The Neurocircuitry of PTSD and Major Depression: Insights into Overlapping and Distinct Circuit Dysfunction - A Tribute to Ron Duman

Abstract

The neurocircuitry that contributes to the pathophysiology of posttraumatic stress disorder (PTSD) and major depressive disorder (MDD), psychiatric conditions that exhibit a high degree of comorbidity, likely involves both overlapping and unique structural and functional changes within multiple limbic brain regions. In this review, we discuss neurobiological alterations that are associated with PTSD and MDD, and highlight both similarities and differences that may exist between these disorders to argue for the existence of a shared neurobiology. We highlight the key contributions based on preclinical studies, emerging from the late Professor Ronald Duman’s research, that have shaped our understanding of the neurocircuitry that contributes to both the etiopathology and treatment of MDD and PTSD.

Introduction

Substantial comorbidity exists between post-traumatic stress disorder (PTSD) and major depressive disorder (MDD) (1,2). The diagnosis of comorbid MDD in PTSD patients is correlated with both a higher degree of disease burden and treatment refractoriness, leading to negative prognosis (3,4). Both PTSD and MDD are debilitating psychiatric disorders that are linked to structural and functional changes in key limbic brain regions (2,5,6), and can be exacerbated/precipitated by exposure to traumatic stress (7). While the prevalence of MDD is almost twice that of PTSD, exposure to stress and early adversity are shared risk factors that enhance vulnerability for both these disorders (7–10). Consistent with the role of stress in precipitating the onset of these diseases, it is notable that virtually every preclinical model of depression (11,12) and PTSD (13–16) includes a psychogenic and/or physical stressor. In this regard, work from Professor Ronald Duman’s laboratory across the past two decades, capitalizing on animal models of chronic stress, has played a seminal role in providing evidence of alterations in key limbic neural circuits. Research from the Duman laboratory has also shaped the working hypothesis that both conventional and fast-acting antidepressants exert their therapeutic actions via a reversal of these alterations in neural circuits that are central to the pathophysiology of MDD and PTSD. We provide a brief overview of the neurobiological phenomena that are associated with PTSD and MDD, and argue for the existence of a shared neurobiology, with the goal to highlight the key contributions of the Duman group to this area of research. Notably, Ron Duman not only published several seminal papers on the neurobiology of depression during his prolific career, but also later in his career contributed to multiple high impact findings from studies on PTSD.

Neural circuits implicated in PTSD and MDD

We focus on key cortical and subcortical circuitry implicated in the etiopathogenesis of PTSD and MDD. This is by no means an exhaustive summary of all brain regions implicated in the neurobiology of these disorders, rather we focus on brain areas that were extensively studied in the Duman laboratory, in predominantly preclinical models for MDD and PTSD, as the purpose of this review is to place in context the critical contributions made by late Professor Duman.

Role of the Ventral Tegmental Area – Nucleus Accumbens circuit

The nucleus accumbens (NAc) is a key nodal brain region for processing rewarding stimuli and integrates inputs from multiple brain areas including the PFC, ventral hippocampus, thalamus, amygdala and ventral tegmental area (VTA) (17,18). Impaired reward processing due to circuit dysfunction in the VTA-NAc pathway is thought to mediate anhedonia, diminished motivational approach behavior and dysphoria; all of which are core dimensional features for MDD (19–22), and are also associated with PTSD (23,24). This suggests that dysfunctional reward processing may be a neurobiological substrate common to both MDD and PTSD.

The Duman laboratory demonstrated robust increases in dynorphin levels within the NAc in diverse animal models of stress (25), and reported that antagonizing the actions of dynorphin in this circuit exerts robust antidepressant-like behavioral effects (25,26). Work from the Duman, Nestler and Carlezon labs, provided the first lines of evidence that altered dynorphin function may be a key driver of stress-evoked dysphoria (27). The role of cyclic-AMP responsive element, (CRE) binding protein (CREB), deltaFosB, brain derived neurotrophic factor (BDNF), dynorphin and other molecular players that modulate VTA-NAc function in animal models of depression, were hypotheses that emerged from work that was first seeded in the joint Nestler-Duman “Laboratory of Molecular Psychiatry” within the Connecticut Mental Health Centre at Yale University (25,26,28–33). The emerging idea indicated that enhanced CREB-BDNF signaling, and expression of CREB target genes like dynorphin in this brain region (34–36), would serve to modulate behavioral responses to sustained stress driving anhedonic/dysphoric responses, hallmark endophenotypes of both MDD and PTSD. One of the interesting translational leads to emerge from this work is the notion that kappa-opioid receptor antagonists, which counteract the actions of dynorphin at the kappa-opioid receptor, may serve as effective antidepressants and be relevant therapeutic targets for both MDD and PTSD (37).

