(p. 465) Resilience
Difficulties break some men but make others. No axe is sharp enough to cut the soul of a sinner who keeps on trying, one armed with the hope that he will rise even in the end.
A letter from Nelson Mandela to Winnie Mandela, written on Robben Island, February 1, 1975
Nelson Mandela (1918–2013) created a lasting legacy of peace and reconciliation in his native South Africa that will continue to inspire people the world over for many years to come. But what source of inner strength did he draw upon to survive the brutal system of apartheid that led to his incarceration in 1962 for his leadership of nonviolent protests by the African National Congress against the racist policies of the South African government? In fact, in the years leading up to his initial sentence of life imprisonment, Mandela came to believe that the only hope for change in South Africa was through armed struggle. As late as 1985, he refused to reject violence for the overthrow of the government. During his 27 years in prison, including 18 years on notorious Robben Island, what allowed him to believe in the possibility of a peaceful transition from a white-minority government to a black-majority government? In 1994, South Africa held democratic elections for the first time, and Mandela was elected its first black president, serving until 1999 (Mandela, 1994). How did he transition from political prisoner to Nobel Peace Prize recipient and president of his country?
For much of this volume, we have focused our attention on the deleterious effects of life stressors on mental health outcomes. But there is a flip side to this coin—many individuals, including Nelson Mandela, are able to demonstrate resilience (physical, mental, and spiritual) in the face of exposure to even the most traumatic and dehumanizing of stressors. In this chapter, we will explore what is known about resilience to the adverse effects of exposure to stressful environments in humans and how we can probe the underlying mechanisms responsible for resilience by studying animal models. Some of these animal models are the same ones we covered in chapters dealing with specific mental disorders, especially on depression (Chapters 12–13) and PTSD (Chapter 14). However, the focus in this chapter will be on those animals that displayed resilience to the stressor paradigm, rather than susceptibility.
(p. 466) The definition of resilience has evolved over time; initially, it was considered the process by which an individual achieved a positive mental health outcome after enduring adversities earlier in life. More recently, resilience has been viewed through a multidisciplinary lens as “the capacity of a dynamic system to withstand or recover from significant challenges that threaten its stability, viability, or development” (Sapienza & Masten, 2011). This systems-level approach to resilience goes beyond a developmental perspective to include the capacity of adults to display stable mental health outcomes during and after exposure to traumatic events, or following prolonged periods of stressful life experiences (American Psychological Association, 2010; Kalisch et al., 2015).
Developmental psychologists and pediatricians have been at the forefront of research on resilience for more than 50 years (Cicchetti et al., 1993; Cowen et al., 1990; Garmezy, 1974, 1985; Luthar, 1991; Rutter, 1979, 1981, 2012). Beginning in the 1960s, longitudinal studies of children subjected to highly stressful environments or traumatic experiences early in life revealed that most, but not all, of these at-risk children developed into healthy and high-functioning adults. What became evident early on from these studies was that children who overcame adverse circumstances and exhibited positive outcomes later in life were the rule rather than the exception (Masten, 2001).
A note of caution on high rates of positive outcomes has been issued by Infurna and Luthar (2016), who challenged the assertion regarding high rates of resilience in adults undergoing stressful life experiences by reanalyzing data from studies of trajectories of change in individuals exposed to spousal loss, divorce, and unemployment. They noted that modest modifications in statistical models could lead to significant variability in trajectories of change in adults exposed to significant life events, such that resilience was the least common outcome.
An obvious next step was to ask how these positive outcomes, however frequent, were possible in the face of adverse circumstances. Armed with this information, researchers then developed strategies for promoting resilience through interventions to enhance positive life outcomes in children as well as adults (Sapienza & Masten, 2011).
Neuroscience Discovers Resilience
Looking on the Bright Side
Beginning in the first decade of the twenty-first century, neuroscientists and biological psychiatrists came to realize that basic and clinical studies involving exposure to stressful stimuli, including traumatic stressors, had been almost totally consumed with psychopathological consequences, with little interest displayed (p. 467) in mechanisms of resilience (Charney, 2003, 2004). With this realization came an opportunity to focus attention on resilience as a novel approach for developing new targets for drug development and as a springboard for new strategies for preventing the onset of stress-related mental illnesses. Several excellent reviews have highlighted advances in understanding the neural, neuroendocrine, and molecular genetic components of a resilient phenotype in humans and in laboratory animals (Chen, 2019; Feder et al., 2009; Hunter et al., 2018; Karatsoreos & McEwen, 2013; Murrough & Charney, 2011; Pfau & Russo, 2015; Russo et al., 2012; Southwick & Charney, 2012).
Insights from twins. What explains the variability that is so evident in individuals’ responses to stressful life events? Amstadter et al. (2014) tackled this critical question by quantifying the genetic and environmental contributions to resilience in a population-based study of 7,500 twins from the Virginia Twin Registry. Data were generated in two waves, with participation rates in the twin subgroups of 72%–92%. Psychiatric resilience was defined in this study as the difference between twins’ internalizing symptoms and their predicted symptoms based upon their cumulative exposure to stressful life events. Statistical modeling was performed using the residual between the actual and predicted internalizing symptom total score. It should be noted that internalizing symptoms (e.g., depressive, anxiety-related, physical symptoms) are frequently reported to be a consequence of exposure to life stressors and formed the basis of a quantitative measure of resilience for this study. Based upon their measure of resilience, Amstadter and coworkers estimated that the heritability of resilience was approximately 31% based upon analyses of both waves of data collected. When error of measurement was incorporated into the model, the heritability for resilience increased to approximately 50%. This left an equal contribution from non-shared environmental effects that were of an enduring nature.
While the study by Amstadter et al. (2014) represents an important step forward, concerns regarding the experimental design have been raised. Wertz and Pariante (2014) noted that experiencing fewer symptoms than expected may still be sufficient to categorize a given individual as in distress. In addition, the magnitude of the difference between observed and expected symptom levels varied greatly across the sample, but all individuals were classified as resilient. Finally, other relevant health data (cardiovascular, immune, or neuroendocrine measures) were not collected from this sample of twins. It is possible that individuals could be classified as resilient in one health domain, but not in others. This is an important issue that should be addressed in future studies.
Unique aspects of resilience. In a comprehensive review regarding the neurobiological underpinnings of resilience in humans and in animal models, Russo et al. (2012) emphasized that resilience includes the presence of key molecular (p. 468) defects in susceptible individuals, as well as the presence of distinct molecular adaptations that are part of the resilient phenotype. More recently, research on resilience has broadened considerably to include studies with animal models that address neural, endocrine, immune, and molecular aspects of resilient phenotypes (Dantzer et al., 2018; Scharf & Schmidt, 2012). In addition, epigenetic mechanisms of resilience have received increased attention (Dudley et al., 2011; Zannas & West, 2014).
Several key issues will be addressed in this chapter, including the following:
• How is resilience assessed in studies with laboratory animals?
• Are there specific neural, endocrine, immune, or molecular systems that contribute to a resilient phenotype?
• Is resilience evident across a range of stressful environments, or is it limited to one or a few of these challenges?
• Do ovarian hormones play a role in the development of a resilient phenotype?
• Is it possible to promote a resilient phenotype by exposure to stressors early in life, or even in adulthood through a process of stress inoculation?
Animal Models of Resilience
Several animal models that have been employed to study the relationship between stress and mental disorders have also been utilized to study resilience to the damaging effects of stressful stimulation (Felger et al., 2015; Liberzon & Knox, 2012; Scharf & Schmidt, 2012). For many of these model systems, a subset of animals exposed to stressful stimulation displayed later evidence of physiological or behavioral disturbances. However, some of the animals came through exposure to stressful stimulation largely unscathed, and were classified as resilient based upon one or several behavioral assessments. Let’s consider what has been learned from this approach to determine if any common resilience mechanisms have emerged across model systems, and then consider their applicability to human resilience research.
Resilience to Predator Scent Stress (PSS)
Overview of PSS. The PSS model was initially developed by Cohen and her group to study biological mechanisms of PTSD, as described in the previous chapter. In this section, I will focus on the applicability of the PSS model specifically for the study of resilience (refer to Wu et al., 2013).
(p. 469) The basic methodology of the predator stress model is as follows. Adult male Sprague-Dawley rats were exposed to the scent of a predator (10 minutes on well-soiled cat litter) and then left undisturbed for 7 days. Controls were exposed to unsoiled cat litter for an equal amount of time. Animals were then run through a battery of tests that included an EPM and the acoustic startle response on day 7, and freezing behavior when placed on clean cat litter (the situational reminder) on day 8. Behavioral cutoff scores were employed to assign animals to the following groups: extreme behavioral response (EBR, approximately 25% of those tested); partial behavioral response (PBR, approximately 50% of those tested); or minimal behavioral response (MBR, approximately 25%). Behavioral scores for the MBR group were similar to behavioral scores for 80% of the rats in an unstressed control group (Cohen & Zohar, 2004).
NPY and resilience. A series of three experiments examined the involvement of brain neuropeptide Y (NPY) systems in resilience to the effects of predator stress (Cohen et al., 2012). NPY was of particular interest because it has been shown to dampen the negative effects of stress and has anxiolytic-like effects in brain (Heilig, 2004). In the first experiment, regional brain NPY levels were compared between controls, EBR rats, and MBR rats. NPY levels did not differ between controls and MBR rats in the anterior cortex, posterior cortex, amygdala, and periaqueductal gray (PAG), but were significantly lower in the hippocampus of MBR rats. In addition, NPY levels were significantly higher in the amygdala and PAG of MBR rats compared to EBR rats.
