The predator odor avoidance model of post-traumatic stress disorder in rats

Abstract

Individuals with post-traumatic stress disorder (PTSD) avoid trauma-related stimuli and exhibit blunted hypothalamic-pituitary-adrenal (HPA) axis response at the time of trauma. Our laboratory uses predator odor (i.e., bobcat urine) stress to divide adult Wistar rats into groups that exhibit high (Avoiders) or low (Non-Avoiders) avoidance of a predator odor-paired context, modeling the fact that not all humans exposed to traumatic events develop psychiatric conditions. Male Avoiders exhibit lower body weight gain after stress, as well as extinction-resistant avoidance that persists after a second stress exposure. These animals also show attenuated HPA axis response to predator odor that predicts subsequent avoidance of the odor-paired context. Avoiders exhibit unique brain activation profiles relative to Non-Avoiders and controls (as measured by Fos immunoreactivity), and higher corticotropin-releasing factor levels in multiple brain regions. Furthermore, Avoider rats exhibit escalated and compulsive-like alcohol self-administration after traumatic stress. Here, we review the predator odor avoidance model of PTSD and its utility for tracking behavior and measuring biological outcomes predicted by avoidance. The major strengths of this model are 1) etiological validity with exposure to a single intense stressor, 2) established approach distinguishing individual differences in stress reactivity, and 3) robust behavioral and biological phenotypes during and after trauma.

Introduction

Post-traumatic stress disorder (PTSD) can develop after exposure to a traumatic event that is more extreme than a typical daily stressor. Events that may lead to PTSD include, but are not limited to, violent personal assaults, natural or human-caused disasters, accidents, combat, and other forms of violence (National Institute of Mental Health, 2017). Even if the majority of people experience at least one traumatic event during lifetime, only a subset ultimately develops PTSD (Breslau, 2009). Based on diagnostic interview data from National Comorbidity Survey Replication, it was estimated that 3.6% of the US adult population (more than 11 million Americans) suffer from PTSD in a given year. Importantly, the prevalence was higher for females (5.2%) in comparison to males (1.8%) (Harvard Medical School, 2007). Settings of perpetual war and military conflict have higher rates of PTSD, for example, in Algeria (37%), Cambodia (28%), Ethiopia (16%) and Gaza (18%) (De Jong et al., 2001).

Although considered a common, costly, and frequently debilitating psychiatric condition, PTSD was not defined as a disorder until the third edition of the Diagnostic and Statistical Manual of Mental Health Disorders (DSM) was published in 1980. This term replaced other labels such as shell shock and war neurosis that had appeared in the literature for years (Trimble, 1985; Crocq and Crocq, 2000). The diagnosis was intended to identify severe and persistent fear responses in individuals who had previously been exposed to a traumatic event, rather than an inherent weakness or failing. A traumatic event was conceptualized as a catastrophic stressor that was outside the range of usual experiences, such as war, torture, rape, the Nazi Holocaust, the atomic bombings of Hiroshima and Nagasaki, natural disasters (earthquakes, hurricanes, volcano eruptions) and human-made disasters (factory explosions, airplane crashes, automobile accidents) (Wilson, 1994).

Over the years, there have been significant changes to the diagnosis of PTSD. For example, in addition to expanding the types of events considered traumatic, the DSM-IV (1994) diagnostic criteria included symptoms from three clusters: intrusive recollections, hyper-arousal and avoidant/numbing symptoms. The DSM-IV also required symptoms 1) to be present for one month for PTSD diagnosis and 2) to be associated with significant distress or functional impairment (American Psychiatric Association, 1994). In the latest revision, the DSM-5 (2013), the definition of PTSD was expanded to separate the avoidance criterion, and PTSD was moved from the category of anxiety disorders to a mental health category titled Trauma- and Stressor-Related Disorders. All of the conditions included in this new classification require exposure to a traumatic or otherwise adverse environmental event as a diagnostic criterion (American Psychiatric Association, 2013).

There are four clusters of PTSD symptoms identified in the DSM-5 (2013): intrusive recollection of the original traumatic event, avoidance of trauma-related reminders, negative alterations in cognition and mood, and alterations in arousal or reactivity (American Psychiatric Association, 2013), each of which must start or be significantly exacerbated after exposure to the traumatic event. Each of these behavioral symptom clusters can result in a PTSD diagnosis, but avoidance may be the most detrimental to psychosocial functioning and, in its most extreme manifestation, can prevent individuals from leaving their home for fear of confronting trauma reminders (Hendrix et al., 1994; Malta et al., 2009). Furthermore, baseline avoidance symptoms are strong predictors of future PTSD severity and chronicity (Bryant et al., 2000; Perkonigg et al., 2005; Shin et al., 2015).

