An ongoing concern for researchers investigating the underlying neurobiology of PTSD is the extent to which the biological markers identified in patients represent a consequence of the traumatic exposure and resultant PTSD versus a vulnerability factor that increases an individual's likelihood of developing PTSD subsequent to trauma exposure. Twin studies have been helpful in exploring this differentiation. In a study of hippocampal volume of Vietnam veterans with PTSD and their unaffected twins without combat exposure, both cohorts displayed smaller hippocampal volume as compared to twin pairs of Vietnam veterans without PTSD and their unaffected siblings (Gilbertson, Shenton et al. 2002). This finding suggests that reduced hippocampal volume may be a heritable, pre-existing vulnerability factor for developing PTSD following trauma exposure. A more recent study included twin pairs in which one twin was combat-exposed in Vietnam and diagnosed with PTSD, while the co-twin was not exposed to combat and not diagnosed with PTSD (“high risk” twin). In contrast, in a second twin pair, one twin was combat-exposed in Vietnam but did not develop PTSD and his sibling who was not exposed to combat and not diagnosed with PTSD (“low risk” twin). The authors found decreased gray matter volume in the pregenual anterior cingulate cortex only in the combat-exposed individuals with a PTSD diagnosis versus the other three groups of twins (Kasai, Yamasue et al. 2008). These data suggested that the lower gray matter volume in this region was a consequence of traumatic combat exposure and the subsequent development of PTSD. The current state of the field is that both of the above findings likely represent different aspects of the true phenomenology, such that: 1) following a significant trauma, PTSD is more likely to occur in those individuals who have a certain level of genetic and neural vulnerability, but that 2) there are almost certainly additional neurological / neuropathological sequelae that result after the onset of PTSD.

Vulnerability to the development of PTSD after trauma exposure has been conceptualized to involve a fear conditioning process in which fear responses are exaggerated and/or resistant to extinction (Keane, Zimering et al. 1985; Amstadter, Nugent et al. 2009; Jovanovic and Ressler 2010). Within this classical conditioning model, a neutral (conditioned) stimulus (CS) that is temporally paired with an aversive (unconditioned) stimulus (US) acquires the ability to elicit a conditioned fear response that can be triggered when the individual subsequently encounters these or similar stimuli. With regard to the clinical features of PTSD, the traumatic event itself serves as the US and the sights, sounds, smells, and other contextual aspects of the ambient environmental stimuli that are present at the time of the trauma serve as CS to elicit conditioned fear responses.

While this process can serve an evolutionarily advantageous role by allowing an individual to identify and respond to threatening stimuli, excessive activation of conditioned fear responses to stimuli that are not accurately predictive of threat may have a role in the basis of PTSD. Consistent with this hypothesis, emotional and physiological reactivity to stimuli resembling the original traumatic event is a central characteristic of PTSD, can persist for an extended period of time following the traumatic exposure, and has been reliably replicated in the laboratory (Blanchard, Kolb et al. 1986; Pitman, Orr et al. 1987; Orr, Metzger et al. 2003).

While published reports are mixed, the majority of findings in patients with PTSD report elevated levels of CRH (Bremner, Licinio et al. 1997; Baker, West et al. 1999), lower basal cortisol levels (Yehuda, Southwick et al. 1990; Yehuda, Teicher et al. 1996; Thaller, Vrkljan et al. 1999; Bremner, Vermetten et al. 2007), and enhanced negative feedback suppression of the HPA axis by dexamethasone, a synthetic glucocorticoid (De Kloet, Vermetten et al. 2006; Yehuda 2009). Despite clear evidence of HPA axis perturbations in PTSD, initial investigations into main effects of candidate genes that impact on this system have not been very revealing. For example, a study of the glucocorticoid receptor gene demonstrated no significant association between two polymorphisms in this gene and a diagnosis of PTSD (Bachmann, Sedgley et al. 2005).