In a second experiment, administration of NPY (5 or 10 μg) bilaterally into the dorsal hippocampus 1 hour following exposure of rats to predator scent stress significantly reduced disruptions in behavior when testing occurred 7 days later. Specifically, rats receiving the higher dose of NPY spent significantly more time in the open arms of an elevated plus-maze (EPM), had reduced startle amplitude scores, and were less likely to engage in freezing behavior during the situational reminder test compared to rats receiving artificial cerebrospinal fluid (a-csf) injections. Administration of exogenous NPY also up-regulated expression of NPY protein and NPY-Y1 receptors as well as expression of BDNF.
In contrast, rats that received micro-injections of an NPY-Y1 receptor antagonist (20μg) bilaterally into the dorsal hippocampus 1 hour following exposure to predator scent displayed extreme disruptions in behavior when tested 7 days later. Blockade of NPY-Y1 receptors resulted in significant decreases in total time spent in the open arms and fewer open-arm and total-arm entries during EPM testing compared to controls that received a-csf injections. In addition, the percentage of drug-treated rats that had EBR scores (approximately 75%) was dramatically higher than control rats that received a-csf injections. These results provided strong evidence that brain NPY systems promote resilience in the PSS model.
(p. 470) The Arc of resilience. In a related study, Kozlovsky et al. (2008) examined the expression of activity-regulated cytoskeleton-associated protein (Arc), an immediate early gene, and its relationship to resilience in the PSS model. Those rats whose plasma CORT levels and behavioral patterns during EPM and ASR testing were most disrupted 7 days after exposure to PSS did not display up-regulation of Arc mRNA levels in areas of the hippocampus and frontal cortex. In contrast, rats that had minimal disruptions in behavior 7 days following exposure to PSS had significant increases in levels of mRNA for Arc. The investigators suggested that Arc protein may be related to resilience and/or facilitated recovery from exposure to traumatic stress.
Diet and resilience. Hoffman et al. (2015) examined the effects of dietary supplementation with β-alanine (BA, 100 mg/kg) for 30 days on the behavioral and neuroendocrine responses of laboratory rats exposed to PSS. Behavioral indices of traumatic stress were measured 7 days after exposure to PSS, and expression of BDNF and brain carnosine concentrations were analyzed on day 8.
Rats fed a normal diet that were exposed to PSS displayed significant behavioral disruptions in the EPM and the ASR tests and higher freezing scores in the situational reminder compared to controls. In contrast, BA-supplemented rats exposed to PSS had patterns of behavior that were somewhat similar to controls. BA supplementation alone did not affect behavioral patterns. Brain carnosine concentrations in the hippocampus and other brain areas were significantly greater in animals supplemented with BA compared to those receiving a control diet. BDNF expression in the CA1 and dentate gyrus subregions of the hippocampus was significantly lower in animals exposed to PSS and fed a normal diet compared to animals exposed to PSS and supplemented with BA, or to controls. Dietary supplementation with BA increased brain carnosine concentrations and reduced trauma-like behaviors, and these actions may be mediated in part by maintaining normal levels of BDNF expression in the hippocampus. The mechanism of action of brain carnosine in enhancing resilient behavior may be related to its role as an antioxidant (Kohen et al., 1988).
Resilience circuitry. Mitra and colleagues (2009) took advantage of the fact that not all animals exposed to intense stressors displayed long-lasting changes in behavior. They focused their efforts on the basolateral amygdala because of its role in affecting fear responses and modulating stress responses. Adult male Long-Evans hooded rats were used in this experiment. EPM and hole board tests were used to assess the behavioral impact of a single 10-minute exposure of rats to a cat. Behavioral tests occurred 2 weeks after exposure to the cat. Cutoff scores were used to select from a total of 71 rats tested a sample of 4 rats that were severely affected by predator exposure, 4 rats that were not significantly affected, and 4 handled controls.
(p. 471) Brief exposure to a cat induced long-lasting increases in anxiety in some rats, although not all animals show increased anxiety. A subset that was defined as well-adapted remained largely unaffected behaviorally by exposure to the cat. Thus, the same stress experience evoked different degrees of behavioral responses among the initial group of rats that was tested. Animals that differed in their behavioral patterns also differed in dendritic architecture of basolateral amygdala (BLA) neurons, which form part of the neural circuitry mediating stress-induced anxiety (Adamec et al., 2005). Well-adapted animals exhibited retracted dendrites and increased packing density of dendrites compared to maladapted animals with high anxiety and, surprisingly, compared to unstressed handled controls. These findings point to a putative neurobiological substrate for resilience to the anxiogenic effects of traumatic stress.
Interestingly, the resilient rats did not differ from their less resilient counterparts in their interactions with the cat. Rather, their differences were limited to the fact that the resilient rats did not generalize the fear they experienced during exposure to the cat. Two possibilities were presented to explain these findings: (1) well-adapted animals underwent dendritic retraction to counteract the effects of predator exposure, or (2) preexisting differences in dendritic morphology in BLA neurons explained the striking behavioral differences between well-adapted and maladapted animals.
Resilience to Chronic Social Defeat Stress (CSDS)
That wasn’t so bad. CSDS has been employed extensively as an animal model of major depressive disorder in humans (refer to Chapters 12 and 13). Similar to studies of predator stress described earlier, there was a range of behavioral responses across animals exposed to CSDS, and this heterogeneity of responses presented an opportunity to explore mechanisms of resilience to the deleterious effects of CSDS. As with the PSS model, employing cutoff behavioral scores provided investigators with the possibility of studying susceptibility and resilience using the same stress paradigm. The following studies clearly indicate that resilience is not simply the absence of susceptibility. Rather, resilience and susceptibility to CSDS are separate and distinct categories, each with its own set of neurobiological underpinnings (e.g., Russo et al., 2012).
In a comprehensive series of experiments from the laboratory of Dr. Eric J. Nestler, C57BL/6J inbred male mice were exposed to 10 consecutive days of CSDS. Test mice were placed individually into the home cages of resident aggressive CD-1 male mice for 5 minutes, after which the resident-intruder pairs were separated by a perforated Plexiglas barrier placed in the middle of the cages for 24 hours. Each day, the C57BL/6J test mice were exposed to different CD-1 (p. 472) resident males. C57BL/6J control mice were housed in pairs with a perforated Plexiglas barrier present and were handled each day.
Control and defeated mice were then tested for social avoidance on day 11, and were scored based upon their tendencies to avoid (susceptible) or approach (resilient) a conspecific. Even though C57BL/6J mice are inbred and therefore genetically identical for all practical purposes, approximately 30%–50% of the males tested did not display avoidance of a social target following CSDS, while the remainder did (Figure 15.1). Social avoidance scores also correlated with other depression-like behavioral measures, including anhedonia, weight loss, and circadian amplitude in body temperature. However, resilient and susceptible mice displayed comparable elevations in plasma levels of CORT following swim stress and similar levels of stress-induced polydipsia and anxiety-like behaviors in the EPM, which are responses consistent with prior exposure of mice to chronic stress (Krishnan et al., 2007).
Reward circuits and resilience. The heterogeneity of responses to CSDS presents an opportunity to investigate neurobiological mechanisms of resilience to depression-like behaviors. Focusing on the subpopulation of mice that had low levels of social avoidance behavior, Krishnan et al. (2007) reported that resilient mice did not have an increase in levels of BDNF in the nucleus accumbens (NAc) following 10 days of CSDS. In addition, resilient mice displayed evidence of regulatory changes in a large number of genes in the ventral tegmental area (VTA) and the NAc compared to susceptible mice.
K+to the rescue. In the VTA, three genes that were up-regulated in resilient mice but not in susceptible mice coded for voltage-gated potassium (K+) channels (Kcnf1, Kcnh3, and Kcnq3). These changes in voltage-gated K+ channels in resilient mice prevented the significant increases in firing rates of VTA DA neurons that were characteristic of susceptible mice following social defeat. (p. 473) Results from related experiments indicated that blunting of increases in firing rates of VTA DA neurons following CSDS also blocked release of BDNF within the NAc, resulting in a resilient phenotype (Krishnan et al., 2007) (refer to Figure 15.2).
Capitalizing on these findings, Friedman et al. (2014) targeted VTA DA neurons in susceptible mice in a novel approach for inducing resilience. Enhancing the excitatory hyperpolarization-activated cation channel current (Ih) in VTA DA neurons of susceptible mice pharmacologically or optogenetically to levels seen in resilient mice reversed the depressive-like behaviors (e.g., reduced social contact) characteristic of a susceptible phenotype. This finding may lead to novel treatment strategies for major depressive disorder based upon enhancing resilience pathways in the brain rather than repairing susceptibility pathways (Han & Nestler, 2017).
Out of the blue. Zhang et al. (2019) explored the role of locus coeruleus (LC) inputs to VTA DA neurons in development of the resilient phenotype following exposure to CSDS. They found that the firing rates of LC neurons that (p. 474) projected to the VTA increased in resilient compared to susceptible mice. Using an optogenetic approach in susceptible mice, these investigators demonstrated that increasing the activity of LC neurons that projected to the VTA resulted in a resilient behavioral phenotype as well as decreased firing rates of VTA → NAc DA neurons. Finally, if α1- and β3-adrenergic receptors on VTA → NAc DA neurons were activated pharmacologically, this also resulted in resilient features at the level of ion channels, neuronal activity, and behavioral characteristics. If both of these receptors were selectively blocked within the VTA over a 10-day period, the effects of optogenetically driven increases in LC → VTA neurons did not result in a resilient phenotype. These exciting findings point to an LC → VTA → NAc circuit that is responsible for the emergence of a resilient phenotype following CSDS. The α1- and β3-adrenergic receptors that were identified as a crucial link in this pro-resilience pathway provide new therapeutic targets to enhance resilience (Zhang et al., 2019).