The ability to examine the neurobiology of PTSD and identify markers of risk and treatment outcomes in humans is limited. Thus, pre-clinical models are needed to study neurobiological alterations after traumatic stress and to identify biomarkers of pathology and treatment response, as well as therapeutic targets that can be leveraged to develop more effective treatment and prevention strategies (Pitman et al., 2012; Gilpin and Weiner, 2017). Several animal models have been developed and used to understand the pathophysiology of PTSD, and each one mimics a particular subset of symptoms associated with the distinctly human disorder (Whitaker et al., 2014).

Here, we will focus on a specific predator odor exposure model, since exposure to predator odor stress produces behavioral, molecular, and physiological alterations that recapitulate many of the same alterations observed in PTSD patients (Cohen et al. 2012a). The goal of this review is to present and discuss findings from a series of studies employing a predator odor (i.e., bobcat urine) stress conditioned place aversion (CPA) procedure to identify individual’s response to trauma-related stimuli in adult male Wistar rats. Over the last several years, we have examined the validity of this model, as well as its utility as a tool for investigating 1) post-trauma adaptations, including avoidance, anxiety-like behavior, hyperarousal, hyperalgesia, and high alcohol drinking, and 2) to what extent, individual differences in stress reactivity are related to neuroendocrine dysfunction and alterations in specific neurobiological substrates.

Predator odor exposure as a model of traumatic stress

PTSD is a highly heterogeneous pathology characterized by broad alterations in arousal, reactivity, mood, social functioning, and cognitive function. Although there is no animal model able to simultaneously capture all PTSD symptomatology, specific behaviors and physiological responses can be assessed in rodents; these include anxiety, hyperarousal, disruption of sleep and circadian cycle, and avoidance or fear of trauma-related cues. In recent years, many pre-clinical models of PTSD have been described (Berardi et al., 2012; Cohen et al., 2012a; Daskalakis et al., 2013b; Goswami et al., 2013). Among the stressors frequently used, exposure to predator threat (Adamec et al., 2004; Zoladz et al., 2008, 2013) or predator odor (Zohar et al., 2008; Mackenzie et al., 2010; Cohen et al., 2012b) has received considerable interest. Predator stress models use an etiologically valid stressor and, in many models, a single stress exposure is sufficient to induce a range of behavioral and physiological responses (Adamec and Shallow, 1993; Adamec, 1997; Adamec et al., 2006a,b; Kozlovsky et al., 2007a,b, 2009; Nanda et al., 2008; Qi et al., 2010). Rodents are exposed to predator stress in a variety of ways including either unprotected exposure, exposure with a physical barrier, or exposure to a predator scent (Deslauriers et al., 2018).

Odors from felines, canines, and other predators elicit innate reactions in rodents, including stereotyped avoidance behavior and stimulation of the hypothalamic-pituitary-adrenal (HPA) axis that coordinates sympathetic stress responses (Apfelbach et al., 2005). Our laboratory has developed a stress model that utilizes bobcat urine as the stressor. Bobcat urine, as well as urine from numerous carnivores, contains the biogenic amine 2-phenylethylamine, a metabolite of phenylalanine, which is an essential amino acid found in dietary protein (Kaufman, 1999; Ferrero et al., 2011). Importantly, 2-phenylalanine activates specific receptors within the rodent olfactory cortex, the trace amine-associated receptor 4 (TAAR4), and can elicit avoidance behavior in rats and mice (Ferrero et al., 2011). In addition, detection of bobcat urine or the component of fox feces 2,3,5-trimethyl-3-thiazoline (TMT) by a specific area of the olfactory cortex, the amygdalo-piriform transition area, is responsible for increases in circulating stress hormones after exposure to volatile predator odor (Kondoh et al., 2016).

It is important to note that TMT exposure produces fear-like behavioral and autonomic responses in rodents (Fendt et al., 2005), but TMT does not generate single-trial conditioning of behavioral responses (e.g., avoidance, freezing) during subsequent re-exposure to TMT-paired stimuli in the absence of odor (Wallace and Rosen, 2000; McGregor et al., 2002; Blanchard et al., 2003); this may be due to the need for repeated TMT contextual conditioning sessions to generate conditioning of behavioral fear responses (Fendt and Endres, 2007). Also, its powerfully acrid nature raises the possibility that TMT may act as a noxious, rather than a specific predator odor (Dielenberg and McGregor, 2001).

