The role of carbonic anhydrases in extinction of contextual fear memory

Significance

Disruption of memories of stressful events may trigger maladaptive responses that are distinctive features of obsessive-compulsive disorders, phobias, posttraumatic stress disorder (PTSD), and generalized anxiety. Drugs available to alleviate these disorders have minimal or inconsistent efficacy. The gold standard treatment, exposure therapy, is based on extinction, which refers to a new memory trace inhibiting original memory retrieval. As not all patients experience beneficial effects of this therapy, drugs potentiating extinction are of utmost importance for improving treatment. Carbonic anhydrases (CAs) are involved in several physiological processes, including memory. The present study indicates that their activation potentiates extinction. Thus, CA activators may have the potential to improve the clinical efficacy of exposure-based treatments of obsessive-compulsive disorders, phobias, generalized anxiety, and PTSD.

Abstract

Carbonic anhydrases (CAs; EC 4.2.1.1) are metalloenzymes present in mammals with 16 isoforms that differ in terms of catalytic activity as well as cellular and tissue distribution. CAs catalyze the conversion of CO₂ to bicarbonate and protons and are involved in various physiological processes, including learning and memory. Here we report that the integrity of CA activity in the brain is necessary for the consolidation of fear extinction memory. We found that systemic administration of acetazolamide, a CA inhibitor, immediately after the extinction session dose-dependently impaired the consolidation of fear extinction memory of rats trained in contextual fear conditioning. d-phenylalanine, a CA activator, displayed an opposite action, whereas C18, a membrane-impermeable CA inhibitor that is unable to reach the brain tissue, had no effect. Simultaneous administration of acetazolamide fully prevented the procognitive effects of d-phenylalanine. Whereas d-phenylalanine potentiated extinction, acetazolamide impaired extinction also when infused locally into the ventromedial prefrontal cortex, basolateral amygdala, or hippocampal CA1 region. No effects were observed when acetazolamide or d-phenylalanine was infused locally into the substantia nigra pars compacta. Moreover, systemic administration of acetazolamide immediately after the extinction training session modulated c-Fos expression on a retention test in the ventromedial prefrontal cortex of rats trained in contextual fear conditioning. These findings reveal that the engagement of CAs in some brain regions is essential for providing the brain with the resilience necessary to ensure the consolidation of extinction of emotionally salient events.

Emotional experiences leave long-lasting traces in the brain, as memories that are formed following stressful events may be initially labile but over time may become insensitive to disruption and may last a lifetime through a process known as consolidation (1–3). This memory persistence is pivotal for individuals to respond adequately to danger (4, 5). Disruption of these memories may trigger maladaptive responses in such mental illnesses as obsessive-compulsive disorders, phobias, posttraumatic stress disorder (PTSD), and generalized anxiety (6, 7). Insight leading to improved treatments of these disorders can be gained by better understanding the neural mechanisms underlying emotional memory. Memory is a multistate process that includes acquisition, consolidation, and retrieval (8). Retrieval reactivates a mnemonic trace, returning it to a labile state and consequently initiating its reconsolidation or extinction. Reconsolidation allows for the updating of the original memory, rendering it persistent (9–11). Conversely, extinction refers to a new memory trace that inhibits retrieval of the original memory (12, 13).

Fear-motivated learning tasks have greatly contributed to the knowledge of extinction memory and are considered a valuable translational model for investigating disorders such as phobias, anxiety, and PTSD (5, 12, 14). Extinction forms the basis of exposure therapy (15–17), currently the gold standard treatment for these disorders (18, 19). However, not all patients experience the beneficial effects of this therapy. Since extinction does not erase the original memory but is a new learning that inhibits its expression (20), extinguished behaviors can recover spontaneously with the passage of time (21, 22). Drugs promoting fear extinction could represent a novel therapeutic strategy to treat these disorders (19, 23–25).

