Nucleus reuniens mediates the extinction of contextual fear conditioning

Abstract

Learning and remembering the context in which events occur requires interactions between the hippocampus (HPC) and medial prefrontal cortex (mPFC). The nucleus reuniens (RE) is a ventral midline thalamic nucleus that coordinates activity in the mPFC and HPC and is involved in spatial and contextual memory. We recently found that the RE is critical for contextual fear conditioning in rats, a form of learning that involves interactions between the HPC and mPFC. Here we examined whether the RE mediates the extinction of contextual fear. After contextual fear conditioning, rats underwent an extinction procedure in which they were merely exposed to the conditioning context; freezing behavior during the extinction procedure and during a retrieval test 24 hours later served as an index of conditioned fear. Muscimol inactivation of the RE prior to extinction impaired the acquisition of both short- and long-term extinction memories. Similarly, inactivation of the RE prior to the extinction retrieval test also impaired the expression of extinction; this effect was not state-dependent. Taken together, these results reveal that the extinction of contextual fear memories requires the RE, which is consistent with a broader role for the RE in forms of learning that require HPC-mPFC interactions.

1. Introduction

Decades of research in both rats and humans indicate that the hippocampus (HPC) and medial prefrontal cortex (mPFC) play essential roles in the encoding and retrieval of episodic memories [1–4]. In recent years it has become apparent that reciprocal interactions between the HPC and mPFC are critical for these memory functions [4,5]. Anatomically, the HPC has robust projections to the medial prefrontal cortex (mPFC) that can support these functions, but there are no direct projections from the mPFC to the HPC [6]. Rather, recent work suggests that the mPFC projects to the HPC indirectly via the nucleus reuniens (RE), a ventral midline thalamic nucleus that synchronizes local field potentials in the HPC and mPFC and mediates forms of learning and memory that depend on HPC-mPFC interactions [5,7–15].

Consistent with a role for RE in hippocampal-dependent learning, we have recently shown that pharmacological inactivation of the RE impairs the acquisition of contextual, but not auditory, fear conditioning in rats [16]. Moreover, inactivation of the RE during retrieval testing impairs the specificity of contextual memory and increases the generalization of fear to novel contexts [16–19]. Interestingly, RE inactivation does not prevent the formation of context memory, but the memories formed under RE inactivation do not require the hippocampus [16]. This suggests that non-hippocampal systems acquire context memory under RE inactivation and that these memories lack the precision of a memory normally encoded by the hippocampus.

The HPC plays a role not only in contextual conditioning, but also in the context-dependent extinction of fear [20–24]. For example, we have recently shown that projections from the HPC to the mPFC mediate the context-dependent renewal of fear that occurs when an extinguished CS is presented outside the extinction context [25–28]. In contrast, either inactivating the RE or pharmacogenetically silencing mPFC projections in RE prevents both the encoding and retrieval of extinction memory while sparing renewal [29]. Together, this work suggests that the RE may be required for processing contextual information during extinction learning, information that is critical for generating context-appropriate defensive responses to a CS that has predicted both danger and safety.

Although our previous work implicates the RE in the contextual processes that support contextual fear conditioning on the one hand, as well as extinction of an auditory CS on the other, it is not known whether the extinction of context fear memories per se requires the RE. To address this question, we employed a contextual fear conditioning paradigm and temporarily inactivated RE using muscimol (MUS; GABA_A agonist) either prior to encoding or retrieval (or both) of contextual fear extinction. Consistent with our previous work [29], we found that inactivation of the RE impaired both the acquisition and expression of contextual fear extinction. Despite this impairment, animals extinguished after RE inactivation exhibited savings during a subsequent drug-free extinction session. Taken together, these results indicate that inactivation of RE is required for the inhibition of freezing behavior after the extinction of contextual fear.