In order for an animal model of traumatic stress to be useful in recapitulating potentially PTSD-related phenotypes, 5 key criteria must be fulfilled (Yehuda and Antelman, 1993): 1) even very brief stressors should be capable of inducing biological and behavioral sequelae of PTSD, 2) the stressor should be capable of producing the PTSD-like sequelae in an intensity-dependent manner, 3) the stressor should produce biological alterations that persist over time or become more pronounced with the passage of time, 4) the stressor should induce biobehavioral alterations that have the potential for bidirectional expression, and 5) there should be inter-individual variability in response to a stressor.

Our predator odor avoidance model satisfies 4 of the 5 key proposed criteria. First, our paradigm induces robust behavioral and biological phenotypes after exposure to a single stressor (Edwards et al., 2013; Whitaker and Gilpin, 2015). Furthermore, the consequences of predator odor exposure persist over time (up to 6 weeks) and are resistant to extinction (Edwards et al., 2013; Schreiber et al., 2018). The effects of predator odor exposure can be either excitatory (e.g., heightened arousal; Roltsch et al., 2014) or inhibitory (e.g., avoidance; Edwards et al., 2013). Finally, our predator odor model shows marked inter-individual variability in stress reactivity (Edwards et al., 2013). We have neither tested yet whether bobcat urine exposure is capable of producing PTSD-like symptoms in an intensity-dependent manner, nor differences between rat strains or rodent species. Finally, our model satisfies a 6^th criterion proposed by Whitaker et al. (2014), which is the ability to produce co-morbid conditions that include escalation of alcohol self-administration, more compulsive-like alcohol responding (Edwards et al., 2013), and increased nociception (Roltsch et al., 2014; Itoga et al., 2016) (Table 1).

Table 1.

Evaluation of the bobcat urine conditioned avoidance model as it relates to PTSD phenomenology, according to the 5 criteria introduced by Yehuda and Antelman (1993) and the 6^th criterion added by Whitaker et al. (2014).

Criterion	Bobcat urine conditioned avoidance model
1. Brief stressor capable of inducing biological and behavioral responses of PTSD	✓
2. Stressor produces PTSD-like symptoms in an intensity-dependent manner	?
3. Stressor produces biological alterations that persist over time	✓
4. Stressor induces biobehavioral alterations that have the potential for bidirectional expression	✓
5. Presence of inter-individual variability in response to the stressor	✓
6. Stressor produce symptoms of comorbid conditions (e.g., high alcohol drinking, hyperalgesia)	✓

The bobcat urine avoidance model of traumatic stress

Individuals exposed to stressful stimuli exhibit varying response; some develop clear psychiatric conditions, whereas others continue to function normally. Retrospective and prospective epidemiological studies indicate that most individuals affected by traumatic experiences adapt within a period of 1 to 4 weeks following exposure and a small proportion develop delayed or prolonged responses to trauma (Bryant, 2006; Foa et al., 2006). Cohen et al. (2006) established cut-off behavioral criteria to classify susceptible versus resilient rats based on a composite of extreme reductions in exploratory behavior and increased arousal 7 to 90 days after exposure to cat urine for 10 minutes. In our laboratory, we use a similar approach to identify high and low stress-reactive individuals based on avoidance (or lack thereof) of a context previously paired with bobcat urine odor. It is important to emphasize that in our model, rats are indexed for avoidance of a predator odor-paired context 24 hours after stress, whereas other traumatic stress and/or avoidance assays often test avoidance (also called escape, darting, etc.) during stress exposure (Wallace and Rosen, 2000; Holmes and Galea, 2002; Gruene et al., 2015; Homiack et al., 2017).

Our conditioning apparatus is constructed from wood and Plexiglas and consists of three large chambers (36 cm length × 30 cm width × 34 cm height) with different types of floor texture (circles, grid or rod floor) and patterned walls (circles, white or stripes), separated by a small triangular connecting chamber (Fig. 1). Three guillotine doors confine rats to individual large chambers on neutral (pseudo-conditioning) and odor conditioning days; these doors can be removed to allow free exploration during pre- and post-conditioning tests. All testing is conducted under indirect dim illumination (one 60W white light facing the wall) during the dark phase of the light-dark cycle (in our lab rats are kept on a reverse light-dark cycle). It is not known whether our observed effects might differ if rats are tested during the light phase and/or if they are housed on a regular light-dark cycle.

Figure 1.