It is important to note that the nature, severity, duration and frequency of the stressor can also play a major role in determining the influence on the VTA-NAc pathway. In animals subjected to chronic mild stress (CMS), optogenetic phasic stimulation of VTA-NAc projections resulted in reduced despair-like behavior and decreased anhedonia on the sucrose preference test, evocative of antidepressant-like behavioral effects, and inhibition of this pathway promoted a depressive phenotype (38). The contextual nature of the effects of VTA-NAc pathway in determining susceptibility or resilience highlights both the differential responses based on nature of stressors, as well as the heterogeneity within the neuronal populations in this circuit, and the distinct neuronal firing patterns that could mediate very different behavioral outcomes. Clinical studies suggest that anhedonia may be a common risk factor for several psychiatric disorders including MDD and PTSD (24,39). Depressed patients exhibit reduced NAc activation in response to reward-associated stimuli (19), along with evidence of lower volume in imaging studies (40). Functional neuroimaging studies also demonstrate lower reward-related activity in the NAc in PTSD patients (23). Independent of the category of psychiatric diagnosis, a reduction in reward responses is associated with a decline in NAc connectivity to the default mode network brain regions implicated in self-focus and introspection (41). This raises the possibility that exposure to stress and trauma could shift the bias of reward processing circuits via enhanced functional connectivity to default mode networks to self-focused thinking, and thus impair the processing of natural rewards and impact motivational behavior. It is of interest to note that bilateral deep brain stimulation of the NAc in treatment refractory MDD patients can ameliorate anhedonia and evoke strong antidepressant effects (42,43). The use of viral tracing approaches, and chemogenetics/optogenetics tools combined with cell-type specific genetic driver lines opens up the possibility of functionally interrogating the role of specific VTA and NAc neuronal subpopulations with distinct efferent-afferent connectivity. See Supplemental File 1 for a discussion on the use of the chronic social defeat stress (CSDS) model to investigate the role of the VTA-NAc pathway in MDD and PTSD.

Role of the HPA Axis

The hypothalamic-pituitary-adrenal (HPA) is a major endocrine system that regulates the body’s response to stress. See Supplemental File 1 for an overview of the HPA axis. Both depression and PTSD are associated with altered HPA axis activity; however, the nature of these differences vary. Studies have reported that PTSD is associated with lower basal levels of cortisol (44,45); however, a meta-analysis of 37 studies reported that blood cortisol levels were not significantly different in PTSD patients compared to controls (46). Cortisol levels can be highly variable due to differences in the time of sample collection, type of sample collected, tobacco and alcohol use and biological sex. Other studies indicate that if individuals exhibit low cortisol levels following acute trauma, they have a greater risk of developing PTSD (47,48). Consistent with the notion that low levels of cortisol could be a risk factor for developing PTSD, acute trauma patients who received hydrocortisone shortly following trauma reported fewer PTSD and depression symptoms during the three months after the trauma compared to placebo controls (49,50). One notable finding is that PTSD patients have been found to be hyper-responsive to glucocorticoid negative feedback in dexamethasone suppression tests (51,52). This phenomenon has been attributed to glucocorticoid receptor (GR) hypersensitivity, potentially due to adaptations as a result of chronically low cortisol levels (53). In contrast to PTSD, MDD has generally been associated with HPA axis hyper-responsiveness and hypersecretion of cortisol. Dexamethasone administration in MDD patients usually leads to an attenuated suppression of cortisol levels compared to controls (54), however, the stage of disease progression may matter as chronic MDD patients have been found not to exhibit significant alterations in functioning of their HPA axis (55).

Both PTSD and MDD have been associated with corticotrophin-releasing hormone (CRH) overactivity, but the data have not always been consistent and are complicated by heterogeneity within the depressive patient population. However, there is a vast amount of evidence that for PTSD and melancholic depressive patients, CRH activity is elevated (56,57). It has been reported that combat veteran PTSD patients exhibit increased basal CSF CRH levels (58,59). Furthermore, studies from post-mortem hypothalamic tissue from depressed patients have found an increase in the number of CRF containing neurons (60). Consistent with these findings, antidepressants including fluoxetine, have been shown to reduce CRH synthesis in the rat hypothalamus and this reduction in expression is associated with alleviation of depressive-like behavior (61,62).

