Pharmacological Treatment of PTSD – Established and New Approaches

Introduction

Posttraumatic stress disorder (PTSD) results from exposure to a traumatic event which evoked fear, helplessness and horror. It is characterized by three symptom clusters, i.e., (1) hypermnesia for the core traumatic event, with frequent re-experiencing of the traumatic event in form of flashbacks and nightmares – aversive memories that can be triggered by sensorimotor cues, for example, a noise that reminds the patient of the traumatic event – and disturbed memory for peritraumatic events, (2) hyperarousal, characterized by exaggerated startle, hypervigilance and irritability, and (3) avoidance behaviour, such as avoidance of reminders associated with the trauma. Symptoms should persist for a minimum of four weeks before a diagnosis is made. PTSD affects a subpopulation (10–15%) of people exposed to traumatic events, with a lifetime prevalence of 6.8% in the US (Kessler et al., 2005).

Neural circuits and substrates implicated in PTSD

Conceptually, PTSD can be considered as a maladaptation to a traumatic stressor, with altered fear-related learning (fear conditioning) and extinction, behavioural sensitisation/kindling, and alterations in brain areas and neurotransmitter systems closely linked to these processes. Here we will review these processes, their interactions and potential treatment strategies to ameliorate them. A large amount of literature now focuses on the corticolimbic circuit in PTSD, with neuroimaging studies reporting abnormalities in the prefrontal cortex (PFC), hippocampus and amygdala in PTSD patients (Milad and Rauch, 2007; Quirk and Mueller, 2008). These neural circuits are implicated in the putative fear learning abnormalities and sensitization reported in PTSD. For example, insufficient top-down control from the PFC to the amygdala has been suggested to play a role in impaired extinction of fear-related memories (Koenigs and Grafman, 2009; Milad et al., 2009) and executive control over fear responses (Aupperle et al 2011, this issue). Poor hippocampal-PFC signalling may also underlie contextual memory deficits in PTSD, resulting in poor contextual control of conditioned fear responses (Acheson et al 2011, this issue). Many of these pathways are involved in different putative phases of PTSD development, either initial fear learning, maintenance of fear memory/responses or extinction. We will discuss the treatment strategies, either prophylactic or therapeutic, targeted at these pathways.

Consideration of these pathways suggests involvement of certain neurotransmitter and - modulator systems: The main projections from the PFC to the amygdala or to dopamine or acetylcholine inputs into the amygdala are glutamatergic in nature (Del Arco and Mora, 2009). Thus, insufficient top-down control from the PFC to the amygdala implies involvement of glutamatergic pathways in PTSD, either directly or indirectly. For example, it is thought that fear extinction requires PFC-activation of intercalated cells in the amygdala, GABAergic interneurons that inhibit local activation and express a unique receptor profile (Likhtik et al. 2008). Hence, at the level of the amygdala, different sub-nuclei can affect each other via glutamatergic or GABAergic interactions (Pitkanen et al., 1997; Amano et al., 2010), bringing the GABAergic system into play as a potential target for PTSD therapeutics. More recently, another functional pathway involved in acute stress responses has been delineated, consisting of an indirect pathway for inhibition of the hypothalamic-pituitary-adrenal (HPA) axis. The PFC inhibits HPA activity via a glutamatergic projection to the bed nucleus of the stria terminalis (BNST), part of the extended amygdala, which activates a GABAergic inhibitory projection from the BNST to the corticotropin-releasing factor (CRF) neurons in the hypothalamic paraventricular nucleus (PVN) (Radley et al., 2009). This pathway may be particularly relevant as PTSD patients exhibit increased cerebrospinal fluid (CSF) levels of CRF (Baker et al., 1999; Bremner et al., 1997) and abnormalities in other HPA axis systems (e.g. pituitary adenylate cyclase-activating polypeptide, PACAP, Ressler et al. 2011) suggests utility of compounds that dampen the CRF system or other HPA axis hormones in the treatment of PTSD (Baker et al., 2009).

Neural circuits and substrates underlying acute stress responding and trauma memory encoding – targets for prevention

A number of interrelated neurochemical systems have been suggested to be involved in the mediation of stress responsivity, formation of traumatic memories and the pathophysiology of PTSD, including glutamate, GABA, CRF and noradrenaline, amongst others. Evidently, there are strong interactions between these systems, giving rise to different therapeutic approaches that could be useful to prevent the development of PTSD. Acute stress exposure, for example, which may mimic the acute traumatic event leading to PTSD, induces increases in glutamate transmission across multiple systems: PFC, amygdala, BNST, hippocampus and noradrenergic locus coeruleus (LC) in rats (Gilad et al., 1990; Moghaddam, 1993; Reznikov et al., 2007; Walker and Davis, 2008). It has been suggested that insufficient top-down control of these circuits from the PFC could lead to stress hyperreactivity. Poor PFC control of the LC could lead to hyperreactivity of the noradrenergic projection from the LC to the basolateral amygdala (BLA), while poor PFC control of the PVN could lead to increased CRF and downstream glucocorticoid signalling (Hurlemann, 2008; Hurlemann et al., 2007). These systems often act in a reciprocal fashion, with altered glucocorticoid signalling in turn affecting acute glutamatergic neurotransmission in cortico-limbic circuits (Moghaddam et al., 1994). There is also evidence for reciprocal modulation across CRF and NE systems, with increased NE driving increased CRF release and vice versa (Gresack and Risbrough, 2010; Dunn et al. 2004). Thus, there are strong interactions between the different neuroanatomical and chemical systems that have been implicated in PTSD and pharmacological manipulations of the glutamatergic or GABAergic systems, the CRF system, the noradrenergic system, or normalization of HPA axis activity by other means could be of utility in the treatment of PTSD, directly or indirectly affecting the different neurochemical systems involved.