Schematic representation of bobcat urine conditioned place aversion procedure. On day 1, rats are allowed free movement between the three chambers for 5 min. The chamber in which each rat exhibits the most deviant time score (i.e., highly preferred or highly avoided) is excluded from all future sessions for that rat. On day 2, rats are allowed to explore the two non-excluded conditioning chambers for 5 min. On day 3, rats are placed in one of the chambers with the guillotine door shut without odor for 15 min. On day 4, rats are placed in the alternative chamber with the guillotine door shut and a bobcat urine soaked-sponge under the floor for 15 min, or no odor for control animals. On day 5, rats are again allowed to explore the two conditioning chambers for 5 min. Avoidance is calculated as a difference score between time spent in odor-paired context (day 5) and pre-conditioning time spent in odor-paired context (day 2). Rats that displayed a > 10 s decrease in time spent in odor-paired context are classified as Avoiders. Rats that displayed a < 10 s decrease in time spent in the predator odor-paired chamber are classified as Non-Avoiders.

Our current protocol consists of a 5-day CPA procedure modified from the classical 4-day CPA paradigm previously described (Edwards et al., 2013). The objective of introducing an additional session was to reduce the effect of spontaneous preference or aversion for one of the compartments before the conditioning phase. Rats are subjected to 1 week of handling prior to initiating the experimental protocol described in Fig. 1. On the first day, the animal is placed in the triangular chamber of the apparatus with all the guillotine doors opened to allow free movement between the three large chambers for 5 minutes (3-chamber pre-test session). The session is videotaped and time spent within each chamber is scored. The apparatus is thoroughly cleaned between animals with Quatricide® PV at a dilution of 1:64 (Pharmacal Research Labs, Waterbury, CT, USA). For each rat, the chamber that exhibits the most deviant time score of the three (i.e., highly preferred or highly avoided) is excluded from all future sessions for that rat. On day 2, the rat is allowed 5 minutes to explore the two non-excluded conditioning chambers (pre-conditioning session). The session is videotaped and time spent in each chamber is scored. After this session, rats are assigned to predator odor stress or unstressed control groups that are counterbalanced for magnitude of baseline preference for one chamber versus the other (i.e., groups are assigned such that mean pre-existing preference for each of the two chambers is approximately zero for stress and unstressed groups). For rats in the stress group, an unbiased and counterbalanced design is used to determine which chamber (i.e., more preferred or less preferred) will be paired with predator odor for each rat. On day 3, the rat is placed in one of the chambers with the guillotine door shut without odor (neutral conditioning session) for 15 minutes. On day 4, the rat is placed in the alternative chamber with the guillotine door shut and a sponge soaked with 3 ml of bobcat urine (Lynx rufus; Maine Outdoor Solutions, Hermon, ME, USA) placed under the floor (odor conditioning session) for 15 minutes. Control rats are treated identically to odor-exposed rats, but the sponges do not contain bobcat urine. Following this session, the room is thoroughly deodorized with PureAyre Odor Eliminator (Clean Earth Inc., Kent, WA, USA). On day 5, rats are again allowed to explore the two conditioning chambers in a 5-minutes video recorded (post-conditioning session), and time spent in each chamber is scored.

Avoidance is calculated as a difference score between time spent in odor-paired context (day 5; post-conditioning session) and time spent in odor-paired context before conditioning (day 2; pre-conditioning session). Rats that display a > 10 seconds decrease in time spent in odor-paired context are classified as Avoiders. Rats that display a < 10 seconds decrease in time spent in the predator odor-paired chamber are classified as Non-Avoiders. This group division criterion was initially chosen because not many animals exhibited avoidance scores approximately equal to 10 seconds (although this certainly happens in some animals in some cohorts). The average change in time spent in odor-paired chamber observed across studies in male Avoider rats is −38.26 ± 1.66 seconds, whereas Non-Avoiders actually spend more time in this context (+17.30 ± 1.18 seconds). Besides that, we intentionally selected a cutoff criterion that represented a data point that would remain constant across experiments and studies rather than a cohort-specific criterion that was likely to change across experiments.

Behavioral effects in the bobcat urine conditioned avoidance model

It is well established that only some individuals exposed to trauma develop PTSD, suggesting that individual differences (i.e., risk and resilience factors) may be critical for determining vulnerability to traumatic experiences (Yehuda et al., 2011). The overall prevalence rate of adult male Wistar rats categorized as Avoiders using our 5-day CPA procedure is approximately 35%. Importantly, post-stress differences in stress reactivity do not appear to be caused by differences in learning and memory (Edwards et al., 2013). For example, when tested in an appetitive classical conditioning procedure in which sweet drinking solution is paired with one context and water with the other, unstressed controls, Avoiders and Non-Avoiders do not systematically differ in their preference for the sweet solution-paired chamber. Moreover, sweet-solution preference does not predict subsequent avoidance of a predator odor-paired chamber (Edwards et al., 2013). Both Avoiders and Non-Avoiders also acquire operant conditioning similarly prior to stress exposure (Edwards et al., 2013). One possibility is that Non-Avoiders may not effectively perceive the odor, but this is unlikely since both Avoiders and Non-Avoiders exhibit enhanced anxiety-like behavior 2 and 5 days after stress, relative to unstressed controls (Whitaker and Gilpin, 2015).

