it might not mediate or moderate the association between the number or severity of objective major stressors and depression, but rather, it might be more a measurement of daily hassles and the degree that one can cope with a stressor (major or minor). A meta-analysis of 212 articles examining psychosocial functioning and CAR conducted by Boggero et al. [41^••] helps to provide some clarity on the discrepancy between studies. Boggero concluded that depression was positively associated with area under the curve relative to the waking period (AUCw), but not with dynamic increases in cortisol that occur during the first waking hour (CAR). Importantly, AUCw is not strictly a measure of CAR because it is influenced by cortisol levels before awakening and not entirely by dynamic increases in cortisol that happen post-wakening. The authors indicate that future studies investigating the CAR and psychosocial functioning should sample the CAR over more days and require larger sample sizes to detect effects with 80% power.

Future directions

Across research that has investigated the relationship between sleep and the HPA Axis findings point to a relationship between both experimentally induced and naturally occurring disturbed sleep and cortisol reactivity and recovery following exposure to a stressor. These findings are consistent with older experimental research among adults with insomnia. For example, Vgontzas et al. [42], in a sample of insomniacs as compared to good sleepers, found an overall increase in cortisol secretion over a 24 hour period in a laboratory setting, which appeared to be primarily caused by a higher number of cortisol pulses throughout the day; these findings suggest insomniacs have sustained arousal and activation of their stress system. This chronic activation of the HPA axis system may be the mechanism by which those with disturbed sleep become particularly vulnerable to stressors and may thereby represent a pathway for risk of depression. Another potential future direction for researchers is to investigate the relationship between cortisol reactivity and recovery in sleep deprivation versus insomnia to gain a deeper understanding of the HPA axis as a potential mechanism.

While literature investigating the relationship between depression and cortisol reactivity and recovery following a stressor has been more consistent [32,43], most of the literature reviewed from the past five years focused on the CAR. The CAR is complex because it interacts with the circadian system. For example, Wright et al. [44^••] compared the effects of acute total sleep deprivation and chronic circadian misalignment on cortisol levels among healthy, drug free adults. They found altered timing of the cortisol rhythm relative to sleep-wakefulness timing for those with circadian misalignment. Importantly, among the studies investigating the relationship between CAR and depression, the influence of sleep and circadian phase on cortisol levels was not considered in the analyses. This is of critical importance given that high rates of insomnia, delayed circadian phase and circadian misalignment are common among patients with depression.

The literature to date comes from two distinct fields, depression and sleep. There were no studies identified that looked at the relationship between the HPA axis, sleep and depression among samples who are both depressed and have insomnia or are at risk for both. Taken together, the literature points to a need to investigate HPA Axis dysregulation as a potential biological pathway in the development of disturbed sleep and depression. Experimental research that aims to determine a causal pathway between sleep, depression and HPA axis functioning is needed.