Hippocampal molecular mechanisms involved in the enhancement of fear extinction caused by exposure to novelty

Significance

Within a restricted time window, a brief exposure to a novel environment enhances the extinction of contextual fear. This can be explained by a hippocampal process of behaviorally induced synaptic tagging and capture. Here, we report that the effect requires glutamate NMDA receptors and L-voltage–dependent calcium channels and involves the activation of calcium/calmodulin-dependent protein kinase II, in addition to both ribosomal and nonribosomal protein synthesis. All these mechanisms operate only when the proteasomal-ubiquitin protein degradation system is intact, which suggests that they depend on synaptic protein turnover. Extinction enhancement by novelty is of great potential importance in the treatment of fear memories, such as those of posttraumatic stress disorder; the treatments of choice for such conditions are based on extinction procedures.

Abstract

Exposure to a novel environment enhances the extinction of contextual fear. This has been explained by tagging of the hippocampal synapses used in extinction, followed by capture of proteins from the synapses that process novelty. The effect is blocked by the inhibition of hippocampal protein synthesis following the novelty or the extinction. Here, we show that it can also be blocked by the postextinction or postnovelty intrahippocampal infusion of the NMDA receptor antagonist 2-amino-5-phosphono pentanoic acid; the inhibitor of calcium/calmodulin-dependent protein kinase II (CaMKII), autocamtide-2–related inhibitory peptide; or the blocker of L-voltage–dependent calcium channels (L-VDCCs), nifedipine. Inhibition of proteasomal protein degradation by β-lactacystin has no effect of its own on extinction or on the influence of novelty thereon but blocks the inhibitory effects of all the other substances except that of rapamycin on extinction, suggesting that their action depends on concomitant synaptic protein turnover. Thus, the tagging-and-capture mechanism through which novelty enhances fear extinction involves more molecular processes than hitherto thought: NMDA receptors, L-VDCCs, CaMKII, and synaptic protein turnover.

Frey and Morris (1, 2) and their collaborators (3–7) proposed a mechanism whereby relatively “weak” hippocampal long-term potentiation (LTP) or long-term depression (LTD) lasting only a few minutes can nevertheless “tag” the synapses involved with proteins synthesized ad hoc, so that other plasticity-related proteins (PRPs) produced at other sets of synapses by other LTPs or LTDs can be captured by the tagged synapses and strengthen their activity to “long” LTPs or LTDs lasting hours or days (8). LTDs and LTPs can “cross”-tag each other; that is, LTDs can enhance both LTDs and LTPs, and vice versa (6, 8). Because many learned behaviors rely on hippocampal LTP or LTD (7–9), among them the processing of novelty (9, 10) and the making of extinction (11–13), interactions between consecutive learnings can also be explained by the “tagging-and-capture” hypothesis (9, 10, 13), whose application to behavior became known as “behavioral tagging and capture” (5, 7, 9, 13). Typically, exposure to a novel environment [e.g., a nonanxiogenic 50 × 50 × 40-cm open field (OF) (5, 7, 9, 10, 14)] is interpolated before testing for another task, which becomes enhanced (4–10, 13). The usual reaction to novelty is orienting and exploration (14), followed by habituation of this response (14–16). Habituation is perhaps the simplest form of learning, and it consists of inhibition of the orienting/exploratory response (14, 16).

We recently showed that the brief exposure of rats to a novel environment (the OF) within a limited time window enhances the extinction of contextual fear conditioning (CFC) through a mechanism of synaptic tagging and capture (13), which is a previously unidentified example of behavioral tagging of inhibitory learning. Fear extinction is most probably due to LTD in the hippocampus (11, 12), although the possibility that it may also involve LTP is not discarded (13). The enhancement of extinction by novelty probably relies on the habituation to the novel environment, which is also probably due to LTD (15, 16). The enhancement of extinction by the exposure to novelty depends on hippocampal gene expression and ribosomal protein synthesis following extinction training and on both ribosomal and nonribosomal protein synthesis caused by the novel experience (13). Nonribosomal protein synthesis that can be blocked by rapamycin is believed to be dendritic (13, 17), so it would be strategically located for tagging-and-capture processes, but it has not been studied in synaptic tagging to date (3–8) or in other forms of behavioral tagging (7–10). As occurs with the interactions between LTPs and/or LTDs (4), the enhancement of extinction by novelty relies on hippocampal but not amygdalar processes (13).