Likewise, the neurochemical systems that mediate acute behavioural and neuroendocrine responses to stress also modulate the neuroplastic events that occur during trauma processing, e.g., increased glutamatergic neurotransmission at the time of exposure to the traumatic event may facilitate encoding of the traumatic memory in PTSD patients. This process may be enhanced by altered glucocorticoid release from the HPA axis as glutamate-induced NMDA receptor activation and stimulation of the glucocorticoid receptor (GR) by glucocorticoids facilitates the activity of common intracellular signalling pathways critical for memory consolidation. Thus, it has been suggested that GR and glutamate signalling may synergistically facilitate activation of the extracellular-signal-regulated kinase (ERK)/mitogen- and stress-activated kinase (MSK), leading to histone phospho-acetylation and chromatin remodelling, a putative molecular substrate of these memories (Reul and Nutt, 2008). NMDA receptor activation has also been suggested to play a role in some of the kindling-like processes that have been associated with the formation of spontaneous intrusive memories (Grillon et al., 1996; Adamec, 1997) and states of high NMDA receptor activity and high glucocorticoid function may serve as risk factors for developing PTSD as it may increase the likelihood for aversive memory encoding (Reul and Nutt, 2008; Mehta and Binder, in press).

Hyperreactivity of the noradrenergic projection from the LC to the BLA, in conjunction with disinhibited glucocorticoid signalling, and the resultant enhanced signalling from the BLA to the anterior hippocampus via the subiculum, has also been suggested to facilitate encoding of the traumatic event at the time of the trauma, thereby contributing to the development of PTSD (Hurlemann, 2008; Hurlemann et al., 2007). Thus, enhanced glucocorticoid signalling and noradrenergic activation may act synergistically at the level of the BLA, leading to potentiation of noradrenaline-induced activation of the cAMP-dependent protein kinase A (PKA) by GR stimulation (Roozendahl et al., 2002b), which in turn will enhance cAMP response element-binding (CREB) protein phosphorylation and consequently chromatin remodelling as well. At the same time, glucocorticoids interact with noradrenergic mechanisms in interfering temporarily with memory retrieval (De Quervain et al., 2007), which could lead to disturbed recollection of peritraumatic events. Finally, high CRF levels at the time of trauma may also facilitate encoding of trauma memory and enduring anxiety effects via direct action at CRF1 receptors (Hubbard et al. 2007, Roozendaal et al. 2008, Adamec et al. 2010).

Thus, there is strong evidence that many of the systems that mediate stress responses also facilitate encoding of aversive memories, which could form the basis for the development of PTSD and open up avenues for the development of novel prevention strategies.

Neural circuits and substrates underlying chronic stress responding and trauma memory retrieval – targets for symptom reduction

Once PTSD is established, the situation may be different: in established PTSD, lower basal 24h circulating cortisol levels have been reported, which may be due to enhanced negative feedback inhibition of the HPA axis by glucocorticoids at the level of the pituitary gland and/or hyporeactivity of the adrenals or the hypothalamus (Yehuda, 2005). These low cortisol levels have been suggested to play a role in re-experiencing of the traumatic event as they may facilitate retrieval of the aversive memories (De Quervain et al., 2009; but see Baker et al., 2005, reporting elevated CSF cortisol levels in PTSD patients despite normal plasma and urinary cortisol levels, suggesting that plasma cortisol is not representative of central cortisol level). At the same time, it seems that there is a greater reactivity of the HPA axis to stressors, which renders the HPA axis maximally responsive to stress-related cues in PTSD (Yehuda, 2005), potentially facilitating re-consolidation of the aversive memories (Taubenfeld et al. 2009).

Recent data generated in an animal model of PTSD, i.e., serial application of three different stressors (called the single prolonged stress model), that recapitulates aspects of established PTSD, i.e., enhanced negative feedback of the HPA axis and enhanced startle reactivity (Khan and Liberzon, 2004; Kohda et al., 2007), also suggests that the glutamatergic system may undergo changes over time: contrary to the increase in glutamate release seen following acute stress exposure, it has been reported that single prolonged stress led to attenuated PFC glutamate levels in rats (Knox et al., 2010). This reduction in glutamate levels could model the putative reduced PFC activity in PTSD patients.

The hyperreactivity of the noradrenergic system seems to persist in PTSD patients and has been suggested to mediate the hyperarousal symptoms seen, as well as the sleep disturbances reported in this disorder (Southwick et al., 1999a; Liberzon et al., 2005; Raskind et al. 2003). Increased NE signalling at the amygdala and hippocampus may also facilitate retrieval of aversive memories (Southwick et al., 1999b).