2. Materials and methods

2.1. Subjects

Sixty-four experimentally naïve adult male Long-Evans rats (Blue-Spruce; 200–224 g) were obtained from a commercial supplier (Envigo, Indianapolis, IN). The rats were individually housed in cages within a temperature- and humidity-controlled vivarium and kept on a 14:10 h light: dark cycle (lights on at 07:00) with ad libitum access to food and water. All experiments took place during the light phase of the cycle. Rats were handled for 1 minute a day for 5 days to habituate them to the experimenter before any surgical procedures or behaviors were carried out. All experiments were conducted at Texas A&M University with approval from its Animal Care and Use Committee.

2.2. Surgical procedure

One week before the behavioral testing, rats were anesthetized with isoflurane (5% for induction, ~2% for maintenance), and placed into a stereotaxic instrument (Kopf Instruments). An incision was made in the scalp, the head was leveled, and bregma coordinates were identified. Small holes were drilled in the skull to affix three jeweler’s screws and to target RE using a cannula (8 mm, 26 gauge; Plastics One) above the RE. A single guide cannula was implanted at a 10° angle on the midline (A/P: −2.15 mm, M/L: −1.0 ~ −1.05 mm, D/V: −6.6 ~ −6.8 mm from dura; coordinates were measured from bregma). The cannula was affixed to the skull with dental cement, and a stainless-steel dummy cannula (30 gauge, 9 mm; Plastics One) was inserted into the guide cannula. Rats were allowed to recover for a 7-d period before behavioral testing during which time the dummy cannulae were replaced twice.

2.3. Behavioral apparatus and procedure

Eight identical rodent conditioning chambers (30 × 24 × 21 cm; Med-Associates, St Albans, VT) were used in all behavioral sessions. Each chamber consisted of two aluminum sidewalls, a Plexiglas ceiling and rear wall, and a hinged Plexiglas door. The floor consisted of 19 stainless steel rods that were wired to a shock source and a solid-state grid scrambler (Med-Associates) for the delivery of footshocks. Additionally, ventilation fans and house lights were provided ambient noise and light, respectively, and the chambers were cleaned with 1% ammonium hydroxide as part of the context. Each conditioning chamber rests on a load-cell platform that is used to record chamber displacement in response to each rat’s motor activity and is acquired online via Threshold Activity software (Med-Associates). For each chamber, load-cell voltages are digitized at 5 Hz, yielding one observation every 200 ms. Freezing was quantified by computing the number of observations for each rat that had a value less than the freezing threshold (load-cell activity = 10). Freezing was only scored if the rat was immobile for at least 1 sec. Rats were transported in white plastic boxes from the vivarium to the conditioning chambers.

2.4. Drug infusions

For RE microinfusions, rats were transported to an isolated room in the laboratory using white buckets (5-gallon) filled with a layer of bedding. Dummies were removed and stainless-steel injectors (33 gauge, 9 mm) connected to tubes was inserted into the guide cannulae for intracranial infusions. Polyethylene tubing connected the injectors to Hamilton syringes (10 μl), which were mounted in an infusion pump (Kd Scientific). Infusions were monitored by the movement of an air bubble that separated the drug or saline solutions from distilled water within the polyethylene tubing. All infusions were made approximately 10 min before the extinction and retrieval sessions. Muscimol (MUS) was diluted in sterile saline (SAL) to a concentration of 0.1 μg/μl. Infusions were made at a rate of 0.1 μl/min for 3 min (0.3 μl total, 0.03 μg muscimol) and the injectors were left in place for 2–3 min for diffusion. After the infusions, clean dummies were inserted into the guide cannula and the animals were transported to chambers for the behavioral sessions.