Although the prevalence of PTSD among adults is higher for females than for males (National Institute of Mental Health, 2017), few animal models of PTSD have considered sex differences and most of them typically use male subjects (Deslauriers et al., 2018). In order to explore sex-related traumatic stress-responses more thoroughly, we have recently started testing female adult Wistar rats (alongside males) in the predator odor stress CPA protocol. Our results indicate that approximately 50% of females become Avoiders, which is consistent with higher PTSD prevalence rates in women (National Institute of Mental Health, 2017). It will be interesting to test for escalated and compulsive-like alcohol drinking in female Avoiders after bobcat urine exposure, since alcohol use disorder (AUD) is actually less prevalent in women than it is in men (Erol and Karpyak, 2015).

In agreement with previous reports (Korte and De Boer, 2003; Cohen et al., 2013), predator-odor exposure increases anxiety-like behavior in stressed male rats. In our model, both Avoiders and Non-Avoiders exhibit reduced time in the open arm of the elevated plus-maze and reduced time in the center of the open field when tested 2 and 5 days after stress (Whitaker and Gilpin, 2015). However, time spent in the open arms of the elevated plus-maze before exposure to stress does not predict the avoidance of the predator odor-paired context (Edwards et al., 2013). Importantly, avoidance of predator odor-paired stimuli is highly persistent and stable; avoidance lasts at least 6 weeks post-stress in Avoiders, and avoidance behavior measured 24 hours post-stress positively and significantly correlates with avoidance of the same chamber 8 days and 6 weeks later (Edwards et al., 2013; Itoga et al., 2016).

PTSD patients usually exhibit persistence of the traumatic memory for months, years or decades after the traumatic event (Careaga et al., 2016), and usually exhibit impairment of fear extinction in conditioning tasks (Blechert et al., 2007). Similarly, avoidance in our model is resistant to extinction, since male Avoiders continue to meet behavioral criterion for classification as Avoiders following 13 re-exposures to the bobcat urine-paired chamber once per day in the absence of odor (Schreiber et al., 2018). Finally, avoidance behavior is stable in the sense that a second exposure to predator odor does not significantly affect magnitude of avoidance or proportion of Avoiders, i.e., Avoiders remain Avoiders and Non-Avoiders remain Non-Avoiders after a second odor conditioning session (Schreiber et al., 2018).

Humans with stress-related disorders exhibit increases in arousal, as well as altered pain processing (Engdahl et al., 1998; Sartor et al., 2010; Norrholm et al., 2011; Roy et al., 2012), and traumatic stress disorders are highly comorbid with chronic pain (Shipherd et al., 2007; Asmundson and Katz, 2009; Lew et al., 2009; Moeller-Bertram et al., 2012). Predator odor stress produces hyper-arousal 24 to 48 hours post odor-exposure (Roltsch et al., 2014). This finding corroborates previous reports showing that predator stress produces lasting increases in startle reactivity in rats and mice (Adamec et al., 2010; Clay et al., 2011). Several studies have reported lower experimental pain thresholds and stress-induced hyperalgesia in rodents (Quintero et al., 2000; Imbe et al., 2010), as well as in humans diagnosed with PTSD (Orr et al, 2000; Defrin et al, 2008). Likewise, in our model, predator odor stress increases thermal nociception specifically in Avoider rats (Roltsch et al., 2014; Itoga et al., 2016) and stress-induced hyperalgesia is predicted by post-stress avoidance. Importantly, pre-stress nociception does not predict post-stress avoidance or post-stress nociception (Itoga et al., 2016).

PTSD is highly comorbid with AUD (Gilpin and Weiner, 2017), and this comorbidity is associated with increased mortality and poor treatment response (Jacobson et al., 2008; Blanco et al., 2013; Debell et al., 2014; Shorter et al., 2015). Consistent with numerous clinical and epidemiological studies (Gilpin and Weiner, 2017), rats that exhibit high avoidance of the bobcat urine-paired chamber also present long-lasting increases in operant alcohol self-administration relative to Non-Avoiders and unstressed controls (Edwards et al., 2013; Schreiber et al., 2017). Moreover, stress history (i.e., predator odor exposure) reduces aversion to a context paired with a moderate dose of alcohol in rats (Schreiber et al., 2018).