Delta force. Using a similar approach, Vialou et al. (2010b) measured ΔFosB in the NAc following CSDS. ΔFosB is a stable transcription factor that was induced following exposure of animals to chronic stress (Perriotti et al., 2004). Resilient mice showed the greatest induction of ΔFosB in both the core and shell subregions of the NAc compared to susceptible mice and controls. In addition, there was a significant positive correlation between amounts of ΔFosB in the NAc and social interaction scores, suggesting that the degree of ΔFosB induction in the NAc may be a critical determinant of whether a mouse displays a susceptible or a resilient pattern of behavior. In a related series of experiments, these investigators reported that induction of ΔFosB following CSDS was dependent upon expression of serum response factor, an up-stream transcriptional activator of several immediate early genes, including c-fos (Vialou et al., 2010a).
Using bi-transgenic mice that inducibly over-expressed ΔFosB in the adult NAc and dorsal striatum, these investigators found that these bi-transgenic mice showed significantly less social avoidance after 4 or 10 days of CSDS, suggesting that ΔFosB served a protective role against the effects of CSDS. In contrast, bi-transgenic mice that inducibly over-expressed ΔcJun, a transcriptionally inactive truncated cJun mutant that antagonizes ΔFosB activity, were more susceptible to CSDS than control littermates, and showed maximal social avoidance behavior after only 4 days of CSDS.
ΔFosB is known to regulate the transcription of many genes in the NAc. One such target gene is the AMPA glutamate receptor subunit, GluR2; mice overexpressing ΔFosB in the NAc had greater amounts of GluR2, but showed no differences in other glutamate receptor subunits. This selective up-regulation of GluR2 in the NAc appeared to play a critical role in resilience to CSDS. The induction of GluR2 seen in resilient mice appeared to reflect a direct effect of (p. 475) ΔFosB on the gene encoding GluR2, as there was increased binding of ΔFosB to the GluR2 promoter.
To identify additional ΔFosB target genes that may contribute to resilience, gene expression array data sets were obtained from the NAc of both bi-transgenic mice overexpressing ΔFosB as well as C57Bl/6J mice 48 hours after CSDS that had a resilient phenotype. There was considerable (>75%) overlap between the genes induced in the NAc by ΔFosB and those induced by CSDS in resilient mice. From the population of induced genes, Sc1 (also known as Sparc-like 1 or hevin) was selected for further study based upon the magnitude of its induction in resilient mice and in mice over-expressing ΔFosB. SC1 is an anti-adhesive matrix molecule that is highly expressed in the adult brain, where it localizes in the postsynaptic density and has been implicated in synaptic plasticity. To assess a possible role for Sc1 in resilience, it was reported that over-expression of Sc1 in the NAc of susceptible mice reversed the high social avoidance scores induced by CSDS.
It’s all about networking. Lorsch et al. (2019) identified a gene network that plays a unique role in the expression of a resilient phenotype following exposure to CSDS. Critical to this gene network effect on resilience is Zfp189, which codes for a zinc finger protein. Overexpression of Zfp189 in PFC neurons activated this network of 281 genes and enhanced resilience. In addition, induction of Zfp189 enhanced resilience whereas suppressing Zfp189 resulted in susceptibility to the effects of CSDS. These exciting results open up the possibility of targeting Zfp189 expression as a novel strategy to upregulate a resilience-related gene network as a means of treating major depressive disorder or preventing its development.
DA plays a role. Medium spiny neurons (MSNs) of the NAc play an important role in resilience to the deleterious effects of CSDS on behavioral responses. MSNs make up the majority of the NAc and come in two types, one expressing D1 receptors and playing a role in reward and the other expressing D2 receptors and playing a role in aversion. Using a variety of optogenetic and pharmacogenetic approaches, Francis et al. (2015) reported that increasing the frequency of excitatory synaptic inputs in D1-expressing MSNs resulted in resilience to the effects of exposure to CSDS. Modulation of activity in D2-expressing MSNs did not affect behavioral outcomes to CSDS.
Taking further advantage of this well-characterized system of D1- and D2-expressing MSNs, Hamilton et al. (2018) examined the effects of cell-type specific targeted epigenetic modifications of the Fosb gene on a resilient behavioral phenotype following exposure of mice to a sub-threshold level of CSDS in mice. Their results indicated that histone acetylation of Fosb in D1-expressing MSNs, which promotes transcription, and histone methylation of Fosb in D2-expressing MSNs, which inhibits transcription, both resulted in a resilient phenotype following sub-threshold CSDS. Behavioral measures of resilience in these (p. 476) experiments included measures of social avoidance and sucrose preference. These powerful findings for the first time have revealed the potential for epigenetic editing in specific neuronal cell types to enhance resilience to a depression-like phenotype in a social stress-induced model of depression.
In summary, these results are consistent with a role for ΔFosB in the NAc in promoting resilience during CSDS by inducing a form of synaptic plasticity that counteracts the strong negative associative learning that occurs in susceptible mice during repeated episodes of social defeat. The dominant role of ΔFosB and its targets in promoting adaptive responses to CSDS revealed new molecular targets for the development of antidepressant treatments (Vialou et al., 2010ab). Identification of novel targets related to resilience may be facilitated by a metabolomics approach, as demonstrated in a series of experiments by Dulka et al. (2017).
Inhibition has its place. GABAB receptors may play a role in stress-related psychiatric disorders such as anxiety and depression (Ghose et al., 2011). O’Leary and colleagues (2014) explored the impact of GABAB receptor subunits on behavioral responses to CSDS in laboratory mice. Compared to wild-type (WT) mice, mice that lacked the GABAB(1b) subunit were resilient to the effects of CSDS, based upon their behavioral scores in tests of social withdrawal and saccharin preference. In addition, GABAB(1b) knockout mice had increased rates of proliferation and survival of newborn neurons in the ventral hippocampus.
Hidden in the genes. A major challenge in studies of animal models of depression is to focus in on transcriptional networks and key driver genes that regulate resilience to the depressive-like behaviors that attend CSDS. Bagot et al. (2016) took on this significant challenge by examining the NAc and its inputs from the ventral hippocampus, amygdala, and PFC in control mice and mice resilient to the effects of CSDS. Mice were tested at 2 and 28 days after the CSDS paradigm as well as at 28 days post-stress, but with an added 1-hour episode of social defeat (stress-primed).
At the 2-day post-stress time point, there were more differentially expressed genes in resilient mice versus control mice in all four brain areas compared to data for susceptible mice versus control mice. However, at 28 days post-stress, the number of differentially expressed genes remained greater in resilient mice only in the ventral hippocampus and PFC. Further analyses revealed that increased synchrony between differentially regulated genes in the PFC and NAc appeared to be especially critical for the development of the resilient phenotype following CSDS. Further analyses revealed 30 clusters of co-expressed genes in resilient mice that were non-overlapping with 52 clusters of co-expressed genes in susceptible mice. Compared to control mice, resilient mice had only two gene clusters that had a gain in connectivity and three that had a loss of connectivity. A gain in connectivity suggested a strengthening of an existing transcriptional (p. 477) network or the emergence of a novel one. In contrast, a loss of connectivity implied a weakening of a basal transcriptional network.
A critical aspect of this complex study was that it provided an approach to study how groups of genes were organized across brain regions into functional clusters. Changes in activity of these functional clusters appeared to underlie susceptibility or resilience to depressive-like behaviors following exposure to CSDS. By targeting hub genes within these functional clusters, it may be possible to promote resilience on the one hand, or to prevent depressive symptoms on the other hand (Bagot et al., 2016).
MRI studies. Anacker et al. (2016) identified resilient and susceptible mice following 10 days of CSDS and their brains, together with those of unstressed controls, were processed for ex vivo structural magnetic resonance imaging and diffusion tensor imaging to reveal any neuroanatomical differences between groups. In addition, correlations were computed for regional brain volumes versus social avoidance scores. The study was correlational and assumptions were made about fractional anisotropy reflecting the extent of dendritic branching.
Social avoidance scores correlated negatively with the local volume of the cingulate cortex, NAc, thalamus, raphe nuclei, and BNST. Social avoidance scores correlated positively with local volume of the VTA, habenula, PAG, cerebellum, hypothalamus, and hippocampal CA3 area. Fractional anisotrophy was increased in the hypothalamus and hippocampal CA3 area. Synchronized anatomical differences between the VTA and the cingulate cortex, hippocampus and VTA, hippocampus and cingulate cortex, and hippocampus and hypothalamus were observed. For resilient mice, a larger cingulate cortex volume predicted a smaller VTA volume.
The approach taken in this study offered several advantages. It revealed novel brain areas involved in resilience versus susceptibility to CSDS (e.g., thalamus, raphe nuclei, BNST) and it allowed for comparisons with complementary MRI studies in humans. Unfortunately, it was not possible to determine if these structural brain differences between resilient and susceptible mice were preexisting, or if they had changed as a result of the CSDS paradigm. This is an important question that should be addressed in the future. With these data in hand, it would be possible to detect anatomical differences in brain areas that predispose some mice to display a resilient phenotype, and others to display a susceptible phenotype following CSDS (Anacker et al., 2016).