2.5. Behavioral procedure

This experiment was run in two separate cohorts with equal representation of groups in each replication (n = 8 per group in each replication). Sixty-four animals were randomly assigned to a 2X2 factorial design with variables of drug condition (SAL or MUS) during extinction and retrieval yielding sixteen subjects in each group. After recovery from surgery, animals were subjected to contextual fear conditioning in conditioning context where they were presented with 5 footshocks (US; 1 mA, 2 sec; 70-sec ISI) after a 3-minute baseline (total length of session: 8 min and 50 sec). Twenty-four hours later, rats received microinfusions of SAL or MUS into the RE and were placed back in the conditioning context for 35 mins to extinguish the contextual fear memory. On the next day, rats again received microinfusions of SAL or MUS into the RE and were placed in the conditioning context for another 35 min session to assess extinction retrieval (i.e., a second extinction session).

2.6. Histology

Rats were overdosed with sodium pentobarbital (Fatal Plus; 100 mg/ml, 0.75 ml) and were transcardially perfused with ice-cold saline and 10% formalin. Brains were then extracted and stored in 10% formalin for up to 24 h and then transferred to 30% sucrose-formalin solution at 4 °C for at least three days. Coronal brain sections of RE (40 μm) were made on a cryostat (−20 °C) and mounted on subbed microscope slides and stained with thionin staining (0.25% thionin) to visualize cannula placements.

2.7. Data analysis

All behavioral data (mean ± SEM) are represented by the average percentage of freezing behavior during 1-min intervals during the conditioning session and 5-min blocks during for the extinction and retrieval tests. All the data in Figures 2 & 4 were analyzed as two-way repeated measures ANOVAs with between-subject factors of drug condition during the extinction and extinction retrieval sessions and a within-subject variable of time. For Figure 3, the data was analyzed as a two-way ANOVAs with between-subject factors of drug condition during extinction and extinction retrieval. Post-hoc comparisons in the form of Fisher’s protected least significant difference tests were performed after a significant overall F-ratio in the ANOVA (p < 0.05 for both main effects and interactions). Statistical analyses were performed with StatView version 5.0.1 (SAS Institute) running under an open-source PowerPC.

Figure 2. Inactivation of RE impairs encoding and expression of contextual fear extinction.

(A) Percentage of freezing during the 3-min baseline period prior to the onset of first conditioning trial (BL) followed by a 70-sec ISI after each conditioning trial. (B) Percentage of freezing during the 35-min stimulus-free extinction session. Animals received infusions of SAL of MUS into the RE 10 mins prior to the extinction session. (C) Percentage of freezing during the 35-min stimulus-free extinction retrieval session. Animals received infusions of SAL of MUS into the RE 10 mins prior to the retrieval session yielding a factorial design with four groups: SAL-SAL, SAL-MUS, MUS-SAL and MUS-MUS. All data are means ± SEMs.

Figure 4. RE inactivation does not yield extinction savings.

Between-session comparison of freezing during a second extinction session in animals that received initial extinction training under either saline [SAL-Ext 1] or muscimol [MUS-Ext 1]; freezing of naïve rats in their first extinction session under SAL [SAL-Naive] are shown to determine whether there was savings in rats extinguished under MUS. All data are means ± SEMs.

Figure 3. Inactivation of RE impairs both encoding and expression of contextual fear extinction.

Average percentage of freezing during the first 15 min of extinction test. Animals received infusions of SAL or MUS into the RE 10 mins prior to both extinction and retrieval session yielding a 2X2 factorial design with four groups: SAL-SAL, SAL-MUS, MUS-SAL and MUS-MUS. All data are means ± SEMs.

3. Results

3.1. Histological analysis

A photomicrograph of a sample thionin stained section is shown in Figure 1A; cannula placements for all the animals included in the study is shown in Figure 1B. Of the 64 animals that started the experiment, four animals either died or had a broken injector and were unable to complete the study. Nineteen animals had cannula placements that missed the RE. This yielded the following group sizes: SAL-SAL (n = 12), SAL-MUS (n = 11), MUS-SAL (n = 9), MUS-MUS (n = 9).

Figure 1. Histology.