Another monoaminergic system linked to PTSD is the serotonergic (5-HT) system (e.g., Krystal and Neumeister, 2009). Polymorphism of the 5-HT transporter (the 5-HTTLPR genotype), in interaction with adult traumatic events and childhood adversity, has been reported to be a susceptibility factor for PTSD (Lee et al., 2005; Grabe et al., 2009; Xie et al., 2009). Furthermore, stress has been reported to increase 5-HT neurotransmission in several forebrain regions, including frontal cortex, hippocampus and amygdala (Linthorst, 2005). Selective serotonin re-uptake inhibitors (SSRIs) are also efficacious in treating the disorder at least in some individuals, suggestive that 5-HT could play a role in the pathogenesis of the disorder (Bandelow et al., 2008).

Thus, neurotransmitter/neuromodulatory systems that have been or could be targeted by a pharmacological approach to treat PTSD could include serotonergic, noradrenergic, GABAergic and glutamatergic mechanisms, manipulations that affect HPA axis reactivity via, for example, glucocorticoid or CRF receptor manipulations, as well as intracellular signalling cascades associated with these systems and that may represent final common pathways.

It is evident that pharmacological approaches may differ, depending on whether treatment focuses on the development aspects of PTSD (preventive intervention around the time of trauma when processes that could lead to PTSD may be initiated) or whether it aims at treating chronic PTSD (symptom reduction). For instance, one may want to prevent the consolidation of trauma-related memories early on by blockade of glutamatergic activity, while one may wish to facilitate extinction of those memories once they have been established by enhancing glutamatergic function, i.e., opposite mechanisms of action may be of utility, depending on the stage of the illness one wants to target.

Prevention

The treatment of chronic PTSD encompasses monotherapies or adjunctive therapies to current pharmacological treatments or psychotherapy once symptoms have developed. Preventive treatment starts prior to symptom development. Symptomatic or preventive approaches raise different ethical and socioeconomic concerns: only a subpopulation of those experiencing a traumatic event will also develop PTSD and indiscriminate treatment should be avoided. Therefore, it would be useful for preventive treatment to be effective to distinguish subjects that are at risk to develop PTSD from those that are not. However, although a number of such markers have been proposed, e.g., lower cortisol levels, increased heart rate dynamics shortly after the traumatic event or increased circulating PACAP (Yehuda, 2004; O’Donnell et al. 2007; Ressler et al. 2011) or other vulnerability factors, such as polymorphism of the 5-HT transporter (see above) or of FkBP5, a co-chaperone that modulates the glucocorticoid receptor (GR) (Binder et al. 2008), none qualify so far as a prognostic tool with sufficiently high accuracy. Another complicating factor of preventive pharmacological approaches is the need of such treatment to effectively counteract the development of PTSD symptoms, while leaving normal function undisturbed, i.e., the normal psychological responses to traumatic events, including cognitive and psychomotor function, should remain unimpaired.

There are some at risk populations, e.g. soldiers facing combat, in which preventive pharmacological treatment before the traumatic event may be feasible. This strategy is called prospective or primary prevention. Alternatively, preventive treatment could be given shortly after the traumatic event, but well before symptoms develop. This is called retrospective or secondary prevention and aims at preventing or blocking the induction or consolidation of processes leading to PTSD. In the latter case, only those that really experienced trauma would need intervention, which opens it up to a wider group of people, including those that faced traumatic experiences under circumstances where trauma is less likely to occur, e.g., following a car accident.

Pharmacological approaches for primary prevention

GR ANTAGONISTS, CRF1 ANTAGONISTS and CCK2 ANTAGONISTS

One strategy for preventative treatment would be to enhance stress coping, i.e., to facilitate stress resilience. A number of preclinical studies have investigated molecular mechanisms involved in the stress response: blockade of the glucocorticoid receptor (GR) prior to exposure to a single prolonged stressor prevented the development of enhanced fear responses in rats (Kohda et al., 2007), CRF1 receptor antagonism prevented the initiation of stress effects in a mouse predator stress model of PTSD (Adamec et al., 2010), and similar findings have been reported with CCK2 antagonism (Adamec et al., 1997). These effects may be mediated via inhibition of the HPA axis or via central effects at limbic circuitry. Indeed, CRF1 receptor antagonism directly at the amygdala alone attenuates fear conditioning (Hubbard et al. 2007; Roozendaal et al. 2002a), which may contribute to the effects of these drugs. For drugs affecting HPA axis activity however, these findings would suggest that interventions that prevent an exaggerated stress response may be beneficial prior to the occurrence of the traumatic event. Following this line of reasoning, it would also be conceivable that other classes of compounds that block HPA activity, such as vasopressin antagonists, e.g., V1b antagonists, could be of utility. Interestingly, vasopressin has been shown to affect consolidation processes, either directly or indirectly (e.g., Ettenberg et al., 1982), further strengthening the case.

GR AGONISTS

However, it should also be noted that individuals with lower peri-traumatic cortisol levels have an increased likelihood for developing PTSD (Yehuda, 2004). This finding in turn suggests that the above mentioned approaches to facilitate stress resilience paradoxically may increase the risk of individuals to develop PTSD. This paradoxical effect may be due to loss of feedback inhibition of the HPA axis, in that lower cortisol levels at the time of the traumatic event may prevent termination of the sympathetic stress response and consequently prolonged noradrenergic activity (Pacak et al., 1995; see also below). Based on this hypothesis it was suggested that increased cortisol during the traumatic event may block development of PTSD. Indeed, beneficial effects of posttrauma hydrocortison have been reported in a few small, randomized clinical trials (Schelling et al., 2001, 2004; Weis et al., 2006). Interestingly, high-, but not low-dose corticosterone administered shortly after a predator stress attenuated stress-related behavioural responses in rats, and it has been suggested that high-dose corticosterone disrupts memory consolidation for the traumatic event, while low-dose corticosterone facilitates memory consolidation (Cohen et al., 2008).