The protein kinase connection. Experiments by Bruchas et al. (2011) focused on the family of mitogen-activated protein kinases (MAPK). p38 MAPK (also referred to as stress-activated protein kinase, SAPK) was known to block stress-induced behavioral responses. The protein kinase is widely distributed in tryptophan hydroxylase-expressing cells of the dorsal raphe nucleus (DRN). The α-isoform of p38 MAPK was selectively inactivated in 5-HT neurons and (p. 478) in astrocytes. A single episode of social defeat stress produced social avoidance in wild-type mice, but not in mice having the selective p38α MAPK deletion in 5-HT neurons of the DRN. Thus, mice with the p38α MAPK deletion in 5-HT neurons were resilient to the effects of acute social defeat stress. Stress-induced activation of p38α MAPK led to translocation of the serotonin transporter to the neuronal membrane, thereby increasing the rate of transmitter uptake at 5-HT nerve terminals. The results of these experiments suggested that CSDS initiated a cascade of molecular and cellular events in which p38α MAPK induced a hypo-serotonergic state underlying stress susceptibility.
Epigenetic changes. Serotonin neurons were also implicated in resilience in a series of experiments by Walsh et al. (2017). These investigators examined the role of PHF8, an X-linked histone demethylase that functions as a transcriptional activator, on depression-like and anxiety-like behaviors in mice. Phf8(–/y) knockout mice (KO) were generated on a mixed C57B6/129SvJae background, and behavioral patterns were assessed using a battery of tests. Compared to wild-type controls, Phf8(–/y) KO mice were more active and spent more time in the center portion of an open field arena and spent more time in the open arms of an EPM. Both of these tests have been used as measures of anxiety-like behaviors, and the results suggest that Phf8(–/y) KO mice are more resilient to anxiety-provoking environments.
Phf8(–/y) KO mice and wild-type controls were also exposed to CSDS each day for 10 consecutive days. Chronically stressed controls displayed high levels of social avoidance behavior following CSDS. In contrast, Phf8(–/y) KO mice spent as much time in social contact as did Phf8(–/y) KO mice and wild-type mice that were not exposed to CSDS. These findings pointed to Phf8(–/y) KO mice being resilient to the depressive effects of CSDS.
What connects downstream effects of Phf8 with the resilient behavioral phenotype observed in Phf8(–/y) KO mice? Further experiments revealed that WT mice had high levels of expression of PHF8 in PFC, ventral striatum, and hippocampus. The absence of PHF8 in Phf8(–/y) KO mice was linked to increased expression of the serotonin receptors 5-HT1A, 5-HT1B, and 5-HT2A, which in turn appeared to explain the increased resilience of KO mice to stressful stimulation (Walsh et al., 2017).
IL-6 looms large. As described in part in Chapter 13, Hodes et al. (2014) examined the role of the peripheral immune system in susceptibility versus resilience of C57BL/6 male mice to CSDS. They found preexisting individual differences in the sensitivity of the peripheral immune system that predicted later resilience to social stress. Cytokine profiles were quantified 20 minutes after the first exposure to social defeat. Of the cytokines regulated by stress, IL-6 was the best predictor of resilience in that levels remained comparable to handled controls, whereas IL-6 was significantly elevated in susceptible mice. In addition, prior to any physical (p. 479) contact in mice, there were individual differences in IL-6 levels from leukocytes stimulated ex vivo with LPS that predicted susceptibility versus resilience to a subsequent social stressor.
IL-6 knockout (IL-6(−/−)) mice were also employed to further link low levels of IL-6 with resilience to CSDS. Bone marrow (BM) chimeras were generated by transplanting hematopoietic progenitor cells from IL-6 knockout (IL-6(−/−)) mice. IL-6(−/−) BM chimeric and IL-6(−/−) mice, as well as mice treated with a systemic IL-6 monoclonal antibody, were resilient to CSDS. These data established that a dampened peripheral IL-6 response to CSDS was associated with a resilient behavioral phenotype (Hodes et al., 2014).
Inflammation is a pain. In a further effort to connect inflammatory responses to depressive-like behaviors, Ren et al. (2016) explored the role of soluble epoxide hydrolase (sEH) in the CSDS paradigm. sEH metabolizes epoxyeicosatrienoic acids, which have potent anti-inflammatory effects. These investigators used an inhibitor of sEH, trifluromethoxyphenyl-3(1-propionylpiperidine-4-yl) urea (TPPU), and sEH knockout (KO) mice to modify levels of inflammation following CSDS. Levels of sEH were increased significantly in brains of chronically stressed mice compared to controls. In addition, TPPU (3.0 mg/kg/day orally) given 1 hour prior to each session of CSDS resulted in a significant increase in the social interaction test compared to chronically stressed animals that received vehicle alone. Similarly, TPPU eliminated the anhedonia characteristic of mice exposed to CSDS. Consistent with these findings, sEH KO mice were resilient to the effects of CSDS, with behavioral scores comparable to unstressed control mice. sEH KO mice also had increased levels of BDNF and phosphorylation of its receptor, TrkB, in the frontal cortex and hippocampus. These investigators suggested that increased BDNF-TrkB signaling in these two brain areas may underlie the resilient phenotype of sEH KO mice (Ren et al., 2016).
Levels of sphingosine-1-phosphate receptor 3 (S1PR3) in the mPFC and in peripheral blood are associated with resilience to CSCS in laboratory rats. S1PR3 is a G protein-coupled receptor that binds to its ligand, S1P, and exerts influences on immune responsiveness, angiogenesis, and cell proliferation in the periphery. Much less is known about the role of S1PR3 signalling in the brain. S1PR3 expression in the mPFC was higher in the mPFC of resilient rats compared to susceptible and control rats. Over-expression of S1PR3 in the mPFC resulted in a resilient phenotype but knockdown of S1PR3 resulted in a susceptible phenotype following the CSDS paradigm. Anxiety- and depression-like behaviors associated with a susceptible phenotype were mediated by TNFα levels in the mPFC. Blockade of TNFα signaling in the mPFC prevented the behavioral changes observed in susceptible rats following CSDS (Corbett et al., 2019).
Small RNAs pack a big punch. Resilience to stress was explored in laboratory rats subjected to a CSDS paradigm (Pearson-Leary et al., 2017). (p. 480) Latencies to display behavioral signs of defeat were employed to generate a resilient group with long latencies (LL) and a stress-susceptible group with short latencies (SL). LL rats were found to have higher levels of miR-455-3p and lower levels of miR-30e-3P in the ventral hippocampus compared to SL rats. Ingenuity Pathway Analysis (IPA) software identified inflammatory and vascular remodeling pathways associated with genes targeted by these two miRNAs. Following CSDS, resilient LL rats did not display the characteristic remodeling of the neurovascular unit within the ventral hippocampus in the way that stress-susceptible rats did. The resilient phenotype was reflected in the ventral hippocampus in a lack of new blood vessel formation, no increase in blood-brain barrier permeability, no increase in concentration of microglia with associated changes in pro-inflammatory cytokines, and no increase in neuronal activity as assessed by levels of FosB/ΔFosB positive cells. Similar changes in resilient and susceptible rats were not observed in the dorsal hippocampus or in the PFC.
To confirm this stress-resilient profile in the ventral hippocampus, these investigators found that daily i.c.v. administration of the proinflammatory cytokine vascular endothelial growth factor (VEGF164) resulted in changes in the ventral hippocampus and behavioral responses to CSDS that were characteristic of the stress-susceptible phenotype. In contrast, administration of the non-steroidal anti-inflammatory drug meloxicam (1 mg/kg, i.p.) enhanced the resilient phenotype in rats exposed to CSDS. Taken together, these findings emphasized the important role played by inflammatory processes within the ventral hippocampus in resilience to CSDS (Pearson-Leary et al., 2017).
Coping and resilience. Two studies utilized a slightly different model of CSDS in which adult male Sprague-Dawley rats were exposed to repeated resident-intruder stress (30 minutes per day for 5 consecutive days), and coping styles were determined. Long-Evans retired breeders were used as resident aggressive males. Passive coping was defined as a short latency (<300 sec) to assume a defensive posture, while active coping was defined as a longer latency to assume a defensive posture (>300 sec).
In the first experiment, Wood et al. (2010) reported that rats exhibiting longer latencies to defeat (i.e., resilient rats) did not develop depressive-like neuroendocrine or behavioral characteristics over the course of CSDS. In these resilient animals, there was habituation of HPA axis responses to CSDS. In addition, behaviors of resilient animals in a forced swim test were similar to behaviors of unstressed controls. An additional correlate of resilience to CSDS in laboratory mice, as reported by Elliott et al. (2010), related to maintenance of DNA methylation of the promoter region of the Crf gene.
In the second experiment (Wood et al., 2015), CSDS was reported to differentially regulate 19 genes in the locus coeruleus (LC) and 26 genes in the dorsal (p. 481) raphe nucleus (DR), and many of these genes coded for inflammatory factors. IL-1β expression was increased in passive coping rats and decreased in active coping rats in both the LC and the DR. Protein changes in the two brain areas were generally consistent with a proinflammatory response in passively coping rats, but not in actively coping rats. Rats that displayed passive coping styles during CSDS also exhibited anhedonia based on a sucrose preference test, and this response was blocked by i.c.v. administration of an antagonist of IL-1β. Resilience to CSDS was associated with a down-regulation of central inflammatory processes. Social stress led to up-regulated neuropeptide Y1 receptor (Npy1r) gene expression in the LC of active and passive coping style rats. However, only in active coping rats was Npy2r and kappa opioid receptor-1 gene expression down-regulated in the LC. These latter two changes may have conferred resilience to stress in actively coping rats.