It is important to emphasize that alcohol-drinking tests in the experiments described above occurred in a separate and distinct environment from the traumatic stress experience, consistent with the human PTSD condition. These findings support the notion that high traumatic-stress reactivity is associated with increased risk for AUD (Gilpin and Weiner, 2017). Avoiders also exhibit more compulsive-like alcohol drinking relative to Non-Avoiders and unstressed controls, as evidenced by the 10-fold higher quinine concentrations needed to suppress alcohol drinking by Avoiders relative to the other two groups (Edwards et al., 2013). Besides, baseline alcohol drinking does not differ between unstressed controls and stressed rats (due to the counterbalanced design), nor does baseline drinking differ between Avoiders and Non-Avoiders, suggesting that a history of alcohol drinking does not predict post-stress avoidance behavior.

In the predator odor stress model, systemic administration of R121919, a selective corticotropin-releasing factor (CRF) receptor 1 (CRFR1) antagonist attenuates thermal nociception, hyper-arousal and alcohol self-administration in stressed but not unstressed male rats (Roltsch et al., 2014), suggesting that CRF-CRFR1 signaling may underlie PTSD-related phenotypes in rats exposed to predator odor stress. In fact, exogenously administered CRF produces anxiogenic effects in rodents (Swerdlow et al.,1986; Arborelius et al., 1999), and alcohol-dependent rats exhibit increases in anxiety-like behavior and escalated alcohol self-administration associated to CRF hyper-function in the extended amygdala (Koob, 2008; Gilpin, 2012).

Endocrine effects in the bobcat urine conditioned avoidance model

HPA axis dysregulation has been implicated in the development of traumatic stress disorders and other psychiatric conditions (Yehuda, 2002; Stephens and Wand, 2012; Daskalakis et al., 2013a). In fact, individuals with lower levels of circulating cortisol immediately after traumatic stress are more likely to develop PTSD symptoms (Yehuda, 2002; 2005; Daskalakis et al., 2013a). To explain this phenomenon some authors propose that traumatic experiences may promote HPA axis dysfunction by enhancing sensitivity of the HPA feedback mechanism. The evidence supporting this are reduced urinary cortisol excretion, increased number of lymphocyte glucocorticoid receptors (GR), and enhanced cortisol suppression in response to dexamethasone in PTSD (Yehuda et al., 1991).

Enhanced sensitivity of the HPA axis would also potentially account for the blunted adrenocorticotropic hormone (ACTH) response to CRF (Yehuda et al., 1991), which seems to be due to a decreased pituitary sensitivity to CRF in PTSD, rather than a decrease in the number of pituitary CRF receptors (Smith et al., 1989). While the paraventricular nucleus of the hypothalamus (PVN) is essential for appropriate initiation and termination of the stress response, limbic system function also modulates HPA axis activity (Herman et al., 2005; Herman and Mueller, 2006; Jankord and Herman, 2008). Forebrain structures such as the hippocampus, amygdala and prefrontal cortex (PFC) can regulate HPA responses to stressors and are implicated in the development of stress-related pathologies (Jankord and Herman, 2008).

In spite of individual differences in avoidance of predator odor-paired stimuli in our CPA paradigm, all rats exposed to stress exhibit significant increases in circulating ACTH and plasma corticosterone levels immediately following bobcat urine exposure (Whitaker and Gilpin, 2015). However, Avoiders show blunted increases in both ACTH and corticosterone immediately following odor exposure relative to Non-Avoiders. Importantly, blunted HPA activation predicts avoidance: rats that display smaller increases in ACTH and corticosterone at the time of stress exhibit greater avoidance of the predator odor-paired chamber 24 hours post-stress (Whitaker and Gilpin, 2015).

Based on these results, we hypothesized that corticosterone administration before stress would reduce the magnitude and/or incidence of avoidance of predator odor-paired chamber. Indeed, pretreatment with corticosterone lowers the percentage of rats that become Avoiders and reduces the magnitude of avoidance in animals that meet the Avoider criterion (Whitaker et al., 2016). In agreement with studies in humans (Mason et al., 1986; Yehuda, 2002; Daskalakis et al., 2013a), these findings suggest that HPA hypo-reactivity immediately after trauma may worsen post-stress sequelae and may dictate disease trajectory and treatment response in humans that experience traumatic stress. To this point, we have not tested the causal relationship between circulating plasma corticosterone and brain CRF changes in Avoiders vs. Non-Avoiders following predator odor stress. Interestingly, predator odor stress increases CRF content in the PVN of Avoiders when compared to unstressed controls 12 days post-stress (Whitaker and Gilpin, 2015).