Thus, the net effect of direct or indirect GR manipulations seems to be dose-dependent and outcome of such manipulations in PTSD patients may depend on exposure achieved at the GR. It will be difficult to predict this response at the individual level as cortisol efficacy may depend on individual differences in cortisol responses to stress and in expression of genes modulating GR signalling (e.g. FKBP5, see Mehta and Binder, in press, this issue). In sum, there seems to be an inherent risk that interventions that inhibit GR signalling, either directly or indirectly, could actually facilitate the development of PTSD. Clearly more studies are required to delineate the complex role of GR signalling effects during and after trauma to develop appropriate prophylactic treatments targeted at this system.

Pharmacological approaches for secondary prevention

GR ANTAGONISTS, CRF1 ANTAGONISTS and CCK2 ANTAGONISTS

However, the design of the preclinical studies mentioned above do not allow unambiguous conclusions that treatment given prior to stress exposure really mirrors primary prevention. Drug effects may be carried over to post-stress conditions and hence could still have effects on consolidation processes, reflecting secondary prevention. In support of this argument, it has been shown that the protein synthesis inhibitor anisomycin, administered either shortly before or after predator stress, also attenuated anxiety-related behaviour in rats (Cohen et al., 2006). Of note, de novo protein synthesis is critical for successful consolidation processes to take place, but not necessarily for stress responsivity. Likewise, it has been shown that CRF1 receptor antagonism or CCK2 receptor blockade also prevented the consolidation of stress effects in rodent stress models of PTSD (Adamec et al., 1997b, 2010; Wang et al., 2010). GR blockade also interferes with aversive memory consolidation at the level of the basolateral amygdala (Roozendaal, 2000). These effects on consolidation support the utility of these compounds in secondary prevention and suggest that their efficacy in models of primary prevention might be confounded with effects on consolidation.

ADRENOCEPTOR AGONISTS and ANTAGONISTS

Other secondary preventive approaches focused on manipulations of the noradrenergic system, for example by prevention of presynaptic noradrenaline release with α2 adrenoceptor agonists or opioids. The α2 adrenoceptor agonist dexmedetomidine indeed blocks fear consolidation (Davies et al., 2004), although this was tested in normal mice not in a PTSD mouse model. However, no preventive clinical PTSD studies using α2 adrenoceptor agonists have been reported. Blocking postsynaptic noradrenaline receptors seems less efficacious as a preventative treatment: the α₁ adrenoceptor antagonist prazosin failed to block increases in stress-related types of behaviour in rats exposed to predator stress (Adamec et al., 1999). Some considered the β adrenoceptor blocker propranolol as being the most promising candidate drug for intervention after a traumatic event (Pitman and Delahanty, 2005) as it has shown efficacy in preventing some trauma-related physiological reactivity (Pitman et al. 2002). However, a subsequent clinical trial found propranolol to be ineffective when given immediately after trauma (e.g. in the hospital) to prevent the development of PTSD as measured by clinical rating scales (Stein et al., 2007). On a side note, it is worth mentioning that although propranolol may not itself show efficacy, these studies do support the feasibility to examine potential prophylactic treatment approaches with future novel targets.

NMDAR AND GABAERGIC COMPOUNDS

NMDA receptor antagonists also interfered with anxiety-related behaviour in rats if given shortly after exposure to predator stress (Adamec et al., 1999), which may not come as a surprise given the involvement of NMDA receptors in memory consolidation processes. In this respect, it is worth noting that in a preliminary, retrospective study, McGhee et al. (2008) found that in a group of burned service men those treated with the NMDA receptor antagonist ketamine during hospitalization had lower incidence of developing PTSD. These preclinical and clinical findings support the utility of novel pharmacological tools targeting NMDA receptor subunits or function could be of benefit while avoiding some of the side effects inherent to NMDA receptor blockade. Some examples of possible targets are metabotropic glutamate receptor (mGluR) 2 positive allosteric modulators (PAMs), which reduce glutamate release via presynaptic negative feedback, antagonists or negative allosteric modulators (NAMs) of postsynaptic mGluR5 receptors (but see Fendt and Schmid, 2002), or NR2B antagonists.

Moreover, the benzodiazepine alprazolam exaggerated stress effects when given shortly after predator stress exposure to rats (Matar et al., 2009). In humans, benzodiazepines have also previously been shown to facilitate memory for events that occurred just prior to treatment (Hinrichs et al., 1984), presumably due to blockade of active interference during consolidation. These findings are in line with clinical reports that secondary prevention with benzodiazepines has no effect (Gelpin et al., 1996; Mellman et al., 2002) or could even increase the likelihood of trauma victims to subsequently develop PTSD. This lack of efficacy may be due to the difference in efficacy of benzodiazepines to induce retrograde vs. anterograde amnesia, in other words benzodiazepines predominantly disrupt active associate processes and only affect consolidation when very high doses are used (Cahill et al. 1986; Jensen et al. 1979). Thus drugs that are efficacious in inducing mild retrograde amnesia may be more fruitful than drugs that facilitate anterograde amnesia only (L. Cahill personal communication).

