TINS-1144; No. of Pages 9
Opinion
Reconsolidation and the regulation of plasticity: moving beyond memory Robert P. Bonin1 and Yves De Koninck1,2 1 2
Institut Universitaire en Sante´ Mentale de Que´bec, Que´bec, Canada Department of Psychiatry and Neuroscience, Universite´ Laval, Que´bec, Canada
Memory reconsolidation is a protein synthesis-dependent process that preserves, in some form, memories that have been destabilized through recall. Reconsolidation is a nearly universal phenomenon, occurring in a diverse array of species and learning tasks. The function of reconsolidation remains unclear but it has been proposed as a mechanism for updating or strengthening memories. Observations of an analog of reconsolidation in vitro and in sensory systems indicate that reconsolidation is unlikely to be a learning-specific phenomenon and may serve a broader function. We propose that reconsolidation arises from the activity-dependent induction of two coincident but opposing processes: the depotentiation and repotentiation of strengthened synapses. These processes suggest that reconsolidation reflects a fundamental mechanism that regulates and preserves synaptic strength. Specific versus general view of reconsolidation The ability to learn and store information in the form of memories is crucial for the survival of many organisms. Equally important is the ability to modify or update these memories to accurately reflect a changing environment. Over the past 15 years there has been a considerable amount of research into one process that is thought to enable the modification and updating of consolidated (see Glossary) memories: memory reconsolidation. The process of memory reconsolidation is triggered by the recall of a memory. Recall transiently destabilizes the reactivated memory and renders it labile, after which the memory restabilizes (‘reconsolidates’) to remain stored. Preventing the reconsolidation process, for example, by blocking protein synthesis, leads to disruption of the memory [1]. This cycle of memory destabilization and reconsolidation appears to provide a window of opportunity for the modification of learned associations [2,3]. Reconsolidation was brought to the fore of neuroscience research in 2000, when it was shown that inhibiting protein synthesis in the amygdala during the recall of a fear memory resulted in disruption of the memory [1]. The necessity of protein synthesis provided evidence for a cellular mechanism of reconsolidation that paralleled Corresponding author: Bonin, R.P. ([email protected]). Keywords: long-term potentiation; heterosynaptic plasticity; homeostatic plasticity; reconsolidation; protein synthesis; protein degradation. 0166-2236/ ß 2015 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.tins.2015.04.007
the well-studied process of memory consolidation [4], triggering an explosion of research into the mechanisms and parameters of memory reconsolidation. Since this revitalizing study, reconsolidation has been demonstrated to be a nearly universal phenomenon. Memory reconsolidation has been observed in more than 10 diverse species, from C. elegans to humans, and has been observed in several regions of the central nervous system (CNS) including the hippocampus, amygdala, and anterior cingulate cortex (reviewed in [3,5]). Despite this wealth of information, the overall function of reconsolidation remains unclear. The scientific origin of reconsolidation has largely branded reconsolidation as a learning-specific phenomenon, and theories of the functional role of reconsolidation have centered on its putative role in the modification and updating of memories [2,3,6–8]. In this article we re-evaluate the functional role of reconsolidation in light of in vitro and in vivo evidence that this phenomenon may exist broadly throughout the CNS [9], and is thus unlikely to serve a learning-specific function. We argue that reconsolidation instead reflects a fundamental mechanism for the regulation of synaptic plasticity, and present evidence that the cellular processes mediating this regulation serve to constrain activity-dependent synaptic potentiation in the nervous system. Functional significance of reconsolidation Reconsolidation has been proposed to serve two general functions: the modification and/or strengthening of memories [2,7,8]. The modification of memories through reconsolidation may maintain memory relevance by allowing
Glossary Consolidation: the stabilization of a memory trace after acquisition. Depotentiation/repotentiation: a reduction/increase in the strength of a synapse that has been potentiated by some stimulus. Fear conditioning: a learning paradigm in which a stimulus becomes associated with fear. Heterosynaptic plasticity: changes in synaptic strength that can be induced through activity at other synapses or through other cellular processes. Homeostatic plasticity: the adjustment of synaptic strength to regulate neuronal activity. Homosynaptic plasticity: changes in synaptic strength that are input-specific and restricted to only the activated synapses. Hyperalgesia: increased sensitivity to painful stimuli. Long-term potentiation (LTP)/depression (LTD): an increase/decrease in synaptic strength that persists for hours or longer in response to an induction stimulus. Synaptic plasticity: the ability to modify the strength of a synapse.
Trends in Neurosciences xx (2015) 1–9
1
TINS-1144; No. of Pages 9
Opinion the encoding of changes in the environment [10]. In humans, memories can be modified through the incorporation of new or conflicting information only after recall of the original memory [11,12]. Similar results were observed in animal studies, where memory recall rendered memory labile and vulnerable to modification [13–16]. One interesting prediction of the memory update function of reconsolidation is that the subsequent updating of memory should be most strongly triggered when recall occurs in a context that is slightly different than the original memory, necessitating an update of the stored memory to reflect the changes [17]. In this case, the induction of reconsolidation should be dependent on some element of novelty during recall. This was precisely the case in a study by Winters et al. [18], in which changing the texture of the flooring during a recall trial was necessary to enable reconsolidation of a well-trained object memory in rodents. Modifying the timing of a shock delivered during re-exposure to a fear-conditioning apparatus was also sufficient to induce reconsolidation of memories that were otherwise resistant to reconsolidation [19]. However, reconsolidation can occur even in the absence of explicitly novel information [1,15,20], indicating that the novelty is not an absolute requirement for reconsolidation. Reconsolidation may also enable the strengthening of memories. This would allow the enhancement of behavioral responses through recall without requiring re-exposure to the original learning situation [7,8]. The strengthening of memory after reconsolidation has been described in both animal models [13,21,22] and in humans [12]. However, the strengthening of memory after recall is not seen in every case. Indeed, animal studies exploring reconsolidation often do not find stronger behavioral responses after memory recall that would indicate the enhancement of memory. For example, the degree of freezing used as a measure of memory in fear-conditioning studies does not typically increase after the recall trial (e.g., [1,15], but see [13,22]). This lack of effect is unlikely to result from a ceiling effect in training because it is possible to pharmacologically manipulate reconsolidation to generate memory strengthening after recall [23–25]. One possible explanation for a lack of memory strengthening after reconsolidation is that the recall of a memory in the absence of a conditioning stimulus such as an electric shock promotes memory extinction, which reduces the behavioral expression of the previously learned association and opposes the effects of memory strengthening induced by reconsolidation. In a recent report, Fukushima et al. [21] explore this possible balance through the elegant use an inhibitory avoidance assay designed to trigger reconsolidation under conditions that do not promote extinction learning. In the absence of extinction learning, the retrieval of a fear memory was shown to promote memory enhancement through reconsolidation. Thus, memory recall may activate multiple processes with opposite effects on memory strength, and whose net balance determines the overall effect of recall on memory strength. The strengthening and modification of memories are not necessarily mutually exclusive outcomes of reconsolidation, and it is plausible that both of these effects can arise from similar underlying mechanisms. Indeed, both memory 2
Trends in Neurosciences xxx xxxx, Vol. xxx, No. x
strengthening and modification can be seen after recall in the same behavioral assay, depending on the conditions of recall [15]. Nonetheless, these interpretations of reconsolidation are still rooted in the idea of reconsolidation as a learning-specific phenomenon. This is reasonable given that the study of behavioral reconsolidation has essentially been entirely restricted to learning tasks [2,3,6]. However, recent in vitro and in vivo work provides evidence that reconsolidation is not restricted to learning tasks and may exist broadly throughout the nervous system, encouraging a re-evaluation of the functional role of reconsolidation. Reconsolidation beyond memory We have recently demonstrated that a reconsolidation-like phenomenon can also be observed in pain-processing circuits of the spinal cord [9]. In this study, mechanical hyperalgesia was induced by the injection of capsaicin into the hind paws of mice. Three hours after the initial plantar injection, capsaicin was reinjected into the hind paw to reactivate the sensitized pain pathways. Although the second capsaicin injection did not modify hyperalgesia on its own, normal mechanosensitivity was quickly restored when this second injection of capsaicin was paired with the spinal administration of the protein synthesis inhibitor, anisomycin, via intrathecal injection. These findings indicate that the reactivation of sensitized spinal pain pathways rendered the hyperalgesia labile and reversible in a manner equivalent to the disruption of fear memories through reconsolidation impairment. The observation of a reconsolidation-like phenomenon in spinal cord pain-processing circuits is not inconceivable given the strong mechanistic parallels between learning and the development of some forms of hyperalgesia (reviewed in [26–28]). Indeed, persistent hyperalgesia has been called a ‘memory trace of pain’ or pain ‘engram’ because of these mechanistic similarities [29,30]. Nevertheless, the observation of reconsolidation in spinal painprocessing pathways encourages a re-examination of the function of reconsolidation on the basis of one key point: given that reconsolidation is not restricted to learning tasks, it is unlikely to be a unique property of learned associations. This shift in the interpretation of reconsolidation does not contradict the proposals of reconsolidation as a means of memory trace strengthening or updating. Rather, we propose that the broad existence of reconsolidation suggests that it may serve a more fundamental role in the regulation of plasticity than possibly predicted from an analysis of reconsolidation solely within learning paradigms. Re-evaluating the functional role of reconsolidation A close examination of the processes triggered by reactivation of a memory trace reveals the existence of two opposing processes that can be recruited to enable the labilization of memory: a destabilizing process and a protein synthesisdependent reconsolidation process (Figure 1A). The activitydependent destabilizing process can be revealed by the inhibition of protein synthesis after reactivation of the memory trace. This typically leads to the disruption of the learned association that is a hallmark of assays designed to study reconsolidation. The labile state of memory has
TINS-1144; No. of Pages 9
Opinion
Trends in Neurosciences xxx xxxx, Vol. xxx, No. x
(A)
Behavioral reconsolidaon Memory formaon (consolidaon)
Labile memory Recall
Stored memory
Synapc reconsolidaon Strong synapc acvaon
Potenated synapse
Destabilizaon and reconsolidaon
Stored memory
Labile plascity Reacvaon
Depotenaon and
repotenaon
Potenated synapse
Blocked depotenaon?