Recent findings indicate that several hippocampal processes related to learning and memory, such as the reconsolidation of spatial learning, are highly dependent on NMDA glutamate receptors, calcium/calmodulin protein kinase II (CaMKII), and long-term voltage channel blockers (L-VDCCs), which, in turn, rely on the proteasomal degradation of proteins (18). Here, we study the effects of an NMDA blocker, 2-amino-5-phosphono pentanoic acid (AP5); the L-VDCC blocker nifedipine (Nife); a CaMKII inhibitor, the autocamtide-2–related inhibitory peptide (AIP); and the irreversible proteasome blocker β-lactacystin (12, 13) on the interaction between novelty and extinction (11). As will be seen, we found that both the setting up of tags by extinction and the presumable production of PRPs by the processing of novelty are dependent on NMDA receptors, CaMKII, and L-VDCCs. This endorses and expands the hypothesis that the novelty–extinction interaction relies on synaptic tagging and capture (13).

Results

Effect of AP5, AIP, and Nife Given into the Hippocampus After the OF or After the Extinction Training Session.

As in a previous paper (13), rats were trained in CFC using three 2-s, 0.5-mA scrambled foot shocks given every 30 s. Twenty-four hours later, they were placed for 5 min in an OF that they had never seen before, immediately after which they were infused intrahippocampally on both sides with vehicle (Veh) or with a drug dissolved in the Veh. Two hours later, they were placed again for 10 min in the CFC compartment and received no stimuli whatsoever (extinction training session). After another 24 h, the animals were submitted to a 3-min extinction test session, again in the CFC apparatus. The schematic drawings above Figs. 1A, 2A, 3A, and 4A illustrate the CFC training-OF-infusion-extinction schedule. The schematic drawings above Figs. 1D, 2D, 3D, and 4C illustrate the time of infusion of Veh or the drugs immediately after the extension training session to study their effect on the consolidation of extinction (13, 19, 20).

Fig. 1.

Effect of AP5, AIP, and Nife given into the hippocampus after the OF or after the extinction training session (Ext). In this and the following figures, the schematic drawings above A and D, respectively, represent the sequence of procedures used for studying the influence of exposure to an OF before the Ext on the extinction test session (Test) and the sequence of procedures used to study the effect of postextinction training treatments on the Test. Animals with infusion cannulae implanted in the CA1 region of the dorsal hippocampus were trained in the CFC task (three 2-s, 0.5-mA scramble foot shocks separated by 30-s intervals). After 24 h, the animals received intrahippocampal infusion of Veh, AP5 (25 nmol per side) (A), AIP (1 nmol per side) (B), or Nife (10 nmol per side) (C) immediately after exposure to an OF or immediately after an Ext (D). When given into the hippocampus AP5, AIP, and Nife blocked the enhancement of extinction caused by the OF. AP5 and Nife, but not AIP, blocked the consolidation of extinction. The figure shows the percentage of time spent freezing in the first 2 min of the CFC training session (Tr), in the first 3 min and last 3 min of the Ext, and in the Test. Data from Veh-treated animals are shown in white, and those from drug-treated animals are shown in gray; data are expressed as mean ± SEM (n = 11–12 animals per group). *P < 0.05; **P < 0.01; ***P < 0.001 vs. control groups in the retention test, Newman–Keuls test after one-way ANOVA.

Fig. 2.