Repeated Exposure to Stress
The pioneering studies of Maier and Seligman (2016) have been adopted by many investigators to study the effects of repeated exposure to aversive stimulation on subsequent behavioral and physiological responses. An additional focus of such studies has been the differential impact of uncontrollable versus controllable delivery of aversive stimuli on adaptive behavioral and physiological responses (Lucas et al., 2014).
Taneja et al. (2011) used 100 unpredictable, inescapable tail shocks (1.2–2.0 mA, 5 seconds duration, every 25–110 seconds for 2 hours) to yield learned helpless (LH) and resilient Sprague-Dawley rats. Controls were restrained and had electrodes affixed to their tails, but no shocks were delivered. The following day, rats were tested in a shuttlebox for escape behaviors. Rats that had escape latencies within 2 standard deviations (SDs) of control rats were defined as resilient, whereas rats with escape latencies greater than 2 SDs of control rats were defined as LH. Measures taken in various brain areas were as follows: α2A-adrenoceptors (α2A-AR), CRF1 receptors, G-protein-coupled receptor kinase 3 (GRK3, phosphorylates receptors and contributes to their down-regulation), GRK2, tyrosine hydroxylase (TH), and carbonylated protein levels.
In resilient rats, α2A-AR and CRF1 receptor levels were significantly down-regulated in the LC after inescapable tail shock. GRK3 was reduced in the LC of LH rats but not resilient rats. In contrast, GRK2 levels were unchanged. In the amygdala, GRK3 but not GRK2 levels were reduced in LH rats but not in resilient rats. Protein carbonylation, an index of oxidative stress, was increased in the LC and amygdala of LH rats but not of resilient rats. Resilient rats appeared to differ from LH rats in their greater ability to handle oxidative stress, maintain levels of (p. 482) GRK3, and regulate receptor signaling more effectively following exposure to inescapable tail shocks (Taneja et al., 2011).
Differences in expression. A genome-wide microarray experiment tracked changes in gene expression in the hippocampus and frontal cortex of male Sprague-Dawley rats that were exposed to acute unavoidable stress (restraint for 50 minutes with 80 tail shocks, 1.5 mA, every 30 seconds) delivered through tail electrodes (Benatti et al., 2012). One day later, both groups of rats were tested for their ability to perform an active avoidance task over 30 trials. Two groups of rats emerged from those exposed to unavoidable stress—one was stress-vulnerable (fewer than 8 successful escapes/30 trials) and the other was stress-resilient (greater than 16 successful escapes/30 trials). These differences in avoidance were associated with specific changes in gene expression, with little overlap between the two groups (approximately 10% of probe sets). In the frontal cortex, there was a down-regulation of the transcripts coding for interferon-b and leukemia inhibitory factor in resilient rats and an up-regulation of the neuroendocrine-related genes coding for growth hormone and prolactin in vulnerable rats. In the hippocampus, the muscarinic M2 receptor was down-regulated in vulnerable rats but up-regulated in resilient rats. These findings demonstrated that vulnerable and resilient rats did not have opposing regulatory changes in the same genes; rather, resilience was associated with specific changes in gene expression, with little overlap, compared to the patterns detected in vulnerable rats (Benatti et al., 2012).
It’s what’s up front that counts. Wang et al. (2014) utilized the learned helplessness procedure in mice to examine the role of the mPFC, a brain region strongly implicated in both clinical and animal models of depression, in adaptive behavioral responses to stress. They reported that uncontrollable, inescapable stress induced changes in the excitatory synapses onto a subset of mPFC neurons that were activated as indicated by their increased levels of the immediate early gene, c-Fos. Whereas synaptic potentiation was linked to learned helplessness, a depression-like response, synaptic weakening, was associated with resilience to stress. If the activity of mPFC neurons was enhanced, it was sufficient to convert resilient mice into mice expressing the learned helplessness phenotype.
Employing the learned helpless paradigm, Yang et al. (2015) examined the roles of BDNF and dendritic spine density in various brain regions of LH and resilient Sprague-Dawley rats. BDNF levels in the mPFC and hippocampus (CA3 and dentate gyrus [DG]) were significantly lower in the LH group than in the control and resilient groups, whereas BDNF levels in the NAc in the LH group but not the resilient group were significantly higher than those in the control group. Furthermore, spine density in the prelimbic cortex, CA3, and DG was significantly lower in LH rats compared to rats in the control and resilient groups, (p. 483) although spine density in the NAc was significantly higher in the LH group than in the control and resilient groups. These results suggested that differential regulation of BDNF levels and spine densities, especially in the mPFC, CA1, DG, and NAc areas, may have contributed to resilience to inescapable stress.
Autoreceptors go off-line. Serotonin (5-HT) neurons in the DRN have been implicated in learned helplessness behaviors, such as poor escape responding and expression of exaggerated conditioned fear, induced by acute exposure to an uncontrollable stressor. DRN 5-HT neurons were hyperactive during exposure to an uncontrollable stressor, resulting in desensitization of 5-HT1A inhibitory autoreceptors in the DRN. 5-HT1A autoreceptor down-regulation was thought to induce transient sensitization of DRN 5-HT neurons, resulting in excessive 5-HT activity in brain areas that controlled the expression of learned helplessness behaviors. Regular physical activity has antidepressant/anxiolytic properties and may promote resilience to stressors, but the neurochemical mediators of these effects are not well known (Maier & Watkins, 2010).
Breaking a sweat. A study by Greenwood et al. (2003) examined the effects of 6 weeks of voluntary free wheel running on learned helpless behaviors, uncontrollable stress-induced activity of 5-HT neurons in the DRN, and basal expression of DRN 5-HT1A autoreceptor mRNA in F344 rats. Free wheel running prevented the shuttle box escape deficits and the exaggerated conditioned fear that occurred in response to uncontrollable tail shock in sedentary rats. Furthermore, double c-Fos/5-HT immunohistochemistry revealed that running wheel activity attenuated tail shock–induced activity of 5-HT neurons in the DRN. Six weeks of free wheel running also resulted in a basal increase in 5-HT1A inhibitory autoreceptor mRNA levels in the DRN. These results suggested that free wheel running prevented learned helpless behaviors and attenuated 5-HT neural activity in the DRN during uncontrollable stress. An increase in 5-HT1A inhibitory autoreceptor expression may have contributed to the attenuation of DRN 5-HT activity and the prevention of learned helplessness in physically active rats (Greenwood et al., 2003).
Additional experiments, reviewed in Greenwood and Fleshner (2011), pointed to plasticity at multiple sites within the central 5-HT system that joined together to facilitate resilience to stress. These central sites included the DRN, with enhanced 5-HT1A autoreceptor-mediated inhibition of 5-HT neurons; DRN afferent systems, such as the habenula, BNST, locus coeruleus, or mPFC, which can modulate DRN 5-HT activity during exposure to stressors; and DRN projection sites, including the basolateral nucleus of the amygdala and the striatum, with reduced expression or sensitivity of 5-HT2C receptors in these brain regions that are critical for the expression of stress-induced behaviors. An increase in 5-HT1A autoreceptors and a reduction in sensitivity of 5-HT2C receptors were two prime examples of exercise-induced neuroplasticity.
(p. 484) Loughridge et al. (2013) extended these findings by examining changes in gene expression in the DRN following a 6-week period of voluntary wheel running and then exposure to inescapable stress. Rats that previously exercised had a greater number of genes that were differentially expressed immediately and 2 hours following exposure to inescapable stress compared to sedentary, stressed rats. In addition, modules composed of genes that were highly co-expressed were activated in a more highly coordinated manner following exposure to inescapable stress in active versus sedentary rats. There were 169 genes that were differentially regulated at both post-stressor time points in active versus sedentary rats, and these were included in further analyses. Finally, many of the stress-responsive genes in the DRN were found to be involved in immune-related pathways, including cytokine signaling and inflammatory processes. These included transforming growth factor-β (tgfβ1), the gene that codes for the cytokine TGF-β1, and Tdo2, a gene that codes for an enzyme, tryptophan 2,3 dioxygenase, involved in metabolism of tryptophan, the precursor of serotonin. Thus, voluntary exercise promoted the development of a resilient phenotype in rats later exposed to inescapable stress.
A natural high. In a departure from the learned helplessness model of inescapable stress, Bluett et al. (2017) explored the relationship between brain endocannabinoid (eCB) systems and resilience to chronic intermittent footshock stress. The brain eCB system consists of a presynaptic cannabinoid receptor (CB1R) and two endogenous ligands, 2-arachidonoylglycerol (2-AG) and anandamide (arachidonoylethanolamine, AEA). AEA and 2-AG are synthesized postsynaptically, and when released, travel in a retrograde fashion to bind to CB1Rs, resulting in a decrease in neurotransmitter release from the presynaptic nerve terminal; 2-AG is metabolized presynaptically by monoacylglycerol lipase (MAGL), whereas AEA is degraded postsynaptically by fatty acid amide hydrolase (FAAH). Administration of drugs that inhibit MAGL and FAAH leads to increases in 2-AG and AEA signaling, respectively.
Initially, these investigators developed a behavioral task for detection of drug treatment effects on stress resilience in outbred ICR mice. This novelty-induced feeding task involved exposing mice to a highly palatable food (vanilla-flavored Ensure) first in their home cages, and then several days later in a novel test chamber. Increased latency to feed and decreased consumption of the Ensure in the novel test chamber provided indices of anxiety-like behaviors before and after exposure to acute or repeated footshock stress.