Synapc strength
(B)
Blocked repotenaon TRENDS in Neurosciences
Figure 1. The ups and downs of reconsolidation. Schematics of the processes involved in reconsolidation at the behavioral and synaptic levels are shown in (A). The activation of memory destabilizing and reconsolidation processes by memory recall places the memory into a labile state, in which it can be disrupted through the application of protein synthesis inhibitors. (B) Synaptic potentiation can be induced by strong activation of a synapse (yellow bolt), leading to an increase of the synaptic strength. This synaptic potentiation is thought to underlie memory and some forms of hyperalgesia [27]. Reactivation of the strengthened synapse, as might occur during memory recall, triggers the induction of two opposing processes: the depotentiation (blue arrow) and repotentiation of synaptic strength (red arrow). These two processes often cancel out to produce no net change in synaptic strength; however, preventing repotentiation through the inhibition of protein synthesis can lead to a net depotentiation of the synapse. Behaviorally, this can be seen as the blockade of reconsolidation and the disruption of memory. It has not yet been conclusively demonstrated whether the selective prevention of depotentiation can lead to additional synaptic strengthening or memory augmentation.
been previously described as arising from two distinct processes of destabilization and reconsolidation [17,31–34], and there is molecular evidence to support such a division. In auditory fear-conditioning experiments, the blockade of GluN2B subunit-containing NMDA receptors in the basolateral amygdala prevented the memory from becoming labile, while blockade of GluN2A-containing NMDA receptors only prevented repotentiation of the memory, indicating that the two processes are activated through separate signaling mechanisms [35]. The destabilization process is necessarily independent of protein synthesis because it persists following inhibition of protein synthesis. The reactivation of a consolidated memory trace triggers the active degradation of proteins necessary for the maintenance of the memory trace, including the synaptic scaffolding proteins Shank (SH3 and multiple ankyrin repeat domains) and GKAP (guanylate kinase-associated protein) [13,19,21,36–38] (Box 1). Protein degradation induced by memory retrieval involves protein polyubiquitination and proteasomal degradation [39]. Protein degradation is necessary for the disruption or modification of a reactivated memory trace because inhibition of the proteasome prevents the disruption of a recalled memory by protein synthesis inhibition [13,19,33,34,36–38]. The polyubiquitination of proteins occurs after memory recall even in the absence of protein synthesis inhibition typically used in reconsolidation
assays [37,38], indicating that the degradation of proteins is not an artifact of protein synthesis inhibition. Memory recall also triggers reconsolidation, which counters the destabilization process and maintains the memory trace. Reconsolidation is dependent on proteinsynthesis and is necessary for the preservation of the memory trace after recall. Several transcriptional control pathways have been implicated in memory repotentiation, including protein kinase A [24,40], protein kinase C [41], ERK/MAP kinase [42,43], CaMKII [34], mTOR activation [44,45], as well as the transcription factors CREB [46], C/ EBPd [47], NF-kB [48,49], Zif268 [19], and NPAS4 [50]. The de novo synthesis of mRNA is also crucial for reconsolidation [51]. Several recent review articles describe the molecular mechanisms of destabilization and reconsolidation processes in detail [2,3,5,52,53], and these will not be further reviewed here. Functional links between destabilization and reconsolidation The destabilization and reconsolidation processes appear to be mechanistically intertwined, whereby both processes are dependent on protein degradation (Box 1). If destabilization and reconsolidation were mechanistically distinct, it is expected that preventing the destabilization process alone by inhibiting protein degradation would enable the strengthening of memories after recall through 3
TINS-1144; No. of Pages 9
Opinion
Trends in Neurosciences xxx xxxx, Vol. xxx, No. x
Box 1. How can protein degradation lead to synaptic strengthening? Protein degradation has been implicated as a key process in both the depotentiation and repotentiation processes of synaptic reconsolidation. At first glance it can seem paradoxical that a destructive process such as protein degradation can result in synaptic strengthening or repotentiation. However, synaptic plasticity is determined by a multitude of proteins, some of which act to negatively modulate synaptic strength or constrain further synaptic potentiation; the removal of these inhibitory controls can produce synaptic facilitation (Figure I). Protein degradation associated with plasticity is mediated, in large part, by the ubiquitin–proteasome system [39]. Functionally, inhibiting the proteasome can enhance the induction, but impair the maintenance of LTP in vitro [56,76], consistent with a facilitating role of protein degradation in plasticity. Inhibition of the proteasome after the induction of LTP leads to the accumulation of the translational repressors PAIP2 and 4E-BP2 [77]. In addition, NMDA receptor activation can induce proteasomal degradation of MOV10, a component of the RNA-mediated gene silencing complex (RISC), in cultured neurons [78]. These findings indicate that the synthesis of proteins associated with long-term plasticity is disinhibited through the activity-dependent proteasomal degradation of repressor proteins. Synaptic strength can also be augmented through the degradation of Arc, a protein that facilitates the internalization of AMPA receptors [79,80]. The mechanisms by which protein degradation contributes to the repotentiation process of reconsolidation are much less clear, but likely involve similar disinhibitory processes. For example, MOV10 is also ubiquitinated and degraded following memory retrieval [36].
Ub Shank
GKAP
PS
D-
95
UPS Ub
Arc Endocytosis
Ub MOV10
+
Translaon
Inseron
Key:
Potenaon process Depotenaon process TRENDS in Neurosciences
Figure I. Protein degradation regulates synaptic strength. Protein degradation mediated by the ubiquitin–proteasome system (UPS) can induce both reductions (blue arrows) and enhancements (red arrows) of synaptic strength. AMPA receptors are stabilized at the membrane by synaptic adhesion molecules including Shank and GKAP. These adhesion molecules are degraded during the depotentiation process of reconsolidation. AMPA receptor endocytosis is regulated by Arc in an activity-dependent manner, and the degradation of Arc can enhance synaptic strength. Notably, the endocytosis of AMPA receptors is a necessary process for the membrane insertion of AMPA receptors during repotentiation. The translation of new synaptic proteins necessary for repotentiation is regulated by translational repressors, such as MOV10. The degradation of these repressors is also activity-dependent and contributes to reconsolidation. Abbreviations: Arc, activity-regulated cytoskeleton-associated protein; GKAP, guanylate kinase-associated protein, also known as DLGAP1 (discs large homolog-associated protein 1); MOV10, Mov10 RISC complex RNA helicase; PSD-95, postsynaptic density 95; Shank, SH3 and multiple ankyrin repeat domains; Ub, ubiquitin.
unbalanced repotentiation. However, this has not yet been directly observed, and inhibition of protein degradation often, but not always [54], prevents the destabilization of memory after recall but does not alter the strength of a recalled memory [13,19,33,34,36–38]. In addition, the inhibition of protein degradation not only failed to produce further memory strengthening but even prevented recallinduced strengthening in an inhibitory avoidance assay designed to reveal an enhancement of memory strength through recall [21]. As a notable parallel with reconsolidation, the initial formation of long-term memories [33,54,55], and the stabilization of late-phase long-term potentiation (LTP) in vitro [56], may also depend on protein degradation (Box 1). Indeed, the inhibition of protein degradation immediately, but not 3 h, after initial learning or recall prevented both consolidation and reconsolidation, respectively [54], suggesting that similar mechanisms may act in both consolidation and reconsolidation. Overall, it remains to be determined whether it is possible to mechanistically separate memory destabilization and reconsolidation, such that 4
preventing destabilization leads to the further potentiation of the memory through unbalanced reconsolidation (Figure 1B and Box 2). Reconsolidation at the synapse Working from the premise that reconsolidation is widely distributed in the nervous system, it becomes essential to question the function of the opposing processes of destabilization and reconsolidation in broader terms by considering their cellular correlates at the level of synaptic plasticity that contributes to mnemonic traces. In a striking parallel to behavioral studies, the reactivation of potentiated synapses can trigger a synaptic depotentiation process that is unmasked by the inhibition of protein synthesis, and a countering, protein synthesis-dependent, synaptic ‘repotentiation’ process (Figure 1B) [9,44,57–59]. In vivo electrophysiological recordings revealed that longterm (but not short-term) facilitation of auditory-evoked field potentials in the lateral amygdala was rendered labile and able to undergo depotentiation after recall following
TINS-1144; No. of Pages 9
Opinion
Trends in Neurosciences xxx xxxx, Vol. xxx, No. x
Box 2. Outstanding questions Can the destabilization and reconsolidation processes activated by memory recall be mechanistically separated? Can the balance of depotentiation and repotentiation processes be manipulated, such that unbalanced repotentiation leads to the strengthening of the memory trace? Are the boundary conditions for memory reconsolidation determined by the extent of synaptic activation (e.g., homosynaptic or heterosynaptic) during recall? Is reconsolidation observed throughout the CNS, beyond memory and pain circuits?
fPSP area (% baseline)
the inhibition of protein synthesis, similarly to memory destabilization [57]. However, the facilitated field potentials did not exhibit additional facilitation upon reactivation of cortical and thalamic inputs to the lateral amygdala after fear conditioning [44]. In the hippocampus, field potentials evoked at CA3 to CA1 synapses were facilitated after novel object recognition training but, again, were not further potentiated after reconsolidation [58]. Notably, reconsolidation in this study was accompanied by transient depotentiation of the field potentials at the CA3 to CA1 synapse immediately after recall, and this was attributed to synapses entering a ‘labile’ state [58]. Studies of reconsolidation in vitro have also failed to reveal a robust increase in synaptic potentiation after reactivation of potentiated synapses (Figure 2). In the spinal cord dorsal horn, long-term enhancement of
synaptic strength was not further increased by a second induction stimulus, but was reversed to baseline levels when the second stimulus was applied in the presence of protein synthesis inhibitors [9]. In addition, the stimulation of hippocampal slices several hours after the initial induction of LTP did not cause further synaptic potentiation, but initiated a reversal of late-phase LTP when applied in the presence of a protein synthesis inhibitor [59]. Reconsolidation has also been studied using the gill- and siphon-withdrawal reflex in the sea slug Aplysia, which serves as a simple model of non-associative memory at both the behavioral and synaptic levels [60–64]. The neuronal circuitry mediating the withdrawal reflex is well characterized and enables the measurement of changes in synaptic strength in the circuit over long periods of time. Facilitation of the sensory-to-motor neuron synapse responsible for the withdrawal reflex is rendered labile after a second presentation of an induction stimulus, and can be reversed through the inhibition of protein synthesis [60– 63]. Similarly to mammalian models, this reconsolidation process is dependent on protein degradation [60]. Crucially, a second induction stimuli alone does not cause further potentiation of the facilitated synapse [60,62], or an increase in the number of synaptic varicosities [63], in agreement with the proposition that synaptic reconsolidation maintains synaptic strength. Overall, these studies reveal a consistent trend in which reactivation of the memory trace fails to produce further synaptic potentiation, despite
200
150
100
50 0
120
240
360
Time (min) Aniso fPSP area (% baseline)
fPSP area (% baseline)
Aniso 200
150
100
200
150
100
X
50
50 0
120
240
Time (min)
360
0
120
240
360
Time (min) TRENDS in Neurosciences
Figure 2. Synaptic reconsolidation outside classical learning pathways. A synaptic correlate of reconsolidation can be seen with an in vitro spinal cord preparation. Strong electrical stimulation of dorsal roots (yellow bolts) induces a long-lasting enhancement of synaptic strength in the superficial dorsal horn, as indicated by the increase in postsynaptic field potentials (fPSP). Delivery of a second stimulus, mimicking the reactivation of a memory trace, produces no net change in synaptic strength (black symbols). However, when the second stimulus is given in the presence of the protein synthesis inhibitor, anisomycin (Aniso, blue rectangle), there is a robust reversal of the synaptic potentiation (blue symbols). This parallels the disruption of memory or hyperalgesia seen when reconsolidation is inhibited. Importantly, in the absence of the second stimulus, the protein synthesis inhibitor did not significantly affect synaptic strength (green symbols). Data modified from [9].
