Keywords

• Fear extinction;
• Exposure therapy;
• Augmented relearning;
• Reconsolidation;
• Drug development;
• Cognitive enhancer

Abbreviations

• 5-HT, 5-hydroxytryptamine = serotonin;
• AC, adenylate cyclase;
• AMPA, α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid;
• AMY, amygdala;
• BA, basal amygdala;
• BDNF, brain-derived neurotrophic factor;
• BLA, basolateral amygdaloid complex;
• BZD, benzodiazepine;
• CaMKII, Ca2+/calmodulin-dependent protein kinase II;
• cAMP, 3′-5′-cyclic adenosine monophosphate;
• Cav 1.2, L-type calcium channel alpha 1C isoform;
• Cav 1.3, L-type calcium channel alpha 1D isoform;
• CBT, cognitive behavioral therapy;
• CCK, cholecystokinin;
• CeA, central amygdala;
• CeM, centromedial amygdala;
• CeL, centrolateral amygdala;
• CNS, central nervous system;
• CREB, cAMP response element binding;
• CS, conditioned stimulus;
• DA, dopamine;
• DCS, D-cycloserine;
• DNA, deoxyribonucleic acid;
• eCB, endogenous cannabinoiod;
• ERK, extracellular regulated kinase;
• ERP, exposure and prevention CBT;
• FGF2, fibroblast growth factor-2;
• GABA, γ-aminobutyric acid;
• GABA_A, GABA_A receptor;
• GAD, generalized anxiety disorder;
• GAD (65/67), glutamate decarboxylase 65/67;
• GluN, NMDA receptor subtype;
• GR, glucocorticoid receptor;
• H3, histone H3 protein;
• H4, histone H4 protein;
• HAT, histone acetyltransferase;
• HDAC, histone deacetylase;
• HDAC2, histone deacetylase 2;
• HPA, hypothalamic-pituitary-adrenal axis;
• HPC, hippocampus;
• icv, intracerebroventricular;
• IL, infralimbic cortex;
• ITC, intercalated cell masses;
• K, lysine;
• KOR, kappa opioid receptor;
• LA, lateral amygdala;
• L-DOPA, levodopa;
• MAPK, mitogen-activated protein kinase;
• MDMA, 3,4-methylenedioxy-N-methylamphetamine;
• Mg2+, magnesium ion;
• mGluR, metabotropic glutamate receptor;
• miRNA, micro-RNA;
• mPFC, medial prefrontal cortex;
• mRNA, messenger ribonucleic acid;
• MS-275, entinostat (Pyridin-3-ylmethyl N-[[4-[(2-aminophenyl)carbamoyl] phenyl]methyl]carbamate);
• NMDA, N-methyl-D-aspartate;
• NMDAR, NMDA receptor;
• ns, not studied;
• NO, nitric oxide;
• NPS, neuropeptide S;
• NPY, neuropeptide Y;
• OCD, obsessive-compulsive disorder;
• OXT, oxytocin;
• PAG, periaqueductal gray;
• PEPA, 4-[2-(Phenylsulphonylamino)ethylthio]-2,6-difluorophenoxyacetamide;
• PKC, as Ca2+ /phospholipid-dependent protein kinase C;
• PTSD, post-traumatic stress disorder;
• PL, prelimbic cortex;
• SAD, social anxiety disorder;
• SAHA, vorinostat (N-hydroxy-N’-phenyl-octanediamide);
• SNP, single nucleotide polymorphism;
• SNRI, selective noradrenaline reuptake inhibitor;
• SSRI, selective serotonin reuptake inhibitor;
• TrkB, tropomyosin-related kinase B;
• US, unconditioned stimulus;
• VGCC, voltage-gated calcium channel;
• VRET, virtual reality exposure;
• Zn2+, zinc ion

1. Introduction

Fear, anxiety and trauma-related disorders are associated with excessive fear reactions triggered by specific objects, situations or internal and external cues in the absence of any actual danger, and often include an inability to extinguish learned fear and to show adequate safety learning [(Michael et al., 2007; Wessa & Flor, 2007; Milad et al., 2009; Jovanovic et al., 2012; Milad et al., 2013) reviewed in (Holmes & Singewald, 2013) and (Kong et al., 2014)]. Pathological fear and anxiety occur in a range of psychiatric conditions, including various types of phobia (e.g. social phobia, agoraphobia or specific phobia), panic disorder with/without agoraphobia, obsessive-compulsive disorder (OCD), generalized anxiety (GAD) and post-traumatic stress disorder (PTSD) (ICD-10, 1994; DSM-5, 2013). These disorders comprise the most common mental disorders and are estimated to have a life-time prevalence of up to 28% amongst western populations (Kessler et al., 2005; Wittchen et al., 2011; Kessler et al., 2012). In addition to the personal suffering of patients, the economic burden caused by anxiety disorders is heavy (Gustavsson et al., 2011).