The FK506-binding protein 51 (FKBP51) protein has a key role in regulating GR mediated gene expression. Interestingly, genetic studies demonstrate that an association exists between single nucleotide polymorphisms (SNPs) that occur in the FKBP51 encoding gene FKBP5 and PTSD (63–65). Four SNP risk alleles in the FKBP5 gene have been identified (rs9470080, rs360780, rs3800373, and rs9296158) that are predictors of adult PTSD onset following childhood trauma. Remarkably these SNPs have been associated with antidepressant response and stress induced onset of major depression (66). Work from the Duman laboratory found that the expression of FKBP5 was decreased in human postmortem tissue from PTSD patients, further implicating aberrant glucocorticoid functioning in PTSD (67).

Role of the Hippocampus

The hippocampus is a temporal lobe structure necessary for the consolidation of contextual and spatial memories (68). In addition, this structure is an important locus for regulating HPA axis activity(69). Abnormalities in hippocampal function have been associated with both PTSD and MDD.

A consistent finding for MDD is an association with alterations in hippocampal volume. Numerous meta-analyses conclude that hippocampal volume is reduced in patients with MDD compared to healthy controls (70–74). Additionally, longitudinal studies indicate that antidepressant treatment leads to an increase in hippocampal volumes (75). However there continues to be debate regarding whether the reduction in hippocampal volume is a cause of the depressive symptoms, or a consequence of depression. However, these possibilities are not entirely mutually exclusive. Although a body of work indicates that reductions in hippocampal volumes could be a risk factor for developing depression (76–79), a meta-analysis found that there are also findings indicating that hippocampal volume is only reduced in depression patients who have been depressed for at least two years and exhibit more than one episode (72). Animal models have shown that hippocampal size is reduced following acute and chronic stress in rodents (80), consistent with the notion that stress leads to hippocampal size reduction. Remarkably, recent findings from rodents indicate that while stress leads to hippocampal reductions, hippocampal reduction is not necessarily sufficient to promote depressive symptoms (81,82), suggesting that there might be other contributing genetic or environmental factors that lead to the development of depression.

Similarly, PTSD patients have been found to possess smaller hippocampi (83–86). They also exhibit impaired performance on hippocampal-dependent tasks (87), consistent with the notion that the hippocampus is compromised. It is noteworthy to mention that a study on monozygotic twins found that reduced hippocampal volume precedes the occurrence of the traumatic event and the onset of PTSD (84,88), indicating that a smaller hippocampus might be a risk factor for developing PTSD. The hippocampus is unique, in that it is one of the few places in which neurogenesis occurs in the adult brain. Specifically, new neurons are created in the subgranular zone of the dentate gyrus, where stem cells develop into mature neurons that integrate into the existing circuitry (89). Hippocampal neurogenesis has been shown to increase following exercise (90), environmental enrichment (91), antidepressant treatment (92) and hippocampal dependent learning (93). Conversely stress (94), stress hormones (95), and drugs of abuse (96), have all been shown to reduce hippocampal neurogenesis.

Dr. Ronald Duman and colleagues made an important observation that led to the development of the neurotrophic hypothesis of depression (97,98). This hypothesis states that depression occurs due to a lack of trophic support leading to the structural and functional changes within the brain that are associated with depression. This is based on the fact that stress and stress hormones lead to reductions in neurogenesis and increases in neuronal loss and atrophy and MDD is associated with reduction of hippocampal volume. Further supporting this hypothesis is the fact that antidepressant treatment leads to an increase in neurotrophin expression such as brain-derived neurotrophic factor (BDNF) (99). Antidepressant-induced neurotrophin expression is believed to lead to a reversal of the structural and functional changes within the brain of MDD patients leading to increases in hippocampal neurogenesis (92), dendritic branching (100), an increase in mossy fiber sprouting (101) and spinogenesis (102).

Given the negative influence of stress on hippocampal function, it is not surprising that hippocampal neurogenesis or lack of, likely contributes an important role in PTSD too. One of the defining features of PTSD is the over-generalization of fear, where the individual has trouble discriminating between safe and dangerous environments and becomes fearful in both. Generalization of fear is an adaptive mechanism to protect oneself, but when exaggerated like in PTSD, it is maladaptive. One important function of the hippocampus and the dentate gyrus specifically is in pattern separation. This is the ability to effectively discern between similar and different environments. When pattern separation is impaired, generalization occurs (102,103). There is accumulating evidence that impairments in hippocampal neurogenesis lead to deficits in pattern separation and increases in generalization (104–106). Additionally, impairments in neurogenesis are associated with exaggerated stress responses. For example, in rats exposed to stressors that induce PTSD-like symptoms it was found that inhibiting hippocampal neurogenesis was associated with prolonged anxious behavioral and endocrine responses compared to stressed animals that possessed intact neurogenesis (107).