5
TINS-1144; No. of Pages 9
Opinion the induction of a protein synthesis-dependent process that counters activity-dependent depotentiation. The lack of further potentiation after the presentation of a second induction stimulus could arise from a ceiling effect in which further potentiation of synapses upon recall is not possible. Nevertheless, there is evidence that an inability to induce further synaptic potentiation in reconsolidation experiments does not reflect synaptic saturation. The long-term enhancement of both auditoryevoked field potentials [57] and persistent synaptic facilitation in Aplysia [62] underwent additional augmentation following the re-presentation of induction stimuli, although in both cases this augmentation was transient and did not persist over the long term. In addition, the difficulties in dissociating the depotentiation and repotentiation processes further complicate the study of saturation in reconsolidation by preventing the selective augmentation of memory traces through the inhibition of depotentiation. In the absence of synaptic facilitation upon recall, it is possible to envisage that the net effect of opposing depotentiation and repotentiation processes would be the recall-induced turnover of synaptic proteins. Such a turnover process can be seen in the exchange of synaptic AMPA receptors induced by memory recall [23,38,65–68]. Changes in synaptic AMPA receptor composition induced by memory recall have been captured through the use of cross-sectional approaches that examine reactivation-induced changes in synaptic responses at various timepoints [23,65]. Memory recall after fear conditioning triggers a transient endocytosis of AMPA receptors that can be observed within 5 minutes of memory recall [65], and may last for up to 4 h [23]. Crucially, preventing AMPA receptor endocytosis and exchange also prevented the destabilization of memory upon recall [23,65]. Hong et al. [65] observed an endocytosis of calcium-permeable, GluA2-containing receptors that was paralleled by the membrane insertion of calcium-impermeable AMPA receptors. Notably, this exchange of AMPA receptors produced no net change in miniature excitatory postsynaptic conductance (mEPSCs), consistent with the lack of reactivation effect seen in previous electrophysiological studies [9,44,60,62]. Conversely, Rao-Ruiz et al. [23] observed a small but significant decrease in mEPSC amplitude 1 h after recall, with restoration of mEPSC amplitude to control levels by 7 h post-recall, indicating that the increase in calcium-impermeable AMPA receptors does not always occur simultaneously with AMPA receptor endocytosis. This brief decrease in synaptic strength is similar to that observed through in vivo hippocampal recordings [58], and suggests that transient synaptic depression induced by recall may depend on a balance of AMPA receptor removal and insertion that, for reasons unknown, can vary between studies. However, this transient decrease in synaptic strength is unlikely to be crucial for reconsolidation because memory was still rendered labile under conditions in which the decrease was absent [65]. Finally, it must be noted that preventing memory destabilization by inhibiting AMPA receptor endocytosis does not necessarily prevent reconsolidation. Blockade of AMPA receptor endocytosis was associated with an enhancement of fear 6
Trends in Neurosciences xxx xxxx, Vol. xxx, No. x
memory that may reflect the modification of memory through recall [23]; however, the enhancement of fear memory following the inhibition of AMPA receptor endocytosis associated with reconsolidation is not always observed [65]. Functional significance of synaptic reconsolidation The activation of a memory trace clearly induces changes in synapse composition but, as summarized above, rarely produces robust changes in synaptic strength. Such findings support the view of synaptic reconsolidation as a product of two opposing processes that largely cancel out in overall effect. These observations raise the question of why two opposing processes would be triggered through the reactivation of a memory trace. Simultaneous depotentiation and repotentiation processes could serve as a means of limiting continual synaptic potentiation in the face of ongoing synaptic activity. As such, synaptic reconsolidation may function as one of several cellular mechanisms that homeostatically maintain synaptic potentiation within a functional range [69,70]. In the absence of these homeostatic mechanisms, the positivefeedback nature of activity-dependent synaptic potentiation can theoretically lead to synaptic saturation, runaway excitation, and loss of network plasticity as activity becomes increasingly concentrated within a select number of potentiated synapses [69,70]. Synaptic reconsolidation would serve a unique role in the homeostatic maintenance of neuronal activity and excitability because depotentiation and repotentiation occur on a much shorter timescale than most identified mechanisms of homeostatic synaptic plasticity that are recruited in response to changes in neuronal activity and occur over hours to days [70]. Specifically, the mechanisms of synaptic depotentiation can be induced almost immediately after synaptic activity [65], making this process much more temporally related to activity-dependent LTP and long-term depression (LTD). In addition, the interpretation of reconsolidation as a property of synaptic plasticity provides a consistent mechanism for the strengthening or modification of memory seen behaviorally after recall because these outcomes would be enabled through the ‘window’ of plasticity at the cellular and network levels provided by synaptic depotentiation and repotentiation. The relative balance of synaptic potentiation and depotentiation triggered by synaptic activity may further determine the overall impact of synaptic reconsolidation on synaptic strength and modification of the memory trace, although evidence for such a regulated balance is currently lacking (Box 2). The putative homeostatic function of synaptic reconsolidation has parallels with the Bienenstock–Cooper– Munro (BCM) model of synaptic metaplasticity, in which the threshold for the induction of LTP or LTD is dependent on the prior activity of the relevant synapses [71]. In the BCM model, a high level of synaptic activity shifts the threshold for the induction of long-term plasticity such that subsequent activity is less likely to produce synaptic potentiation, and may even induce synaptic depotentiation. In a similar manner, a second presentation of strong stimuli as may occur during recall does not necessarily
TINS-1144; No. of Pages 9
Opinion
Trends in Neurosciences xxx xxxx, Vol. xxx, No. x
produce further strengthening of a previously-potentiated synapse owing to the activation of a depotentiation process in synaptic reconsolidation. The net effects of synaptic reconsolidation on network activity and learning will depend on the precise mechanisms underlying synaptic depotentiation and repotentiation, and how broadly these mechanisms impact upon neuronal function. Intriguingly, a recent study in Aplysia found that reconsolidation was triggered only by stimuli that generate heterosynaptic but not homosynaptic plasticity (Figure 3A–C) [62]. In this study, persistently facilitated synapses were reactivated with either electrical stimulation that normally induces transient homosynaptic potentiation, or by serotonin application that transiently facilitates many synapses (heterosynaptic potentiation). While electrical stimulation failed to induce synaptic depotentiation in the presence of the translation inhibitor, rapamycin, depotentiation was induced by the combination of serotonin stimulation and rapamycin. The observation that heterosynaptic facilitation is required for reconsolidation in Aplysia [62] may be relevant to behavioral observations where reconsolidation is only triggered by specific conditions, such as novelty [18,19] or prediction error [72] during recall. Specifically, these observations raise the questions of whether the boundary conditions for
memory reconsolidation are determined by the extent of synaptic activation during recall, and whether memory reconsolidation requires behavioral stimuli that activate heterosynaptic plasticity (Figure 3D and Box 2). In a similar vein, the use of ‘weak’ reactivation protocols in vitro and in vivo was not sufficient to render memory traces labile [61,63,73,74], suggesting that reconsolidation requires strong reactivation of the facilitated memory traces. It is also unknown whether the processes of synaptic potentiation and depotentiation are restricted to a subset of synapses, such as those synapses that have undergone some degree of potentiation or were directly activated by reactivation of the memory trace, or whether the mechanisms of synaptic reconsolidation globally affect synaptic strength in the neuron. There is compelling evidence that activity-dependent synaptic depotentiation can occur on a local scale. The selective activation of a subset of AMPAcontaining synapses led to the removal and proteasomedependent degradation of AMPA receptors exclusively at activated synapses, without affecting neighboring synapses [75]. Determining whether synaptic reconsolidation is constrained by a similar synaptic selectivity will have important implications for its effects on information storage and processing [69,70].
(A) Potenaon
(B) No effect Homosynapc reacvaon (C) Synapc reconsolidaon
?
Heterosynapc reacvaon (D) Stored memory
?
No effect
Homosynapc acvaon?
Destabilizaon and reconsolidaon
Heterosynapc acvaon?
Recall
TRENDS in Neurosciences
Figure 3. Synaptic reconsolidation induced by heterosynaptic activity. The reactivation of facilitated synapses does not always induce synaptic reconsolidation, similarly to observations that memory recall does not always lead to behavioral reconsolidation. In Aplysia, reconsolidation was triggered only by stimuli that generate heterosynaptic but not homosynaptic plasticity [62]. (A) The initial stimulation of a synapse can induce long-lasting synaptic potentiation. (B) The reactivation of the potentiated synapse alone (homosynaptic activation) does not necessarily trigger synaptic reconsolidation. (C) Reactivation of several synapses, including the potentiated synapses (heterosynaptic activation), can trigger the depotentiation and repotentiation processes of synaptic reconsolidation. However, the net effect of reactivation stimuli on unpotentiated synapses is unknown. (D) The putative requirement of heterosynaptic activation for synaptic reconsolidation may underlie observations that recall does not always trigger memory reconsolidation – the behavioral stimuli that fail to trigger memory reconsolidation may not cause sufficient synaptic activation.
7
TINS-1144; No. of Pages 9
Opinion Concluding remarks The interpretation of reconsolidation as a fundamental property of plasticity greatly expands the relevance of this process to overall CNS function. Reconsolidation has already been observed to occur in a wide variety of learning tasks [3,5], while the demonstration that reconsolidation exists in spinal pain-processing networks reveals a role for reconsolidation in sensory function [9]. We anticipate that, similarly to LTP, which was originally described in memory circuits, reconsolidation will continue to be observed in increasingly diverse CNS regions and cognitive functions, reflecting its fundamental regulatory role in plasticity. Acknowledgments We thank Jimena Perez-Sanchez for her assistance with figures. This work was supported by a Pfizer–Fonds de Recherche Que´bec–Sante´ (FRQS) Innovation Fund Award to Y.D.K., a Canadian Institutes of Health Research grant (MOP 12942) to Y.D.K., and a Catherine Bushnell Pain Research Fellowship from the Louise and Alan Edwards Foundation to R.P.B.