Selective inhibition of MAGL but not FAAH resulted in a resilient phenotype as assessed following footshock stress. In addition, administration of an inhibitor of MAGL enhanced resilience in mice previously classified as stress-susceptible. In contrast, depletion of brain 2-AG levels or blockade of eCB-1Rs increased (p. 485) susceptibility to footshock stress in mice previously classified as stress-resilient. Stress resilience was associated with 2-AG-mediated synaptic suppression at ventral hippocampal-amygdaloid glutamatergic synapses. Finally, depletion of 2-AG selectively within the amygdala prevented the development of resilience to footshock stress. Taken together, the results of this series of experiments clearly pointed to 2-AG signaling within the amygdala as a critical component of the development of a resilient phenotype following exposure to repeated footshock stress (Bluett et al., 2017).
Stress-Enhanced Fear Learning (SEFL)
SEFL has been advanced as a valid animal model of PTSD (Maren & Holmes, 2016). This model combines acute exposure of laboratory animals to a stressor, followed by cued fear conditioning and then tests of extinction. Sillivan et al. (2017) described their exciting results of experiments with the SEFL model in which they employed inbred male and female C57BL/6 mice. Mice exposed to 2 hours of restraint stress 7 days prior to fear conditioning displayed elevated levels of freezing behavior during extinction testing, and this effect was consistently observed in females but was more variable in males. A subgroup of males was identified based upon extinction testing that appeared resilient to the effects of the SEFL paradigm. In contrast to stress-susceptible male mice, stress-resilient male mice did not display elevations in the acoustic startle response (ASR) or in plasma levels of CORT at 30 days post-testing.
To characterize further neuromolecular differences between stress-susceptible and stress-resilient male mice, animals were exposed to restraint stress for 2 hours, followed 7 days later by cued fear conditioning. Thirty days after fear conditioning, mice were exposed to a 5-tone remote memory test, and brains were processed for Fos immunohistochemistry and RNA sequence analysis. Fos was increased in the posterior infralimbic cortex of stress-resilient mice compared to stress-susceptible mice, while the opposite pattern was observed in the basolateral amygdala. These two brain areas are critically involved in the extinction of fear-related memories. The results of RNA sequence analyses in the basolateral amygdala revealed 61 differentially expressed genes between stress-susceptible and stress-resilient mice. Of this total, 9 were decreased and 52 were increased in stress-resilient male mice, including several that have previously been associated with PTSD. The SEFL paradigm and the results from these experiments with C57BL/6 male mice provide another avenue to approach the neural and molecular underpinnings of resilience to traumatic stressors (Sillivan et al., 2017).
(p. 486) Chronic Mild Stress (CMS)
The CMS paradigm varies slightly from laboratory to laboratory, but basic elements are shared in common. It involves exposure of laboratory mice or rats to various mildly stressful environments for extended periods of time across multiple days. Stressors include intermittent illumination, exposure to strobe lights, housing with a strange conspecific, food or water deprivation, exposure to soiled bedding, and so on.
Varying sensitivities. Delgado y Palacios et al. (2011) used a CMS paradigm with laboratory rats, and then characterized the animals as resilient or susceptible based upon their behavior in a two-bottle choice test (water versus 1.5% sucrose). Noninvasive magnetic resonance imaging and spectroscopy were employed to study structural changes in the hippocampus. The results revealed that CMS did not reduce hippocampal volume or alter glutamate metabolism in resilient mice, but did so in susceptible mice.
Bergström et al. (2007) used the CMS paradigm over a 5-week period with adult male Wistar rats. Stress-sensitive and stress-resilient rats were characterized based upon their preference for sucrose in a two-bottle choice test. Approximately 67% of animals displayed a decrease in sucrose preference, while 33% of animals had sucrose preference scores that were comparable to unstressed controls.
Resilient genes. These investigators sought to identify hippocampal gene expression pathways that were associated with resilience to CMS. Based upon microarray data, they found 155 genes that were more than 2-fold differentially regulated between stress-sensitive and stress-resilient rats. In addition, approximately 44 genes were more than 2-fold differentially regulated between unstressed controls and stress-resilient rats. From these totals, 78 of the 155 genes were deemed eligible for generating 6 networks, while 30 of the 44 genes were deemed eligible for generating 3 networks, both by using Ingenuity software. Several cellular functions were revealed by these network analyses, including signal transduction, molecular transport, growth and cellular proliferation, apoptosis, and multiple links to immune functions. The authors emphasized that the stress-resilient phenotype was associated with the capacity to maintain hippocampal neurogenesis (Bergström et al., 2007).
In an earlier study from this same group (Bisgaard et al., 2007), ventral hippocampal proteomes were compared between stress-sensitive and stress-resilient rats and unstressed controls following CMS exposure. Two forms of soluble NSF attachment protein, α-SNAP and β-SNAP, which are involved in vesicular fusion and neurotransmitter release, were not up-regulated in stress-resilient rats but were in stress-sensitive rats following CMS.
Using a 4-week CMS paradigm, Taliaz et al. (2011) examined the role of hippocampal BDNF on resilience in male Sprague-Dawley rats. Lentiviral vectors (p. 487) were employed to induce BDNF over-expression or knockdown within the hippocampus. The findings revealed that hippocampal BDNF expression was strongly associated with resilience to the behavioral effects of exposure to CMS, but had no effect on basal or post-CMS levels of plasma CORT.
DA joins in. Żurawek et al. (2013) compared mesolimbic DA receptors in stress-susceptible and stress-resilient rats after 2 or 5 weeks of CMS. Compared to stress-sensitive and control rats, resilient rats exhibited a down-regulation of D2 receptors as measured by 3H-domperidone autoradiography in the lateral and medial aspects of the striatum, in the shell and core of the NAc, and in the lateral but not the medial portion of the VTA. In contrast, D2 receptor mRNA levels did not differ across brain areas in rats of the three groups. After 5 weeks of CMS, D2 receptor levels of resilient rats were comparable to controls, with the exception of the core of the NAc. In addition, D2 receptor mRNA levels were elevated in several brain areas of resilient rats. These results pointed to an initial blunting, followed by an up-regulation of D2 receptors in stress-resilient rats exposed to 5 weeks of CMS.
Using inbred C57BL/6 mice, Nasca et al. (2015) identified a subpopulation of mice that displayed resilience to the effects of CMS for 28 days and were also distinguishable from stress-susceptible mice after a single 2-hour period of restraint stress. Resilient mice were similar to unstressed controls in their preference for sucrose and in their behavior during a forced swim test. In addition, resilient mice had significantly higher expression of hippocampal mGlu2 metabotropic receptors compared to stress-susceptible mice. In the PFC, mGlu2 expression was reduced significantly compared to unstressed controls in both stress-susceptible and stress-resilient mice (mGlu2 is a presynaptic receptor that inhibits glutamate release and it is regulated in part by CORT signaling via hippocampal MRs). Resilient mice displayed reductions in hippocampal MRs that were key to maintaining high levels of presynaptic mGlu2 receptor signaling in the hippocampus. This hippocampal MR effect is related to inhibition of acetylation of H3K27, an epigenetic transcriptionally active mark bound to the GRM2 promoter gene, which regulates expression of the mGlu2 receptor. Taken together, the different pattern of glutamate signaling in the hippocampus of resilient mice makes possible a more adaptive coping strategy, perhaps by perceiving stressors in less threatening ways (Nasca et al. (2015).
Can you hear me now? Drugan and his collaborators (Drugan et al., 2009; Stafford et al., 2015) have raised the possibility that one can predict in advance which animals will have a resilient phenotype based upon their ultrasonic vocalizations (USVs). USVs occur under a variety of conditions, including aversive and stressful ones. In an initial study with very small sample sizes (Drugan et al., 2009), USVs were recorded from rats exposed to intermittent cold water swim (ICWS) stress, and these same rats were later evaluated for their (p. 488) performance in an instrumental swim escape test (SET). In the SET, rats exposed to ICWS fell into two categories, resilient or vulnerable, based upon good or poor learning, respectively. Four of 16 rats exposed to ICWS emitted far more USVs during the stressor than the remaining 12 rats. Interestingly, in the SET these USV-emitting rats appeared resilient, with escape performance scores comparable to those of controls, while on average the non-USV-emitting rats failed to learn the task.
In a follow-up study (Stafford et al., 2015), rats were tested in a social exploration test of anxiety following exposure to ICWS. The results revealed that those rats that emitted USVs during repeated swim stress were resilient to the anxiety-like behaviors in the social interaction test. Taken together, these two studies demonstrate that USVs may serve as a predictor of stress resilience.
Female Sex Hormones and Resilience
Males and females differ in their resilience profiles as a function of time of exposure to stressors during development, developmental stage when resilience is assessed, and the hormonal state of the individual at the time of assessment (Hodes & Epperson, 2019). Bredemann and McMahon (2014) explored the influence of female sex hormones on resilience to inescapable stress in adult Sprague-Dawley rats. Female rats were ovariectomized (OVX) at 6–8 weeks of age. At 14 days post-OVX, rats received 2 injections of estradiol (E2, 10 µg/250 g, s.c.), separated by 24 hours. Control rats received vehicle alone. For learned helplessness, male and female rats were exposed to 60 inescapable footshocks (0.65 mA, 25–35 sec durations and 15–35 sec intervals) on each of 2 consecutive days. On day 3, rats were placed in a novel shuttle box and tested for escape behavior. A light cued the opening of a door that allowed for escape from a shock. Helplessness was defined as failing on at least 5 of the 10 escape trials. Resilience was defined as more than 5 successful escapes out of 10 trials. Based upon these criteria, 55% of males and 56% of OVX females were “helpless.” OVX rats that were not exposed to inescapable footshocks never performed so poorly in the shuttle box task as to be defined as helpless. Inescapable shock produced a significantly greater incidence of helpless behavior in vehicle-treated compared to E2-treated OVX female rats. In the vehicle-treated females, LTP was absent at CA3–CA1 synapses in hippocampal slices only from helpless rats, and spine density of neurons in CA1 was decreased compared to resilient rats. In contrast, significant LTP was observed in hippocampal slices from E2-treated helpless females. Spine density did not differ between E2-treated helpless and resilient rats, clearly dissociating spine density from magnitude of LTP. E2 replacement also reversed previously established helpless behavior in the shuttle box. These (p. 489) results may be especially relevant for humans, where women are at significantly greater risk of depression than men, especially following menopause.
