Abstract
Chronic stress accounts for billions of dollars of economic loss annually in the United States alone, and is recognized as a major source of disability and mortality worldwide. Robust evidence suggests that chronic stress plays a significant role in the onset of severe and impairing psychiatric conditions, including major depressive disorder, bipolar disorder, and posttraumatic stress disorder. Application of molecular imaging techniques such as positron emission tomography and single photon emission computed tomography in recent years has begun to provide insight into the molecular mechanisms by which chronic stress confers risk for these disorders. The present paper provides a comprehensive review and synthesis of all positron emission tomography and single photon emission computed tomography imaging publications focused on the examination of molecular targets in individuals with major depressive disorder, posttraumatic stress disorder, or bipolar disorder to date. Critical discussion of discrepant findings and broad strengths and weaknesses of the current body of literature is provided. Recommended future directions for the field of molecular imaging to further elucidate the neurobiological substrates of chronic stress-related disorders are also discussed. This article is part of the inaugural issue for the journal focused on various aspects of chronic stress.
Keywords
Introduction
The experience of chronic psychological stress is associated with a variety of serious physical, financial, and emotional consequences at both individual and societal levels. 1 The World Health Organization has labeled stress as a “worldwide epidemic” in recognition of the magnitude of its deleterious effects. 2 The American Institute of Stress estimates that the effects of chronic stress cost US companies over 300 billion dollars annually. 3 More gravely, stress and resultant changes in affect are associated with increased morbidity and mortality, 4 including increased rates of mental illness and suicide. 5 Robust evidence has linked the experience of chronic stress to onset of major depressive disorder (MDD), which is the leading cause of disability worldwide.1,6,7 Likewise, onset of bipolar (BD) and posttraumatic stress (PTSD) disorders, both of which are associated with substantial emotional, physical, and financial burden,8,9 has been linked to chronic stress. Unfortunately, the molecular mechanisms by which chronic stress confers risk for mental health problems are poorly understood, hindering the development of critical interventions to resolve and prevent chronic stress-related disorders.
Application of Positron Emission Tomography and Single Photon Emission
Computed tomography (SPECT) has begun to provide information concerning how the brain changes following chronic stress on a molecular level. Positron emission tomography (PET) and SPECT molecular imaging techniques allow for the quantification of biochemical processes, including those known to play a pivotal role in the pathophysiology of psychiatric disorders in the living human brain. 10 Both techniques rely on the injection of a “radiotracer,” a physiologically active compound with affinity for a specific target, which has been tagged with a small amount of radioactive isotope. This review is part of the journal’s inaugural issue, which is designed to comprehensively describe the neurobiological and neuroclinical effects of chronic stress. The role of this review is to summarize the existing molecular imaging work in psychiatric conditions known to be associated with chronic stress—MDD, BD, and PTSD—and recommend directions for future research.
Methods
Separate searches of the PubMed database were conducted for each disorder. In each case, the following standard prompts were used: “positron emission tomography” OR “PET” OR“single photon emission computed tomography” OR “SPECT” OR “single photon emission tomography” OR “SPET”). Distinct specifiers were then added for each of the three disorders: AND (“depression” OR “MDD”), AND (“bipolar disorder” OR “mania”), AND (“posttraumatic stress disorder” OR “PTSD”). Searches for MDD, BD, and PTSD initially yielded 2, 302, 447, and 158 PubMed article results, respectively. Studies were included in this review if they satisfied the following criteria: (1) articles focused on the use of PET or SPECT techniques, (2) articles describing original research, (3) articles presenting in vivo results in human participants, and (4) articles focused on the imaging of specific molecule/receptor types (i.e. excluding glucose metabolism and cerebral blood flow). Following careful review, 126 MDD, 22 BD, and 15 PTSD articles were retained.
Results
Major Depressive Disorder
Radiotracers used for in vivo molecular imaging in individuals with major depressive disorder.
Note. MDD, major depressive disorder; PET, positron emission tomography; MAO, monoamine oxidase; HC, healthy control.
The majority of existing PET and SPECT research in MDD populations has focused on the functioning of two monoaminergic neurotransmitters: serotonin and dopamine. Focus on monoamine functioning in depression resulted in part from the monoaminergic hypothesis, 97 which postulates that alterations in the function of monoaminergic neurotransmission play a causal role in the development of MDD.