Effect of intrahippocampal Ani, Rapa, or βLac after the OF or after the Ext. Animals with infusion cannulae implanted in the CA1 region of the dorsal hippocampus were trained in the CFC task (three 2-s, 0.5-mA scramble foot shocks separated by 30-s intervals). After 24 h, the animals received intrahippocampal infusion of Veh, Ani (300 nmol per side) (A), Rapa (20 pmol per side) (B), or βLac (200 pmol per side) (C) immediately after exposure to an OF or immediately after an Ext (D). When given into the hippocampus, Ani and Rapa blocked the Ext enhancement caused by the OF and the consolidation of extinction. βLac had no effect on both variables. The figure shows the percentage of time spent freezing in the first 2 min of the Tr, in the first 3 min and last 3 min of the Ext, and in the Test. Data from Veh-treated animals are shown in white, and those from drug-treated animals are shown in gray; data are expressed as mean ± SEM (n = 11–12 animals per group). **P < 0.01 and ***P < 0.001 vs. control groups in the retention test, Newman–Keuls test after one-way ANOVA. (A and D, Upper) Schematic representation of the behavioral protocol used.

Fig. 3.

AP5, AIP, and Nife inhibit both the effect of exposure to an OF on extinction and the consolidation of extinction only if the proteasome-ubiquitin system is unimpaired. Animals with infusion cannulae implanted in the CA1 region of dorsal hippocampus were trained in the CFC task (three 2-s, 0.5-mA scramble foot shocks separated by 30-s intervals). After 24 h, the animals received intrahippocampal infusion of Veh, AP5 plus βLac (25 nmol per side and 200 pmol per side, respectively) (A), AIP plus βLac (1 nmol per side and 200 pmol per side, respectively) (B), or Nife plus βLac (10 nmol per side and 200 pmol per side, respectively) (C) immediately after exposure to an OF or immediately after an Ext (D). When given into the hippocampus, AP5, AIP, or Nife lost the effect it had when given alone both on the enhancement caused by the OF and on its consolidation. The figure shows the percentage of time spent freezing in the first 2 min of the Tr, in the first 3 min and last 3 min of the Ext, and in the Test. Data from Veh-treated animals are shown in white, and those from drug-treated animals are shown in gray; data are expressed as mean ± SEM (n = 11–12 animals per group), Newman–Keuls test after one-way ANOVA. (A and D, Upper) Schematic representation of the behavioral protocol used.

Fig. 4.

Ani and Rapa also require integrity of the proteasome-ubiquitin system to interfere with extinction and the influence of the OF thereupon. Animals with infusion cannulae implanted in the CA1 region of the dorsal hippocampus were trained in the CFC task (three 2-s, 0.5-mA scramble foot shocks separated by 30-s intervals). After 24 h, the animals received intrahippocampal infusion of Veh, Ani plus βLac (300 nmol per side and 200 pmol per side, respectively) (A) or Rapa plus βLac (20 pmol per side and 200 pmol per side, respectively) (B) immediately after exposure to an OF or immediately after an Ext (C). When given into the hippocampus, Ani and Rapa lost the effect they had on their own on the enhancing effect of the OF, but only Ani lost the effect on the consolidation of extinction. The figure shows the percentage of time spent freezing in the first 2 min of the Tr, in the first 3 min and last 3 min of the Ext, and in the Test. Data from Veh-treated animals are shown in white, and those from drug-treated animals are shown in gray; data are expressed as mean ± SEM (n = 11–12 animals per group). **P < 0.01 vs. control groups in the retention test, Newman–Keuls test after one-way ANOVA. (A and C, Upper) Schematic representation of the behavioral protocol used.

In Fig. 1, the drugs studied were the antagonist of glutamate NMDA receptors, AP5 (25 nmol per side; Fig. 1A); the inhibitor of CaMKII, AIP (1 nmol per side; Fig. 1B); and the blocker of L-VDCCs, nifedipine (10 nmol per side; Fig. 1C). As can be seen, the three drugs, when given immediately after exposure to the OF, impaired extinction of the CFC task.

Fig. 1D shows that an impairment of the consolidation of extinction was seen in the test session when AP5 or Nife, but not AIP, was given right after the extinction training session.

Therefore, the three drugs were able, at the doses used, to block the enhancing effect of the exposure to the OF, which is presumably due to a tagging-and-capture process (Fig. 1 A–C), and AP5 and Nife, but not AIP, were capable of blocking memory consolidation of the extinction itself (Fig. 1D).

Effect of Intrahippocampal Anisomycin, Rapamycin, or β-Lactacystin After the OF or After the Extinction Training Session.