Mahmoud et al. (2016) utilized female Sprague-Dawley rats to examine the effects of ovarian hormones on behavioral responses to chronic unpredictable stress. OVX or sham surgery was performed at 5 months of age. Four months after surgery, rats were exposed to chronic unpredictable stress (2 sessions per day) for 6 weeks. Compared to sham-operated controls, OVX rats displayed increased immobility with less struggling and swimming in the forced swim test, increased anxiety-like behavior in the novelty-suppressed feeding test, and increased anhedonia based upon a sucrose preference test. OVX also impaired glucocorticoid-dependent negative feedback on the HPA axis. Based upon these behavioral and endocrine findings, the authors concluded that ovarian hormones were key determinants of stress resilience.
Stress Inoculation and Resilience
Early efforts. As summarized in Chapter 7, Seymour Levine and his collaborators undertook a line of research to explore the long-term effects of early life stressors, including maternal separation and exposure to electric shock, on adult patterns of behavioral and physiological responses in laboratory rats (Levine, 1957, 1962; Levine et al., 1956; Levine et al., 1957). These publications and others that followed from the Levine laboratory influenced a generation of researchers and culminated in studies of mechanisms relating to mother-pup interactions and epigenetic changes in the HPA axis (Francis et al., 1999; Liu et al., 1997; Meaney & Szyf, 2005a, 2005b).
From this initial body of research with laboratory rats between birth and weaning at 21 days of age, studies have broadened to include stress inoculation paradigms with juvenile animals and with adults. A strength of these studies is the variety of stressor paradigms that have been employed, adding to the generality of the findings on stress inoculation (Ashokan et al., 2016). These studies in animal models also have great relevance for interventions that might be employed in at-risk children, adolescents, and adults to enhance levels of resilience to later life stressors (Davidson & McEwen, 2012).
Infancy. Mildly stressful stimuli during the first 10 days after birth in mice and rats have been shown to promote resilience to stressful disturbances in adulthood. Moderately increased HPA activity in lactating mothers exposed to stressful stimuli result in elevations in circulating CORT, which can in turn be transmitted to the progeny via lactation. In an experiment by Macrì et al. (2009), lactating CD-1 mice were given low (33 mg/liter) or high (100 mg/liter) doses of CORT in drinking water for the first 10 days after giving birth to their litters. (p. 490) Offspring were then tested in adulthood (approximately 7 months of age). Low levels of CORT delivered to lactating mothers via the drinking water resulted in improved cognitive function and increased levels of natural auto-antibodies directed to the serotonin transporter in their adult offspring. In contrast, high levels of CORT provided to the mother resulted in reduced hippocampal BDNF levels and high levels of natural auto-antibodies directed toward the DA transporter. These results indicated that low to moderate levels of maternal CORT during the first 10 days after birth may enhance resilience in their adult offspring. In contrast, high levels of maternal CORT during the first 10 days after birth may result in greater susceptibility of adult offspring to stressful stimulation.
Match or mismatch? Branchi et al. (2013) tested the match-mismatch hypothesis relating to early life stress as described by Schmidt (2011). Briefly, early life adversity prepares animals and humans for a hostile environment in adulthood. If the early environment matches with characteristics of the adult environment, animals will be resilient. In contrast, if the early environment and the adult environment are a mismatch, then animals will be susceptible to the deleterious effects of a stressful adult environment.
C57BL/6 mice were reared under standard laboratory conditions or in a communal nest (CN) where three mothers and their litters were housed together. Communal nesting provided for increased social interactions between mothers and pups as well as between pups from different litters. In adulthood, control and CN mice were exposed for 4 weeks to either forced swim stress (10 min per day at 21oC) or a social stressor (social group disruption each day). CN mice were more resilient to the social stressor compared to control mice based upon a lack of anhedonia and lower plasma CORT levels. In contrast, mice of both groups were susceptible to the negative effects of swim stress. Resilience to the social stressor in adulthood was matched to the quality of the CN environment, which was also in the social domain (Branchi et al., 2013).
In two further tests of the match-mismatch hypothesis from Schmidt’s laboratory, the first with BALB/c female mice (Santarelli et al., 2014) and the second with BALB/c male mice (Santarelli et al., 2017), a 2 × 2 experimental design was employed such that litters were reared under positive or negative conditions and adult animals from these litters experienced favorable or unfavorable environments prior to testing of behavioral, endocrine, and molecular responses to a battery of tests.
The results from these studies supported the match-mismatch hypothesis in that an impoverished early environment (limited nesting material) only exerted negative effects on stressful responses when these same male or female animals were reared in a favorable environment in adulthood. Similarly, a stressful adult environment exerted greater disruptions in male and female mice that had experienced a favorable early rearing environment (neonatal handling). Mice from (p. 491) these two mismatch conditions were clearly at a disadvantage compared to mice from the two matched conditions, including the group that experienced an unfavorable early rearing environment plus an unfavorable environment in adulthood (Santarelli et al., 2014, 2017).
Strained comparisons. Binder et al. (2011) took advantage of the fact that inbred mouse strains vary in their responses to early life stressors. Their experiments included litters of four inbred mouse strains (129S1/SvlmJ, C57BL/6J, DBA/2J, and FVB/NJ) that were exposed to a single 24-hour period of maternal separation beginning on PND 9 or were left undisturbed (controls). Separated litters were kept together and the temperature of the home cage was maintained with a heating pad. All animals were tested using a forced swim test and a hole board test. Despite exposure to a 24-hour period of maternal separation (MS), most animals seemed to be resilient to this early life stressor. However, one compelling finding was the long-lasting, strain-specific effect of maternal separation on 129S1/SvImJ mice, which resulted in increased depression-like behaviors in a forced swim test and elevated anxiety-related behavior in the hole board test. In contrast, mice of the C57LB/6J, DBA/2J, and FVB/NJ strains were largely unaffected by maternal separation.
Handle with care. Stiller et al. (2011) examined the effects of early weaning or daily handling of litters on resilience to stress in adulthood. Litters of Sprague-Dawley rats were exposed to one of three rearing conditions: (1) unhandled controls, (2) early weaning at postnatal day 16, with individual housing thereafter, or (3) maternal separation for 3 hours/day from postnatal days 2 to 14. Behavioral tests included: (1) an emergence test at approximately 60 days of age, where individuals were placed into a black Plexiglas box for 1 hour the day before testing, and on the day of testing, after 15 minutes in the box, the holding box door was opened and the rat could explore an open field arena for 30 minutes; and (2) at least 2 weeks later, rats were exposed to 80 5-second swim stress bouts with water maintained at 23oC, and the following day were given swim escape trials using a warning tone, with a lever available to terminate the swim stress. After the swim escape trial, rats were decapitated and blood was collected for measurement of CORT and spleens were collected for measurement of con-A stimulation of lymphocyte proliferation in vitro. Resilient rats were defined as the top 33% of the distribution of the swim escape scores (i.e., shortest escape latencies: 30 males and 25 females) and vulnerable rats were defined as the bottom 33% of the distribution of swim escape scores (i.e., longest escape latencies: 30 males and 25 females).
Results revealed that stress-vulnerable rats were characterized by increased anxiety-like behaviors, greater post-stress plasma levels of CORT, and a higher con-A-induced T-cell proliferation response compared to resilient rats. The early weaning and social isolation treatment was a contributing factor in predicting (p. 492) total escape time, but maternal separation was not. Offspring that were separated from their mothers did have significantly elevated levels of plasma CORT compared to control offspring. There was no evidence that handling with maternal separation enhanced resilience when animals were tested in adulthood (Stiller et al., 2011).
Inoculating monkeys. A captive colony of squirrel monkeys (Saimiri sciureus) maintained at Stanford University was utilized to explore the effects of stress inoculation in a nonhuman primate species. Squirrel monkeys live in large social groups that include adult males and females as well as infants and juveniles. The gestation period in this species is 5 months and weaning occurs between 6 and 8 months of age (Zimbler-Delorenzo & Stone, 2011).
Parker et al. (2004) tested whether exposure to moderate stress in infancy produced later stress resilience in squirrel monkeys. Twenty squirrel monkeys were randomized to intermittent stress inoculation (removal from their mothers and natal group for 1 hour per week for 10 weeks) or the control condition (monkeys remained with their mothers in their natal groups throughout the 10-week period). At postnatal week 35, each mother-offspring dyad underwent testing in a moderately stressful novel environment and measures of offspring anxiety (i.e., maternal clinging, mother-offspring interactions, object exploration, and food consumption) and stress hormone concentrations were obtained. At postnatal week 50, after acclimation to an initially stressful wire-mesh box attached to the home cage, independent young monkeys underwent testing for measures of anxiety. In the novel environment, previously stressed monkeys demonstrated diminished anxiety as measured by decreased maternal clinging, enhanced exploratory behavior, and increased food consumption compared to controls. Mothers of stressed offspring accommodated offspring-initiated exploration, and served as a secure base of attachment more often than control mothers. Compared to control offspring, previously stressed offspring had lower basal plasma levels of ACTH and CORT, and both hormones remained lower following stress. In the wire-box test in the home cage, previously stressed offspring exhibited enhanced exploratory and play behaviors compared with control offspring. These results provided prospective evidence that moderately stressful early experiences strengthen socio-emotional and neuroendocrine responses to stressors later in life.
