Fig. 1. Top panel. Animals exposed to the SPS paradigm show significantly lower plasma ACTH response than naive animals when both groups are pretreated with hydrocortisone and receive restraint stress. These data suggest enhanced negative feedback in the SPS group. Middle panel. MR/GR ratio changes within hippocampal subfields, 24h, 7 and 14 days following the SPS, compared to controls. SPS-7 and SPS-14 groups are significantly lower than the controls, while SPS-24h group is significantly higher. These data suggest a net upregulation of hippocampal glucocorticoid receptors following SPS. Bottom panel: Mean startle amplitude, in response to 108 dB, for control and SPS rats. SPS-exposed rats (grey bars) show a heightened startle response post-SPS (Time 2) as compared to pre-SPS (Time 1). For control animals, the startle response remains unchanged or slightly decreased over time. indicates significantly higher startle response compared to SPS (time 1) (/><0.01). # Indicates significantly higher startle response compared to control (time 2) (P < 0.05).
hippocampal glucocorticoid receptors. Specifically, they observed that CRF administered intracerebro-ventricularly (ICV) produced an upregulation of hippocampal MR receptors in adrenalectomized rats, an effect dependent on the presence of corticosterone (preadministered), although not the GR agonist dexamethasone. The authors suggested that CRF might exert positive effects on glucocorticoid receptor expression either through direct pharmacological effects, or indirect activation of serotonergic or noradrenergic systems. A potential role for hippocampal GR receptors in the response to predator exposure has also been suggested from studies showing that stress-induced increases in hippocampal serotonin release (produced by rat exposure) is enhanced in glucocorticoid receptor deficient mice (Linthorst et al., 2000). Thus, there is potential for interactions between CRF, corticosterone, and/or glucocorticoid receptors in mediating the response to predator exposure. It would be of interest to examine whether predator exposure produces long-term changes in HPA axis functioning similar to that observed in PTSD, and whether these effects might be mediated by CRF and glucocorticoid receptor activation at the amygdala or hippocampus.
Research into the neurobiology of predator exposure has also yielded other interesting results potentially relevant to human PTSD. Recent studies have shown that NMDA antagonists administered prior to predator exposure produced a significant reduction in anxiety. Adamec et al. (1998) suggested that a lasting effect of predator exposure includes long-term changes in NMDA-mediated LTP activity in limbic structures (such as the amygdala and hippocampus), which are closely involved in defensive behaviors and anxiety. LTP of excitatory transmissions from the amygdala to the ventromedial hypothalamus has been specifically linked to defensive behaviors and kindling activity, and LTP activity appears to correlate well with defensive postures taken by animals following predator exposure.
Cohen et al. (2003) introduced a variation of predator exposure models, which may offer greater ethological validity with respect to clinical application. Rather than studying all animals exposed to trauma, this approach focused only on those that developed a maladaptive response. Since PTSD affects roughly 20-30% of those exposed to traumatic events, this approach essentially incorporates inclusion criteria for animals analogous to inclusion criteria in clinical studies. In their studies, following a 10-min cat exposure, rats were subdivided into "maladapted" and "well-adapted" based on the intensity of their behavioral response. Criteria for maladapted included a higher acoustic startle response (as compared to control or "well-adapted" rats) and little or no time spent in the open arm of the elevated plus maze. The first criterion was considered indicative of hyperarousal symptoms, while the second indicative of heightened avoidant responses. Maladapted rats, which comprised 25% of all rats exposed, also exhibited significantly higher plasma corticosterone and ACTH concentrations, increased sympathetic activity, and diminished vagal tone. Interestingly, it was also observed that following stress exposure, animals responded initially with a widespread, presumably normative acute response, but only about 25% go on to suffer from enduring effects. Again, this shows considerable analogy with regard to human PTSD development. It would be of considerable interest for future research to expand the characterization of the relative expression of other behavioral and biological signs of PTSD between these two groups.
In addition to assessing the effects of environmental stressors on the development of PTSD symptoms, it is also important to characterize the role of genetic and predisposing factors. True et al. (1993) reported that up to 30% of the variance in PTSD symptoms may be based on genetic factors. More importantly, not all people exposed to similar traumas develop PTSD, suggesting that differential sensitivity to stress and trauma contributes to individual differences in vulnerability. Genetic makeup is one of the potential sources of this differential sensitivity. Finally, genetic models also provide an opportunity to identify specific biological mechanisms underlying certain PTSD symptoms, and to investigate the link between biological abnormalities and behavioral changes.