OPIODS

There is some evidence that morphine administration shortly after the traumatic event reduces the likelihood for trauma victims to develop PTSD (Saxe et al., 2001; Bryant et al., 2009; Holbrook et al., 2010). Thus far studies have been purely naturalistic and further random controlled studies are needed to confirm these intriguing findings. The potential mechanism of these effects is unknown, however it is possible that it could be via an indirect reduction in noradrenergic activity after morphine treatment, or a direct action at intercalated cells in the amygdala that are critical for fear extinction processes (Likhtik et al. 2008).

Treatment of Established PTSD – Non-cognitive symptoms

Once the disorder is established, one could consider targeting the emotional response, i.e., the expression of fear or other non-cognitive symptoms associated with PTSD, such as hyperarousal, or the cognitive processes associated with PTSD, such as retrieval of aversive memories or extinction of fear-related memories. Of note, treatments that suppress non-cognitive PTSD symptoms are the only currently approved pharmacotherapeutic strategy.

SSRIs

Selective serotonin re-uptake inhibitors (SSRIs) have shown efficacy in reducing symptom severity and in relapse prevention in PTSD patients (e.g., Van der Kolk et al., 1994; Connor et al., 1999; Brady et al., 2000; Martenyi et al., 2002; McRae et al., 2004; Davidson et al., 2006; Onder et al., 2006), although only approximately 60% of patients respond to the treatment and only about 20 – 30% of patients will achieve full remission (Stein et al., 2002; Zohar et al., 2002). However, a recent report from the Institute of Medicine concluded that current evidence to determine efficacy of SSRIs is at best suggestive (Committee on Treatment of Posttraumatic Stress Disorder, 2008) and more recent guidelines on the treatment of PTSD question the use of SSRIs for veterans with combat-related PTSD relative to their therapeutic benefit in patients with non-combat-related PTSD (Benedek et al., 2009). Thus, while SSRIs can be considered relatively well tolerated and safe, further studies with higher power are needed before conclusions can be drawn. Moreover, SSRIs still suffer a number of shortcomings, including delayed onset of action, partial response with residual symptoms, or non-response, and undesirable side effects (e.g., loss of sexual drive, gastrointestinal effects, changes in body weight) which limits their utility and indicates a major unmet medical need for novel treatment approaches in PTSD.

OTHER ANTIDEPRESSANTS

Besides SSRIs, a number of other pharmacological approaches have been investigated in the clinic for treating PTSD patients, including other antidepressants, adrenoceptor antagonists, anticonvulsants, atypical antipsychotics and benzodiazepines (see Ravindran and Stein, 2010, for a review). Antidepressant drugs include dual serotonin and noradrenaline re-uptake inhibitors, such as venlafaxine (Davidson et al., 2006), tricyclic antidepressants such as amitriptyline (Davidson et al., 1990) and imipramine (Frank et al., 1988), monoamine oxidase inhibitors (MAOIs) like phenelzine (Frank et al., 1988), reversible monoamine oxidase A inhibitors (RIMAs) such as moclobemide (Onder et al., 2006), as well as drugs with other mechanism of action, like the 5-HT_2A/2C antagonist/5-HT re-uptake inhibitor nefazodone (McRae et al., 2004), the mixed α_2A/2C adrenoceptor antagonist/5-HT_2A/2C/3 antagonist mirtazapine (Davidson et al., 2003) and 5-HT re-uptake enhancer/glutamate modulator tianeptine (Onder et al., 2006). While these drugs showed therapeutic utility in clinical trials and some of them seem to be equally effective as SSRIs, they have not become first line treatment for PTSD, partly also because they are less well tolerated (Bandelow et al., 2008). Although the primary mechanism of action differs amongst these antidepressant drugs, it is noteworthy that all of them interact with monoaminergic (serotonergic and noradrenergic) systems. In addition, antidepressants of various classes have been shown to normalize HPA axis activity in response to stress and to enhance hippocampal neurogenesis (Reul et al., 1994; Gold et al., 1995; Matheson et al., 1997; Stout et al., 2002; Xu et al., 2006; Kasper and McEwen, 2008; Szymanska et al., 2009; McEwen et al., 2010), which may represent a final common pathway.