Carbonic anhydrases (CAs; EC 4.2.1.1) are enzymes present in mammals with 16 isoforms that differ in terms of catalytic activity as well as cellular and tissue distribution (26). CAs catalyze the conversion of CO₂ to bicarbonate and protons and are involved in a number of physiological processes (27), including memory formation (28). CA activation was found to ameliorate spatial memory in rats (29). In keeping with these findings, mice genetically deficient in the CA IX isoform performed more poorly in the same task than wild-type littermates (30). More recently, it was reported that administration of the widely used CA inhibitor acetazolamide (ACTZ) (31) to CD1 mice reduced CA activity in the brain and caused amnesia in the object recognition (OR) test, whereas treatment with d-phenylalanine (d-phen) enhanced CA activity and potentiated OR memory as a result of extracellular signal-regulated kinase (ERK) activation (31, 32). In line with these results, inhibition of CAs also impaired fear memory consolidation in rats through inhibition of ERK phosphorylation (33). Thus, it is conceivable that CAs are involved in fear extinction memory as well.

The present study was specifically designed to address this question in rats using contextual fear conditioning (CFC), an associative learning paradigm that has greatly contributed to the understanding of extinction processes (34). CA activity was modulated pharmacologically with inhibitors and activators of the enzymes that were administered systemically or locally in brain areas involved in extinction memory. c-Fos expression was also evaluated in these brain regions after the retention test. The results clearly indicate that CA modulation in selected brain regions affects the consolidation of fear extinction memory. This may represent a previously unexplored mechanism to develop drugs for improved treatment of mental illnesses.

Results

Effect of Systemic Administration of CA Inhibitors on Fear Extinction Memory.

The experimental design is shown in Fig. 1A. At 24 h after CFC training, animals were exposed to a 30-min extinction training session, immediately followed by i.p. administration of ACTZ (10 mg/kg or 30 mg/kg) or 1-N-(4-sulfamoylphenyl-ethyl)-2,4,6-trimethylpyridinium perchlorate (C18; 30 mg/kg). Controls received comparable injections of vehicle. Since extinction is essentially a new learning, treatment applied after the extinction training session will affect consolidation but not acquisition (35). A 3-min retention test was delivered at 24 h after the extinction session; the results are displayed in Fig. 1B. We observed that freezing was extinguished over the course of the extinction session, as indicated by a decrease in freezing time during the final 3 min (27-30) compared with the initial 3 min (0-3) (Fig. 1B). This is a clear indication that all rats learned the extinction of CFC. Moreover, extinction memory lasted at least 24 h, since rats given vehicle spent significantly less time freezing in the retention test than in the initial 3-min interval (0-3) of the extinction session (Fig. 1B). However, freezing behavior displayed at the retention test by rats injected with 30 mg/kg of ACTZ did not differ significantly from that expressed in the initial 3 min (0-3) of the extinction session (Fig. 1B), revealing that ACTZ at this dosage impaired the consolidation of CFC extinction. Accordingly, rats injected with 30 mg/kg of ACTZ spent significantly more time freezing than all other group of rats in the retention test (Fig. 1B).

Fig. 1.

Brain CA inhibition impaired the consolidation of extinction memory. (A) Schematic drawing showing the sequence of behavior procedures and treatments. Rats were trained in the CFC and after 24 h subjected to a 30-min extinction training session. Immediately after this session, they received an i.p. injection of vehicle, ACTZ (10 mg/kg), ACTZ (30 mg/kg), or C18 (30 mg/kg). A 3-min retention test was performed at 24 h after the extinction training session. (B) Freezing time shown by rats given vehicle, ACTZ (10 mg/kg), ACTZ (30 mg/kg), or C18 (30 mg/kg) during training, extinction, and the retention test. Data are expressed as mean ± SEM of 6 to 10 rats for each treated group. ^####P < 0.0001, rat groups during training vs. corresponding groups in the 0 to 3-min interval of the extinction session; ***P < 0.001, ****P < 0.0001, rat groups in the 0 to 3-min interval of the extinction session vs. corresponding groups in the 27 to 30-min interval of the extinction session and in the retention test; °°°P < 0.001, °°°°P < 0.0001, ACTZ 30 mg/kg vs. vehicle, ACTZ 10 mg/kg, and C18 30 mg/kg in the retention test. Statistical details are provided in SI Appendix, Table S1.