References 1 Nader, K. et al. (2000) Fear memories require protein synthesis in the amygdala for reconsolidation after retrieval. Nature 406, 722–726 2 Finnie, P.S. and Nader, K. (2012) The role of metaplasticity mechanisms in regulating memory destabilization and reconsolidation. Neurosci. Biobehav. Rev. 36, 1667–1707 3 Besnard, A. et al. (2012) Reconsolidation of memory: a decade of debate. Prog. Neurobiol. 99, 61–80 4 Dudai, Y. (2012) The restless engram: consolidations never end. Annu. Rev. Neurosci. 35, 227–247 5 Almeida-Correa, S. and Amaral, O.B. (2014) Memory labilization in reconsolidation and extinction – evidence for a common plasticity system? J. Physiol. 108, 292–306 6 Kroes, M.C. and Fernandez, G. (2012) Dynamic neural systems enable adaptive, flexible memories. Neurosci. Biobehav. Revs. 36, 1646–1666 7 Forcato, C. et al. (2014) Strengthening a consolidated memory: the key role of the reconsolidation process. J. Physiol. 108, 323–333 8 Alberini, C.M. and Ledoux, J.E. (2013) Memory reconsolidation. Curr. Biol. 23, 746–750 9 Bonin, R.P. and De Koninck, Y. (2014) A spinal analog of memory reconsolidation enables reversal of hyperalgesia. Nat. Neurosci. 17, 1043–1045 10 Lee, J.L. (2009) Reconsolidation: maintaining memory relevance. Trends Neurosci. 32, 413–420 11 Hupbach, A. et al. (2007) Reconsolidation of episodic memories: a subtle reminder triggers integration of new information. Learn. Mem. 14, 47–53 12 Forcato, C. et al. (2010) Reconsolidation in humans opens up declarative memory to the entrance of new information. Neurobiol. Learn. Mem. 93, 77–84 13 Lee, J.L. (2008) Memory reconsolidation mediates the strengthening of memories by additional learning. Nat. Neurosci. 11, 1264–1266 14 Haubrich, J. et al. (2015) Reconsolidation allows fear memory to be updated to a less aversive level through the incorporation of appetitive information. Neuropsychopharmacology 40, 315–326 15 De Oliveira Alvares, L. et al. (2013) Reactivation enables memory updating, precision-keeping and strengthening: Exploring the possible biological roles of reconsolidation. Neuroscience 244, 42–48 16 Olshavsky, M.E. et al. (2013) Updating appetitive memory during reconsolidation window: critical role of cue-directed behavior and amygdala central nucleus. Front. Behav. Neurosci. 7, 186 17 Exton-McGuinness, M.T. et al. (2014) Updating memories – the role of prediction errors in memory reconsolidation. Behav. Brain Res. 278, 375–384 18 Winters, B.D. et al. (2009) Older and stronger object memories are selectively destabilized by reactivation in the presence of new information. Learn. Mem. 16, 545–553
8
Trends in Neurosciences xxx xxxx, Vol. xxx, No. x
19 Lee, J.L. (2010) Memory reconsolidation mediates the updating of hippocampal memory content. Front. Behav. Neurosci. 4, 168 20 de Oliveira Alvares, L. et al. (2008) Opposite action of hippocampal cb1 receptors in memory reconsolidation and extinction. Neuroscience 154, 1648–1655 21 Fukushima, H. et al. (2014) Enhancement of fear memory by retrieval through reconsolidation. Elife 3, e02736 22 Inda, M.C. et al. (2011) Memory retrieval and the passage of time: from reconsolidation and strengthening to extinction. J. Neurosci. 31, 1635–1643 23 Rao-Ruiz, P. et al. (2011) Retrieval-specific endocytosis of GluA2AMPARs underlies adaptive reconsolidation of contextual fear. Nat. Neurosci. 14, 1302–1308 24 Tronson, N.C. et al. (2006) Bidirectional behavioral plasticity of memory reconsolidation depends on amygdalar protein kinase A. Nat. Neurosci. 9, 167–169 25 Lee, J.L. et al. (2006) Reconsolidation and extinction of conditioned fear: Inhibition and potentiation. J. Neurosci. 26, 10051–10056 26 Latremoliere, A. and Woolf, C.J. (2009) Central sensitization: a generator of pain hypersensitivity by central neural plasticity. J. Pain 10, 895–926 27 Sandkuhler, J. and Lee, J. (2013) How to erase memory traces of pain and fear. Trends Neurosci. 36, 343–352 28 Ruscheweyh, R. et al. (2011) Long-term potentiation in spinal nociceptive pathways as a novel target for pain therapy. Mol. Pain 7, 20 29 Drdla-Schutting, R. et al. (2012) Erasure of a spinal memory trace of pain by a brief, high-dose opioid administration. Science 335, 235–238 30 Asiedu, M.N. et al. (2011) Spinal protein kinase Mz underlies the maintenance mechanism of persistent nociceptive sensitization. J. Neurosci. 31, 6646–6653 31 Nader, K. and Hardt, O. (2009) A single standard for memory: the case for reconsolidation. Nat. Rev. Neurosci. 10, 224–234 32 Kaang, B.K. et al. (2009) Synaptic protein degradation as a mechanism in memory reorganization. Neuroscientist 15, 430–435 33 Sol Fustinana, M. et al. (2014) Protein degradation by ubiquitin– proteasome system in formation and labilization of contextual conditioning memory. Learn. Mem. 21, 478–487 34 Da Silva, W.C. et al. (2013) Memory reconsolidation and its maintenance depend on L-voltage-dependent calcium channels and CaMKII functions regulating protein turnover in the hippocampus. Proc. Natl. Acad. Sci. U.S.A. 110, 6566–6570 35 Milton, A.L. et al. (2013) Double dissociation of the requirement for GluN2B- and GluN2A-containing NMDA receptors in the destabilization and restabilization of a reconsolidating memory. J. Neurosci. 33, 1109–1115 36 Jarome, T.J. et al. (2011) Activity dependent protein degradation is critical for the formation and stability of fear memory in the amygdala. PLoS ONE 6, e24349 37 Lee, S.H. et al. (2008) Synaptic protein degradation underlies destabilization of retrieved fear memory. Science 319, 1253–1256 38 Ren, Z-Y. et al. (2013) A critical role for protein degradation in the nucleus accumbens core in cocaine reward memory. Neuropsychopharmacology 38, 778–790 39 Hegde, A.N. et al. (2014) Local ubiquitin–proteasome-mediated proteolysis and long-term synaptic plasticity. Front. Mol. Neurosci. 7, 96 40 Arguello, A.A. et al. (2014) Involvement of amygdalar protein kinase A, but not calcium/calmodulin-dependent protein kinase II, in the reconsolidation of cocaine-related contextual memories in rats. Psychopharmacology 231, 55–65 41 Bonini, J.S. et al. (2007) On the participation of hippocampal PKC in acquisition, consolidation and reconsolidation of spatial memory. Neuroscience 147, 37–45 42 Wells, A.M. et al. (2013) Extracellular signal-regulated kinase in the basolateral amygdala, but not the nucleus accumbens core, is critical for context-response-cocaine memory reconsolidation in rats. Neuropsychopharmacology 38, 753–762 43 Duvarci, S. et al. (2005) Activation of extracellular signal-regulated kinase- mitogen-activated protein kinase cascade in the amygdala is required for memory reconsolidation of auditory fear conditioning. Eur. J. Neurosci. 21, 283–289
TINS-1144; No. of Pages 9
Opinion 44 Li, Y. et al. (2013) Learning and reconsolidation implicate different synaptic mechanisms. Proc. Natl. Acad. Sci. U.S.A. 110, 4798–4803 45 Stoica, L. et al. (2011) Selective pharmacogenetic inhibition of mammalian target of rapamycin complex I (mTORC1) blocks longterm synaptic plasticity and memory storage. Proc. Natl. Acad. Sci. U.S.A. 108, 3791–3796 46 Tronson, N.C. et al. (2012) Distinctive roles for amygdalar creb in reconsolidation and extinction of fear memory. Learn. Mem. 19, 178–181 47 Arguello, A.A. et al. (2013) CAAT enhancer binding protein delta plays an essential role in memory consolidation and reconsolidation. J. Neurosci. 33, 3646–3658 48 Boccia, M. et al. (2007) Activation of hippocampal nuclear factor-kappa B by retrieval is required for memory reconsolidation. J. Neurosci. 27, 13436–13445 49 Merlo, E. et al. (2005) Activation of the transcription factor NF-kappaB by retrieval is required for long-term memory reconsolidation. Learn. Mem. 12, 23–29 50 Ploski, J.E. et al. (2011) The neuronal PAS domain protein 4 (Npas4) is required for new and reactivated fear memories. PLoS ONE 6, e23760 51 Duvarci, S. et al. (2008) De novo mRNA synthesis is required for both consolidation and reconsolidation of fear memories in the amygdala. Learn. Mem. 15, 747–755 52 Jarome, T.J. and Helmstetter, F.J. (2013) The ubiquitin–proteasome system as a critical regulator of synaptic plasticity and long-term memory formation. Neurobiol. Learn. Mem. 105, 107–116 53 Flavell, C.R. et al. (2013) Mechanisms governing the reactivationdependent destabilization of memories and their role in extinction. Front. Behav. Neurosci. 7, 214 54 Artinian, J. et al. (2008) Protein degradation, as with protein synthesis, is required during not only long-term spatial memory consolidation but also reconsolidation. Eur. J. Neurosci. 27, 3009–3019 55 Lopez-Salon, M. et al. (2001) The ubiquitin–proteasome cascade is required for mammalian long-term memory formation. Eur. J. Neurosci. 14, 1820–1826 56 Fonseca, R. et al. (2006) A balance of protein synthesis and proteasomedependent degradation determines the maintenance of LTP. Neuron 52, 239–245 57 Doyere, V. et al. (2007) Synapse-specific reconsolidation of distinct fear memories in the lateral amygdala. Nat. Neurosci. 10, 414–416 58 Clarke, J.R. et al. (2010) Plastic modifications induced by object recognition memory processing. Proc. Natl. Acad. Sci. U.S.A. 107, 2652–2657 59 Fonseca, R. et al. (2006) Neuronal activity determines the protein synthesis dependence of long-term potentiation. Nat. Neurosci. 9, 478–480 60 Lee, S.H. et al. (2012) A cellular model of memory reconsolidation involves reactivation-induced destabilization and restabilization at the sensorimotor synapse in Aplysia. Proc. Natl. Acad. Sci. U.S.A. 109, 14200–14205 61 Hu, J.Y. et al. (2011) Persistent long-term synaptic plasticity requires activation of a new signaling pathway by additional stimuli. J. Neurosci. 31, 8841–8850
Trends in Neurosciences xxx xxxx, Vol. xxx, No. x
62 Hu, J.Y. and Schacher, S. (2014) Persistent long-term facilitation at an identified synapse becomes labile with activation of short-term heterosynaptic plasticity. J. Neurosci. 34, 4776–4785 63 Chen, S. et al. (2014) Reinstatement of long-term memory following erasure of its behavioral and synaptic expression in Aplysia. Elife 3, e03896 64 Cai, D. et al. (2012) Reconsolidation of long-term memory in Aplysia. Curr. Biol. 22, 1783–1788 65 Hong, I. et al. (2013) AMPA receptor exchange underlies transient memory destabilization on retrieval. Proc. Natl. Acad. Sci. U.S.A. 110, 8218–8223 66 Yu, Y-J. et al. (2013) Ampa receptor endocytosis in the amygdala is involved in the disrupted reconsolidation of methamphetamineassociated contextual memory. Neurobiol. Learn. Mem. 103, 72–81 67 Descalzi, G. et al. (2012) Rapid synaptic potentiation within the anterior cingulate cortex mediates trace fear learning. Mol. Brain 5, 6 68 Clem, R.L. and Huganir, R.L. (2010) Calcium-permeable AMPA receptor dynamics mediate fear memory erasure. Science 330, 1108–1112 69 Turrigiano, G. (2012) Homeostatic synaptic plasticity: local and global mechanisms for stabilizing neuronal function. Cold Spring Harb. Perspect. Biol. 4, a005736 70 Lee, K.F. et al. (2014) Tuning into diversity of homeostatic synaptic plasticity. Neuropharmacology 78, 31–37 71 Cooper, L.N. and Bear, M.F. (2012) The BCM theory of synapse modification at 30: interaction of theory with experiment. Nat. Rev. Neurosci. 13, 798–810 72 Sevenster, D. et al. (2013) Prediction error governs pharmacologically induced amnesia for learned fear. Science 339, 830–833 73 Rodriguez-Ortiz, C.J. et al. (2008) Intrahippocampal anisomycin infusions disrupt previously consolidated spatial memory only when memory is updated. Neurobiol. Learn. Mem. 89, 352–359 74 Morris, R.G. et al. (2006) Memory reconsolidation: sensitivity of spatial memory to inhibition of protein synthesis in dorsal hippocampus during encoding and retrieval. Neuron 50, 479–489 75 Hou, Q. et al. (2011) Homeostatic regulation of AMPA receptor trafficking and degradation by light-controlled single-synaptic activation. Neuron 72, 806–818 76 Dong, C. et al. (2008) Proteasome inhibition enhances the induction and impairs the maintenance of late-phase long-term potentiation. Learn. Mem. 15, 335–347 77 Dong, C. et al. (2014) Proteasome modulates positive and negative translational regulators in long-term synaptic plasticity. J. Neurosci. 34, 3171–3182 78 Banerjee, S. et al. (2009) A coordinated local translational control point at the synapse involving relief from silencing and MOV10 degradation. Neuron 64, 871–884 79 Greer, P.L. et al. (2010) The Angelman syndrome protein Ube3A regulates synapse development by ubiquitinating Arc. Cell 140, 704–716 80 Mabb, A.M. et al. (2014) Triad3A regulates synaptic strength by ubiquitination of Arc. Neuron 82, 1299–1316