(A) A representative thionin-stained coronal section of midline thalamus showing cannula placement (arrow) in RE. (B) Cannula placements of all the subjects included in the analysis. The cannula placements were similar across the groups.

3.2. Inactivation of RE impairs encoding and retrieval of fear extinction

To explore the contribution of the RE to the acquisition and expression of context extinction memories, we reversibly inactivated RE with intracranial infusions of MUS either before the extinction session, a subsequent retrieval test, or both. This yielded a 2X2 factorial design that enabled us to determine whether the effects of RE inactivation on extinction are state-dependent, as has previously been observed [16].

During fear conditioning (Figure 2A), animals exhibited low levels of freezing during the 3-min baseline and subsequently showed an increase in freezing across 5 conditioning trials. There were no differences between the levels of fear acquisition across groups indicating that all groups have acquired fear conditioning to a similar extent. These results were confirmed by a two-way repeated measures ANOVA which revealed a significant main effect of conditioning trial [F(5,195) = 41.58, p < 0.0001], but no effects of either the prospective extinction [F(1,37) = 0.08, p = 0.77] or retrieval [F(1,37) = 0.549, p = 0.46] drug conditions or their interaction [F(1,37) = 0.07, p = 0.78].

Twenty-four hours after fear conditioning, animals received microinfusions of either SAL or MUS into the RE 10 min prior to a 35-min context extinction session. For clarity, we collapsed animals in the SAL-SAL and SAL-MUS groups (“SAL”) and animals in the MUS-SAL and MUS-MUS groups (“MUS”), because the drug assignment for the subsequent extinction retrieval test did not interact with the extinction effects. As shown in Figure 2B, SAL-infused rats showed high levels of freezing early in the extinction session that decreased across the session. In contrast, rats receiving intra-RE MUS infusions exhibited impaired freezing early in the session and an increase in freezing across the extinction session. These observations were confirmed by a two-way repeated measures ANOVA which revealed a significant main effect of drug [F(1,39) = 15.76, p = 0.0003] and a significant drug*time interaction [F(6,222) = 15.82, p < 0.0001]. This indicates that inactivation of the RE impaired within-session (short-term) extinction memories.

Twenty-four hours after the extinction session, animals again received microinfusions of SAL or MUS to RE 10 min prior to the 35-min extinction retrieval session to assess the strength of their extinguished fear memories. As shown in Figure 2C, animals that received SAL during extinction and SAL during the retrieval test (SAL-SAL) exhibited the lowest levels of freezing during the retrieval test. In contrast, animals that received MUS during the extinction session and SAL during the retrieval test (MUS-SAL) exhibited an impairment in extinction retrieval early in the test, but came to reduce their freezing across the retrieval test. RE inactivation prior to the retrieval test also impaired both extinction retrieval and extinction learning during the retrieval test (i.e., the second extinction session). Rats that underwent extinction after SAL infusions and received MUS prior to the test session (SAL-MUS) exhibited an increase in freezing across the test. The impairments in extinction encoding and retrieval were not due to shifts in drug state in the SAL-MUS and MUS-SAL groups, because matching drug state across the two sessions (MUS-MUS) did not restore extinction performance. In fact, MUS-MUS animals exhibited the largest impairments in extinction retrieval. These observations were confirmed by a two-way repeated measures ANOVA which revealed significant main effects of drug condition during extinction [F(1,37) = 6.65, p = 0.014] and drug condition during the retrieval test [F(1,37) = 17.05, p = 0.0002] on conditioned freezing during the test session; there was no interaction between the drug conditions across the extinction and retrieval sessions [F(1,37) = 0.067, p = 0.81]. In other words, RE inactivation impaired the formation of long-term extinction memory and inhibited extinction retrieval and extinction learning during the retrieval session (i.e., the second extinction session).

