Keywords

Introduction

Criterion A of post-traumatic stress disorder (PTSD) in the fifth edition of the American Psychiatric Association Diagnostic and Statistical Manual of Mental Disorders (DSM-V) identifies the trigger to PTSD as exposure to actual or threatened death, serious injury, or sexual violation by witnessing an event, learning of an event, or experiencing repeated indirect exposures.

This view holds a hidden assumption – that PTSD is induced by exposure to a traumatic event. Various attempts to develop a translational model focused on the question of what may be an effective experimental trauma that could induce a PTSD state in the animal. However, this view ignored the well-known fact about PTSD, i.e., that only a small fraction of people exposed to a traumatic experience will eventually develop PTSD. Reported numbers vary substantially according to gender, the type of traumatic event, and community contexts (e.g., Santiago et al. 2013; Ditlevsen and Elklit 2012). But the reality remains that most people exposed to a traumatic event will not develop PTSD. This is a significant characteristic of the phenomenon, which indicates that the trauma may be a necessary but not a sufficient condition to induce PTSD and that additional factors should be considered in the effort to elucidate the core of the disorder. Thus, when attempting to develop an effective translational animal model of PTSD, these factors should be taken into consideration. Several factors have been proposed as contributing to vulnerability or to resilience in face of an exposure to a significant traumatic challenge. These include genetic background, developmental factors, and proximal factors, such as the quality and amount of sleep, social relations, and stress load just before exposure to the trauma. Until recently, relatively little was done to incorporate such factors in animal models of PTSD.

Notably, a history of childhood adversities was suggested as major risk factor for subsequent depression and anxiety disorders, including PTSD later in life (Anda et al. 2006; Bremner 2003; Penza et al. 2003). In the last decade, we set out to develop a rodent model of PTSD that would include exposure to stress in a comparable developmental period – the postweaning, prepuberty, or early adolescence period – and reexposure to stress in adulthood. The assumption was straightforward; if indeed a risk factor, animals exposed to childhood adversities would exhibit stronger and longer-lasting responses to a stressful experience in adulthood.

This chapter summarizes several years of research in developing this model as an effective platform for studying the neurobiology of PTSD.

Stress in Childhood and Adolescence

Freud was among the first to highlight the importance of early life experiences in shaping an individual’s mental functions for the rest of his/her life. This emphasis on the importance of early developmental experiences still persists and with it the notion that stress early in life may induce a vulnerability to the effects of stress later in life, possibly by inducing a persistent sensitization in stress-responsive neural circuits, which augments the consequences of later adverse experiences (Agid et al. 2000; Heim and Nemeroff 2001; Nemeroff 2004a).

Across species, including humans, the early adolescent or juvenile brain is considered to be in a transition phase, differing markedly both anatomically and neurochemically from that of newborns, weanlings, or adults. Preteens and adolescent humans have enhanced stress perception and responses. Stressful life events during this period have been suggested as being associated with later socio-emotional maladaptive behaviors and to represent a significant risk factor for the later development of stress-related psychopathologies (Maughan and McCarthy 1997; Spear 2004).

There is increasing evidence that the adolescent brain is particularly vulnerable to effects of stress. Casey et al. (2008) addressed human adolescent brain development and suggested that during adolescence, a unique imbalance exists between levels of activity in subcortical regions and cortical regions. Based on both preclinical evidence and brain imaging studies of children, adolescents, and adults, it appears that during adolescence, subcortical activity is relatively higher than cortical activity. It was suggested that these findings may indicate that during the processing of negative emotional information, this bias between unregulated and prominent subcortical activity (e.g., limbic regions) and a relatively reduced activity in the prefrontal cortex may relate to the emergence of affective disorders during adolescence (Casey et al. 2008).

Many studies in humans point to late childhood and early adolescence as periods of particular vulnerability development of psychopathologies later in life (Maercker et al. 2004; Pynoos et al. 1999). The net consequence of stress in the juvenile/adolescent brain that has been suggested is the sensitization of the emotional brain to the effects of later life stress, increasing the likelihood of depression and anxiety (Agid et al. 2000; Anisman and Matheson 2005; Costello et al. 2002; Gregory et al. 2007; Heim and Nemeroff 2001; Levine 2005; Maughan and McCarthy 1997; Nemeroff 2004b; Nemeroff et al. 2006; Spear 2004). However, it is important to note that some epidemiological studies suggest that stress early in life, but at different developmental stages, may result in the development of different psychopathologies. Childhood trauma (under the age of 12 year) was found to increase the risk of developing major depression, while trauma during adolescence was associated with a greater predisposition to PTSD (Maercker et al. 2004), suggesting that emotional challenges at different developmental phases affect different neural systems (Heim and Nemeroff 2001; Maercker et al. 2004; Pynoos et al. 1999).