9
Opinion
Reconsolidation and the regulation of plasticity: moving beyond memory Robert P. Bonin1 and Yves De Koninck1,2 1 2
Institut Universitaire en Sante´ Mentale de Que´bec, Que´bec, Canada Department of Psychiatry and Neuroscience, Universite´ Laval, Que´bec, Canada
Memory reconsolidation is a protein synthesis-dependent process that preserves, in some form, memories that have been destabilized through recall. Reconsolidation is a nearly universal phenomenon, occurring in a diverse array of species and learning tasks. The function of reconsolidation remains unclear but it has been proposed as a mechanism for updating or strengthening memories. Observations of an analog of reconsolidation in vitro and in sensory systems indicate that reconsolidation is unlikely to be a learning-specific phenomenon and may serve a broader function. We propose that reconsolidation arises from the activity-dependent induction of two coincident but opposing processes: the depotentiation and repotentiation of strengthened synapses. These processes suggest that reconsolidation reflects a fundamental mechanism that regulates and preserves synaptic strength. Specific versus general view of reconsolidation The ability to learn and store information in the form of memories is crucial for the survival of many organisms. Equally important is the ability to modify or update these memories to accurately reflect a changing environment. Over the past 15 years there has been a considerable amount of research into one process that is thought to enable the modification and updating of consolidated (see Glossary) memories: memory reconsolidation. The process of memory reconsolidation is triggered by the recall of a memory. Recall transiently destabilizes the reactivated memory and renders it labile, after which the memory restabilizes (‘reconsolidates’) to remain stored. Preventing the reconsolidation process, for example, by blocking protein synthesis, leads to disruption of the memory [1]. This cycle of memory destabilization and reconsolidation appears to provide a window of opportunity for the modification of learned associations [2,3]. Reconsolidation was brought to the fore of neuroscience research in 2000, when it was shown that inhibiting protein synthesis in the amygdala during the recall of a fear memory resulted in disruption of the memory [1]. The necessity of protein synthesis provided evidence for a cellular mechanism of reconsolidation that paralleled Corresponding author: Bonin, R.P. ([email protected]). Keywords: long-term potentiation; heterosynaptic plasticity; homeostatic plasticity; reconsolidation; protein synthesis; protein degradation. 0166-2236/ ß 2015 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.tins.2015.04.007
the well-studied process of memory consolidation [4], triggering an explosion of research into the mechanisms and parameters of memory reconsolidation. Since this revitalizing study, reconsolidation has been demonstrated to be a nearly universal phenomenon. Memory reconsolidation has been observed in more than 10 diverse species, from C. elegans to humans, and has been observed in several regions of the central nervous system (CNS) including the hippocampus, amygdala, and anterior cingulate cortex (reviewed in [3,5]). Despite this wealth of information, the overall function of reconsolidation remains unclear. The scientific origin of reconsolidation has largely branded reconsolidation as a learning-specific phenomenon, and theories of the functional role of reconsolidation have centered on its putative role in the modification and updating of memories [2,3,6–8]. In this article we re-evaluate the functional role of reconsolidation in light of in vitro and in vivo evidence that this phenomenon may exist broadly throughout the CNS [9], and is thus unlikely to serve a learning-specific function. We argue that reconsolidation instead reflects a fundamental mechanism for the regulation of synaptic plasticity, and present evidence that the cellular processes mediating this regulation serve to constrain activity-dependent synaptic potentiation in the nervous system. Functional significance of reconsolidation Reconsolidation has been proposed to serve two general functions: the modification and/or strengthening of memories [2,7,8]. The modification of memories through reconsolidation may maintain memory relevance by allowing
Glossary Consolidation: the stabilization of a memory trace after acquisition. Depotentiation/repotentiation: a reduction/increase in the strength of a synapse that has been potentiated by some stimulus. Fear conditioning: a learning paradigm in which a stimulus becomes associated with fear. Heterosynaptic plasticity: changes in synaptic strength that can be induced through activity at other synapses or through other cellular processes. Homeostatic plasticity: the adjustment of synaptic strength to regulate neuronal activity. Homosynaptic plasticity: changes in synaptic strength that are input-specific and restricted to only the activated synapses. Hyperalgesia: increased sensitivity to painful stimuli. Long-term potentiation (LTP)/depression (LTD): an increase/decrease in synaptic strength that persists for hours or longer in response to an induction stimulus. Synaptic plasticity: the ability to modify the strength of a synapse.
Trends in Neurosciences xx (2015) 1–9
1
TINS-1144; No. of Pages 9
Opinion the encoding of changes in the environment [10]. In humans, memories can be modified through the incorporation of new or conflicting information only after recall of the original memory [11,12]. Similar results were observed in animal studies, where memory recall rendered memory labile and vulnerable to modification [13–16]. One interesting prediction of the memory update function of reconsolidation is that the subsequent updating of memory should be most strongly triggered when recall occurs in a context that is slightly different than the original memory, necessitating an update of the stored memory to reflect the changes [17]. In this case, the induction of reconsolidation should be dependent on some element of novelty during recall. This was precisely the case in a study by Winters et al. [18], in which changing the texture of the flooring during a recall trial was necessary to enable reconsolidation of a well-trained object memory in rodents. Modifying the timing of a shock delivered during re-exposure to a fear-conditioning apparatus was also sufficient to induce reconsolidation of memories that were otherwise resistant to reconsolidation [19]. However, reconsolidation can occur even in the absence of explicitly novel information [1,15,20], indicating that the novelty is not an absolute requirement for reconsolidation. Reconsolidation may also enable the strengthening of memories. This would allow the enhancement of behavioral responses through recall without requiring re-exposure to the original learning situation [7,8]. The strengthening of memory after reconsolidation has been described in both animal models [13,21,22] and in humans [12]. However, the strengthening of memory after recall is not seen in every case. Indeed, animal studies exploring reconsolidation often do not find stronger behavioral responses after memory recall that would indicate the enhancement of memory. For example, the degree of freezing used as a measure of memory in fear-conditioning studies does not typically increase after the recall trial (e.g., [1,15], but see [13,22]). This lack of effect is unlikely to result from a ceiling effect in training because it is possible to pharmacologically manipulate reconsolidation to generate memory strengthening after recall [23–25]. One possible explanation for a lack of memory strengthening after reconsolidation is that the recall of a memory in the absence of a conditioning stimulus such as an electric shock promotes memory extinction, which reduces the behavioral expression of the previously learned association and opposes the effects of memory strengthening induced by reconsolidation. In a recent report, Fukushima et al. [21] explore this possible balance through the elegant use an inhibitory avoidance assay designed to trigger reconsolidation under conditions that do not promote extinction learning. In the absence of extinction learning, the retrieval of a fear memory was shown to promote memory enhancement through reconsolidation. Thus, memory recall may activate multiple processes with opposite effects on memory strength, and whose net balance determines the overall effect of recall on memory strength. The strengthening and modification of memories are not necessarily mutually exclusive outcomes of reconsolidation, and it is plausible that both of these effects can arise from similar underlying mechanisms. Indeed, both memory 2
Trends in Neurosciences xxx xxxx, Vol. xxx, No. x
strengthening and modification can be seen after recall in the same behavioral assay, depending on the conditions of recall [15]. Nonetheless, these interpretations of reconsolidation are still rooted in the idea of reconsolidation as a learning-specific phenomenon. This is reasonable given that the study of behavioral reconsolidation has essentially been entirely restricted to learning tasks [2,3,6]. However, recent in vitro and in vivo work provides evidence that reconsolidation is not restricted to learning tasks and may exist broadly throughout the nervous system, encouraging a re-evaluation of the functional role of reconsolidation. Reconsolidation beyond memory We have recently demonstrated that a reconsolidation-like phenomenon can also be observed in pain-processing circuits of the spinal cord [9]. In this study, mechanical hyperalgesia was induced by the injection of capsaicin into the hind paws of mice. Three hours after the initial plantar injection, capsaicin was reinjected into the hind paw to reactivate the sensitized pain pathways. Although the second capsaicin injection did not modify hyperalgesia on its own, normal mechanosensitivity was quickly restored when this second injection of capsaicin was paired with the spinal administration of the protein synthesis inhibitor, anisomycin, via intrathecal injection. These findings indicate that the reactivation of sensitized spinal pain pathways rendered the hyperalgesia labile and reversible in a manner equivalent to the disruption of fear memories through reconsolidation impairment. The observation of a reconsolidation-like phenomenon in spinal cord pain-processing circuits is not inconceivable given the strong mechanistic parallels between learning and the development of some forms of hyperalgesia (reviewed in [26–28]). Indeed, persistent hyperalgesia has been called a ‘memory trace of pain’ or pain ‘engram’ because of these mechanistic similarities [29,30]. Nevertheless, the observation of reconsolidation in spinal painprocessing pathways encourages a re-examination of the function of reconsolidation on the basis of one key point: given that reconsolidation is not restricted to learning tasks, it is unlikely to be a unique property of learned associations. This shift in the interpretation of reconsolidation does not contradict the proposals of reconsolidation as a means of memory trace strengthening or updating. Rather, we propose that the broad existence of reconsolidation suggests that it may serve a more fundamental role in the regulation of plasticity than possibly predicted from an analysis of reconsolidation solely within learning paradigms. Re-evaluating the functional role of reconsolidation A close examination of the processes triggered by reactivation of a memory trace reveals the existence of two opposing processes that can be recruited to enable the labilization of memory: a destabilizing process and a protein synthesisdependent reconsolidation process (Figure 1A). The activitydependent destabilizing process can be revealed by the inhibition of protein synthesis after reactivation of the memory trace. This typically leads to the disruption of the learned association that is a hallmark of assays designed to study reconsolidation. The labile state of memory has
TINS-1144; No. of Pages 9
Opinion
Trends in Neurosciences xxx xxxx, Vol. xxx, No. x
(A)
Behavioral reconsolidaon Memory formaon (consolidaon)
Labile memory Recall
Stored memory
Synapc reconsolidaon Strong synapc acvaon
Potenated synapse
Destabilizaon and reconsolidaon
Stored memory
Labile plascity Reacvaon
Depotenaon and
repotenaon
Potenated synapse
Blocked depotenaon?