A subsequent series of experiments (Parker et al., 2006) revealed that the effects of early exposure of squirrel monkey infants to brief episodes of stress were not mediated by differences in maternal care. Taken together, these findings indicate that stress inoculation in squirrel monkeys included repeated exposure to mild stressors early in life that in turn stimulated mild levels of anxiety and activated the HPA axis. These early experiences then promoted behavioral and physiological resilience to stressors that endured well into adulthood (Lee et al., 2016; Lyons & Parker, 2007; Lyons et al., 2009, 2010).
(p. 493) Adolescence. Suo et al. (2013) used a predictable chronic mild stress procedure (5 minutes of restraint stress daily for 28 days) in adolescent rats (postnatal days 28–55) to assess its ability to enhance resilience to stressful stimuli in adulthood (postnatal days 63–83). Behavioral assessments in adulthood included a sucrose preference test, a novelty-suppressed feeding test, an EPM test, and a forced swim test. In addition, the role of mammalian target of rapamycin (mTOR) signaling in brain was measured during behavioral testing of adult animals. Compared to controls, rats repeatedly exposed to a mild stressor during adolescence exhibited no change in sucrose preference, a reduced latency to feed in the novelty-suppressed feeding test, increased time spent in the open arms of the EPM, and less time spent immobile in a forced swim test. Either systemic administration or intra-PFC infusion of the mTOR inhibitor, rapamycin, completely blocked the behavioral effects produced by chronic mild stress in adolescence. Finally, chronic mild stress during adolescence blocked the effects of exposure to chronic unpredictable stress in adulthood.
Kendig et al. (2011) exposed groups of four male Wistar rats (33–57 days of age) to cat fur for 30 minutes in a large arena once every other day for 24 days. Controls were placed in an empty arena without cat fur. Testing occurred from days 58 to 77. When exposed to cat fur, 3–4 rats would huddle together in the corner of the arena. This huddling behavior and suppression of activity occurred throughout the 12 exposures. In adulthood, predator-stressed rats displayed significantly less immobility in a forced swim test and exhibited increased sociability with a novel conspecific in a social interaction test. These findings suggest greater resilience in adulthood following stress exposure in adolescence.
Adulthood. The ability to control the onset, duration, intensity, frequency, or offset of a stressful stimulus may have long-lasting effects on reducing the negative impact of subsequent exposure to stressful stimuli, including those that cannot be controlled.
The illusion of control. Remarkably, as reviewed by Maier and his colleagues (Maier, 2015; Maier et al., 2006; Maier & Watkins, 2010), laboratory rats that have experienced control over electric shocks to the tail have reduced neural and behavioral responses to an uncontrollable stressor presented up to two months later. Indeed, the animals responded as if the subsequent stressor was actually controllable. It appeared that prior exposure to a controllable stressor subserved a stress inoculation function, even when the subsequent stressor was dramatically different from the initial controllable stressor, such as exposure to CSDS.
Exposure to a controllable stressor activated neurons in the ventral-medial PFC (vmPFC), which exerted inhibitory control over limbic and brainstem structures, including the dorsal raphe nucleus. In an elegant series of experiments from their laboratory, Maier and Watkins (2010) demonstrated that neural activation of the vmPFC during controllable stress was required for (p. 494) inhibiting activity of serotonin neurons within the DRN, and inoculation to the deleterious behavioral effects of exposure to a subsequent inescapable stressor 1 week later. Similarly, pharmacological activation of vmPFC input to the DRN during exposure of laboratory rats to inescapable tail shock resulted in an inoculation effect as if the inescapable tail shock was actually controllable. Finally, pharmacological blockade of vmPFC input to the DRN prevented the stress inoculation effect following exposure of rats to a controllable stressor. These results have far-reaching implications for understanding mechanisms of stress inoculation, particularly as it relates to resilience to the effects of traumatic stressors. Lucas et al. (2014) have made similar arguments regarding the impact of stressor controllability on enhancing stress resilience in laboratory rats. In addition, Kerr et al. (2012) supported the work of Maier and Watkins (2010) in their functional magnetic resonance imagining (fMRI) studies of the human vmPFC and anticipation of control over presentation of aversive video clips.
The social scene. Brockhurst et al. (2015) employed a stress inoculation paradigm that entailed intermittent exposure of C57BL/6 adult male mice to a mild social stressor. Mice were randomized to stress inoculation or a control treatment condition and were assessed for plasma CORT responses and behavior during open field, object-exploration, and tail-suspension tests and repeated restraint stress. Every other day for 21 days, mice in the stress inoculation condition were removed from their home cages and individually placed for 15 minutes behind a wire mesh barrier in the cage of a retired Swiss-Webster male breeder. Each subject was repeatedly exposed to the same Swiss-Webster breeder. The wire mesh barrier prevented fighting and wounding during all 11 stress inoculation sessions, but allowed visual and olfactory interactions. Mice were immediately returned to their home cages after each of the 11 sessions. Control mice remained undisturbed except for intermittent human handling during routine cage cleaning.
Stress inoculation training sessions that acutely increased plasma levels of CORT diminished subsequent immobility as a measure of behavioral despair on tail-suspension tests. Stress inoculation also decreased subsequent freezing in an open field arena despite comparable levels of thigmotaxis in mice from both treatment conditions. Stress inoculation subsequently decreased novel object exploration latencies and reduced plasma CORT responses to repeated restraint stress. These results demonstrated that stress inoculation acutely stimulated glucocorticoid signaling, and then enhanced subsequent indications of active coping behavior in mice. Stress inoculation training reflected an experience-dependent process in adulthood that resembled in part interventions designed to build resilience in humans (Brockhurst et al., 2015).
An enriched experience. Lehmann and Herkenham (2011) took a novel approach to promote resiliency in C57BL/6 male mice prior to exposure to CSDS. They housed individual mice in a standard cage (SE) or an enriched (EE) or (p. 495) impoverished environment (IE) for 3 weeks, followed by 2 weeks of CSDS as described previously. SE was a polycarbonate cage with bedding chips and a cardboard tunnel. An EE consisted of a polycarbonate cage with bedding chips, nest material, running wheels, and tubes of different shapes and sizes. An IE consisted of a polycarbonate cage with bedding chips. EE-housed mice but not IE- or SE-housed mice exhibited increased levels of FosB/ΔFosB immunostaining in the infralimbic, prelimbic, and anterior cingulate cortices; amygdala; and NAc following CSDS. EE mice also exhibited a resilient behavioral phenotype following CSDS. Discrete lesions of the infralimbic cortex prior to EE blocked behavioral resilience to CSDS and reduced FosB/ΔFosB expression in the NAc and amygdala while increasing expression in the paraventricular nucleus. These findings were consistent with enhanced output of neurons in the vmPFC playing a critical role in stress resilience.
Charney and his colleagues made the astute observation that resilience could be studied in laboratory animals using the tools of molecular biology and neuroscience. Prior to this, many studies of resilience had focused on psychological characteristics of children that survived and even thrived under adverse circumstances.
A consistent finding from different laboratories has been the wide range of behavioral and biological responses of laboratory animals following exposure to an acute stressor or to a chronic intermittent stressful paradigm. The variability of responses has been described in outbred as well as inbred strains of laboratory mice and rats. These findings match well with observations of humans who survived highly stressful or traumatic experiences and came through relatively unscathed psychologically.
Several stress paradigms have been employed to study neurobiological mechanisms of resilience, including two that score high on ethological validity: PSS and CSDS. Experiments employing these and other stressor paradigms have revealed a host of molecular alterations in stress-sensitive brain circuits that contribute to a resilient phenotype. A subset of these changes are not merely polar opposite changes from animals that displayed susceptibility to the stressors; rather, there are specific changes that are essential for a resilient phenotype. These studies provide new approaches to the etiology of stress-related mental disorders, especially depression and PTSD. Instead of focusing solely on the development of therapies that target stress-sensitive brain circuits, an alternative approach in the near future may be to develop therapies that up-regulate stress-resilient brain circuits.
(p. 496) Another major research area relates to inoculation of animals against the deleterious effects of a later intense stressor. Experiments were reviewed that included stress inoculation of animals during infancy, adolescence, or adulthood. There was compelling evidence of the positive effects of stress inoculation on the capacity of animals to adapt favorably to stressful stimulation later in life. These experiments have a direct bearing on approaches to training military personnel for deployment to combat zones and to training first-responders for the challenges of dealing effectively with casualty events following accidents or terrorist attacks.
Successful adaptation in the face of adverse circumstances involves a complex and interacting array of alterations in brain circuits, peripheral hormonal systems, innate and adaptive immune cells and signaling molecules present in the circulation, alterations in gut microbiota, and subtle alterations in the blood-brain barrier. A thorough understanding of how these stress-responsive peripheral and central systems interact in susceptible and resilient animals and humans may lead to the development of novel therapeutic strategies to treat a host of mental disorders, including depression and PTSD (Cathomus et al., 2019).