The Postweaning Prepuberty/ Juvenile Stress Animal Model

Many of the rodent early life stress animal models focus on the perinatal preweaning period and involve some form of maternal deprivation or separation (reviewed in Sanchez et al. 2001; Harrison and Baune 2014), producing acute and long-term effects that vary with the pups’ age. For example, prolonged early maternal separation attenuated rates of synaptic development in the hippocampus, which was evident only after sexual maturation (Andersen and Teicher 2004), and maltreatment of rat pups results in a lifelong, and even a trans-generational, effect on behavior and DNA methylation (Roth et al. 2009; Franklin et al. 2010).

Accumulating evidence has thus raised the possibility that early life stress induces a sensitization of stress response mechanisms that may be due to altered limbic functioning, thereby augmenting the consequences of stressors later in life (Nemeroff 2004a, b).

However the brain’s development continues well after the preweaning period, and substantial maturation processes like myelination continue with varying dynamics well into puberty (Hamano et al. 1998). The ongoing maturational changes render the postweaning brain susceptible to the harmful effects of stress.

While most early life stress rodent models focus on the perinatal to preweaning periods, recent work in our laboratory and others focused on an alternative period in the rat ontology, “juvenility” (~28 days), the earlier phase of the adolescent/postweaning to the prepubertal period. This period is likely to relate closely to human childhood, a developmental period known to be relevant to the pathogenesis of a range of psychiatric disorders. The rationale behind the “juvenile stress” model is that this is expected to induce long-term alterations in stress responsiveness by exposing rats to stressors during juvenility, thus augmenting the consequences of additional exposure to stressors in adulthood. For a depiction of a typical experimental design using this model, see Fig. 1.

Fig. 1
figure 1

Example of a typical experimental design for the juvenile + adulthood stress procedure. Rats are randomly assigned to one of three experimental groups: control group, adulthood stress group, and juvenile + adulthood stress group. Rats in the juvenile + adulthood stress group are exposed to three different stressors applied during juvenility (postnatal days 27–29; see practice and procedure methods). Four weeks later, rats are exposed to adulthood trauma (UWT; see practice and procedure methods). Rats are then assessed for anxiety-related measures in the elevated plus maze and open field arena for further behavioral profiling. Following behavioral assessments, rats are sacrificed, and tissue is harvested. Animals in the adulthood stress group are exposed only to UWT in adulthood, and control animals are unexposed to stress (JVS juvenile stress, UWT underwater trauma)

Full size image

The “juvenile stress ” paradigm involves exposure to a combination of three stressors in juvenility. A series of experiments using this paradigm revealed altered behavioral responses to challenges in adulthood (see Tables 1 and 2). These include high anxiety scores in the elevated plus maze, open field test and startle response test following exposure to adulthood stress (Avital and Richter-Levin 2005), impaired avoidance learning (Tsoory and Richter-Levin 2006), increased cued fear response following auditory fear conditioning and increased aversive vocalization (Yee et al. 2012), and altered decision-making (Brydges et al. 2012). The exacerbation of the effects of the adulthood stress by the “juvenile stress” was also evident at the physiological (Cohen et al. 2007; Yee et al. 2011), biochemical (Bazak et al. 2009; Jacobson-Pick et al. 2008; Tsoory et al. 2008a, b; Jacobson-Pick and Richter-Levin 2012; Brydges et al. 2014a), molecular (Brydges et al. 2014b) and electrophysiological levels (Maggio and Segal 2011; Grigoryan et al. 2015).

Table 1 Examples of behavioral findings using the juvenile stress model
Full size table
Table 2 Key facts of juvenile stress
Full size table

Individual Differences in the Response to Stress

A significant difference between the way psychiatric disorders are defined in humans (individual diagnosis) and the way they are modeled in the animals (group averages) is evident. In the case of PTSD, this discrepancy is of particular significance: following exposure to a life-threatening event, only a subset of individuals will eventually develop the full range of PTSD symptoms. The heterogeneous response to the trauma observed in humans is often disregarded in animal models of PTSD (Goswami et al. 2013). To that end, several studies began to address such individual responses by forming “behavioral profiles ” of individual animals (Cohen et al. 2004; Matar et al. 2013). This approach resembles the clinical diagnosis procedure, where characteristics of different behavioral categories are assessed at the individual level and clusters of symptoms in affected patients are compared to the healthy norm.