Serotonin
Between 1991 and 2016, 71 molecular imaging studies meeting the criteria specified above were published. Strong evidence supports the role of the serotonergic (5HT) system, which plays a role in the regulation of sleep, stress responses, and affective cognition 98 in the development and continuation of depressive symptoms. 99 However, the exact nature of the relationship between 5HT functioning and MDD is still under debate. Both preclinical and postmortem evidence suggests that serotonergic dysfunction and specifically deficits in serotonin are central to the pathophysiology of MDD.100,101 In line with such findings, evidence from PET work suggests lower serotonin synthesis in MDD individuals relative to healthy control,15–19 and that treatment with selective serotonin reuptake inhibitor (SSRI) resulted in increased serotonin synthesis. 16
The serotonin transporter (SERT) plays a key role in modulation of brain 5HT levels via reuptake into presynaptic neurons, 98 and is the primary target of action for many commonly prescribed antidepressant medications. 102 Interest in SERT has been motivated in part by evidence suggesting that expression of the SERT gene (5HTT) may moderate emotional and behavioral responses to stress, 103 such that individuals displaying a specific 5HTT polymorphism (associated with decreased serotonergic functioning in preclinical studies103,104) were more likely to develop MDD and suicidal behavior following stress. 105 Evidence from a recent SPECT study similarly found lower SERT availability in the thalamus and that high levels of life stress interacted to predict depressive symptom severity, suggesting that SERT may play a specific role in the development of depression following chronic stress. 106 Miller et al. 25 observed lower SERT availability across brain regions in individuals reporting a history of child abuse who went on to develop MDD, but not PTSD, relative to MDD individuals without an abuse history. Further, support for the relationship between SERT and stress comes from an experimental study of HPA axis dysfunction, as measured by the dexamethasone suppression test, 107 showing an association with reduction in SERT levels. Evidence points to lower SERT availability in multiple brain areas, most commonly the thalamus,20,29 and midbrain21,22,30,31,67 or brainstem,63,69 areas responsible for regulation of sleep/wake cycles which are frequently disrupted in MDD.
Importantly, SERT levels appear to be related to antidepressant treatment outcome and illness progression.32,108,109,110,111 Increase in midbrain SERT availability is associated with depressive symptom remission following antidepressant treatment,109,111,112 and Amsterdam et al. 110 showed a significant increase in both midbrain and medial temporal lobe SERT availability in treatment responders following 12 weeks of cognitive behavioral therapy. Conversely, in participants whose MDD symptoms did not remit following a year of antidepressant use, lower SERT availability was detected in several brain regions in individuals with MDD relative to healthy controls. 27
Of note, in vivo evidence for low SERT in MDD is strong but not universal. Some studies have reported no significant differences in SERT availability between MDD and healthy control groups,23,28,64,113,114 while others reported higher SERT availability in individuals with MDD relative to controls.65,113,115 No clear explanation for these discrepant findings is available, though low sample size due to cost in PET and SPECT studies, difficulty in recruiting these patients, and cross-sample variability (e.g. symptom severity, sex, race, medication) likely play a role. Notably, characteristics of some specific radiotracers may have biased findings in some cases. For example, both SPECT tracers, [123I]nor-β-CIT and [123I]β-CIT, have been shown to bind not only to SERT but to norepinephrine (NET) and dopamine (DAT) transporters. 116 Thus, depending on the region, quantification of SERT might represent SERT and DAT/NET densities.
A large focus in the serotonergic literature has also been the 5HT1A receptor, which is a post-synaptic G-coupled protein receptor and the most common serotonergic receptor in the brain. 105 The 5HT1A receptor is located both in brain areas with projections from the raphe nuclei (RN), where the serotonergic system is centralized, 117 and in the RN itself. Preclinical research has shown that animals with higher numbers of 5HT1A receptors are more vulnerable to stress 118 and display more “depressive” behavior (e.g. helplessness 119 ). Similarly, some postmortem studies have shown higher 5HT1A density in select brain regions (e.g. midbrain dorsal RN) of individuals with MDD relative to healthy controls. 49 Based on evidence from other modalities, it is reasonable to hypothesize observation of similarly increased 5HT1A availability in vivo.