Synapc strength
(B)
Blocked repotenaon TRENDS in Neurosciences
Figure 1. The ups and downs of reconsolidation. Schematics of the processes involved in reconsolidation at the behavioral and synaptic levels are shown in (A). The activation of memory destabilizing and reconsolidation processes by memory recall places the memory into a labile state, in which it can be disrupted through the application of protein synthesis inhibitors. (B) Synaptic potentiation can be induced by strong activation of a synapse (yellow bolt), leading to an increase of the synaptic strength. This synaptic potentiation is thought to underlie memory and some forms of hyperalgesia [27]. Reactivation of the strengthened synapse, as might occur during memory recall, triggers the induction of two opposing processes: the depotentiation (blue arrow) and repotentiation of synaptic strength (red arrow). These two processes often cancel out to produce no net change in synaptic strength; however, preventing repotentiation through the inhibition of protein synthesis can lead to a net depotentiation of the synapse. Behaviorally, this can be seen as the blockade of reconsolidation and the disruption of memory. It has not yet been conclusively demonstrated whether the selective prevention of depotentiation can lead to additional synaptic strengthening or memory augmentation.
been previously described as arising from two distinct processes of destabilization and reconsolidation [17,31–34], and there is molecular evidence to support such a division. In auditory fear-conditioning experiments, the blockade of GluN2B subunit-containing NMDA receptors in the basolateral amygdala prevented the memory from becoming labile, while blockade of GluN2A-containing NMDA receptors only prevented repotentiation of the memory, indicating that the two processes are activated through separate signaling mechanisms [35]. The destabilization process is necessarily independent of protein synthesis because it persists following inhibition of protein synthesis. The reactivation of a consolidated memory trace triggers the active degradation of proteins necessary for the maintenance of the memory trace, including the synaptic scaffolding proteins Shank (SH3 and multiple ankyrin repeat domains) and GKAP (guanylate kinase-associated protein) [13,19,21,36–38] (Box 1). Protein degradation induced by memory retrieval involves protein polyubiquitination and proteasomal degradation [39]. Protein degradation is necessary for the disruption or modification of a reactivated memory trace because inhibition of the proteasome prevents the disruption of a recalled memory by protein synthesis inhibition [13,19,33,34,36–38]. The polyubiquitination of proteins occurs after memory recall even in the absence of protein synthesis inhibition typically used in reconsolidation
assays [37,38], indicating that the degradation of proteins is not an artifact of protein synthesis inhibition. Memory recall also triggers reconsolidation, which counters the destabilization process and maintains the memory trace. Reconsolidation is dependent on proteinsynthesis and is necessary for the preservation of the memory trace after recall. Several transcriptional control pathways have been implicated in memory repotentiation, including protein kinase A [24,40], protein kinase C [41], ERK/MAP kinase [42,43], CaMKII [34], mTOR activation [44,45], as well as the transcription factors CREB [46], C/ EBPd [47], NF-kB [48,49], Zif268 [19], and NPAS4 [50]. The de novo synthesis of mRNA is also crucial for reconsolidation [51]. Several recent review articles describe the molecular mechanisms of destabilization and reconsolidation processes in detail [2,3,5,52,53], and these will not be further reviewed here. Functional links between destabilization and reconsolidation The destabilization and reconsolidation processes appear to be mechanistically intertwined, whereby both processes are dependent on protein degradation (Box 1). If destabilization and reconsolidation were mechanistically distinct, it is expected that preventing the destabilization process alone by inhibiting protein degradation would enable the strengthening of memories after recall through 3
TINS-1144; No. of Pages 9
Opinion
Trends in Neurosciences xxx xxxx, Vol. xxx, No. x
Box 1. How can protein degradation lead to synaptic strengthening? Protein degradation has been implicated as a key process in both the depotentiation and repotentiation processes of synaptic reconsolidation. At first glance it can seem paradoxical that a destructive process such as protein degradation can result in synaptic strengthening or repotentiation. However, synaptic plasticity is determined by a multitude of proteins, some of which act to negatively modulate synaptic strength or constrain further synaptic potentiation; the removal of these inhibitory controls can produce synaptic facilitation (Figure I). Protein degradation associated with plasticity is mediated, in large part, by the ubiquitin–proteasome system [39]. Functionally, inhibiting the proteasome can enhance the induction, but impair the maintenance of LTP in vitro [56,76], consistent with a facilitating role of protein degradation in plasticity. Inhibition of the proteasome after the induction of LTP leads to the accumulation of the translational repressors PAIP2 and 4E-BP2 [77]. In addition, NMDA receptor activation can induce proteasomal degradation of MOV10, a component of the RNA-mediated gene silencing complex (RISC), in cultured neurons [78]. These findings indicate that the synthesis of proteins associated with long-term plasticity is disinhibited through the activity-dependent proteasomal degradation of repressor proteins. Synaptic strength can also be augmented through the degradation of Arc, a protein that facilitates the internalization of AMPA receptors [79,80]. The mechanisms by which protein degradation contributes to the repotentiation process of reconsolidation are much less clear, but likely involve similar disinhibitory processes. For example, MOV10 is also ubiquitinated and degraded following memory retrieval [36].
Ub Shank
GKAP
PS
D-
95
UPS Ub
Arc Endocytosis
Ub MOV10
+
Translaon
Inseron
Key:
Potenaon process Depotenaon process TRENDS in Neurosciences
Figure I. Protein degradation regulates synaptic strength. Protein degradation mediated by the ubiquitin–proteasome system (UPS) can induce both reductions (blue arrows) and enhancements (red arrows) of synaptic strength. AMPA receptors are stabilized at the membrane by synaptic adhesion molecules including Shank and GKAP. These adhesion molecules are degraded during the depotentiation process of reconsolidation. AMPA receptor endocytosis is regulated by Arc in an activity-dependent manner, and the degradation of Arc can enhance synaptic strength. Notably, the endocytosis of AMPA receptors is a necessary process for the membrane insertion of AMPA receptors during repotentiation. The translation of new synaptic proteins necessary for repotentiation is regulated by translational repressors, such as MOV10. The degradation of these repressors is also activity-dependent and contributes to reconsolidation. Abbreviations: Arc, activity-regulated cytoskeleton-associated protein; GKAP, guanylate kinase-associated protein, also known as DLGAP1 (discs large homolog-associated protein 1); MOV10, Mov10 RISC complex RNA helicase; PSD-95, postsynaptic density 95; Shank, SH3 and multiple ankyrin repeat domains; Ub, ubiquitin.
unbalanced repotentiation. However, this has not yet been directly observed, and inhibition of protein degradation often, but not always [54], prevents the destabilization of memory after recall but does not alter the strength of a recalled memory [13,19,33,34,36–38]. In addition, the inhibition of protein degradation not only failed to produce further memory strengthening but even prevented recallinduced strengthening in an inhibitory avoidance assay designed to reveal an enhancement of memory strength through recall [21]. As a notable parallel with reconsolidation, the initial formation of long-term memories [33,54,55], and the stabilization of late-phase long-term potentiation (LTP) in vitro [56], may also depend on protein degradation (Box 1). Indeed, the inhibition of protein degradation immediately, but not 3 h, after initial learning or recall prevented both consolidation and reconsolidation, respectively [54], suggesting that similar mechanisms may act in both consolidation and reconsolidation. Overall, it remains to be determined whether it is possible to mechanistically separate memory destabilization and reconsolidation, such that 4
preventing destabilization leads to the further potentiation of the memory through unbalanced reconsolidation (Figure 1B and Box 2). Reconsolidation at the synapse Working from the premise that reconsolidation is widely distributed in the nervous system, it becomes essential to question the function of the opposing processes of destabilization and reconsolidation in broader terms by considering their cellular correlates at the level of synaptic plasticity that contributes to mnemonic traces. In a striking parallel to behavioral studies, the reactivation of potentiated synapses can trigger a synaptic depotentiation process that is unmasked by the inhibition of protein synthesis, and a countering, protein synthesis-dependent, synaptic ‘repotentiation’ process (Figure 1B) [9,44,57–59]. In vivo electrophysiological recordings revealed that longterm (but not short-term) facilitation of auditory-evoked field potentials in the lateral amygdala was rendered labile and able to undergo depotentiation after recall following
TINS-1144; No. of Pages 9
Opinion
Trends in Neurosciences xxx xxxx, Vol. xxx, No. x
Box 2. Outstanding questions Can the destabilization and reconsolidation processes activated by memory recall be mechanistically separated? Can the balance of depotentiation and repotentiation processes be manipulated, such that unbalanced repotentiation leads to the strengthening of the memory trace? Are the boundary conditions for memory reconsolidation determined by the extent of synaptic activation (e.g., homosynaptic or heterosynaptic) during recall? Is reconsolidation observed throughout the CNS, beyond memory and pain circuits?
fPSP area (% baseline)
the inhibition of protein synthesis, similarly to memory destabilization [57]. However, the facilitated field potentials did not exhibit additional facilitation upon reactivation of cortical and thalamic inputs to the lateral amygdala after fear conditioning [44]. In the hippocampus, field potentials evoked at CA3 to CA1 synapses were facilitated after novel object recognition training but, again, were not further potentiated after reconsolidation [58]. Notably, reconsolidation in this study was accompanied by transient depotentiation of the field potentials at the CA3 to CA1 synapse immediately after recall, and this was attributed to synapses entering a ‘labile’ state [58]. Studies of reconsolidation in vitro have also failed to reveal a robust increase in synaptic potentiation after reactivation of potentiated synapses (Figure 2). In the spinal cord dorsal horn, long-term enhancement of
synaptic strength was not further increased by a second induction stimulus, but was reversed to baseline levels when the second stimulus was applied in the presence of protein synthesis inhibitors [9]. In addition, the stimulation of hippocampal slices several hours after the initial induction of LTP did not cause further synaptic potentiation, but initiated a reversal of late-phase LTP when applied in the presence of a protein synthesis inhibitor [59]. Reconsolidation has also been studied using the gill- and siphon-withdrawal reflex in the sea slug Aplysia, which serves as a simple model of non-associative memory at both the behavioral and synaptic levels [60–64]. The neuronal circuitry mediating the withdrawal reflex is well characterized and enables the measurement of changes in synaptic strength in the circuit over long periods of time. Facilitation of the sensory-to-motor neuron synapse responsible for the withdrawal reflex is rendered labile after a second presentation of an induction stimulus, and can be reversed through the inhibition of protein synthesis [60– 63]. Similarly to mammalian models, this reconsolidation process is dependent on protein degradation [60]. Crucially, a second induction stimuli alone does not cause further potentiation of the facilitated synapse [60,62], or an increase in the number of synaptic varicosities [63], in agreement with the proposition that synaptic reconsolidation maintains synaptic strength. Overall, these studies reveal a consistent trend in which reactivation of the memory trace fails to produce further synaptic potentiation, despite
200
150
100
50 0
120
240
360
Time (min) Aniso fPSP area (% baseline)
fPSP area (% baseline)
Aniso 200
150
100
200
150
100
X
50
50 0
120
240
Time (min)
360
0
120
240
360
Time (min) TRENDS in Neurosciences
Figure 2. Synaptic reconsolidation outside classical learning pathways. A synaptic correlate of reconsolidation can be seen with an in vitro spinal cord preparation. Strong electrical stimulation of dorsal roots (yellow bolts) induces a long-lasting enhancement of synaptic strength in the superficial dorsal horn, as indicated by the increase in postsynaptic field potentials (fPSP). Delivery of a second stimulus, mimicking the reactivation of a memory trace, produces no net change in synaptic strength (black symbols). However, when the second stimulus is given in the presence of the protein synthesis inhibitor, anisomycin (Aniso, blue rectangle), there is a robust reversal of the synaptic potentiation (blue symbols). This parallels the disruption of memory or hyperalgesia seen when reconsolidation is inhibited. Importantly, in the absence of the second stimulus, the protein synthesis inhibitor did not significantly affect synaptic strength (green symbols). Data modified from [9].