One approach which proved to be of high translational value is the cutoff behavioral criteria (CBC) analysis approach, developed by Cohen et al. (2003). By integrating different levels of response patterns, classification criteria are formed in a similar manner to that used in clinical diagnosis procedures to form psychopathological symptom clusters; thus sets of classification criteria, representing inclusion and exclusion criteria, produce distinct patterns of stress-induced indices. The CBC analysis maximizes the accuracy of the animals’ classifications and minimizes the likelihood of including “false positives,” by making sure that each animal that meets both sets of criteria (inclusion and exclusion) is defined as “affected” or “unaffected.” The validity of the criteria is affirmed by ascertaining that the vast majority of “unexposed” animals are found within the “unaffected” category and only a minority with the “affected.” Cohen et al. (2003) were able to show that in comparison with unexposed and “unaffected” exposed rats, “affected” exposed rats exhibited significantly higher plasma corticosterone and corticotropin concentrations, increased sympathetic activity, diminished vagal tone, and increased sympathovagal balance. These differences were present only when individual behavior was taken into account.

In recent studies conducted in our lab, we apply an adaptation to the original CBC approach. In these studies, we use the “underwater trauma” (UWT) protocol as a brief traumatizing experience. UWT has been previously shown to increase anxiety-like behavior (Cohen et al. 2004; Richter-Levin 1998) and impair spatial memory as well as long-term potentiation in the dentate gyrus of the hippocampus (Wang et al. 2000). We further extended this model by including an exposure to UWT reminders, thus providing a platform for studying traumatic reexperiencing (Ritov and Richter-Levin 2014; Ardi et al. 2014). In addition, we examined also the impact of exposure to juvenile stress to model childhood adversity. Here, testing whether juvenile stress will increase the likelihood of developing PTSD-like symptoms will further validate our model.

In our model, each of the experimental groups (control-unexposed, adulthood trauma exposure only, and juvenile + adulthood trauma exposure) is tested for several anxiety-related symptoms (see depiction of experimental design in Fig. 1). While in the original CBC methodology the lowest 25th percentile of the trauma-exposed group was compared with the highest 25th percentile (Cohen et al. 2004), the comparison group we believe should be used is the averaged performance of the control group for each behavioral variable. The performance of each individual animal in the experimental groups is then compared to the distribution in controls. Animals are identified as “affected” when they fell into the lower or upper 25 % percentile of the control group in five of six parameters. Thereby, instead of using a fixed cutoff value for each parameter as suggested initially (Cohen et al. 2003, 2004), the matching toward the normal distribution in the control group appears to translate the diagnostic procedure more closely to that used in humans (Horovitz et al. 2014).

Using this approach, the distribution of “affected” and “unaffected” animals differs significantly between the experimental groups: while only 12 % of control rats and 24 % of adulthood trauma-exposed rats are considered “affected,” 61 % of juvenile + adulthood trauma-exposed rats are considered “affected” (unpublished results, Fig. 2). Our results reveal that previous exposure to juvenile stress increased the prevalence of affected animals in our model, thereby matching epidemiological data from PTSD patients.

Fig. 2
figure 2

Distribution of “affected” and “unaffected” animals in the different stress-exposure groups according to the behavioral profiling approach. In order to create the behavior profiles, the performance of the control group was referred to as behavior of a normal population. We focused on six anxiety-related behavioral parameters. We first determined the distribution of values in the control group and calculated the upper and lower 25th percentages for each chosen parameter. As a second step, the performances of each animal that participated in the experiment were compared to the distribution curve of the control group. In order to be classified as affected, an animal must exhibit values that are under the 25th or above the 75th percentages in at least five of the six measurements (control, n = 34; adult stress, n = 42; juvenile + adult stress, n = 38; unpublished data)

Full size image

Gender/Sex Differences

Women are more prone than men to mood disorders, particularly depression (Gater et al. 1998; Noble 2005) and anxiety disorders, including PTSD and generalized anxiety disorder (Bekker and van Mens-Verhulst 2007; Gater et al. 1998; Tolin and Foa 2006). In addition to the different prevalence’s of mood and anxiety disorders in men and women, there is evidence for gender-dependent differences in the response to psychotropic medications between men and women as well (Gorman 2006; Yonkers et al. 1992). These differences suggest that gender-dependent differences in neuroanatomy and neurophysiology might underlie some of the observed differences in the prevalence and symptom profiles of mood and anxiety disorders between men and women. Consistent with this hypothesis, differences between the genders were reported in several components of the central nervous system. Cahill (2006), for example, noted gender differences in hippocampal structure and functions. These gender differences included the adrenergic, serotonergic, cholinergic, and cholecystokinin systems; anatomical structure; relative size (adjusted for total brain size); reactivity to stressors; as well as the effects of CORT and benzodiazepines.