In Fig. 2, data are presented as in Fig. 1; the only difference is that, here, the drugs studied are the inhibitor of ribosomal protein synthesis, anisomycin (Ani, 300 nmol per side; Fig. 2A); the inhibitor of mammalian target of rapamycin (mTOR)-mediated protein synthesis, rapamycin (Rapa, 20 pmol per side; Fig. 2B); and the irreversible inhibitor of proteasome-ubiquitin–mediated protein catalysis, β-lactacystin (clastolactacystin β-lactone) (βLac, 200 pmol per side; Fig. 2C).

When given into the hippocampus immediately after exposure of the animals to the OF, both Ani (Fig. 2A) and Rapa (Fig. 2B) markedly inhibited extinction, in agreement with a previous finding (13), but βLac had no effect (Fig. 2C). When given after an extinction training session (Fig. 2D), both Ani and Rapa inhibited the expression of extinction in the test session but, again, βLac had no effect.

Thus, both ribosomal protein synthesis and nonribosomal protein synthesis are involved in the enhancing effect of OF on extinction learning, attributable to a tagging-and-capture process (13), and in the consolidation of extinction; inhibition of the proteasome-ubiquitin system by itself is ineffective on both variables.

AP5, AIP, and Nife Inhibit both the Effect of Exposure to an OF on Extinction and the Consolidation of Extinction only if the Proteasome-Ubiquitin System Is Unimpaired.

As shown in Fig. 1, in Fig. 3, AP5, AIP, and Nife inhibit the enhancing effect of exposure to the OF on extinction. Here, we show that this only happens if the proteasome-ubiquitin protein degradation system is not inhibited by βLac. As was shown in Fig. 2, βLac at a dose of 200 pmol per side has no effect of its own on the consolidation of extinction or on the influence of the OF upon it, but Fig. 3 shows that its intrahippocampal administration simultaneously with AP5 (Fig. 3A), AIP (Fig. 3B), or Nife (Fig. 3C) blocks the effect of the other three compounds. Fig. 3D shows that it also blocks the effect of AP5 or Nife on the consolidation of extinction, as reflected in the test in Fig. 1D.

Thus, the inhibitory action of NMDA receptor blockade, CaMKII inhibition, or L-VDCC blockade on consolidation of extinction and on the influence of tagging-and-capture processes thereupon requires ongoing protein turnover to become manifest.

Ani and Rapa also Require Integrity of the Proteasome-Ubiquitin System to Interfere with Extinction and on the Influence of the OF Thereupon.

Fig. 2 shows that Ani and Rapa, but not βLac, inhibit the enhancing effect of the OF on extinction and the consolidation of extinction itself. In Fig. 4, we show that the impairment of the influence of the OF on extinction by both Ani (Fig. 4A) and Rapa (Fig. 4B) is not seen in animals that receive intrahippocampal βLac (Fig. 4 A and B). As to the consolidation of extinction itself, βLac blocked the inhibitory effect of Ani in the test session performance but, seemingly, not that of Rapa (Fig. 4C), suggesting that the former is more sensitive to interference with the ongoing catabolism of proteins by the proteasome-ubiquitin system than the latter.

Discussion

The present findings point to several important molecular mechanisms underlying both the consolidation of extinction learning and the enhancement of fear extinction induced by an interpolated novelty. They are compatible with, and complement, those suggesting that this interaction is due to a tagging-and-capture process (13). That process applies to a variety of behaviors that rely on LTP or LTD (4–10). Based on our present result, the tags would be set by the purposefully incomplete (13) extinction of CFC and the PRPs would result from the processing of novelty by hippocampal synapses. The processing of novelty initially involves its detection, followed by habituation to it (14–16), two processes that are well known to involve the hippocampus (21, 22).