Freezing behavior was not affected by changes in locomotion, as no significant difference in the general motor activity was detected comparing vehicle-treated animals with those given 30 mg/kg of ACTZ or C18 (SI Appendix, Fig. S2), in agreement with earlier reports from our and other laboratories (32, 33). ACTZ at a lower dose (10 mg/kg) or C18 did not impair the consolidation of CFC extinction, as rats administered ACTZ 10 mg/kg or C18 spent significantly less time freezing in the retention test than during the first 3 min (0-3) of the extinction session (Fig. 1B). C18 is a CA inhibitor that, unlike ACTZ, does not cross the blood-brain barrier (36). Therefore, inhibition of CA peripheral activity elicited by C18 left the consolidation of CFC extinction intact.

d-Phen Potentiated CFC Extinction.

We tested the effects of CA activation on CFC extinction by administering d-phen, an activator of several CA isoforms (37). d-phen (300 mg/kg i.p.) did not influence locomotion (SI Appendix, Fig. S2). The experimental design is shown in Fig. 2A. At 24 h after CFC training, animals were exposed to a 15-min extinction training session, followed immediately by i.p. administration of d-phen (300 mg/kg) with or without ACTZ (30 mg/kg). Controls received comparable injections of vehicle.

Fig. 2.

CA activation enhanced the consolidation of extinction memory. (A) Schematic drawing showing the sequence of behavior procedures and treatments. Rats were trained in the CFC and after 24 h subjected to a 15-min extinction training session, then immediately given an i.p. injection of vehicle, d-phen (300 mg/kg), or d-phen (300 mg/kg) + ACTZ (30 mg/kg). A 3-min retention test was performed at 24 h after the extinction training session. (B) Rats treated with d-phen showed fear extinction memory, as they spent significantly less time freezing than vehicle-treated rats in the retention test. ACTZ prevented the promnesic effect of d-phen. Data are expressed as mean ± SEM of eight or nine rats for each treated group. ^####P < 0.0001, rat groups during training vs. corresponding groups in the 0 to 3-min interval of the extinction session; ***P < 0.001, d-phen 300 mg/kg vs. corresponding groups in the 0 to 3- and 12 to 15-min intervals of the extinction session; °P < 0.05, °°P < 0.01, d-phen 300 mg/kg vs. vehicle and d-phen 300 mg/kg + ACTZ 30 mg/kg in the retention test. Statistical details are provided in SI Appendix, Table S1.

To evaluate a potentiating effect, a weaker extinction protocol was induced by shortening the extinction session to 15 min, according to previous studies (38). Results are shown in Fig. 2B. Vehicle-treated rats showed no significant differences in freezing duration in the initial (0-3) and final (12-15) 3 min of the extinction session or in the retention test (Fig. 2B). Conversely, rats treated with d-phen spent significantly less time freezing than vehicle-treated rats in the retention test (Fig. 2B). They also demonstrated significantly less freezing behavior in the retention test than during the initial 3 min (0-3) or the final 3 min (12-15) of the extinction training session (Fig. 2B). Thus, administration of d-phen at 300 mg/kg enhanced extinction memory, as a short extinction session that normally does not produce long-term extinction memory was effective.

To assess whether the improved learning in d-phen–treated rats may depend on CAs activation, we tested the effects of ACTZ (30 mg/kg i.p.) on the procognitive effects elicited by 300 mg/kg d-phen (Fig. 2B). Rats of this group received both d-phen and ACTZ through separate i.p. injections at the end of the extinction training session (Fig. 2A). These rats did not differ significantly in freezing time in the retention test or in the initial (0-3) and final (12-15) 3 min of the extinction session. Moreover, freezing time at retention was not significantly different from that of vehicle-treated animals (Fig. 2B). d-phen/ACTZ-treated rats spent significantly more time freezing than rats given d-phen alone (Fig. 2B). Therefore, inhibition of CAs abolished the extinction memory enhancement elicited by d-phen.