Amongst the critical roles of the hippocampus is the regulation of the HPA axis. When circulating glucocorticoids rise, they will bind to GRs on hippocampal neurons and induce the hippocampus to repress the HPA axis at the level of the hypothalamus (108). Chronic stress can lead to maladaptation of this regulation where the hippocampus becomes impaired in its ability to dampen the HPA axis causing heightened activity of the HPA axis, leading to a breakdown in this negative feedback mechanism and higher levels of glucocorticoids. For example, reduction in GR expression within hippocampal neurons will impair the ability of the hippocampus to effectively respond to changing levels of glucocorticoids. Notably, early life stress in rodents and humans has been associated with epigenetic modifications of the GR gene promoter, resulting in reduced expression of GR. This is associated with heightened anxiety-like behavior due to maternal neglect in rodents (109,110) and suicide due to childhood abuse in humans (111).

Role of the Prefrontal cortex

The prefrontal cortex (PFC) is considered to be an extensively interconnected, “integrative hub” for executive function, emotional processing, attention and decision making, memory, social cognition, and assigning salience to self-related information processing (112–114). The PFC can be broadly segregated into two divisions, namely the ventromedial PFC (vmPFC) thought to be present in all mammals including rodents, and the dorsolateral prefrontal cortex (dlPFC), thought to be restricted to only primates (115). While there still remains controversy over the precise attribution of specific functions to distinct PFC subdivisions, overall the vmPFC has been linked to the regulation of “affective” behaviors, whereas the dlPFC is predominantly associated with “executive” functions (116). Consistent across several studies is the finding of lower PFC volume in both MDD (117) and PTSD (118) patients, thought to arise through neuronal atrophy, synaptic loss, and a reduction in neuronal and glial numbers (119,120). Further, neuroimaging studies indicate dysfunctional activity of the vmPFC and dlPFC in both MDD and PTSD patients (116,121–124), providing compelling evidence that PFC dysfunction could establish a key neural substrate for perturbed cognitive control and associated mood-related abnormalities. Preclinical studies in rodent models spearheaded by the Duman lab indicate that mPFC synaptic atrophy and loss, induced via a disruption of local protein synthesis, can evoke enhanced depressive-behavior (125). Converging evidence supports the notion that exposure to sustained stress impinges on key pathways that regulate synapse plasticity in multiple brain regions, including the PFC (126). Lower PFC volume, altered activity and connectivity, and synaptic loss in the vmPFC and dlPFC have been correlated with the severity of negative affect both in MDD and PTSD patients (6,116,127–129). It still remains a source of debate whether this lower volume in the PFC can be thought of as a cause or a consequence of major depressive episodes (204). In this regard, it is of interest to note that a life-history of early stress is linked to lower PFC volume, and this is noted also in the absence of MDD (130,131), highlighting this as a potential predisposing risk factor. Stress has been considered as one of the key environmental factors that contributes to PFC damage and the establishment of vulnerability for MDD or PTSD (204). Neuroinflammation evoked within PFC circuitry due to stressful life-events could then promote the progression into psychopathology. Imaging studies indicate heightened inflammation within the PFC of depressed patients (132), as well as reports of enhanced pro-inflammatory cytokines in the periphery, and perturbed gene expression of inflammation associated regulatory networks in MDD (133).

Based on both preclinical and clinical studies, it has been postulated that the PFC exerts powerful top-down regulation over mood related behavior (113), via pathways that regulate multiple brain regions, including the amygdala and the raphe. The top-down inhibitory control exerted by the vmPFC on the amygdala is thought to be key to the ability to suppress negative emotion (134–136). Enhanced vmPFC activity is correlated with a decline in neuronal activity in the amygdala, and with a decline in negative affective behavioral states (135,137). This inverse relationship of vmPFC-amygdala activity is thought to be disrupted in both MDD and PTSD (135), which are both characterized by high levels of negative affect. PTSD patients exhibit significant vmPFC hypoactivity linked with amygdala hyperactivity (138), suggesting that the impaired affective state regulation and emotional distress that are characteristic of PTSD patients may arise due to a disrupted top-down inhibitory control of the amygdala due to vmPFC hypoactivity (127,139,140). A working model for the neurocircuitry that could mediate pathogenesis of PTSD posits that enhanced amygdala activity could promote the enhanced fear and amplified startle response that is characteristic of PTSD patients when provoked by trauma-associated cues and contexts. The reduced ability of the PFC to inhibit the amygdala has been suggested to play a nodal role in failing to rein in an overactive amygdala (137,141). While the preponderance of literature supports a role for the vmPFC in the inhibition of negative affective states, it is noteworthy that lesions studies suggest the opposite adding to the complexity of our understanding of the role of the vmPFC in emotion processing. Lesions of the vmPFC in humans are linked to blunting of affect and a decline in vulnerability to MDD and PTSD (142,143). It has been suggested that since the PFC is associated with the processing of self-referential information, psychiatric disorders that are characterized by self-focus or distress associated with past traumatic events may be attenuated due to PFC lesions likely due to a “loss of self-insight or self-reflection that would diminish the core symptoms of the disorder” (137,141,144).