To date all published 5HT1A studies have used PET, with all but one using the 5HT1A radiotracer [11C]WAY-100635 with somewhat mixed findings. Unlike cited preclinical and postmortem studies, there does not appear to be a consensus on whether 5HT1A is up or downregulated in depression: about half the published studies observed lower 5HT1A availability in MDD participants across various brain regions,41–46 whereas the other half found evidence for higher levels of 5HT1A in MDD, particularly in treatment-resistant individuals,26,36–39,47,48,120,121 and with only one study reporting no 5HT1A differences between MDD participants and controls. 40 Interestingly, downregulation in 5HT1A appears to be associated with treatment response following SSRI treatment.39,121
Importantly, a specific, testable explanation for the observed variability in 5HT1A findings has been proposed. In molecular imaging, there are various ways to quantify receptor availability, one of which is by the use of a reference tissue, which should be a region that is devoid of the target molecule (commonly the cerebellum). This method has been implemented in the studies that report lower 5HT1A availability in MDD; however, it appears that the cerebellum is not devoid of 5HT1A and there is evidence of lower cerebellar 5HT1A availability in the control relative to the MDD groups, 98 which would bias the results and lead to an underestimation of the true 5HT1A density in MDD. Illustrating this point, Hesselgrave and Parsey 122 showed that the same data set showed lower 5HT1A availability in MDD when analyses were completed using the cerebellum as reference, and the opposite finding when the outcome measure was a blood derived input function. It has also been suggested that 5HT1A level varies temporally over the course of MDD, possibly in part as a function of antidepressant exposure. 122 At present, all that can be concluded is that the preponderance of evidence suggests 5HT1A is elevated in MDD.
Like 5HT1A, the 5HT1B receptor is an autoreceptor present in the terminal regions of serotonergic neurons and plays a role in regulating, specifically decreasing when activated, the amount of serotonin in the synapse. 98 Both clinical 123 and preclinical 124 studies have shown that 5HT1B agonists have antidepressant properties, suggesting that unlike 5HT1A, lower levels of 5HT1B may be associated with MDD symptom experience. Although limited, evidence is somewhat consistent with this hypothesis with two reports of lower 5HT1B availability in the ventral striatum and pallidum 50 and the anterior cingulate cortex (ACC), subgenual prefrontal cortex (PFC), and hippocampus, 51 respectively, but another report of lower 5HT1B availability in the dorsal brainstem following cognitive behavioral therapy for MDD. 52
The 5HT2A receptor has received significant molecular imaging research attention, largely because SSRIs are known to act directly on 5HT2A receptors, leading to receptor downregulation. 125 Importantly, the 5HT2A receptor has proven responsive to stress exposure; 5HT2A levels have been shown to increase following exposure to chronic stress in preclinical studies.126,127 Depressive behavior following stress has also been shown to increase in response to 5HT2A agonists,128,129 further implicating upregulated 5HT2A in the pathophysiology of MDD. In vivo molecular imaging studies, however, have shown mixed results. There are several reports of downregulated53–55,57,58,60 5HT2A in MDD across several brain regions. However, there are also studies of no significant differences between untreated MDD participants and controls,59,130 and two showed higher levels of 5HT2A in the basal ganglia 131 and frontal, parietal, and occipital cortices 56 of unmedicated MDD participants relative to controls. The reports of higher 5HT2A density in MDD are in agreement with studies showing a reduction in cortical 5HT2A availability following treatment with tricyclic antidepressants.61,130 In attempting to account for mixed findings concerning 5HT2A in MDD, Ruhe et al. 98 suggested that individuals with MDD may have lower 5HT2A availability in the hippocampus due to HPA axis dysregulation and higher 5HT2A availability in other areas (e.g. the frontal cortex), a pattern that has been shown in animals exposed to chronic stress. However, this proposed explanation has not yet been directly tested. 132
Finally, a single PET study investigated the relationship between MDD risk and availability of the 5HT4 receptor, 62 which has been implicated in recent preclinical studies as a potential target for rapidly acting antidepressant agents. 133 The authors reported that healthy individuals with a family history of MDD have lower 5HT4 availability in the striatum, which was inversely correlated with the number of first degree relatives with MDD. These findings implicate the 5HT4 receptor as a potential target subserving familial risk for MDD, though replication and exploration of this receptor system in individuals with MDD is warranted.