5
TINS-1144; No. of Pages 9
Opinion the induction of a protein synthesis-dependent process that counters activity-dependent depotentiation. The lack of further potentiation after the presentation of a second induction stimulus could arise from a ceiling effect in which further potentiation of synapses upon recall is not possible. Nevertheless, there is evidence that an inability to induce further synaptic potentiation in reconsolidation experiments does not reflect synaptic saturation. The long-term enhancement of both auditoryevoked field potentials [57] and persistent synaptic facilitation in Aplysia [62] underwent additional augmentation following the re-presentation of induction stimuli, although in both cases this augmentation was transient and did not persist over the long term. In addition, the difficulties in dissociating the depotentiation and repotentiation processes further complicate the study of saturation in reconsolidation by preventing the selective augmentation of memory traces through the inhibition of depotentiation. In the absence of synaptic facilitation upon recall, it is possible to envisage that the net effect of opposing depotentiation and repotentiation processes would be the recall-induced turnover of synaptic proteins. Such a turnover process can be seen in the exchange of synaptic AMPA receptors induced by memory recall [23,38,65–68]. Changes in synaptic AMPA receptor composition induced by memory recall have been captured through the use of cross-sectional approaches that examine reactivation-induced changes in synaptic responses at various timepoints [23,65]. Memory recall after fear conditioning triggers a transient endocytosis of AMPA receptors that can be observed within 5 minutes of memory recall [65], and may last for up to 4 h [23]. Crucially, preventing AMPA receptor endocytosis and exchange also prevented the destabilization of memory upon recall [23,65]. Hong et al. [65] observed an endocytosis of calcium-permeable, GluA2-containing receptors that was paralleled by the membrane insertion of calcium-impermeable AMPA receptors. Notably, this exchange of AMPA receptors produced no net change in miniature excitatory postsynaptic conductance (mEPSCs), consistent with the lack of reactivation effect seen in previous electrophysiological studies [9,44,60,62]. Conversely, Rao-Ruiz et al. [23] observed a small but significant decrease in mEPSC amplitude 1 h after recall, with restoration of mEPSC amplitude to control levels by 7 h post-recall, indicating that the increase in calcium-impermeable AMPA receptors does not always occur simultaneously with AMPA receptor endocytosis. This brief decrease in synaptic strength is similar to that observed through in vivo hippocampal recordings [58], and suggests that transient synaptic depression induced by recall may depend on a balance of AMPA receptor removal and insertion that, for reasons unknown, can vary between studies. However, this transient decrease in synaptic strength is unlikely to be crucial for reconsolidation because memory was still rendered labile under conditions in which the decrease was absent [65]. Finally, it must be noted that preventing memory destabilization by inhibiting AMPA receptor endocytosis does not necessarily prevent reconsolidation. Blockade of AMPA receptor endocytosis was associated with an enhancement of fear 6
Trends in Neurosciences xxx xxxx, Vol. xxx, No. x
memory that may reflect the modification of memory through recall [23]; however, the enhancement of fear memory following the inhibition of AMPA receptor endocytosis associated with reconsolidation is not always observed [65]. Functional significance of synaptic reconsolidation The activation of a memory trace clearly induces changes in synapse composition but, as summarized above, rarely produces robust changes in synaptic strength. Such findings support the view of synaptic reconsolidation as a product of two opposing processes that largely cancel out in overall effect. These observations raise the question of why two opposing processes would be triggered through the reactivation of a memory trace. Simultaneous depotentiation and repotentiation processes could serve as a means of limiting continual synaptic potentiation in the face of ongoing synaptic activity. As such, synaptic reconsolidation may function as one of several cellular mechanisms that homeostatically maintain synaptic potentiation within a functional range [69,70]. In the absence of these homeostatic mechanisms, the positivefeedback nature of activity-dependent synaptic potentiation can theoretically lead to synaptic saturation, runaway excitation, and loss of network plasticity as activity becomes increasingly concentrated within a select number of potentiated synapses [69,70]. Synaptic reconsolidation would serve a unique role in the homeostatic maintenance of neuronal activity and excitability because depotentiation and repotentiation occur on a much shorter timescale than most identified mechanisms of homeostatic synaptic plasticity that are recruited in response to changes in neuronal activity and occur over hours to days [70]. Specifically, the mechanisms of synaptic depotentiation can be induced almost immediately after synaptic activity [65], making this process much more temporally related to activity-dependent LTP and long-term depression (LTD). In addition, the interpretation of reconsolidation as a property of synaptic plasticity provides a consistent mechanism for the strengthening or modification of memory seen behaviorally after recall because these outcomes would be enabled through the ‘window’ of plasticity at the cellular and network levels provided by synaptic depotentiation and repotentiation. The relative balance of synaptic potentiation and depotentiation triggered by synaptic activity may further determine the overall impact of synaptic reconsolidation on synaptic strength and modification of the memory trace, although evidence for such a regulated balance is currently lacking (Box 2). The putative homeostatic function of synaptic reconsolidation has parallels with the Bienenstock–Cooper– Munro (BCM) model of synaptic metaplasticity, in which the threshold for the induction of LTP or LTD is dependent on the prior activity of the relevant synapses [71]. In the BCM model, a high level of synaptic activity shifts the threshold for the induction of long-term plasticity such that subsequent activity is less likely to produce synaptic potentiation, and may even induce synaptic depotentiation. In a similar manner, a second presentation of strong stimuli as may occur during recall does not necessarily
TINS-1144; No. of Pages 9
Opinion
Trends in Neurosciences xxx xxxx, Vol. xxx, No. x
produce further strengthening of a previously-potentiated synapse owing to the activation of a depotentiation process in synaptic reconsolidation. The net effects of synaptic reconsolidation on network activity and learning will depend on the precise mechanisms underlying synaptic depotentiation and repotentiation, and how broadly these mechanisms impact upon neuronal function. Intriguingly, a recent study in Aplysia found that reconsolidation was triggered only by stimuli that generate heterosynaptic but not homosynaptic plasticity (Figure 3A–C) [62]. In this study, persistently facilitated synapses were reactivated with either electrical stimulation that normally induces transient homosynaptic potentiation, or by serotonin application that transiently facilitates many synapses (heterosynaptic potentiation). While electrical stimulation failed to induce synaptic depotentiation in the presence of the translation inhibitor, rapamycin, depotentiation was induced by the combination of serotonin stimulation and rapamycin. The observation that heterosynaptic facilitation is required for reconsolidation in Aplysia [62] may be relevant to behavioral observations where reconsolidation is only triggered by specific conditions, such as novelty [18,19] or prediction error [72] during recall. Specifically, these observations raise the questions of whether the boundary conditions for
memory reconsolidation are determined by the extent of synaptic activation during recall, and whether memory reconsolidation requires behavioral stimuli that activate heterosynaptic plasticity (Figure 3D and Box 2). In a similar vein, the use of ‘weak’ reactivation protocols in vitro and in vivo was not sufficient to render memory traces labile [61,63,73,74], suggesting that reconsolidation requires strong reactivation of the facilitated memory traces. It is also unknown whether the processes of synaptic potentiation and depotentiation are restricted to a subset of synapses, such as those synapses that have undergone some degree of potentiation or were directly activated by reactivation of the memory trace, or whether the mechanisms of synaptic reconsolidation globally affect synaptic strength in the neuron. There is compelling evidence that activity-dependent synaptic depotentiation can occur on a local scale. The selective activation of a subset of AMPAcontaining synapses led to the removal and proteasomedependent degradation of AMPA receptors exclusively at activated synapses, without affecting neighboring synapses [75]. Determining whether synaptic reconsolidation is constrained by a similar synaptic selectivity will have important implications for its effects on information storage and processing [69,70].
(A) Potenaon
(B) No effect Homosynapc reacvaon (C) Synapc reconsolidaon
?
Heterosynapc reacvaon (D) Stored memory
?
No effect
Homosynapc acvaon?
Destabilizaon and reconsolidaon
Heterosynapc acvaon?
Recall
TRENDS in Neurosciences
Figure 3. Synaptic reconsolidation induced by heterosynaptic activity. The reactivation of facilitated synapses does not always induce synaptic reconsolidation, similarly to observations that memory recall does not always lead to behavioral reconsolidation. In Aplysia, reconsolidation was triggered only by stimuli that generate heterosynaptic but not homosynaptic plasticity [62]. (A) The initial stimulation of a synapse can induce long-lasting synaptic potentiation. (B) The reactivation of the potentiated synapse alone (homosynaptic activation) does not necessarily trigger synaptic reconsolidation. (C) Reactivation of several synapses, including the potentiated synapses (heterosynaptic activation), can trigger the depotentiation and repotentiation processes of synaptic reconsolidation. However, the net effect of reactivation stimuli on unpotentiated synapses is unknown. (D) The putative requirement of heterosynaptic activation for synaptic reconsolidation may underlie observations that recall does not always trigger memory reconsolidation – the behavioral stimuli that fail to trigger memory reconsolidation may not cause sufficient synaptic activation.
7
TINS-1144; No. of Pages 9
Opinion Concluding remarks The interpretation of reconsolidation as a fundamental property of plasticity greatly expands the relevance of this process to overall CNS function. Reconsolidation has already been observed to occur in a wide variety of learning tasks [3,5], while the demonstration that reconsolidation exists in spinal pain-processing networks reveals a role for reconsolidation in sensory function [9]. We anticipate that, similarly to LTP, which was originally described in memory circuits, reconsolidation will continue to be observed in increasingly diverse CNS regions and cognitive functions, reflecting its fundamental regulatory role in plasticity. Acknowledgments We thank Jimena Perez-Sanchez for her assistance with figures. This work was supported by a Pfizer–Fonds de Recherche Que´bec–Sante´ (FRQS) Innovation Fund Award to Y.D.K., a Canadian Institutes of Health Research grant (MOP 12942) to Y.D.K., and a Catherine Bushnell Pain Research Fellowship from the Louise and Alan Edwards Foundation to R.P.B.