ADRENOCEPTOR AGONISTS AND ANTAGONISTS

There have also been attempts to normalize the noradrenergic hyperreactivity suggested to underlay PTSD hyperarousal symptoms. Blockade of the α₁ adrenoceptor with the α₁ adrenoceptor antagonist prazosin has been reported to improve various PTSD symptoms, but in particular sleep and ameliorate nightmares (Peskind et al., 2003; Raskind et al., 2003; 2007; Taylor et al., 2008). A recent comparison study of prazosin and quetiapine for ameliorating night-time sleep disturbance indicates better overall tolerability of prazosin (Byers et al 2010). Ligands acting at the α₂ receptor have been less promising. Mirtazapine has shown positive results in one study of PTSD (Davidson et al., 2003), which suggests that α_2A/C adrenoceptor blockade also has beneficial effects, although mirtazapine’s effects could of course also be mediated via serotonergic mechanisms (Yamamura et al. 2011). In this respect it is of note that mirtazapine increases NE release in various brain areas via α_2A autoreceptor blockade (Haddjeri et al., 1996), thereby facilitating α₁ activity. Stimulation of presynaptic α_2A autoreceptors with the α_2A adrenoceptor agonist guanfacine, which should lead to reduced noradrenaline release, on the other hand, failed to reveal therapeutic benefit (Neylan et al., 2006; Davis et al., 2008). However agonist activity at post-synaptic α_2A receptors may mask effects of presynaptic blockade of NE release, as guanfacine has potent agonist activity at post-synaptic receptors (Arnsten et al., 1988). Non-selective β adrenoceptor blockade with propranolol was ineffective when given to prevent the development of PTSD in a recent study by Stein et al. (2007), although there is some resurgence of interest in using propranolol in conjunction with therapy (see below). Thus, while there is some evidence that manipulations of noradrenergic activity may have utility specifically for sleep disturbances, efficacy for overall symptom reduction is not supported thus far.

ANTICONVULSANTS

Anticonvulsant drugs have been proposed to be of benefit in treating PTSD due to their anti-kindling effects (Hageman et al., 2001; Berlin, 2007). This is a very heterogeneous group of drugs and often their mechanism of action is poorly understood. However, some anticonvulsants, such as lamotrigine and topiramate, have downstream effects that include inhibition of glutamate neurotransmission (Ahmad et al., 2004; Sitges et al., 2007), which could also be a mechanism through which tianeptine affects PTSD (Reznikov et al., 2007). As will be discussed below, glutamatergic approaches offer potential for the development of novel pharmacological treatments for PTSD. However, while some authors consider treatment with anticonvulsant drugs to be a promising approach for PTSD (e.g., Hageman et al., 2001; Adamou et al., 2007; Berlin, 2007), others conclude that the use of anticonvulsants in PTSD has only very limited support based on recent clinical trials (Berger et al., 2009; Ravindran and Stein, 2010). Likewise, only a limited number of randomized clinical trials have evaluated the effects of benzodiazepines in PTSD, with no or modest beneficial effects in PTSD patients (Braun et al., 1990; Gelpin et al., 1996; Mellman et al., 2002; Cates et al., 2004).

ANTIPSYCHOTICS

Atypical antipsychotic drugs are largely used as adjunctive therapy to e.g. antidepressant drugs in the treatment of PTSD. Only a limited number of randomized, double-blind, placebo-controlled clinical trials have been reported with risperidone and olanzapine, leading to mixed results (e.g., Butterfield et al., 2001; Stein et al., 2002; Hamner et al., 2003; Monnelly et al., 2003; Reich et al., 2004; Padala et al., 2006).

Thus, although there is evidence that pharmacological approaches using psychoactive drugs that are currently in the clinic have some beneficial effects in PTSD, with the most convincing data generated for antidepressant drugs, this evidence must be considered mixed. Almost all of the drug classes examined for efficacy in PTSD suffer from a dearth of adequately powered studies to support definitive conclusions either for or against efficacy. Along these lines, the Committee on Treatment of Posttraumatic Stress Disorder (2008) concluded that for all the drug classes mentioned above, the evidence is inadequate to determine efficacy in the treatment of PTSD. Overall the field clearly requires more efficacious pharmacological approaches to treat this disorder, as well as a more concentrated effort to adequately test potential therapeutics in large randomized clinical trials (Leon and Davis, 2009).

Treatment of Established PTSD – Cognitive symptoms

Once a memory about the traumatic event is formed, a number of processes take place that could still be amenable to pharmacological intervention. Retrieval refers to the activation of an aversive memory trace of the traumatic event that led to PTSD, for example, in flashbacks, nightmares or intrusive recollections of the traumatic event, with an external or internal stimulus that triggered the recollection. Preventing retrieval of such memories would be one possible strategy to improve PTSD symptoms.

Memory retrieval, consolidation and re-consolidation

GR AGONISTS

Memory retrieval in animals (De Quervain et al., 1998) and humans (Newcomer et al., 1999) has been shown to be impaired by the administration of glucocorticoids (but see Tollenaar et al., 2009, for negative results measuring physiological responses to aversive emotional memories in healthy volunteers following cortisol administration). Preliminary data from a case-control study with three PTSD patients seem to support the idea that low-dose cortisol treatment reduced the ratings of the severity of traumatic memories (Aerni et al., 2004). These preliminary findings need to be substantiated in an appropriately powered vehicle controlled, randomized, double-blind clinical trial before any conclusions can be drawn. However, the data would suggest that cortisol treatment may be beneficial both as secondary prevention and to interfere with the retrieval of aversive memories once PTSD is established.

Another therapeutic option that has been more widely investigated would be to modify reconsolidation processes. Reconsolidation refers to the fact that a memory is “re-consolidated” after reactivation/retrieval. In PTSD, flashbacks or intrusive memories about the traumatic event represent a retrieval of that aversive memory trace, that will subsequently be reconsolidated (Charney 2004). Prevention of reconsolidation, which over time should lead to a weakening of the aversive memory trace, may represent another window of opportunity for pharmacological intervention. Not surprisingly, some of the same pharmacological mechanisms that have been suggested to play a role in the consolidation of traumatic memories have also been suggested to be of relevance for reconsolidation processes, including GR receptors (Tronel and Alberini, 2007; Taubenfeld et al., 2009). For example, a recent double-blind placebo controlled study found that glucocorticoid administration after imagery-driven reactivation of trauma memories had a temporary (<1 mo) effect on PTSD symptom severity (Suris et al. 2010).