The attenuation by βLac of the effects of Ani, Rapa, AIP, AP5, and Nife suggests that those effects rely on proteasome-ubiquitin–mediated protein turnover at the synapses, which is specifically and irreversibly blocked by βLac (23–25). The lack of influence of βLac on the inhibition of the consolidation of extinction by Rapa suggests that this particular effect does not need ongoing protein turnover to the same extent as the others. Roles for the proteasome-ubiquitin system in consolidation and reconsolidation (18, 23–25), as well as in LTP (26) and in synaptic tagging (27), have been suggested. Recently, Rapa has been reported to inhibit proteasome activity allosterically (28), but the present findings suggest that this might not be a factor in its influence on the effects of novelty on extinction or in that on extinction itself. The former was actually blocked by βLac, and the latter was neither shared nor potentiated by the irreversible proteasome blocker.

The effects of Nife indicate that the processing of both tasks and the processing of their interaction via behavioral tagging are two more learning variables in which L-VDCCs play a necessary role (18); other such variables are discussed elsewhere (29–32).

The role of CaMKII-dependent processes in consolidation has been amply documented (32–35), including the consolidation of the extinction of inhibitory avoidance learning (34). For that reason, the lack of effect of AIP on CFC extinction was surprising (Fig. 1D). It might be attributable to an insufficient dose of the inhibitor [which is unlikely, considering that the intrahippocampal dose used here has strong effects in other tasks (18)] or to a peculiarity of this specific task. However, AIP clearly inhibited the novelty–extinction interaction dependent on behavioral tagging (Fig. 1B), which indicates a role for CaMKII in at least one of the two underlying NMDA receptor-dependent events studied. The proteasome-ubiquitin system is also involved in this role of CaMKII, as indicated by the effect of βLac on the inhibitory action of AIP on the tagging-and-capture–dependent event. A role for CaMKII in synaptic tagging has been demonstrated in LTP experiments (36, 37). The present findings endorse the idea that fear extinction and the processing of novelty rely on NMDA receptor-dependent plasticity processes in the hippocampus (4–7, 10, 13, 19). Convincing recent evidence suggests that extinction (11, 12) and habituation (16, 38) rely on LTD but leaves the door open for a possible additional role of LTP in both forms of learning (13, 15). Both “direct” and “cross”-tagging between LTP and LTD have been described (6, 8, 39). It is well documented that retrieval requires the participation of the hippocampus (40, 41); there is no certainty that the inhibition of retrieval, which is what both habituation and extinction consist of (13, 14, 40, 41), does require LTD instead of LTP (42). The hippocampus is not the only part of the brain in charge of extinction and habituation; in both forms of learning, it is part of complex circuits involved in their processing (43), and its LTD could be secondary to LTP elsewhere. In addition, it is not really known if novelty facilitates extinction because of the habituation that it usually entails (14–16) or merely because of its detection, which is also a function of the hippocampus (21, 22) and might also rely on hippocampal plasticity for all we know. Of the various types of LTD (that triggered by NMDA or metabotropic glutamate receptors or that triggered by muscarinic cholinergic receptors) (44, 45), the findings reported here on the participation of NMDA receptors both in the consolidation of extinction and in its enhancement by the OF (see above) support the possibility that NMDA-triggered LTD may be involved in extinction and linked to the participation of L-VDCCs, CaMKII, and protein turnover in both behavioral processes.

The data on the effect of exposure to the OF on extinction may be remindful of those described by LeDoux and coworkers (46, 47) on the effect of reconsolidation of the fear task on the reinterpretation of its fear content at the time of extinction. The latter are methodologically quite different, and therefore interpretable through other mechanisms (labilization of the fear trace). However, both sets of data, those on interpolation of a labilization/reconsolidation session and ours on the interpolation of an OF (13), can lead to behaviorally similar results through different mechanisms.

Materials and Methods

Mice.

Male Wistar rats (3–4 mo old, 300–350 g) were used, purchased from Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil, our regular provider. They were housed four to a cage with water and food (Purina laboratory pellets) ad libitum, under a 12-h light/dark cycle (lights on at 7:00 AM). The temperature of the animal room was maintained at 22–24 °C. All procedures were in accordance with the National Institutes of Health’s Guide for the Care and Use of Laboratory Animals (48) and were approved by Pontifícia Universidade Católica do Rio Grande do Sul’s Bioethics Committee.

Surgery.