Amongst the abnormal behavioral changes in MDD and PTSD is biased attribution of negative valence to socioaffective stimuli in the environment (145–147). The CSDS model in rodents has been postulated to be an ethologically relevant behavioral model that captures endophenotypes of socioaffective bias, and perturbs social approach-avoidance behavior (148). CSDS results in enhanced vmPFC-evoked inhibition of raphe serotonergic neurons, resulting in decreased 5-HT release in key target brain regions, including the amygdala, thus driving social avoidance behavior in animals susceptible to social defeat (147). Work from the Dzirasa group indicates that heightened PFC-amygdala phase locking is strongly associated with susceptibility to social defeat (149). Accumulating evidence suggests that antidepressant treatments may exert therapeutic effects on mood behavior by reversing the negative bias in socioaffective processing in both MDD and PTSD patients (150,151). This restoration of socioaffective processing is thought to involve neuroplasticity within the PFC (147), such that it promotes the ability to determine positive valence in social and emotional associations, thus facilitating an improvement in mood.

Work from the Duman lab and colleagues indicates that fast-acting antidepressant treatments, such as ketamine and scopolamine, could reverse synaptic atrophy and loss in the mPFC through the recruitment of mechanistic target of rapamycin complex 1 (mTORC1) signaling, and thus serve to normalize PFC control over negative affective states (152). These fast acting antidepressants promote enhanced glutamate release within the PFC and AMPA receptor activation mediated increases in BDNF release and trkB signaling, thus resulting in the recruitment of mTOR and downstream effects on synaptic plasticity and architecture (153). Clinical studies demonstrate that there is a rapid-onset increase in PFC connectivity in response to infusion of ketamine, and this enhanced PFC functionality could then serve to restore the impaired top-down control exerted by the PFC in depressed patients (154,155). There is relatively limited clinical data for the use of ketamine in PTSD patients but the evidence available thus far indicates cautious optimism for the beneficial effects of ketamine in PTSD patients in particular those refractory to other antidepressant treatments (156). The PFC plays a pivotal role in regulating mood-related behavioral states, and PFC dysfunction is implicated to contribute to the pathogenesis of several psychiatric disorders, including MDD and PTSD (116,137). The PFC is also thought to be a key target for deep brain stimulation and transcranial magnetic stimulation-based therapeutics (6). In this regard, several clinical studies indicate that PFC volume, functional connectivity and activity may be particularly relevant in predicting treatment response to antidepressants (114,116,157).

See Supplemental File 1 for a discussion on the role of the amygdala and the dorsal raphe in MDD and PTSD pathology.

Conclusions

The available data clearly indicate the existence of a shared neurobiology between MDD and PTSD (Figure 1). There are many shared features at the levels of gene regulation, neuronal and glial architectural and physiological alterations, and at the level of functional connectivity and activity of neuronal networks and systems between these disorders. As is common with most psychiatric disorders, the heterogeneity within patient populations due to stage of illness, age, existing comorbidities, state and nature of treatment can make it difficult to ascertain the exact neurobiological phenomena that exist in any psychiatric disorder, in addition to differences in clinical manifestations. Despite these confounds, the past several decades of clinical and preclinical research clearly highlight important commonalities in brain structure and dysfunction, and neurohormonal dysregulation between these disorders.

Figure 1:

A summary of the major overlapping changes noted in preclinical models of MDD and PTSD, and in specific cases also observed in clinical studies.

The work emerging from the Duman laboratory has played a critical role in shaping this field and provided the impetus to translate key preclinical observations. Moving forward, we predict the coming decades will substantially refine our understanding, by leveraging a deeper insight of existent genetic and epigenetic variation across patient populations, overlaid with consequences of specific environmental circumstances to provide a mechanistic framework for the shared neurobiology of PTSD and MDD.