Given that PTSD is a psychiatric disorder that, by definition, requires exposure to a specific environmental event (the criterion A trauma), and that individuals vary in their phenotypic response to the event, with only a minority of individuals subsequently developing PTSD, the genetic basis of this disorder is likely to be better elucidated by studies of gene-environment (GxE) interactions rather than studies of the main effects of genes. Two studies have recently demonstrated an interaction between polymorphisms in FKBP5, a gene which impacts on HPA axis function by regulating GR activity, and childhood environment to predict severity of PTSD. FKBP5 is a co-chaperone protein that interacts with another molecular chaperone, hsp90, and is part of the mature GR heterocomplex (Hubler and Scammell 2004). It regulates GR sensitivity and nuclear translocation of GRs such that reduced activity of FKBP5 increases GR sensitivity (Figure 2). Polymorphisms in FKBP5 have previously been reported to be associated with peritraumatic dissociation in medically ill children, a symptom which is predictive of the development of PTSD (Koenen, Saxe et al. 2005). In a study of over 700 highly traumatized, inner city, African-American subjects recruited from primary care clinics (Binder, Bradley et al. 2008), four polymorphisms of the FKBP5 gene were demonstrated to interact with severity of childhood abuse to predict severity of adult PTSD symptoms (Table 1). Further, these polymorphisms were demonstrated to be functional, in that those individuals with both probable PTSD and the risk alleles demonstrated enhanced suppression of cortisol in response to dexamethasone (Binder, Bradley et al. 2008). A more recent study partially replicated these findings, reporting a significant interaction for one of the polymorphisms in the African-American population, such that the TT genotype was associated with the highest risk for PTSD amongst those individuals with a history of childhood adversity, while it was associated with the lowest risk for PTSD amongst those with no history of childhood adversity (Xie, Kranzler et al. 2010). While this GxE interaction was not significant in their European-American study population, alcohol dependence was found to interact with FKBP5 polymorphisms and childhood adversity to increase the risk for a diagnosis of PTSD. These data, along with a number of similar studies, suggest that a number of subject population characteristics as well as similar measurements of childhood and adult trauma will be critical factors in the robust identification of GxE effects.

As with the HPA axis, studies of the main effects of candidate genes that impact on the LC-noradrenergic system have been largely negative. There was no significant association between polymorphisms in the gene for dopamine beta hydroxylase (DBH), an enzyme that catalyzes the conversion of dopamine to norepinephrine, and PTSD in combat veterans, although there was an association between one genotype and lower DBH activity in those individuals diagnosed with PTSD versus those without the diagnosis (Mustapic, Pivac et al. 2007). Neuropeptide Y (NPY) is colocalized with norepinephrine and has been reported to reduce the release of norepinephrine. It has been demonstrated to have anxiolytic properties in animal models, and levels of NPY have been found to be reduced in those with PTSD (Morgan III, Wang et al. 2000; Morgan 2002). However, in a 2002 study of a polymorphism within the NPY gene, no association was found with a diagnosis of PTSD (Lappalainen, Kranzler et al. 2002).

Given the extent and complexity of this limbic-frontal neurocircuitry, many different genes can impact on this system. Dopamine has been demonstrated to innervate the amygdala and animal models have reported alterations of this dopaminergic neurotransmission in response to stress (Puglisi-Allegra and Cabib 1997). Two studies reported a significant main effect of the D2A1 allele of the D2 dopamine receptor (DRD2) gene in association with a diagnosis of PTSD (Comings, Comings et al. 1991; Comings, Muhleman et al. 1996), while another study of this allele did not find a significant association (Gelernter, Southwick et al. 1999). However, the latter study did not assess for trauma exposure in the control group which may have had the effect of not identifying susceptible controls who would have developed PTSD if exposed to a traumatic event. Another study reported a significant association between the D2A1 allele and PTSD only amongst those individuals with PTSD who engaged in harmful drinking (Young, Lawford et al. 2002). A more recent study revealed a significant association between a different polymorphism in the DRD2 gene and PTSD in combat veterans (Voisey, Swagell et al. 2009). Finally, another study reported a significant association between a polymorphism in the dopamine transporter SLC6A3 gene and chronic PTSD versus trauma-exposed controls (Segman, Cooper-Kazaz et al. 2002).