The nature of the extinction impairment produced by RE inactivation is more clearly shown in the early part of the extinction retrieval session before new extinction learning had occurred. Figure 3 shows the average freezing in each of the four groups during the first 15-min of the retrieval test. In this graph it is clear that RE inactivation prior to extinction, extinction retrieval, or both increased freezing relative to SAL-SAL controls. This observation was confirmed by a two-way factorial ANOVA that revealed a main effect for drug condition during extinction [F(1,37) = 6.15, p = 0.018] and drug condition during retrieval [F(1,37) = 4.99, p = 0.032] and no interactions between these conditions [F(1,37) = 2.01, p = 0.16]. Planned comparisons revealed that freezing in SAL-SAL group was significantly lower than MUS-MUS (p = 0.0018), MUS-SAL (p = 0.008), SAL-MUS (p = 0.001), which did not differ from each other (all p’s >0.47). Taken together, these data reveal that temporary inactivation of RE impairs both encoding and retrieval of contextual fear extinction. Furthermore, these deficits were not state-dependent.

3.3. Inactivation of RE during extinction does not result in savings of extinction

We determined whether animals extinguished under RE inactivation exhibited savings when extinguished once again during the retrieval test (the test constituted a second extinction session). To this end, we compared freezing during the retrieval test session (which constituted a second extinction session) for animals originally extinguished under saline (SAL-Ext 1) or muscimol (MUS-Ext 1) with freezing in the saline controls during the first extinction session [SAL-Naive]. As shown in Figure 4, animals extinguished under RE inactivation (MUS-Ext 1) exhibited a substantial extinction impairment relative to controls (SAL-Ext 1). Importantly, the rate of extinction in animals extinguished under MUS was similar to that of naïve rats (SAL-Ext1) suggesting an absence of savings in animals extinguished under RE inactivation. These observations were confirmed by a two-way repeated measures ANOVA, which revealed a significant main effect of group [F(2,41) = 4.25, p = 0.021] and group*time interaction [F(12,246) = 9.15, p < 0.0001]. This main effect of group was driven by differences between the SAL-Naive and SAL-Ext 1 groups (post-hoc comparisons p = 0.006). Furthermore, the group*time interaction reveals that the rate of extinction was different among the groups, and animals extinguished after intra-RE infusions (MUS-Ext 1) showed a slower rate of extinction than naïve (SAL-Naive) rats. To estimate the rate of extinction in animals undergoing extinction, we determined the block in which animals reached 50% of their initial freezing level during the second extinction session. There was no difference in the average time to half-max in the MUS-Ext 1 (11.33 ± 2.032 mins) and SAL-Naive (10.39 ± 1.13 mins) groups [F(1,30) = 0.184, p=0.67]. This confirms that RE inactivation during the initial extinction session does not result in behavioral savings expressed during the second extinction session.

4. Discussion

The present study reveals that RE inactivation impairs both acquisition and expression of contextual fear extinction. These results are consistent with recent reports indicating that the RE plays an important role both acquisition and expression of both contextual fear conditioning and auditory fear extinction [16,19,29,30]. Specifically, we have recently shown that RE inactivation prevents the acquisition and expression of extinction to an auditory CS [29]. The effects of RE inactivation are not due to performance effects, such as a nonspecific increase in freezing, because neither expression of freezing to the CS (prior to extinction) nor renewal of fear to the extinguished CS were affected by RE inactivation [29]. Furthermore, we have previously shown that MUS infusions are confined to the ventral midline nuclei, particularly RE and rhomboid nucleus, limiting the possibility that the spread of drug to adjacent thalamic nuclei accounts for these effects [29]. We now show that RE inactivation also impairs acquisition and expression of contextual fear extinction, an effect that was not due to a state-dependent generalization deficit in extinction.