References 1 Nader, K. et al. (2000) Fear memories require protein synthesis in the amygdala for reconsolidation after retrieval. Nature 406, 722–726 2 Finnie, P.S. and Nader, K. (2012) The role of metaplasticity mechanisms in regulating memory destabilization and reconsolidation. Neurosci. Biobehav. Rev. 36, 1667–1707 3 Besnard, A. et al. (2012) Reconsolidation of memory: a decade of debate. Prog. Neurobiol. 99, 61–80 4 Dudai, Y. (2012) The restless engram: consolidations never end. Annu. Rev. Neurosci. 35, 227–247 5 Almeida-Correa, S. and Amaral, O.B. (2014) Memory labilization in reconsolidation and extinction – evidence for a common plasticity system? J. Physiol. 108, 292–306 6 Kroes, M.C. and Fernandez, G. (2012) Dynamic neural systems enable adaptive, flexible memories. Neurosci. Biobehav. Revs. 36, 1646–1666 7 Forcato, C. et al. (2014) Strengthening a consolidated memory: the key role of the reconsolidation process. J. Physiol. 108, 323–333 8 Alberini, C.M. and Ledoux, J.E. (2013) Memory reconsolidation. Curr. Biol. 23, 746–750 9 Bonin, R.P. and De Koninck, Y. (2014) A spinal analog of memory reconsolidation enables reversal of hyperalgesia. Nat. Neurosci. 17, 1043–1045 10 Lee, J.L. (2009) Reconsolidation: maintaining memory relevance. Trends Neurosci. 32, 413–420 11 Hupbach, A. et al. (2007) Reconsolidation of episodic memories: a subtle reminder triggers integration of new information. Learn. Mem. 14, 47–53 12 Forcato, C. et al. (2010) Reconsolidation in humans opens up declarative memory to the entrance of new information. Neurobiol. Learn. Mem. 93, 77–84 13 Lee, J.L. (2008) Memory reconsolidation mediates the strengthening of memories by additional learning. Nat. Neurosci. 11, 1264–1266 14 Haubrich, J. et al. (2015) Reconsolidation allows fear memory to be updated to a less aversive level through the incorporation of appetitive information. Neuropsychopharmacology 40, 315–326 15 De Oliveira Alvares, L. et al. (2013) Reactivation enables memory updating, precision-keeping and strengthening: Exploring the possible biological roles of reconsolidation. Neuroscience 244, 42–48 16 Olshavsky, M.E. et al. (2013) Updating appetitive memory during reconsolidation window: critical role of cue-directed behavior and amygdala central nucleus. Front. Behav. Neurosci. 7, 186 17 Exton-McGuinness, M.T. et al. (2014) Updating memories – the role of prediction errors in memory reconsolidation. Behav. Brain Res. 278, 375–384 18 Winters, B.D. et al. (2009) Older and stronger object memories are selectively destabilized by reactivation in the presence of new information. Learn. Mem. 16, 545–553
8
Trends in Neurosciences xxx xxxx, Vol. xxx, No. x
19 Lee, J.L. (2010) Memory reconsolidation mediates the updating of hippocampal memory content. Front. Behav. Neurosci. 4, 168 20 de Oliveira Alvares, L. et al. (2008) Opposite action of hippocampal cb1 receptors in memory reconsolidation and extinction. Neuroscience 154, 1648–1655 21 Fukushima, H. et al. (2014) Enhancement of fear memory by retrieval through reconsolidation. Elife 3, e02736 22 Inda, M.C. et al. (2011) Memory retrieval and the passage of time: from reconsolidation and strengthening to extinction. J. Neurosci. 31, 1635–1643 23 Rao-Ruiz, P. et al. (2011) Retrieval-specific endocytosis of GluA2AMPARs underlies adaptive reconsolidation of contextual fear. Nat. Neurosci. 14, 1302–1308 24 Tronson, N.C. et al. (2006) Bidirectional behavioral plasticity of memory reconsolidation depends on amygdalar protein kinase A. Nat. Neurosci. 9, 167–169 25 Lee, J.L. et al. (2006) Reconsolidation and extinction of conditioned fear: Inhibition and potentiation. J. Neurosci. 26, 10051–10056 26 Latremoliere, A. and Woolf, C.J. (2009) Central sensitization: a generator of pain hypersensitivity by central neural plasticity. J. Pain 10, 895–926 27 Sandkuhler, J. and Lee, J. (2013) How to erase memory traces of pain and fear. Trends Neurosci. 36, 343–352 28 Ruscheweyh, R. et al. (2011) Long-term potentiation in spinal nociceptive pathways as a novel target for pain therapy. Mol. Pain 7, 20 29 Drdla-Schutting, R. et al. (2012) Erasure of a spinal memory trace of pain by a brief, high-dose opioid administration. Science 335, 235–238 30 Asiedu, M.N. et al. (2011) Spinal protein kinase Mz underlies the maintenance mechanism of persistent nociceptive sensitization. J. Neurosci. 31, 6646–6653 31 Nader, K. and Hardt, O. (2009) A single standard for memory: the case for reconsolidation. Nat. Rev. Neurosci. 10, 224–234 32 Kaang, B.K. et al. (2009) Synaptic protein degradation as a mechanism in memory reorganization. Neuroscientist 15, 430–435 33 Sol Fustinana, M. et al. (2014) Protein degradation by ubiquitin– proteasome system in formation and labilization of contextual conditioning memory. Learn. Mem. 21, 478–487 34 Da Silva, W.C. et al. (2013) Memory reconsolidation and its maintenance depend on L-voltage-dependent calcium channels and CaMKII functions regulating protein turnover in the hippocampus. Proc. Natl. Acad. Sci. U.S.A. 110, 6566–6570 35 Milton, A.L. et al. (2013) Double dissociation of the requirement for GluN2B- and GluN2A-containing NMDA receptors in the destabilization and restabilization of a reconsolidating memory. J. Neurosci. 33, 1109–1115 36 Jarome, T.J. et al. (2011) Activity dependent protein degradation is critical for the formation and stability of fear memory in the amygdala. PLoS ONE 6, e24349 37 Lee, S.H. et al. (2008) Synaptic protein degradation underlies destabilization of retrieved fear memory. Science 319, 1253–1256 38 Ren, Z-Y. et al. (2013) A critical role for protein degradation in the nucleus accumbens core in cocaine reward memory. Neuropsychopharmacology 38, 778–790 39 Hegde, A.N. et al. (2014) Local ubiquitin–proteasome-mediated proteolysis and long-term synaptic plasticity. Front. Mol. Neurosci. 7, 96 40 Arguello, A.A. et al. (2014) Involvement of amygdalar protein kinase A, but not calcium/calmodulin-dependent protein kinase II, in the reconsolidation of cocaine-related contextual memories in rats. Psychopharmacology 231, 55–65 41 Bonini, J.S. et al. (2007) On the participation of hippocampal PKC in acquisition, consolidation and reconsolidation of spatial memory. Neuroscience 147, 37–45 42 Wells, A.M. et al. (2013) Extracellular signal-regulated kinase in the basolateral amygdala, but not the nucleus accumbens core, is critical for context-response-cocaine memory reconsolidation in rats. Neuropsychopharmacology 38, 753–762 43 Duvarci, S. et al. (2005) Activation of extracellular signal-regulated kinase- mitogen-activated protein kinase cascade in the amygdala is required for memory reconsolidation of auditory fear conditioning. Eur. J. Neurosci. 21, 283–289
TINS-1144; No. of Pages 9
Opinion 44 Li, Y. et al. (2013) Learning and reconsolidation implicate different synaptic mechanisms. Proc. Natl. Acad. Sci. U.S.A. 110, 4798–4803 45 Stoica, L. et al. (2011) Selective pharmacogenetic inhibition of mammalian target of rapamycin complex I (mTORC1) blocks longterm synaptic plasticity and memory storage. Proc. Natl. Acad. Sci. U.S.A. 108, 3791–3796 46 Tronson, N.C. et al. (2012) Distinctive roles for amygdalar creb in reconsolidation and extinction of fear memory. Learn. Mem. 19, 178–181 47 Arguello, A.A. et al. (2013) CAAT enhancer binding protein delta plays an essential role in memory consolidation and reconsolidation. J. Neurosci. 33, 3646–3658 48 Boccia, M. et al. (2007) Activation of hippocampal nuclear factor-kappa B by retrieval is required for memory reconsolidation. J. Neurosci. 27, 13436–13445 49 Merlo, E. et al. (2005) Activation of the transcription factor NF-kappaB by retrieval is required for long-term memory reconsolidation. Learn. Mem. 12, 23–29 50 Ploski, J.E. et al. (2011) The neuronal PAS domain protein 4 (Npas4) is required for new and reactivated fear memories. PLoS ONE 6, e23760 51 Duvarci, S. et al. (2008) De novo mRNA synthesis is required for both consolidation and reconsolidation of fear memories in the amygdala. Learn. Mem. 15, 747–755 52 Jarome, T.J. and Helmstetter, F.J. (2013) The ubiquitin–proteasome system as a critical regulator of synaptic plasticity and long-term memory formation. Neurobiol. Learn. Mem. 105, 107–116 53 Flavell, C.R. et al. (2013) Mechanisms governing the reactivationdependent destabilization of memories and their role in extinction. Front. Behav. Neurosci. 7, 214 54 Artinian, J. et al. (2008) Protein degradation, as with protein synthesis, is required during not only long-term spatial memory consolidation but also reconsolidation. Eur. J. Neurosci. 27, 3009–3019 55 Lopez-Salon, M. et al. (2001) The ubiquitin–proteasome cascade is required for mammalian long-term memory formation. Eur. J. Neurosci. 14, 1820–1826 56 Fonseca, R. et al. (2006) A balance of protein synthesis and proteasomedependent degradation determines the maintenance of LTP. Neuron 52, 239–245 57 Doyere, V. et al. (2007) Synapse-specific reconsolidation of distinct fear memories in the lateral amygdala. Nat. Neurosci. 10, 414–416 58 Clarke, J.R. et al. (2010) Plastic modifications induced by object recognition memory processing. Proc. Natl. Acad. Sci. U.S.A. 107, 2652–2657 59 Fonseca, R. et al. (2006) Neuronal activity determines the protein synthesis dependence of long-term potentiation. Nat. Neurosci. 9, 478–480 60 Lee, S.H. et al. (2012) A cellular model of memory reconsolidation involves reactivation-induced destabilization and restabilization at the sensorimotor synapse in Aplysia. Proc. Natl. Acad. Sci. U.S.A. 109, 14200–14205 61 Hu, J.Y. et al. (2011) Persistent long-term synaptic plasticity requires activation of a new signaling pathway by additional stimuli. J. Neurosci. 31, 8841–8850
Trends in Neurosciences xxx xxxx, Vol. xxx, No. x
62 Hu, J.Y. and Schacher, S. (2014) Persistent long-term facilitation at an identified synapse becomes labile with activation of short-term heterosynaptic plasticity. J. Neurosci. 34, 4776–4785 63 Chen, S. et al. (2014) Reinstatement of long-term memory following erasure of its behavioral and synaptic expression in Aplysia. Elife 3, e03896 64 Cai, D. et al. (2012) Reconsolidation of long-term memory in Aplysia. Curr. Biol. 22, 1783–1788 65 Hong, I. et al. (2013) AMPA receptor exchange underlies transient memory destabilization on retrieval. Proc. Natl. Acad. Sci. U.S.A. 110, 8218–8223 66 Yu, Y-J. et al. (2013) Ampa receptor endocytosis in the amygdala is involved in the disrupted reconsolidation of methamphetamineassociated contextual memory. Neurobiol. Learn. Mem. 103, 72–81 67 Descalzi, G. et al. (2012) Rapid synaptic potentiation within the anterior cingulate cortex mediates trace fear learning. Mol. Brain 5, 6 68 Clem, R.L. and Huganir, R.L. (2010) Calcium-permeable AMPA receptor dynamics mediate fear memory erasure. Science 330, 1108–1112 69 Turrigiano, G. (2012) Homeostatic synaptic plasticity: local and global mechanisms for stabilizing neuronal function. Cold Spring Harb. Perspect. Biol. 4, a005736 70 Lee, K.F. et al. (2014) Tuning into diversity of homeostatic synaptic plasticity. Neuropharmacology 78, 31–37 71 Cooper, L.N. and Bear, M.F. (2012) The BCM theory of synapse modification at 30: interaction of theory with experiment. Nat. Rev. Neurosci. 13, 798–810 72 Sevenster, D. et al. (2013) Prediction error governs pharmacologically induced amnesia for learned fear. Science 339, 830–833 73 Rodriguez-Ortiz, C.J. et al. (2008) Intrahippocampal anisomycin infusions disrupt previously consolidated spatial memory only when memory is updated. Neurobiol. Learn. Mem. 89, 352–359 74 Morris, R.G. et al. (2006) Memory reconsolidation: sensitivity of spatial memory to inhibition of protein synthesis in dorsal hippocampus during encoding and retrieval. Neuron 50, 479–489 75 Hou, Q. et al. (2011) Homeostatic regulation of AMPA receptor trafficking and degradation by light-controlled single-synaptic activation. Neuron 72, 806–818 76 Dong, C. et al. (2008) Proteasome inhibition enhances the induction and impairs the maintenance of late-phase long-term potentiation. Learn. Mem. 15, 335–347 77 Dong, C. et al. (2014) Proteasome modulates positive and negative translational regulators in long-term synaptic plasticity. J. Neurosci. 34, 3171–3182 78 Banerjee, S. et al. (2009) A coordinated local translational control point at the synapse involving relief from silencing and MOV10 degradation. Neuron 64, 871–884 79 Greer, P.L. et al. (2010) The Angelman syndrome protein Ube3A regulates synapse development by ubiquitinating Arc. Cell 140, 704–716 80 Mabb, A.M. et al. (2014) Triad3A regulates synaptic strength by ubiquitination of Arc. Neuron 82, 1299–1316
9