Serotonin is another catecholamine neurotransmitter which plays an important role in regulating activity within the limbic-frontal system and has been reported to be involved in the inhibition of the development of fear memories (Rainnie 1999). A polymorphism of the serotonin 2A receptor gene has been found to be significantly associated with PTSD in Korean women (Lee, Kwak et al. 2007). More scientific attention has been focused on a polymorphism in the promoter region of the serotonin transporter (5-HTTLPR), in which the short (s) allele has been demonstrated to be less efficient than the long (l) allele and is also associated with decoupling of the circuit between the prefrontal cortex and the amygdala responsible for extinction of fear conditioning (Lesch, Bengel et al. 1996; Pezawas, Meyer-Lindenberg et al. 2005). An excess of the s/s genotype has been reported in a Korean population of PTSD patients relative to controls (Lee, Lee et al. 2005). Recent GxE studies of this polymorphism have been more revealing. Kilpatrick et al. (2007) reported that in a study of adults exposed to the 2004 Florida hurricanes, the s/s genotype was associated with an increased risk for PTSD only in those exposed to an environment of high stress (high exposure to hurricanes and low social support). An additional study in this population found a significant interaction between genotype of 5-HTTLPR and county-level environment, such that the s allele was associated with decreased risk of PTSD in a low-risk environment (low county-level crime and unemployment), but increased risk of PTSD in a high-risk environment (Koenen, Aiello et al. 2009). Replicating this finding in a different population (survivors of the Rwandan genocide), Kolassa et al. (2010) reported that those with the s/s genotype were at high risk for developing PTSD, even in response to a small number of traumatic events, whereas those with the l allele demonstrated an S-shaped dose-response curve with increasing rates of PTSD in response to an increasing number of traumatic life experiences. All three studies indicate risk associated with the s allele of 5-HTTLPR in association with exposure to environments of high stress/trauma.

A well described example of epigenetic modification in an animal model is provided by the work of Meaney and colleagues. In this rat model, maternal care characterized by higher amounts of licking and grooming produced lower levels of cortisol in the rat pups, but only if this variation in maternal care occurred during a specific, early developmental window. These rat pups also demonstrated enhanced suppression of cortisol in response to dexamethasone, as well as greater expression of the GR gene and a greater number of hippocampal GRs as a result of hypomethylation in promoter region of the hippocampal GR gene (Francis, Diorio et al. 1999; Weaver, Szyf et al. 2002). Further, by producing stable changes in stress response and behavior, these epigenetic changes and the associated phenotype were transmitted from one generation of female offspring to the next (Seckl and Meaney 2006; Weaver 2007). It is interesting to note that the neuroendocrine alterations described in this animal model parallel those of PTSD, with low basal cortisol and enhanced suppression of cortisol in response to a synthetic glucocorticoid. In fact, lower levels of GR mRNA have been demonstrated in the hippocampus of suicide victims who had a history of childhood abuse (McGowan, Sasaki et al. 2009). One caveat of these studies to date, is that it has not been possible to follow the state of methylation over time in a given individual or animal. Thus, although the interpretation is that there are early developmental methylation changes that are long-lasting and enduring, it cannot yet be ruled out that methylation may represent a state-dependent effect. In this alternative view, other causal factors in the adult animals and suicide victims (e.g. ongoing adult stress or depression) may lead to similar methylation effects as seen during development or in the previous generation, even if they were not maintained consistently over time in these individuals.

It can be hypothesized that the epigenetic-mediated changes in HPA axis reactivity could be associated with increased vulnerability to PTSD following subsequent traumatic exposure, perhaps by failing to constrain stress-induced elevations in noradrenergic neurotransmission leading to overconsolidation of fear memories, as well as resulting in increased physiologic arousal and distress. These findings could also help explain the mechanism by which early life trauma is a strongly validated risk factor for the subsequent development of PTSD in adulthood by recalibrating the set point and stress-responsivity of the HPA axis in an enduring manner that increases vulnerability to PTSD (Yehuda, Flory et al. 2010).

