Valproic acid but not d-cycloserine facilitates sleep-dependent offline learning of extinction and habituation of conditioned fear in humans

Abstract

The effectiveness of d-cycloserine (DCS), an N-methyl-d-aspartate glutamate receptor partial agonist, and valproic acid (VPA), a histone deacetylase inhibitor, in facilitating the extinction of fear-conditioned memory has been explored in humans and animals. Here, we confirmed whether DCS (100 mg) and VPA (400 mg) act in off-line learning processes during sleep or waking, for further clinical application to anxiety disorders and posttraumatic stress disorder (PTSD). We performed a randomized, blind, placebo-controlled clinical trial in 90 healthy adults. Visual cues and electric shocks were used as the conditioned stimulus (CS) and unconditioned stimulus (US), respectively. The extinction effect was observed not in simple recall after the extinction of coupled CS–US, but was observed in the post-re-exposure phase after unexpected re-exposure to reinstatement CS–US coupling. Newly acquired conditioned fear was also eliminated or habituated by DCS and VPA administration, in line with previous findings. Furthermore, VPA facilitated the off-line learning process of conditioned fear extinction and habituation during sleep, while DCS facilitated this process during waking. These novel findings suggest that DCS and VPA might enhance exposure-based cognitive therapy for anxiety disorders and PTSD by reducing the vulnerability to reinstatement and preventing relapses of fear-conditioned responses, and provide evidence for a peculiarity of the sleep-dependent off-line learning process for conditioned fear extinction.

This article is part of a Special Issue entitled ‘Cognitive Enhancers’.

Introduction

Fear conditioning involves learning that certain environmental stimuli predict aversive events to promote survival in the face of present and future threats. The experimental fear-conditioning paradigm, in which a neutral conditioned stimulus (CS) acquires aversive properties as a consequence of being paired with an aversive unconditioned stimulus (US), implicitly promotes association between the CS and US. Although fear-conditioning is an essential component of defensive behavior systems (Fanselow, 1998), it is also involved in the pathophysiology of fear circuit disorders, including posttraumatic stress disorder (PTSD) and anxiety disorders (Shin and Liberzon, 2010). Fear extinction, a relearning paradigm of CS–US dissociation in which fear-conditioned subjects are repeatedly exposed to a CS without a US under the same context, overwhelms established conditioned fear (Myers and Davis, 2002; Quirk and Mueller, 2008). Thus, fear extinction has been considered a psychophysiological model of cognitive behavioral therapy (CBT) for those disorders (Quirk et al., 2006).

Learning occurs with repetitive training on-line, but it can also strengthen off-line between training-retest sessions, including sleep periods (Diekelmann and Born, 2010; Stickgold, 2005). N-methyl-d-aspartate (NMDA) receptor involvement has been indicated in off-line learning of conditioned fear extinction, rather than in encoding (on-line learning) with repeated exposure to CS–US dissociation. NMDA glutamate receptor activation mediates a long-lasting increase in intracellular synaptic potentiation (Artola and Singer, 1987; Rumpel et al., 1998), which is thought to elicit synaptic plasticity and off-line learning (Kirkwood et al., 1996; Miller et al., 1989). Off-line learning of conditioned fear extinction was impaired by NMDA receptor antagonist administration during the post-training period (Santini et al., 2001). In contrast, the NMDA receptor partial agonist d-cycloserine (DCS) enhances off-line learning of conditioned fear extinction (Ledgerwood et al., 2003). DCS also enhances the effects of exposure-based cognitive-behavioral therapies for anxiety disorders (Guastella et al., 2008; Hofmann et al., 2006; Kushner et al., 2007; Otto et al., 2010; Ressler et al., 2004; Wilhelm et al., 2008).

Although its clinical efficacy has not yet been clarified, valproic acid (VPA), a histone deacetylase (HDAC) inhibitor (Göttlicher, 2004), also has a positive effect on off-line learning of conditioned fear extinction in rats (Bredy et al., 2007; Bredy and Barad, 2008). VPA is used as an anticonvulsant but also strongly inhibits HDAC and increases mRNA. HDAC inhibitors and histone acetyltransferases (HATs), which cause chromatin structure relaxation and tension, lead to enhanced and diminished DNA transcription, respectively (Varga-Weisz and Becker, 1998; Yang and Seto, 2007). Since histone acetylation regulation is critical for fear conditioning (Levenson et al., 2004), VPA could accelerate delayed consolidation of conditioned fear extinction.

Several questions remain regarding whether the pharmacological effects of DCS and VPA on extinction and habituation learning of human fear conditioning predominantly act in on-line (acquisition) or off-line (delayed consolidation) processes, and whether DCS and VPA activate off-line learning during sleep or wakefulness. Although sleep is crucial to the off-line development of various learning domains (Diekelmann and Born, 2010; Stickgold, 2005), off-line learning can also occur during wakefulness (Sheth et al., 2009; Song et al., 2007), suggesting that differences in the delayed learning process during sleep and wakefulness result from corresponding differences in background mechanisms such as changes in synaptic signaling efficiency and gene transcription levels (Carr et al., 2011; Diekelmann and Born, 2010; Tononi and Cirelli, 2006).

A previous study suggested that both DCS and VPA enhance extinction learning of human fear conditioning and simultaneously prevent new acquisition of fear conditioning through so-called “habituation learning” (Kuriyama et al., 2011a). Although DCS exhibited a similar effect on reinstatement in rats (Ledgerwood et al., 2004), the effects of DCS and VPA on extinction and habituation learning were robust when reinstatement stimuli were exposed, rather than during the extinction and habituation learning per se, in humans (Guastella et al., 2007; Kuriyama et al., 2011a). When these drugs are applied clinically during exposure therapy for anxiety disorders or PTSD, we do not require enhancement of acquisition learning but rather enhancement of extinction learning of the fear (Hofmann, 2008). Therefore, in the current study, we explored whether DCS and VPA facilitate on-line or off-line learning of conditioned fear extinction and habituation in humans. We also examined whether the effects of DCS and VPA on off-line extinction and habituation learning of conditioned fear are observed during sleep or wakefulness.