Available pharmacological and psychotherapeutic treatments (Bandelow et al., 2007) which aim to reduce fear and anxiety are associated with decreased symptom severity, but up to 40% of anxiety patients show only partial long-term benefit, and a majority of them fail to achieve complete remission (Hoffman & Mathew, 2008; Stein et al., 2009; Bandelow et al., 2012) clearly underlining the need for further improvement. Current pharmacological approaches either induce rapid anxiolytic effects (e.g. benzodiazepines, some antipsychotics) or require prolonged, chronic treatment (e.g. antidepressants) to attenuate symptoms of pathological fear and anxiety. Commonly employed psychotherapeutic interventions apply cognitive behavioral strategies and exposure techniques to help patients overcome the maladaptive beliefs and avoidance behaviors that reinforce the pathology related to fear-eliciting cues. Meta-analyses show that cognitive behavioral therapy (CBT) does have efficacy for several anxiety disorders, including PTSD, but patients have difficulty bearing the demanding and exhausting process of therapy and many who do manage to cope with it respond only partially and often relapse with time (Choy et al., 2007).

One strategy to improve CBT is to augment psychotherapy with adjunctive pharmacological treatments. Early attempts at combining ‘CBT’ with anxiolytic medications [e.g. benzodiazepines (BZD)] showed that the combination was no more effective [in some instances even counterproductive (Marks et al., 1993; Wilhelm & Roth, 1997)] than psycho- or pharmacotherapy alone [for details see (Otto et al., 2010a; Otto et al., 2010b; Rodrigues et al., 2011; Dunlop et al., 2012; Hofmann, 2012)]. However, at least in some cases, this failure may have reflected idiosyncratic effects of the drugs tested (especially BZDs) rather than utility of the strategy itself, and there has been an intense search to identify agents that serve as more effective adjuncts to CBT. The preclinical assay most frequently used in this search is fear extinction – the focus of this current review. Extinction of fear following Pavlovian fear learning (Pavlov, 1927) in animals is procedurally similar to exposure-based CBT (Milad & Quirk, 2012). We will briefly outline different aspects of Pavlovian fear learning [(which is thought to be involved in the etiology and maintenance of anxiety disorders, e.g. (Amstadter et al., 2009)] and extinction highlighting the key processes that could be targeted to augment fear extinction (see Figure 1 for an overview).

Fear and fear extinction

Experimentally, fear conditioning occurs when a previously neutral stimulus [conditioned stimulus (CS) – such as a tone or light-] is paired with an aversive, unconditioned stimulus (US – e.g. electric shock to the forearm in humans, mild foot shock in rodents), resulting in a CS-US association whereby the CS alone elicits a conditioned fear response (e.g. freezing in rodents or increased skin conductance in humans). Following a successful CS-US association, fear memories require consolidation, a process involving a cascade of molecular and cellular events that alter synaptic efficacy, as well as a prolonged systems level interaction between brain regions, to stabilize the memory (McGaugh, 2000). Once consolidated, fear memories, reactivated by presenting the CS, are de-stabilized to render the original fear memory liable to pharmacological/behavioral interference (see overview in Figure 1) and this is then followed by a second phase of molecular and cellular events to re-stabilize (re-consolidate) the (adapted) memory (Nader et al., 2000).

Fear memories can be attenuated by various processes and interventions (including pharmacological and psychological approaches), some producing temporary blunting of fear behaviors and others causing more long-lasting relief. Interfering with the re-consolidation by inhibiting molecular and cellular events supporting fear memory re-stabilization [Figure 1C, (Nader et al., 2000; Lee et al., 2006)], for example with ß-adrenoceptor blockers (Debiec & Ledoux, 2004; Kindt et al., 2009), has been proposed as a clinical approach to alleviating fear memories. For discussion on this and other means of reducing fear (e.g. via US habituation) we refer the reader to some excellent prior reviews (Graham et al., 2011; Schwabe et al., 2014). Other potential ways to relieve fear include safety learning (Rogan et al., 2005; Kong et al., 2014) and erasure-like mechanisms such as destruction of erasure-preventing perineuronal nets (Gogolla et al., 2009).