Interestingly, in the current report we observed that RE inactivation produced a decrement in contextual freezing early in the context extinction session, suggesting that RE inactivation impaired retrieval of the context memory formed during conditioning. This finding contrasts with a recent report from our laboratory in which we found no effect of RE inactivation on freezing in the conditioning context [16]. The reason for this disparity is not clear, although animals in our previous report had prior experience with the infusion procedure before the context retrieval test, whereas the animals in the present experiment first experienced the infusion procedure before this test. It is possible that the novelty of the infusion procedure and RE inactivation produce some impairments in memory retrieval that are insufficient alone to yield a deficit in contextual freezing, but together produce a retrieval deficit. Alternatively, there may have been differences in the distribution of cannula placements within the RE across the studies that yielded different effects on context freezing.

The involvement of the RE in the encoding and retrieval of context memories [16], as well as in fear extinction [29] reveals a broad role for RE in learning and memory processes that involve interactions between the HPC and mPFC. For example, inactivating either the infralimbic region of the mPFC (a homologue of human subgenual anterior cingulate cortex [31]), the ventral HPC [24], or lesions of the dorsal HPC [23], produce robust impairments in extinction learning with an auditory CS. Several studies show that both the HPC and mPFC are similarly involved in encoding contextual fear extinction [20,32–34]. In addition, the RE has a critical role in coordinating information flow between the mPFC and HPC [35,36]. Hence, the effects of RE inactivation on extinction learning and retrieval may be mediated by a loss of coordinated neuronal activity in the HPC and mPFC. Consistent with this idea, inactivation of RE leads to impairments in tasks that require coordinated activity between mPFC and HPC [13,14,37], such as spatial working memory [3–5]. Moreover, lesions or functional inactivation of RE causes deficits in spatial maze tasks [7,11], acquisition and specificity of contextual fear [16–19] and context-dependent retrieval of extinguished fear memories [29]. Furthermore, recent literature from both anesthetized and freely behaving rodents shows that RE plays a critical role in mediating mPFC-HPC synchrony, which is critical for various memory related processes [11,38,39]. The present results supplant this literature by showing that RE is also critically involved in both acquisition and expression of contextual fear extinction.

The involvement of the RE extinction learning and retrieval suggests that it may have an important role in the inhibition of fear. When animals are presented with an extinguished CS or placed in a context that has undergone extinction, they actively suppress the original fear memory and consequently exhibit reduced freezing. Given the role for the RE in coordinating information flow between the HPC and mPFC [8], the RE may function to suppress retrieval of fear memories in the extinction context thereby reducing freezing behavior. Support for this hypothesis comes work in humans that shows that mPFC coordinates the suppression of memory retrieval by the hippocampus [40]. That is, when subjects are asked to actively suppress retrieval of a memory, it results in increased BOLD activity in dorsolateral prefrontal cortex that correlates with decreased HPC activity. Because the dorsolateral prefrontal cortex is connected to the HPC through RE, it was concluded that the RE mediates this top-down inhibition—a phenomenon termed retrieval-induced suppression [40–42]. This contrasts with the role for monosynaptic projections from the HPC to the mPFC, which are critical for the renewal of fear responses outside of the extinction context [25–28].

The role for the RE in regulation of Pavlovian fear extinction adds to a growing body of literature that the midline thalamic nuclei are critical hubs for regulating emotional learning and memory. For instance, it has previously been shown that the paraventricular thalamic nucleus and its projections to the central nucleus of the amygdala are involved in the expression and consolidation of conditional fear [43–46]. The mediodorsal nucleus of the thalamus also plays a role in both fear conditioning and extinction [47–50]. Beyond the midline nuclei, several other thalamic nuclei including the zona incerta and reticular thalamic nuclei have been implicated in associative memory processes [51–54]. Importantly, the current findings extend this and other work [16–19,29] and reveal a critical role for RE in mediating memory retrieval processes that underlie fear extinction.

Abbreviations:

HPC

Hippocampus

mPFC

medial prefrontal cortex

RE

Nucleus reuniens

MUS

Muscimol

SAL

Saline

CS

conditioned stimulus

US

Unconditioned stimulus