Dopamine
Another monoaminergic neurotransmitter implicated in the pathophysiology of MDD is dopamine, which is known to be associated with emotion regulation, attention, motivation, and reward. 134 Dopaminergic circuits contribute to the regulation of concentration and memory functioning, as well as the ability to experience pleasure, all of which are impaired in MDD. 135 Dopaminergic dysfunction (i.e. reduced D2 receptor function), specifically in the nucleus accumbens, has also been linked directly to the experience of chronic stress, and subsequent development of depressive symptoms (anhedonia) in preclinical studies.80,136
As with serotonin, early molecular imaging of the dopamine system sought to examine the neurotransmitter’s synthesis in the brain, speculating in part based on results from clinical studies, 137 that less dopamine synthesis would be observed in depressed individuals. 98 This was verified by some PET studies in limited brain regions.19,138 A much larger number of studies have examined the functioning of dopamine transmitters (DAT) in MDD but have produced incongruent results. Specifically, genetic research has suggested that a DAT gene polymorphism, which increases DAT availability, is associated with higher risk for MDD.70,139 Several postmortem studies, however, have observed lower brain-wide DAT availability in MDD individuals. Findings in the molecular imaging literature are similarly mixed. Although there are reports of higher DAT availability in the striatum,71,140 basal ganglia, 72 and other brain areas 73 of MDD participants, others reported no significant differences between MDD and healthy control (HC),74,75 and three reported lower DAT availability in MDD in the striatum.76,78,141 Furthermore, Meyer et al. 76 reported that a downregulation in DAT availability was related to more severe depressive symptoms. However, they did not detect a significant change in DAT availability before and after bupropion antidepressant therapy, 76 suggesting that bupropion treatment did not appear to target this system or this system may not be significantly involved in MDD pathophysiology.
Measuring DAT with SPECT has also yielded some mixed results.28,63,67–69,77,108,109 However, not all of these studies characterized their results as indicative of DAT availability because, as noted above, SPECT DAT tracers are recognized to have affinity for SERT and NET as well. Results of these studies should therefore be interpreted with caution. Furthermore, Camardese et al. 66 highlighted the heterogeneity in presentation of MDD symptoms that can meet diagnosis and the potential effects of pharmacological treatment history and drug abuse as potential confounds affecting the observed variability in findings. As with SERT, inconsistent findings make it difficult to form a conclusion concerning the relationship between DAT and MDD at present.
The majority of neuroimaging studies investigating dopamine receptor availability have focused primarily on D2/D3 receptors, in part due to these receptors being targets for antidepressant and antipsychotic therapy. 136 Evidence from other modalities implicates possible downregulation of D2/3 receptors in MDD. Epigenetic studies have presented evidence for a link between chronic stress, D2 downregulation, and the development of depression. 66 Similarly, preclinical studies have shown downregulation of D3 in animals experiencing stress and depression. However, of the 14 studies which have examined D2/3 availability using PET and SPECT in individuals with MDD, all but 3 failed to observe differences in both medicated44,81,142,143 and unmedicated26,43,73,79,142 MDD participants relative to HCs. The remaining three showed higher D2/3 availability in MDD in the striatum,82,131 and in the case of Lehto et al., 77 a positive correlation between MDD symptom severity and D2/3 availability in the temporal cortex. However, although in vivo evidence suggests no specific D2/3 dysregulation in MDD, alterations treatment with SSRIs appears to influence D2/3 availability. For example, Montgomery et al. 24 reported lower D2/3 availability in the striatum of MDD participants using SSRIs, while two other studies observed increased D2/3 in the basal ganglia 136 and striatum 83 of SSRI treatment responders but not non-responders. Of note, while only two published studies have examined the D1 receptor in vivo in MDD, both found lower D1 availability in the striatum of MDD relative to control groups.85,144
Glutamate
More recently, there has been a major focus on the glutamatergic system in MDD. Glutamate is the primary excitatory neurotransmitter in the brain, with 80%–90% of synapses in the human brain being glutamatergic. Dysfunction of the glutamate system is thought to play a role in MDD 84 and there is currently an intensive focus on the glutamate system as a target for novel treatments for MDD. 84 Part of this focus has been on metabotropic glutamate receptors (mGluRs), G-protein-coupled receptors that mediate neuromodulatory effects of glutamate. 145 Indeed, mGluR5 antagonists have consistently shown antidepressant effects in animal models of depression.146–151 Lower levels of mGluR5 protein expression was also found in postmortem PFC tissue of individuals with MDD; 152 however, another study did not detect differences using autoradiography. 86 To date, there are four PET imaging studies that report on mGluR5 availability in MDD. Two smaller scale studies found lower levels of mGluR5 in various brain regions of those with unmedicated MDD including the PFC, ACC, and insula.152,153 A recent PET study focusing on late-life MDD found no differences in mGluR5 availability between older adults with MDD and matched controls. 87 Similarly, a recent study by our group found no differences in mGluR5 availability in a large group of unmedicated MDD individuals compared to HCs. As part of the same study, we used MRS to assess measures of glutamate and found a negative correlation between mGluR5 availability and MRS measures of glutamate in the ACC. 154 It is possible, therefore, that higher levels of glutamate are associated with a downregulation of mGluR5 in MDD. Furthermore, we recently observed that rapid downregulation of mGluR5 is associated with a reduction of somatic anxiety symptoms in MDD participants, suggesting this might be a treatment target for anxiety in MDD. 153 Further work will be required to definitively determine whether mGluR5 is a viable treatment target for symptoms of MDD.