While there have been no published studies providing specific proof of the role of epigenetic modification in association with risk for PTSD, the results of several studies on intergenerational transmission of risk for PTSD could plausibly be explained by epigenetic mechanisms. Yehuda and colleagues (2007; 2008) have reported increased risk for PTSD, as well as low cortisol levels, in the offspring of female, but not male, Holocaust survivors with PTSD. As with the rat model, this increased risk may be produced by alternations in maternal care associated with the mother's PTSD, which may disrupt child-maternal attachment, a known risk factor for PTSD (Charuvastra and Cloitre 2008). These effects may also produce enduring changes in HPA axis responsivity to stressors that may influence risk for PTSD following subsequent traumatic exposure (Brand, Brennan et al. 2010).

The exaggerated startle response that is characteristic of PTSD can be studied in the context of the fear conditioning model by examining fear inhibition – the ability to suppress fear responses in the presence of safety cues. One of the primary laboratory models used for behavioral analysis of fear inhibition is the conditioned inhibition paradigm, in which the pairing of a conditioned stimulus (CS+) with an aversive stimulus (US) is intermingled with a separate stimulus (CS-) that is never paired with the US, and thus represents a safety signal. The acoustic startle response provides a useful behavioral outcome measure of fear in these paradigms as it is easily measured in the lab, is biologically relevant to PTSD, and is mediated by the output of the amygdala (LeDoux, Iwata et al. 1988; Davis 1992; Grillon 2008). This response is typically reported as fear-potentiated startle, which is the relative increase in the acoustic startle response elicited in the presence of the reinforced conditioned stimulus.

Myers and Davis (2004) developed an animal model using a conditional discrimination procedure that allows for the independent evaluation of excitation and inhibition of fear conditioning. A version of this paradigm studied in humans contains a danger signal (AX+), a safety signal (BX-), and a safety transfer test (combination of A and B, where B reduces fear to A). The A stimulus elicits fear potentiation of startle while the B stimulus elicits reduced startle compared to A (Jovanovic and Ressler 2010).

Conditioned inhibition of the acoustic startle response as a result of learning the safety signal has been demonstrated in healthy controls and in combat veterans with low levels of current PTSD symptoms (Jovanovic, Norrholm et al. 2009). In contrast, study subjects with high levels of PTSD symptoms were unable to reduce startle to AB trials. This has now been demonstrated in combat veterans from the United States, as well as veterans from Croatia (Jovanovic, Norrholm et al. 2007), and in a traumatized civilian population in urban Atlanta (Jovanovic, Norrholm et al. 2010). As these individuals with high levels of PTSD symptoms did not demonstrate fear potentiation to the danger signal (AX+) along, the results suggest a selective deficit in fear inhibition as a key trait of PTSD. It is unclear whether an impaired ability to inhibit fear responses is a risk factor for PTSD versus an acquired trait associated with the disorder.

As reviewed above, the neurocircuitry of fear inhibition involves the limbic-frontal neuronal systems. The amygdala is required for associative learning of conditioned fear and elaboration of the fear response. Specifically, output from the CeA to the pons in the midbrain mediates the fear-induced increase in the acoustic startle response. Descending inhibitory input from the mPFC to the amygdala is required for extinction or inhibition of conditioned fear responses. The key role of this circuit in regulating fear potentiated startle suggests that researchers should investigate genes known to impact on this system to assess the genetic basis of conditioned fear. As noted above a polymorphism in 5-HTTLPR is known to be associated with relative decoupling of mPFC inhibition from the amygdala (Pezawas, Meyer-Lindenberg et al. 2005) and interactions between this polymorphism and a history of trauma have been found to be predictive of PTSD. In addition, a recent study found great fear-potentiated startle in individuals with the s (risk) allele of this gene (Lonsdorf, Weike et al. 2009). It would be of interest to determine if those individuals with the s allele also demonstrate deficits in conditioned inhibition of fear.