NMDAR AND GABAERGIC COMPOUNDS

NMDA receptors also play an important role in reconsolidation processes (Suzuki et al., 2004; Lee et al., 2006) and it can be suggested that manipulations that (indirectly) attenuate NMDA receptor function (such as treatment with mGluR2 PAMs or mGluR5 NAMs) may also be of benefit. Conversely, facilitation of GABAergic function by the benzodiazepine midazolam disrupted reconsolidation processes of fear memory (Bustos et al., 2006; Zhang and Canney, 2008), suggesting that this class of drugs may have utility in the treatment of PTSD, although possibly not for preventive treatment. The question is how these effect on reconsolidation can be readily translated to clinical research, at what times does reconsolidation normally occur post-trauma, either naturalistically or how can it be induced during clinical intervention.

It has also been reported that inhibition of the mammalian target of rapamycin (mTOR) inhibits reconsolidation of fear memory (Blundell et al., 2008). This finding is interesting for two reasons: First, it could open another new avenue of treatment for PTSD using mTOR kinase inhibitors, although it is of note that those compounds also modulate the immune response, which may prevent their use for the indication of PTSD. Second, it has recently been reported that mTOR-dependent synapse formation underlies the rapid antidepressant effects of NMDA antagonists (Li et al., 2010), and it is tempting to speculate that the same signalling cascades may also play a role in the potential therapeutic effects of NMDAR blockade in PTSD.

CANNABINOIDS

Another interesting pharmacological approach could be the manipulation of the cannabinoid system, as it was shown that bilateral infusion of CB1 receptor agonists into the amygdala after memory reactivation blocked reconsolidation of fear memory (Lin et al., 2006), while bilateral hippocampal CB1 blockade facilitated reconsolidation of fear memory (De Oliveira Alvares et al., 2008). Clearly, treating PTSD patients with CB1 agonists would be problematic, not at least because of the abuse risk of such compounds. However, one could consider indirect manipulation of the endocannabinoid system, for example, by fatty acid amide hydrolase (FAAH) inhibition, which prevents degradation of endogenous endocannabinoids such anandamide. Interestingly, anandamide administration into the hippocampus blocked reconsolidation of fear conditioning (De Oliveira Alvares et al., 2008), suggesting increasing endogenous levels may be a useful strategy to attenuate reconsolidation processes in PTSD.

ADRENOCEPTOR ANTAGONISTS

The effects of the β-adrenoceptor antagonist propranolol were studied by Debiec and LeDoux (2004), reporting that propranolol injected into the amygdala blocked reconsolidation, but not consolidation. These findings suggest that propranolol may be of utility in PTSD patients during reconsolidation, while being less efficacious for secondary prevention. However, other studies suggest that the efficacy of systemically administered propranolol to affect reconsolidation is limited and depends on the specific preclinical test used - inhibitory avoidance versus fear conditioning in that particular study (Muravieva and Alberini, 2010). In humans, there is preliminary evidence from a recent small, placebo-controlled, randomized double-blind clinical trial by Brunet et al. (2008) suggesting that propranolol may be beneficial when given to patients with chronic PTSD after they had to retrieve the traumatic memory, i.e., during reconsolidation. While encouraging, this finding must be confirmed in a larger trial.

Limitations to the strategy of disrupting consolidation or re-consolidation are that the treatment must not disrupt other critical processes such as fear extinction learning, and must not disrupt normal, non-trauma related cognitive processes. One practical way to get around this issue is to treat the patient only in the clinic during specific re-consolidation based therapy (e.g. Brunet et al. 2008). For chronic use pre-clinical studies of such target compounds will need to be conducted to evaluate the potential side effects of these drugs on other cognitive domains. Legal/ethical concerns have also been voiced for drugs that alter memory of traumatic events that result in legal actions (e.g. rape). For a consideration of the legal aspects of therapeutic strategies that interfere with memory consolidation see a recent review by Fletcher et al. (2010).

Fear extinction

Patients suffering from PTSD are also impaired in extinction of learned fear (Guthrie and Bryanth, 2006; Blechert et al., 2007), with impairment predicting symptom severity (Norrholm et al 2011). Extinction is a process whereby a learned fear response is reduced via repeated presentation of the conditioned stimulus (CS, i.e., a trauma-related cue, for example, a noise of a horn previously associated with a car crash or a tone previously associated with foot shock in a rodent fear conditioning experiment) in the absence of the unconditioned stimulus (US; in our examples, the traumatic event of a car crash or the foot shock). It is considered to be a process whereby new memories are formed, i.e., the patient learns that the CS is not necessarily associated with the US. As such, extinction memory is encoded, consolidated and expressed as are other types of memory.