Inflammation
Convergent evidence indicates that inflammation plays a significant role in MDD.88,155,156 Inflammatory mechanisms have been proposed to underlie the link between chronic stress and depression, 33 and meta-analyses have confirmed elevated inflammatory markers in individuals with MDD.157–160 However, the specific role of inflammation in the brain is less clear. It is possible to measure neuroinflammation with PET and radioligands that bind to the translocator protein (TSPO), which is upregulated on activated microglia, the immune cells of the brain. 161 To date, there are two published PET studies investigating TSPO in depression. The first found no difference in TSPO availability between depressed individuals with symptoms of mild severity and controls. 162 The second larger study showed elevated microglial activation in medication-free depression of moderate to severe severity, specifically in the PFC, ACC, and insula. 89 This finding has recently been replicated in a similar sample of medication-free depressed individuals. 90 Interestingly, this study found elevated TSPO in depressed individuals with versus without current suicidal thinking, suggesting a role for neuroinflammation in suicidal ideation specifically. Indeed, the heterogeneity in MDD symptomatology may play a role in the expression of neuroinflammation in MDD and account for divergent findings. PET and the 40+ existing TSPO radioligands provide a unique opportunity for clarifying the role of neuroinflammation in the pathophysiology of MDD and evaluating the use of anti-inflammatory medications in treating its symptoms.
Other Neurotransmitters
In addition to those reviewed above, PET and SPECT have been used to measure other systems such as monoamine oxidase (MAO-A), β-Amyloid, nicotinic acetylcholine receptors (β2*-nAChR), and γ-aminobutyric acid (GABA)—benzodiazepine receptors (GABAA-BZR). MAO-A is an enzyme known to be responsible for the regulation of brain monoamine levels. 91 Higher MAO-A availability is observed in MDD participants relative to HC,91,92 which would be in line with the studies showing lower monoamine neurotransmission in MDD.
β-Amyloid deposition in the brain, which is associated with the development of dementia, has been shown to correlate with history of lifetime MDD episode in postmortem work. 93 A single PET study by Wu et al. 93 showed higher β-Amyloid levels in depressed, but cognitively normal, older adults relative to age-matched HCs in multiple brain regions. A recent study by Yasuno et al. 94 looking at older adults without clinical MDD also reported a positive correlation between β-Amyloid availability averaged across the brain and depressive symptom severity. The cholinergic system has long been identified to be dysregulated in MDD,95,163–166 with the hypothesis of excesses in acetylcholine levels contributing to the experiences of depression. Specific to the neuroimaging literature, the β2*-nAChR has been identified as potentially relevant to MDD symptom expression, 167 although clinical trials with targets to downregulate β2*-nAChR did not prove efficacious for treatment of MDD. Combined in vivo/postmortem study from our group verified that β2*-nAChR does not appear to be dysregulated in MDD; however, elevations in acetylcholine levels might indeed play a role in depression. 168 Finally, a single study examined GABAA-BZR availability in MDD and found no differences in availability between MDD and control groups. 96
PET and SPECT Literature: Suicidal behavior
Importantly, while suicidal behavior is most frequently investigated as an associated feature of MDD, evidence from clinical research suggests the existence of meaningful neurobiological differences between suicide attempters and non-attempters. 169 For example, platelet studies showed that suicidal individuals had higher number for 5HT2A binding sites compared both to MDD non-attempters and HCs. 169 Similarly, variation in the DRD2 gene affecting D2 receptor expression has been consistently associated with risk for suicidal behavior, but not for MDD. 35 To date, seven published studies, all focused on the serotonergic system, have explicitly examined associations with current or historical suicidal behavior (ideation or attempt). Five studies focused on SERT availability in MDD participants with suicidal behavior with interesting if mixed results. For example, Yeh et al. observed a positive correlation between SERT availability and suicidal ideation across brain regions 33 while two other studies observed lower SERT availability in the midbrain 170 and putamen 170 of MDD participants with a history of suicide attempts compared to both healthy controls and MDD non-attempters. In combination, these findings suggest that lower midbrain SERT availability might distinguish between MDD individuals with and without risk for suicide. Notably, Oquendo et al. 171 failed to confirm midbrain SERT availability as a predictor of attempts in a recent longitudinal study. They did, however, observe that increased 5HT1A availability in the RN predicted both suicidal ideation and lethality of suicidal behavior. 171 In a cross-sectional study, Sullivan et al. also found that 5HT1A availability in the RN correlated positively with the lethality of attempt. 90 Thus, 5HT1A availability in the RN serve as a marker for suicide risk. Of note, a single PET study in high-lethality suicide attempters across psychiatric diagnoses found lower serotonin synthesis in the orbital and ventromedial PFC of attempters relative to HCs. 172 These promising results underscore the need for more molecular imaging work in suicidal individuals.