Neurobiological effects in the bobcat urine conditioned avoidance model

Low cortisol levels and low glucocorticoid signaling predict a corresponding enhanced GR responsiveness in PTSD (Daskalakis et al., 2013a). In the HPA axis, activation of GR can repress transcription of the crf gene as part of negative feedback, whereas stress-related noradrenergic and glutamatergic excitatory signals can increase the expression of the crf gene in the PVN (Herman et al., 1996; Palkovits et al., 1999), in part via activation of the transcription factor cAMP response element binding (CREB) protein (Watts, 2005; Liu et al., 2008). GR activation and acute stress each increase CRF transcript in the central nucleus of the amygdala (CeA) of rats (Kalin et al., 1994; Makino et al., 1994).

The GR is part of a molecular complex that includes the FK506 binding protein 5 (FKBP5). This chaperone acts as a GR inhibitor by reducing ligand binding and the translocation of the bound GR to the nucleus (Touma et al., 2011). In our model, we observed significantly lower FKBP5 protein levels in the CeA of male Avoiders when compared to unstressed controls at 48 hours post-stress (Whitaker et al., 2016). Likewise, in the aftermath of World Trade Center attacks, humans that went on to develop PTSD exhibited reductions in fkbp5 gene expression (Yehuda et al., 2009; Matić et al., 2013). Reduced FKBP5 expression has also been linked with peri-traumatic dissociation, which is related to PTSD risk (Koenen et al., 2005). Genetic and epigenetic variation of fkbp5 confer risk for PTSD in adults that endured abuse during childhood (Binder et al, 2008; Klengel et al., 2013), and specific risk alleles located within fkbp5 genetic loci are associated with lifetime and early onset of the disease (Boscarino et al., 2012). Additionally, clinical improvement in PTSD patients is associated with increased expression of FKBP5 and increased hippocampal volume, which are positively correlated (Levy-Gigi et al., 2013).

An additional important co-regulator for GR-mediated gene transcription is the steroid receptor co-activator-1 (SRC-1) (Winnay et al., 2006; Lachize et al., 2009). SRC-1 is highly expressed in the brain (Meijer et al., 2000), and is involved in crf gene expression in response to glucocorticoids and stress-related stimuli (Lachize et al., 2009). SRC-1 negatively regulates crf expression in the hypothalamus but positively regulates expression of this gene in the CeA, likely due to brain region differences in the expression of specific SRC-1 splice variants (Lachize et al., 2009). In the bobcat urine model, Avoider rats exhibit lower SRC-1 expression in the PVN and CeA that is negatively correlated with avoidance, and higher SRC-1 expression in the ventral hippocampus (VH) that is positively correlated with avoidance (Whitaker et al., 2016).

In the CeA, CRF promotes inhibitory synaptic transmission, thereby gating the activity of neurons that project to downstream effector regions (Fu and Neugebauer, 2008; Ji et al., 2013; Gilpin et al., 2015). For example, CeA projections to periaqueductal gray modulate nociception, presumably by gating the activity of descending pain inhibition pathways (Oliveira and Prado, 2001; Avegno et al., 2018). Avoiders have higher extrahypothalamic CRF content in the CeA for 2 to 3 weeks after bobcat urine exposure relative to Non-Avoiders and unstressed controls, and this effect is not rescued by post-stress alcohol drinking (Itoga et al., 2016; Schreiber et al., 2018). Avoiders exhibit hyperalgesia after bobcat urine exposure, and this hyperalgesia is mimicked by intra-CeA CRF infusion in experimentally naïve rats (Itoga et al., 2016). Both systemic and intra-CeA CRFR1 antagonism reverse stress-induced thermal hyperalgesia in Avoiders without affecting nociception in Non-Avoiders or unstressed controls (Itoga et al., 2016). These findings suggest that stress-induced increases in extrahypothalamic CRF signaling in the CeA mediate post-stress hyperalgesia in Avoider rats, and implicate this circuitry as a potential mediator of hyperalgesia in humans diagnosed with PTSD.