Animal studies reflect the suggested human gender differences in responding to stress. Sex differences were found in learning tasks under both “controllable” and “uncontrollable” stress conditions (Shors et al. 2001; Steenbergen et al. 1990; Wood and Shors 1998). Furthermore, female rats showed higher baseline corticosterone levels, larger corticosterone secretion after adrenocorticotropic hormone injections, and enhanced responses to stressors (Campbell et al. 2003; Kirschbaum et al. 1992). Brotto et al. (2000) reported that male rats quickly habituate their corticosterone response to a chronic stressor, whereas female rats showed little hormonal habituation even after 21 days of restraint stress. Dalla et al. (2008) noted that overall females consistently show higher level of active behavior than males. In the forced swim test, melatonin administration increased struggling in males during the first session but did not affect females; on the contrary, it increased immobility in females but not in males (Brotto et al. 2000). Yet, following exposure to chronic stress, male rats showed more immobility than female rats in the forced swim (Dalla et al. 2005).

While numerous studies indicated a gender differences in rate of anxiety disorders, little is known about whether childhood exposure to stressful events has differential effects on males or females. Horovitz et al. (2014) examined the differential effects of combined juvenile and adulthood stress on male and female rats, while taking individual differences into account. Results reveal that while both adult male and female rats were affected by exposure to juvenile stress, they were affected in different ways. Sex differences were evident in anhedonia measures, coping with the stressful challenge of an avoidance learning task, and in “adulthood stress”-induced changes in anhedonia measures. Furthermore, individual profiling of altered behavioral responses revealed sex differences in the prevalence of the different categories of “affected” rats. Although the prevalence of “affected” males and females among “juvenile + adulthood stress” rats was similar, the distribution of behavioral profiles (“depressed,” “anxious,” “comorbid”) was significantly different (Horovitz et al. 2014). In another study, Brydges et al. (2014c) utilized the “juvenile stress ” paradigm to examine whether pubertal stress in females and in males induces alterations in hippocampal function in adulthood. The hippocampus is of special interest in this case since it plays a central role in stress reactivity (Romeo and McEwen 2006; Pervanidou and Chrousos 2012). The results of this study revealed that exposure to juvenile stress alters hippocampal-dependent behaviors in adulthood, but it does so in a sex-specific manner.

These studies support the hypothesis that there are sex-based differences in the manner in which prepubertal stress induces alterations in hippocampal function in adulthood and further highlight the fact that males and females should be considered separately in preclinical models of neuropsychiatric disorders (Cahill 2006).

To conclude, it is suggested that in preclinical research that attempts to model stress-related psychopathologies, the establishment of different control groups for each of the sexes may be instrumental in setting “sex-/gender-specific” criteria in order to better profile the animals. A similar gender-specific profiling approach may also be beneficial for human pathology diagnosis. This possibility should at least be considered, in association with the consideration of promoting also “sex-/gender-specific” therapeutic interventions.

Practice and Procedures

Some of the experimental procedures mentioned above are described here.

Juvenile Stress Protocol

This protocol involves a 3-day exposure to different stressors applied during juvenility (postnatal days 27–29) and serves as an animal model for childhood adversity (Tsoory et al. 2008a). Each of the stressors was chosen for its well-documented effects on adult rats.

  • Day 1 (postnatal day 27) – Forced swim. Ten minutes forced swim in an opaque circular water tank with water temperature at 22 ± 2 °C .

  • Day 2 (postnatal day 28) – Elevated platform. Three 30 min trials; intertrial interval: 60 min in the home cage. Elevated platform: 70 cm above floor level, located in the middle of a small closet-like room.

  • Day 3 (postnatal day 29) – Restrain. Rats are placed in a metal mesh restraining box that prevents forward-backward movement and limits side-to-side mobility. Rats remain in the restraining box for 2 h.

Underwater Trauma Protocol (UWT)

Underwater trauma (as described in Richter-Levin 1998) is a tangible life-threatening situation which is designed to model sudden, brief traumatizing experiences. This procedure is currently used in our lab as a traumatic stress event in adulthood. In this paradigm, rats are given 1 min to swim in a circular pool, containing water at and then held under water for 30 s using a metal net. Control rats are given 1 min to swim and then are put in a resting cage.

For an example of an experimental design, see Fig. 1.

Summary Points

  • An effective translational animal model of PTSD should take predisposing factors into consideration.

  • A history of childhood adversities is suggested as major risk factor for subsequent depression and anxiety disorders, including PTSD later in life.

  • The “juvenile stress model” serves as an animal model for childhood adversity and was shown to induce long-term alterations in stress responsiveness.

  • The behavioral profiling approach enables grouping animals according to characteristics of different behavioral categories and bears resemblance to human clinical diagnosis.

  • This approach, along with gender-specific profiling, may clarify some of the pending issues in PTSD research.