Bipolar Disorder
Relationship Between BD and Chronic Stress
Radiotracers used for in vivo molecular imaging in individuals with bipolar disorder.
Note. BP, bipolar disorder.
PET and SPECT Literature: BD
Serotonin
Serotoninergic function has long been thought to be central to the development of bipolar disorder. 58 Prange et al. 195 proposed what they termed the “permissive hypothesis of 5HT function” in bipolar disorder, which stated that deficient serotonergic functioning underlies bipolar disorder, and that manic and depressive episodes specifically would be marked by low 5HT. Supporting this proposal, a recent meta-analysis confirmed two gene SNPs related to the production of tryptophan hydroxylase-2, a rate-limiting enzyme for 5HT in the brain, are consistently associated with the development of bipolar disorder. 196 However, existing PET and SPECT imaging evidence has not yet substantiated the permissive hypothesis. It appears that there is a variability in SERT availability across mood states in BD. 98 During euthymia, there is lower SERT availability consistently measured across studies.179–182,184,197 During depression, studies appear to diverge in their findings, with reports of no differences in SERT availability in BD compared to HC, 184 lower SERT availability in thalamus, putamen, amygdala of depressed BD participants, 34 and higher SERT availability in the thalamus compared to MDD participants, but lower availability in the RN relative to controls. 178 No studies investigating SERT availability in manic BD participants have been published to date.
Both 5HT1A and 5HT2A are target receptors for medications used to treat BD (e.g. lamotrigine 198 and ziprasidone, 199 respectively). Further, genetic studies have found that the presence of specific SNPs in the 5HT1A gene predict both MDD and BD, 200 suggesting a possible similar pattern of 5HT1A upregulation in both disorders. By contrast, a postmortem study observed decreased 5HT2A mRNA levels in the hippocampal region of BD individuals relative to MDD and HC. 201 Quantitation of 5HT1A and 5HT2A receptors in BD in vivo has led to divergence in findings,187,188,202 including lower and higher receptor availability during depression, and no differences during euthymia. 187 5HT2A availability has been examined in bipolar mania only, with observations of lower availability relative to HC in all cortical regions, 185 and no alterations in 5HT2A availability following three to five weeks treatment with mood stabilizers. 186 More research across mood states is needed to increase understanding of the potential involvement of both receptors in BD. Of note, researchers have observed that while serotonin modulation plays a role in antidepressant action in bipolar depression, based on examination of the most effective agents for BD depression, it is not sufficient for symptom relief. 198 It is therefore crucial that molecular imaging research in BD examine other receptor systems.
Dopamine
Dopaminergic dysfunction is also implicated in the development and expression of BD. 98 Researchers have suggested that increased dopaminergic function underlies mania, while the opposite is true for depressive episodes in BD. 203 Many antipsychotic drugs are commonly prescribed to treat mania block dopamine receptors, decreasing levels of dopamine in the brain. 204 Indeed drugs known to increase levels of dopamine (e.g. some tricyclic antidepressants) have also been shown to induce mania in BD individuals.205,206 Two postmortem studies have observed upregulation of D2 in the dlPFC of BD.207,208 However, to date, only one molecular imaging study has examined D2 in BD during mania, finding no differences in D2 receptor availability in the striatum of manic BD participants relative to HCs. 191 The remainder of molecular imaging studies have focused on DAT, dysfunction of which, genetic research suggests, is associated with risk for BD.209,210 Interestingly, DAT findings in BD diverge based on imaging modality used. More specifically, two studies using SPECT and [99mTc]TRODAT-1 (a non-selective DAT tracer 98 ) observed higher DAT availability in the striatum of both depressed 189 and euthymic 190 BD participants relative to HC. By contrast, using PET and a DAT selective tracer ([11C]CFT), Anand et al. 208 observed lower striatal DAT in BD across mood states relative to HC. As with serotonin, more research is needed to further clarify the relationship between dopaminergic neurotransmission and BD.