NMDAR COMPOUNDS

Extinction of learned fear has been shown to be susceptible to NMDA receptor blockade in the amygdala (Falls et al., 1992), while enhancement of NMDA receptor function (e.g., indirectly by the glycine receptor partial agonist d-cycloserine either systemically administered or infused into the amygdala) facilitates extinction learning (e.g., Walker et al., 2002; Ledgerwood et al., 2003; see also Myers et al., 2011, for a recent review on glutamatergic mechanisms involved in extinction processes). Likewise, d-cycloserine attenuates impaired fear extinction induced by single prolonged stress (Yamamoto et al., 2008), providing evidence from a preclinical PTSD model that d-cycloserine in support of potential beneficial effects on extinction in PTSD patients. Clinically, beneficial effects of d-cycloserine in combination with exposure therapy have been reported in agoraphobia (Ressler et al., 2004), social anxiety disorder (Guastella et al., 2008) and panic disorder (Otto et al., 2010), although controlled clinical trials looking specifically at the potentially beneficial effects of d-cycloserine on extinction processes in PTSD are outstanding. Interestingly extinction of learned “fear” produced in the laboratory using shock stimuli in healthy controls does not seem to be affected by d-cycloserine treatment (Guastella et al 2007). This somewhat surprising finding may be due to differing neural substrates underlying trauma-related clinical symptoms versus fear conditioning in the laboratory (Grillon 2009). Along similar lines to dcycloserine as a putative adjunctive treatment to exposure-based therapy, it can be argued that other drugs that facilitate NMDA receptor function, such as glycine transporter inhibitors, mGluR2 NAMs or mGluR5 PAMs, or that enhance other aspects of glutamatergic neurotransmission, such as AMPA PAMs, should also enhance extinction of fear responses to trauma memories. In rats, the AMPA receptor potentiator PEPA and glycine transporter inhibitor NFPS facilitated extinction learning for contextual and cued fear respectively (Zushida et al., 2007, Mao et al. 2009). PAMs of mGluR5 have been shown to facilitate extinction of cocaine contextual memory (Gass and Olive, 2009), although no data showing similar effects of mGluR5 modulation on fear conditioning or investigation of these mechanisms in PTSD models have been published yet.

Histone deacetylase (HDAC) inhibition, which prevent the deacetylation of histones, thereby affecting the same intracellular signalling cascade that is also susceptible to NMDA receptor modulation (Reul and Nutt, 2008), has also shown promise in facilitating extinction in preclinical models (Lattal et al., 2007). A limitation to this target however is that HDAC inhibitors, especially non-subtype selective ones, may be associated with a safety profile that again would prevent their use in this indication (Menegola et al., 2005).

CANNABINOIDS

Facilitation of extinction of fear conditioning was also seen following administration of a CB1 receptor agonist (Pamplona et al., 2006; but see Lin et al., 2008, showing that chronic CB1 agonism impaired fear conditioning extinction) or inhibitor of endocannabinoid breakdown and reuptake (Chhatwal et al. 2005). However, the abuse potential associated with CB1 agonism is likely to make this approach for treating PTSD patients undesirable due to an increased risk/benefit ratio.

ADRENOCEPTOR ANTAGONISTS

Activation of norepinephrine has also shown promise in preclinical models of fear extinction as well as treatment with the non-specific alpha2 adrenoceptor antagonist yohimbine (Cain et al., 2004; Morris and Bouton, 2007; Holmes and Quirk, 2010, but see Mueller et al., 2009). However, yohimbine itself can cause panic attacks in PTSD patients (Southwick et al. 1997; 1999b).

A complicating factor is that many of the potential therapeutic approaches that facilitate extinction also enhance other forms of learning and memory, at least preclinically. For example, NMDA receptor blockade not only interferes with extinction of aversive memories, but also with reinstatement of conditioned fear, i.e., also prevents the reinstatement of aversive memories, which occurs when the US is presented alone following extinction of the CS (Johnson et al., 2000). Thus facilitation of NMDA receptor function might not only enhance extinction or habituation, but also reinstatement of conditioned fear, depending on which process predominated at the time of adjunctive treatment to exposure therapy, thereby worsening rather than improving the clinical condition. However, this potential issue has not yet been sufficiently investigated clinically to allow firm conclusions.

Summary

Because PTSD involves a precipitating traumatic event that often leads to medical evaluation there is a significant potential window for prophylactic treatment that should not be ignored. Preclinical models of PTSD and some early clinical trials suggest that prophylactic treatment approaches are feasible. Clearly such treatments, as with most preventive treatments (e.g. aspirin for heart attack prevention or statin drugs for cholesterol reduction), must be extremely safe for use across patients with varied degrees of physical injury. This risk/benefit ratio will be a high bar for drug development to clear. The impact of an effective prophylactic treatment would be vast however, especially in socially critical and high risk personnel such as police, fire fighters and the military. We have also reviewed a number of potential targets from preclinical models that could modulate conditioned fear processes after PTSD has developed. In the clinic, the efficacy of such putative adjunctive treatments with trauma-focused therapies will greatly depend not only on the efficacy of the compound itself, but also the protocol of the psychological therapy (e.g. being focused on extinction learning or being targeted towards reconsolidation). Nonetheless, pharmacological treatments that aid specific therapies in mental health, such as learning new skills (e.g. oxytocin to facilitate social interaction training in autism, Hollander et al. 2007) and remodulating memories or behaviours (e.g. exposure therapy in PTSD) is an exciting avenue of research that could represent a paradigm shift in pharmacological treatment of PTSD. Such therapeutic approaches may also circumvent some safety issues as they will be taken only under therapy supervision, given over limited periods of time, thus reducing issues of tolerance, abuse potential, and side effects linked to chronic administration.

Acknowledgments