Alternatively, fear memories can also be extinguished. Fear extinction, a process originally described by Pavlov (Pavlov, 1927), entails repeated exposure to anxiety-provoking cues to establish a new memory that counters the original fear memory. The process is highly relevant to fear, anxiety and trauma-related disorders which are associated with negative emotional reactions triggered by specific objects, situations or internal and external cues that are excessive to the actual danger posed. Moreover, extinction in animals is procedurally similar to forms of CBT that rely on exposure to anxiety-provoking cues (see Figure 1) (Milad & Quirk, 2012), and anxiety disorders are associated with an inability to extinguish learned fear and to respond adequately to safety signals (Michael et al., 2007; Wessa & Flor, 2007; Milad et al., 2009; Jovanovic et al., 2012; Milad et al., 2013) [reviewed in (Holmes & Singewald, 2013)]. Thus, fear extinction has considerable translational utility. The key processes that can be targeted to pharmacologically augment fear extinction are summarized in Figure 1.

Extinction is a learning process driven by violation of the original CS=US contingency (termed ‘prediction error’ for review see (Pearce & Bouton, 2001; McNally & Westbrook, 2006)], but it also contains other elements including habituation and desensitization some also say erasure/destabilization (Lin et al., 2011). Some authors therefore favor the term ‘relearning’ over extinction [for discussion see (Riebe et al., 2012)]. As with fear learning, extinction occurs in two phases: extinction acquisition and extinction consolidation retrieval. The decrement in the fear response during ’extinction training’ is also termed ‘within-session extinction’. Similar to fear memory, stabilization of the extinction memory requires both a cascade of overlapping, but dissociable, molecular and cellular events [for review see (Myers & Davis, 2007; Orsini & Maren, 2012)] that alter synaptic efficacy and brain systems-level interactions (Pape & Pare, 2010). The strength of extinction memory can be assessed at some interval (usually >1 day) after extinction training in “extinction retrieval” sessions (also termed ‘extinction retention’ or ‘extinction expression’, but note that we use the term “extinction retrieval” throughout this review).

That extinction memories are prone to re-emergence (due to insufficient ‘longterm extinction’) indicates that the original fear memory is still in place and the extinction memory is weaker/more labile than the fear memory. This may be particularly true of older (‘remote’) fear memories (Tsai & Graff, 2014). The re-emergence of extinguished fear occurs under multiple circumstances: (i) renewal, when the CS is presented in a different context to that in which extinction training occurred; (ii) reinstatement, when the original US or another stressor as is given unexpectedly; and (iii) spontaneous recovery, when a significant period of time has elapsed following successful extinction training (Myers & Davis, 2007; Herry et al., 2010). The likelihood of fear re-emergence is dependent on the strength of the extinction memory which is also determined by the type of extinction protocol used see e.g. (Li & Westbrook, 2008; Laborda & Miller, 2013). Spontaneous recovery (Rowe & Craske, 1998b, a; Schiller et al., 2008), reinstatement (Schiller et al., 2008) and renewal (Effting & Kindt, 2007) are all observable in clinical settings, and can be readily exploited in the laboratory to identify drug- and other interventions that can prevent or fear re-emergence in animals and relapse in humans (Vervliet et al., 2013).

Our goal in the current review is to offer a comprehensive overview of preclinical work on the possible pharmacological approaches (see Box 1 for an overview) to augment fear extinction and protect against the re-emergence of fear, and to discuss the potential translational value of these candidates as adjuncts to exposure-based CBT in anxiety patients.

2. Neuronal substrates of fear extinction

Using a number of complementary techniques (including electrophysiology, immediate-early gene mapping, tracing studies, lesioning/inactivation approaches and optogenetics) research is revealing the complex and interconnected brain circuitry mediating fear extinction. Several key brain areas including the amygdala (AMY), hippocampus (HPC), medial prefrontal cortex (mPFC), periaqueductal grey (PAG), bed nucleus of the stria terminalis and others have been implicated in extinction [for recent detailed reviews see (Myers & Davis, 2007; Ehrlich et al., 2009; Herry et al., 2010; Pape & Pare, 2010; Knapska et al., 2012; Orsini & Maren, 2012; Duvarci & Pare, 2014)]. Of these, we focus on the AMY, HPC and mPFC as major, well-defined components of the fear circuitry (see Figure 2).