Two genetically distinct strains of rats, Lewis and F344 (Fischer), differ in terms of their basal and stress-induced HPA function. The Lewis rats show an attenuated diurnal corticosterone level as compared to Fischer (Griffin and Whitacre, 1991; Dhabhar et al, 1993), and a lower ACTH and corticosterone response to a variety of stressors (Sternberg et al, 1992; Dhabhar et al, 1993) and to cytokine administration (Sternberg et al, 1989). For instance, in response to immobilization stress, Lewis rats show a lower ACTH response. This is combined with decreased POMC mRNA concentrations at the pituitary, a build up of corticosterone in the adrenals, but decreased circulating corticosterone levels (Moncek et al, 2001). These findings are intriguing in that they suggest possible mechanisms for low HPA responsivity, such as a decrease in the precursor for ACTH and beta-endorphin. The authors also suggest that increased circulating corti-costerone binding protein (CBG) might account for high corticosterone in the adrenals, but low circulating levels detected in plasma. Behaviorally, Lewis rats also show a heightened acoustic startle response to acoustic and tactile stimuli (Glowa et al, 1992; Glowa and Hansen, 1994). In attempting to find a relationship between circulating corticosterone levels and the startle response, Glowa et al. (1992) observed a negative association between the corticosterone response to startle and the startle amplitude. Pavkovich and Valentino (1997) have also found that the discharge rate of LC neurons were increased in adrenalectomized Sprague-Dawley rats. The authors suggested that the absence of glucocorticoids might lead to an increase in basal and stress-induced CRF release at the LC, thus possibly contributing to a heightened arousal response. It may also provide an explanation for the paradoxical finding noticed in PTSD patients, whereby low circulating glucocorticoid levels are combined with high central CRF levels. With respect to other behavioral measures relevant to PTSD, the differences between Lewis and Fischer strains are less clear. Studies have shown that Lewis rats demonstrate greater (Rex et al, 1996), less (Chaouloffet al, 1995), or equal amounts (Kosten et al, 1994) of activity levels in a novel open field. Also, in contrast to what may be expected in a model of PTSD, Lewis rats show a lower fear-conditioning response as assessed, for instance, in a model of fear-conditioned suppression of drinking (Stohr et al, 2000). Nonetheless, the differences in basal and stress-induced HPA function provides an intriguing opportunity to assess a major physiological symptom of PTSD and how it relates to particular behavioral responses.
Thus, research using these strains seems to suggest that low HPA responsivity may be linked to mechanisms underlying the acoustic startle response, although they may be less relevant to symptoms of neophobia and certain types of fear conditioning. Further research into the specific mechanisms which link the HPA axis and hyperarousal symptoms could prove particularly useful in understanding the physiology behind heightened startle in PTSD patients.
In the congenital learned helplessness model, animals are preselected based on their responses in the learned helplessness paradigm, and then selectively bred for 30-40 generations. The offspring of rodents with a propensity to develop an escape deficit are labeled "congenital learned helpless" (CLH), and those with resistance to developing escape deficits are called "congenital nonlearned helpless" (nCLH). CLH animals show some physiological differences in hippocampal, hypothalamic, and prefrontal cortex gene expression when compared to nCLH (Kohen et al, 2000) as well as differences in behavioral and neuroendocrine responses potentially relevant to PTSD (King et al, 2001). CLH animals show a higher shock-induced opiate-analgesia (as assessed with the tail-flick test) as well as spatial memory deficits, assessed using the Morris water maze, possibly resulting from hippocampal dysfunction. These animals also have a decreased ACTH and corticosterone response to intermittent conditioned and unconditioned stressors, which may be the result of an enhanced negative feedback system. Interestingly, in CLH animals receiving early-life maternal separation or early-life cold stress, glucocorticoid responsiveness was also attenuated (King and Edwards, 1999). Thus, although initially bred as model for the psychoge-nomic study of depression, this model may also provide interesting insights into PTSD. Further research examining startle responsiveness, exploratory behaviors, and conditioned and unconditioned fear responses would be of considerable interest.
As discussed elsewhere, the upregulation of glucocorticoid receptors in lymphocytes has been reported in PTSD patients and an upregulation of hippocampal GR has been reported in some PTSD animal models. Alterations in GR receptor concentrations are particularly interesting since they are the primary site of action for dexamethasone in the HPA axis. Recently Reichardt et al. (2000) developed a mouse model which actually over express the GR glucocorticoid receptor (called YGR mice). This was accomplished by breeding transgenic mice carrying two additional copies of the glucocorticoid receptor gene. In YGR mice, GR mRNA is elevated in the brain and pituitary by 60 and 45% respectively. Although behavioral data on these mice is still lacking, in response to restraint stress and endotoxic shock produced by lipopolysaccharide administration, glucocorticoid release was significantly attenuated. It is of interest to note that many of the physiological features of these mice are indicative of a supersensitive HPA negative feedback system. Although YGR mice show decreased hypothalamic CRF levels, they also show decreased POMC levels in the anterior pituitary and lowered plasma corticos-terone levels in plasma. Further research investigating behavioral performance in these mice may provide insight into particular deficits linked to higher GC receptor concentrations.
Given the prevalence of PTSD and its profound debilitating effects, there is a strong motivation toward further understanding neurobiological and behavioral mediators and risk factors. In this context, animal models can play a critical role as they provide unique opportunities for delineating specific biological mechanisms, identifying genetic factors, and for efficiently testing potential therapeutic agents. The models described in this chapter (summarized in Table 1) provide good opportunities for starting to address this goal. For instance, models of IES can be used to delineate mechanisms or evaluate treatments for opioid-mediated analgesia, noradrenergic sensitization as well as fear-conditioning effects. Single prolonged stress offers an opportunity to characterize the neurobehavioral modulators of trauma-induced HPA axis dysfunction as well as hyperarousal symptoms. Similarly predator exposure models can be used to investigate causes and treatments for changes in the startle response, amygdala sensitization, and LTP effects. Given the variety of factors influencing the development of PTSD and the broad range of symptoms, it is perhaps unlikely that any single model will fully capture all of
Table 1. Putative and PTSD-specifie animal models
Was this article helpful?