Inflammation
Mounting evidence suggests that inflammation plays a key role in BD as well as MDD. 211 For example, pro-inflammatory cytokines have been shown to be elevated in the blood of BD individuals, during periods of mania, depression, and euthymia, suggesting that low-grade inflammation may be a trait marker of BD. 212 Preliminary evidence suggests the presence of inflammation in the brain as well as the periphery in BD, with one postmortem study showing upregulation of markers of neuroinflammation in the frontal cortex in BD. 213 One study has investigated neuroinflammation in vivo using PET to date. Using [11C](R)PK11195, Haarman et al. reported elevated TSPO levels in the right hippocampus of euthymic-medicated BD participants compared to controls. 214 Whether these medications impact neuroinflammation and whether level of measurable neuroinflammation differs across mood states remains to be determined.
Other Neurotransmitters
Cholinergic neurotransmitter systems have also been implicated in the pathophysiology of BD. 193 Both muscarinic agonists and acetylcholinesterase inhibitors increase cholinergic neurotransmission, and have been shown to increase depressive and decrease manic symptoms. 193 They have also been found to elicit erratic emotional responding—similar to that manifested in BD—when administered to HCs. 215 Additionally, CHRM2, which encodes for muscarinic receptors (M2) is one of several candidate genes shown to be expressed more in the ACC of BD individuals relative to HC in a recent postmortem study. Three molecular imaging studies have examined the availability of muscarinic (mACh) and nicotinic (nACh) receptors in BD. Cannon et al. first showed lower M2 availability in the ACC of depressed BD participants relative to HC. 193 They then showed that variation in the gene that encodes for M2, specifically, presence of the TT allele, was associated with lower M2 availability in the ACC of BD participants, and with more severe lifetime BD symptoms. 194 However, we did not find lower β2*-nAChR in depressed BD participants compared to euthymic BD and HC individuals, although variability in ACh levels between depression and euthymia may interfere with receptor availability quantification. 192 Alterations in ACh levels were not taken into account in the Cannon et al. studies; therefore, the role of the cholinergic receptors in the pathophysiology of BD needs further research.
Posttraumatic Stress Disorder
Relationship Between PTSD and Chronic Stress
Radiotracers used for in vivo molecular imaging in individuals with posttraumatic stress disorder.
PET and SPECT Literature: PTSD
Serotonin
As was the case with MDD and BD, altered serotonergic function has been implicated in the development of PTSD. 203 More specifically, in contrast to reductions observed in MDD models, preclinical studies have reported increased 5HT synthesis in PTSD in multiple brain regions.231–233 As in MDD, presence of the short SERT gene allele is a risk factor for PTSD, although a recent meta-analysis suggests that risk only extends to highly trauma exposed individuals. 234 Published imaging data suggest lower SERT availability in PTSD as well as correlations between SERT and severity of PTSD-related symptomatology.218–220
Focusing on presynaptic serotonin receptors, preclinical evidence suggests that alterations in 5HT1A function are associated with attentional bias to threat-related stimuli, 235 and exposure to stress reduces functioning of the 5HT1B receptor. 236 In line with preclinical findings, Sullivan et al. 222 observed upregulation of 5HT1A in the amygdala, RN, and all cortical regions of individuals with PTSD with and without comorbid MDD relative to HCs. Of note, an earlier study by Bonne et al. 221 observed no differences in 5HT1A availability in PTSD participants. However, Sullivan et al. 222 argued that characteristics of the PET tracer used and issues with the analytic approach used might have contributed to the Bonne’s null result. Further, consistent with preclinical evidence, there is a finding of lower 5HT1B availability in trauma-exposed control participants relative to HCs with no previous trauma exposure. 50 Earlier age of trauma exposure appears to be associated with both lower 5HT1B availability and PTSD symptom severity, suggesting that 5HT1B availability may be affected by trauma exposure, particularly during development.