At least 1 wk after their arrival, the animals were implanted under deep anesthesia (75 mg/kg of ketamine plus 10 mg/kg of xylazine) with bilateral 27-gauge guide cannulae 1 mm above the dorsal CA1 area of the hippocampus (anterior, −4.2 mm; lateral, ±3.0 mm; ventral, −1.8 mm) [coordinates according to Paxinos and Watson (49)]. Seven days after surgery, rats were submitted to daily handling for 3 consecutive days and then trained in the CFC procedure between 8:00 and 11:00 AM (see below).

CFC.

The training apparatus was a 35 × 35 × 35-cm aluminum box with a floor made of parallel caliber bronze bars spaced 0.8 mm apart. This training box was within another, larger box made of soundproof walls so as to attenuate external sounds. The percentage of time spent freezing in the apparatus was measured automatically by a counter connected to photocells (Panlab). On the day of training, the animals were left to explore the apparatus freely for 2 min, after which they received three 2-s, 0.5-mA scrambled foot shocks with a 30-s interval between them. Animals were left in the conditioning chamber for another 30 s and then placed back into their home cages. Basal freezing behavior was registered prior to the administration of the shocks. Twenty-four hours later, the animals were exposed to the same apparatus without any stimulation for 10 min (extinction training session). Animals were submitted to a retention test 24 h later. In all sessions, the percentage of time spent freezing (i.e., with no movement) was measured (13, 19, 20).

Exposure to an OF.

The OF was a 50 × 50 × 40-cm wooden box painted white, whose front wall was made of glass. The animals were exposed to it for 5 min 2 h before the extinction training session. Because the animals had never seen the apparatus before, this constituted an exposure to a novel environment (13).

Drug Treatments.

One minute or less after the end of the exposure to the OF or the extinction training session, the 30-gauge infusion cannula was tightly fitted into the guides so that its tip protruded 1 mm beyond that of the guide, first on one side and then on the other (13, 19, 20), and drug or Veh infusions were carried out at a rate of 0.5 μL over 30 s. The infusion cannula was left in place for an additional 60 s to minimize backflow. It was then carefully withdrawn and placed on the other side, where the procedure was repeated. The entire bilateral infusion procedure took about 90 s. The drugs used, at the doses stated in each case, were the inhibitor of ribosomal translation, Ani (300 nmol per side; Sigma) (13); the extraribosomal protein synthesis inhibitor, Rapa (20 pmol per side; Sigma) (13, 17, 33); the L-VDCC blocker, Nife (10 nmol per side; Sigma) (18); the proteasome-ubiquitin blocker, βLac (200 pmol per side; Sigma) (18, 23, 24); the CaMKII inhibitor, AIP (1 nmol per side; Sigma); and the glutamate NMDA receptor antagonist, AP5 (25 nmol per side; Sigma). The doses were chosen among those found to be effective in previous papers from our group or others (13, 17). Rapa acts on an extraribosomal, presumably dendritic protein synthesis system that uses preexistent mRNA molecules, which includes as a central component the protein kinase mTOR (17). The doses used of all these drugs have effects on consolidation, reconsolidation, or extinction when given into the hippocampus in rats but have no effects on locomotion, exploration, or performance in an elevated plus maze (17, 19). The volume of the drugs infused was 1 μL per side into the dorsal CA1 area of the hippocampus.

Correct Cannula Placements.

Correct cannula placement was verified by infusing a 4% (wt/vol) methylene blue solution over 30 s into the CA1 region of the dorsal hippocampus (1 μL per side) at the coordinates mentioned above at 2 d after the last behavioral procedure. The spread of the dye was taken as an estimate of that of the drug infusions in the same animals. Placements were considered correct when the spread was 1 mm³ or less (13, 19) from the intended infusion sites; this occurred in 98% of the animals. As explained elsewhere (17), despite the uncertainties given by the unknown rate of solubility of the drugs used relative to methylene blue, this is an improvement over the mere determination of the cannula tip location (Fig. S1).

Statistics.

For data analysis, one-way ANOVA, followed by Newman–Keuls tests, was applied to the CFC data. Differences between groups were considered significant if P < 0.05. Data were analyzed using GraphPad Prism software.