The CeA can be divided into centromedial (CeM) and centrolateral (CeL) sub-regions that differ in anatomical inputs and outputs, neuropeptide expression (Cassell et al., 1999; Gilpin et al., 2015), and regulation of fear-related behavior (Ciocchi et al., 2010; Keifer et al., 2015). First or second exposure to predator odor (i.e., bobcat urine) increases Fos immunoreactivity in CeM of both Avoiders and Non-Avoiders compared to unstressed controls. However, in the CeL, Avoiders exhibit increases in Fos immunoreactivity after a 2^nd stress exposure relative to Non-Avoiders and unstressed controls (Schreiber et al., 2018). These data suggest that predator odor stress activates the CeA of all exposed rats, but CeL is differentially activated based on stress reactivity. Accordingly, humans exhibit increases in amygdala activity in response to a threat, as measured by increased oxygenation in a functional magnetic resonance imaging study (Mobbs et al., 2010), but patients with PTSD exhibit hyper-reactivity of the amygdala that is predicted by avoidance symptoms (Sripada et al., 2013). These results also agree with data from our group showing that Avoiders exhibit higher ERK phosphorylation (pERK) in CeA than Non-Avoiders and unstressed controls after re-exposure to predator odor-paired stimuli in the absence of odor (Schreiber et al., 2018).

The ventromedial PFC (vmPFC) facilitates the hypothalamic stress response via indirect projections to the PVN and regulates anxiety and fear via direct projections to extended amygdala (Radley et al., 2006; Sotres-Bayon and Quirk, 2010; Adhikari et al., 2015; Keifer et al., 2015; Motzkin et al., 2015). Following re-exposure to predator odor-paired context, Avoiders and Non-Avoiders exhibit unique patterns of neuronal activation in vmPFC (Edwards et al., 2013). Trauma reminder (i.e., re-exposure to bobcat urine-paired chamber in the absence of odor) produces higher pERK levels in the vmPFC of Avoiders relative to Non-Avoiders. Moreover, post-reminder vmPFC pERK levels are predicted by avoidance behavior. In humans, combat veterans exhibited heightened vmPFC activity while exposed to emotional combat-related scenes, which was highly positively correlated with level of PTSD symptoms (Morey et al., 2008; Pannu et al., 2009). Moreover, veterans with bilateral vmPFC damage reported significantly lower depression severity than did veterans with damage involving other brain areas or patients with no brain damage (Koenigs et al., 2008). Avoiders also display higher CRF cell density in the vmPFC after exposure to predator odor, and CRF cell counts in vmPFC correlate with avoidance behavior (Schreiber et al., 2017). It is worth noting that CRF can increase CRFR1 expression through the ERK pathway (Meng et al., 2011), and intra-vmPFC CRFR1 antagonism reduces avoidance behavior in Avoider rats (Schreiber et al., 2017). Altogether, these findings support a role for dysregulation of vmPFC signaling in stress-related disorders.

Conclusions

Development of a reliable and valid animal model of PTSD is challenging due to the heterogeneity of the disease and the fact that not all humans exposed to traumatic stress develop psychiatric conditions. Like other animal models of PTSD symptoms, predator odor stress has inherent limitations: diagnosis in human patients relies primarily on verbally reported symptoms, particularly nightmares, intrusive thoughts, grief, and trauma-related guilt, which cannot be studied in rodents. In addition, our bobcat urine conditioned avoidance model does not recapitulate some of the complex features of trauma experienced by humans that may influence the onset and progression of PTSD, including multiple types of traumatic events and history of childhood abuse or neglect. Moreover, there is no clearly effective pharmacological treatment for PTSD, which makes it difficult to test the predictive (i.e., pharmacological) validity of the model (Cohen et al., 2012a).

Despite all these limitations, animal models permit a prospective follow-up design, in which the disorder can be triggered at a specified time and in a uniform manner in controllable population samples. In addition, pre-clinical models enable not only the assessment of PTSD-like behavioral, physiological, and molecular changes, but also provide insights into the influence of variety of factors that may determine the susceptibility of subsets of individuals to traumatic experiences. Thus, there is no doubt that animal models are of great value for understanding PTSD pathophysiology and possibly for the development of new and improved treatments and prevention strategies.

The bobcat urine conditioned avoidance model of traumatic stress developed by our group is a well-characterized and translationally relevant paradigm that facilitates investigation of the neurobiology underlying stress reactivity and post-trauma sequelae. Our model captures critical aspects of PTSD symptomatology including individual variability, avoidance of stress-paired stimuli, higher anxiety-like behavior, hyper-arousal, altered HPA axis reactivity, hyperalgesia, and increased alcohol drinking. Moreover, this model is a useful tool for investigating the neurobiology underlying variable response to traumatic stress and potentially for screening therapeutic targets and drugs that may reduce PTSD symptom severity in humans.