Dopamine
Dopaminergic hyper-functioning has also been implicated in the development of PTSD, and specifically in the experience of hyperarousal symptoms. 237 More specifically, a DAT gene polymorphism has been shown to be associated with risk for developing PTSD following trauma exposure. 238 In keeping, researchers found higher availability of DAT in the striatum of PTSD participants relative to trauma-exposed control subjects. 223
GABAA-BZR
Evidence has shown that exposure to trauma/acute stress is associated with reduction in GABAA benzodiazepine receptor (GABAA-BZR) density, 239 which plays an important role in modulating nervous stress response in the central nervous system. 225 In vivo quantification is in line with the animal studies reporting lower PFC GABAA-BZR availability in individuals with combat-related or nonrelated PTSD relative to HCs.224,225
Other Neurotransmitters
Other neurotransmitters have also been examined in trauma-related psychopathology: cannabinoid receptor 1 (CB1), NET, β2*-nAChR, and mGluR5. In line with preclinical reports of an association between chronic stress and altered CB1,240,241 Neumeister et al. 226 reported elevated CB1 availability brain-wide in PTSD participants relative to both trauma-exposed controls and HCs. In an research domain criteria approach to examine the pathophysiology of PTSD, we reported that trauma and PTSD-related symptomology was associated with higher amygdalar CB1 availability. 227 NET has also been implicated in the development of anxiety-related disorders like PTSD. 242 A single PET study by Pietrzak et al. 243 found lower NET availability in PTSD in the locus coeruleus relative to healthy controls. However, no significant differences in NET availability were observed in the trauma-exposed control group (compared to PTSD and HC groups). 228 β2*-nAChRs availability, which has been implicated in regulation of memory and arousal which function abnormally in PTSD, 244 appears to be upregulated in the mesiotemporal cortex of PTSD individuals relative to HCs. This receptor also appears to be implicated in the re-experiencing symptoms in PTSD. 229 Finally, mGluR5 has emerged as a target of interest in PTSD 245 due to its role in fear conditioning and emotion regulation. In a recent pilot study, we observed increased mGluR5 availability in PTSD participants relative to HC across cortical regions. MGluR5 availability was also correlated with severity of PTSD avoidance symptoms. The cause of this dysregulation is not clear; however, postmortem examination and preclinical data suggest deficits in glucocorticoid functioning may be impacting mGluR5 density. Based on this preliminary evidence, CB1, NET, β2*-nACh, and mGluR5 may all play roles in the pathophysiology of PTSD; however, the limited numbers of studies preclude firm conclusions.
Discussion
In the last 20 years, molecular imaging has made a substantial contribution to the understanding of the neurobiological alterations underlying chronic stress-related disorders. The reviewed studies have enriched our understanding of how and why existing psychotherapeutic agents affect neurotransmitter function and have implicated novel targets for optimizing treatment efficacy. 246 To begin, they have confirmed the presence of in vivo alterations in serotonergic and dopaminergic functioning in at least some subgroups of MDD, BD, and PTSD individuals and provided preliminary evidence that other neurotransmitter systems play a role in these disorders. 247 Areas of overlap between molecular imaging evidence and other research modalities (e.g. lower SERT in MDD) lend confidence to assertions concerning the role of some receptors and psychiatric diagnoses. Areas of discrepancy, however, encourage critical examination of explanations for such incongruity like that proposed and tested by Hesselgrave and Parsey 122 to explain discrepant in vivo 5HT1A results in MDD.
While PET and SPECT have been confirmed as powerful and highly useful clinical research tools, 228 as it stands, the molecular imaging literature on MDD, BD, and PTSD raises as many questions as it answers. One exciting novel application of molecular imaging with the potential to reduce variability in findings is transdiagnostic research. As noted, the experience of chronic stress conveys risk for complex and highly comorbid psychiatric diagnoses (e.g. MDD and PTSD). Heterogeneous symptom presentation within wide and overlapping diagnostic categories makes the systematic study of these disorders by obtaining a “representative sample” extremely challenging. Transdiagnostic designs could eliminate a significant source of error by basing analyses on the presence or absence of a specific symptoms in lieu of diagnosis (i.e. using research domain criteria 248 ). In a recent example, Pietrzak et al. 249 examined the availability of Kappa-opioid receptors (KORs) in individuals with a range of trauma-related pathology (including MDD and PTSD). Instead of examining the relationship between diagnostic status and receptor density, they used principal components analysis to group symptoms into transdiagnostic “threat” and “loss” categories, and observed that low KOR availability in the amygdalar-ACC-ventral striatal circuit was associated with severe loss, but not threat symptoms, suggesting the use of molecular imaging to investigate the relationship between receptor availability and transdiagnostic aspects of chronic stress-related psychopathology may be an important next step.
In summary, molecular imaging has provided unique insight into the neurobiological underpinnings of chronic stress-related disorders. Despite inconsistencies in specific findings, the wide-reaching impact of chronic stress, and subsequent disorders like MDD, BD, and PTSD, is evident. The ability to confirm preclinical and postmortem findings, and to quantify the action of pharmacological agents in vivo has allowed for identification of biomarkers to improve risk assessment, progress toward optimization of pharmacotherapy, and identification of promising new treatment targets.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