While the AMY and the mPFC are crucial for the formation and maintenance of fear extinction memories, the HPC, linked with the mPFC and the AMY (Pape & Pare, 2010), processes contextual information linked with extinction (Orsini & Maren, 2012). Altering these hippocampal contributions to extinction memory or mimicking the hippocampal response within the extinction context (see below) may be mechanisms that render extinction context-independent. The AMY is a core hub in fear extinction processing. Data in rodents and humans suggest sustained AMY activity in extinction-impaired individuals, possibly resulting from a failure to engage pro-extinction circuits in cortical areas and subregions of the AMY [reviewed in (Holmes & Singewald, 2013)]. Different AMY subregions and neuronal populations have differential contributions to extinction. The centromedial AMY (CeM) is the major output station of the AMY that drives fear via its connections to the hypothalamus and brainstem regions (LeDoux et al., 1988; Fendt & Fanselow, 1999; Maren, 2001), and its responding is modulated following extinction via intra-amygdala and remote inputs. Extinction training has been shown to cause a rapid reduction of CS-evoked responses of lateral AMY (LA) neurons possibly via depotentiation of thalamic inputs (Duvarci & Pare, 2014). The basal AMY (BA) contains extinction-encoding neurons which drive GABAergic cells in the medial intercalated cell masses (ITC) and neurons in the centrolateral AMY (CeL) to inhibit the centromedial AMY (CeM) and the expression of fear (Herry et al., 2008; Duvarci & Pare, 2014). Following from these initial findings, additional studies involving optogenetic approaches, have shown that a subpopulation of the BA pyramidal neurons which express the Thy1 gene may mark the extinction neuron subpopulation (Jasnow et al., 2013).

It is also becoming apparent that different ITCs are part of a GABAergic feed-forward relay station interconnecting AMY nuclei with distinct networks within the ITCs that are engaged in particular fear stages exerting different influences on extinction (Whittle et al., 2010; Busti et al., 2011; Duvarci & Pare, 2014). Separate neuronal populations in the CeL have been found to have contrasting roles in fear and fear extinction. CeL-OFF neurons (PKCδ+ phenotype) inhibit fear-promoting CeM neurons, while CeL-ON (PKCδ- phenotype) cells stimulate fear-promoting CeM neurons (Ciocchi et al., 2010; Haubensak et al., 2010). Additionally, a very recent finding suggests that a subpopulation of neurons within the CeM –that of the Tac2 peptide-expressing cells -, is critically involved in fear learning and fear expression (Andero et al., 2014).

Another important node within the extinction circuitry is the mPFC, which shares strong interconnections with the HPC and the AMY. Fear extinction recruits the ventral segment of the mPFC, the infralimbic (IL) subdivision [the rodent correlate of the human ventromedial PFC (vmPFC)] which projects to BA extinction and ITC neurons and can thereby inhibit the activity of CeM fear output neurons [reviewed in (Amir et al., 2011; Senn et al., 2014)]. When a conditioned cue following extinction is presented in the extinction (therapy) context, the HPC activates the IL (an activation which is absent or less in a novel context), supporting subsequent CeM inhibition (Herry et al., 2010). The IL subdivision is involved in the consolidation of extinction memory and synaptic plasticity in this region and is crucial for successful extinction retrieval (Mueller & Cahill, 2010). Impaired extinction has been associated with relatively low activity in IL neurons, and successful extinction with relatively high IL neuronal activity [reviewed in (Holmes & Singewald, 2013)]. The region of the vmPFC neighboring the IL [the prelimbic cortex (PL) in rodents and the dorsal anterior cingulate in humans] plays an opposite role to the IL in regulating fear and extinction [reviewed in (Milad & Quirk, 2012)]. The PL interacts with BA fear neurons, augmenting output of the basolateral amygdaloid complex (BLA) and hence CeM activity (Senn et al., 2014).

There is now strong evidence from both animal and human studies that these various components of the neural circuit underlying extinction are functionally disturbed in individuals with impaired extinction (reviewed in (Holmes & Singewald, 2013) and importantly that successful extinction in animals (Whittle et al., 2010) and exposure therapy in humans (Hauner et al., 2012) can correct these abnormalities by triggering a lasting reorganization of this network.