Effect of ACTZ Local Administration into Selected Brain Regions on CFC Extinction.

We next asked whether local infusions of ACTZ into the ventromedial prefrontal cortex (vmPFC), the basolateral amygdala (BLA), the CA1 region of the dorsal hippocampus (CA1), or the substantia nigra pars compacta (SNpc) affected the extinction of CFC. All animals, distributed into four experimental subsets according to the brain region investigated, were subjected to a CFC training session and, 24 h later, a 30-min extinction training session, immediately followed by bilateral infusions of ACTZ (10 nmol/side) or vehicle into separate brain areas (Fig. 3A). A 3-min retention test was performed at 24 h after the extinction session (Fig. 3A). The effects of drug infusion into the vmPFC are shown in Fig. 3B.

Fig. 3.

CA inhibition in selected brain regions hindered the consolidation of extinction memory. (A) Schematic drawing showing the sequence of behavior procedures and treatments. Rats were trained in the CFC and after 24 h subjected to a 30-min extinction training session. Immediately after this session, the rats received bilateral infusions of vehicle or ACTZ (10 nmol/side) into the vmPFC (B), BLA (C), CA1 (D), or SNpc (E). A 3-min retention test was delivered at 24 h after the extinction training session. The freezing time observed in the retention test indicates that bilateral infusions of ACTZ into the vmPFC (B), BLA (C), and CA1 (D), but not in the SNpc (E), impaired the consolidation of fear extinction memory. Data are expressed as mean ± SEM of 8 to 13 rats for each treated group. For all regions, ^####P < 0.0001, rat groups during training vs. corresponding groups in the 0 to 3-min interval of the extinction session; ****P < 0.0001, rat groups in the 27 to 30-min interval of the extinction session and in the retention test vs. corresponding groups in the 0 to 3-min interval of the extinction session; °°°°P < 0.0001 vehicle, vs. ACTZ 10 nmol/side in the retention test. Statistical details are provided in SI Appendix, Table S1.

Freezing extinguished over the course of the extinction session in both groups of animals, as its duration in the final 3 min (27-30) was significantly shorter than that in the initial 3 min (0-3) (Fig. 3B). Extinction memory lasted at least 24 h, since in the retention test, vehicle-treated rats displayed significantly reduced freezing behavior compared with that observed during the first 3 min (0-3) of the extinction session (Fig. 3B). However, the freezing time of ACTZ-infused rats in the retention test was significantly longer than that during the final 3 min (27-30) of the extinction training session (Fig. 3B). Furthermore, in the retention test, ACTZ-treated rats froze for a significantly longer period than vehicle-treated rats (Fig. 3B).

The effects of ACTZ infusion into the BLA are shown in Fig. 3C. All animals learned extinction, since freezing duration in the final 3 min (27-30) was significantly shorter than that in the initial 3 min (0-3) of the extinction session (Fig. 3C). Extinction memory was long-lasting, as vehicle-treated rats exhibited significantly reduced freezing behavior in the retention test than during the first 3 min (0-3) of the extinction training session (Fig. 3C). Conversely, rats that received ACTZ infusion into the BLA exhibited a significantly longer freezing time in the retention test than during the final 3 min (27-30) of the extinction training session (Fig. 3C). Moreover, freezing behavior significantly longer in these rates compared with vehicle-treated rats in the retention test.

Fig. 3D shows the effects of ACTZ infusion into the hippocampal CA1 region. Freezing was extinguished in both groups of animals over the course of the extinction session, with a significantly decreased duration in the final 3 min (27-30) compared with the initial 3 min (0-3) (Fig. 3D). In the retention test, vehicle-treated rats exhibited a significantly shorter freezing time compared with that observed during the initial 3 min (0-3) of the extinction training session; however, ACTZ-infused rats had a significantly longer freezing time in the retention test than in the final 3 min (27-30) of the extinction session (Fig. 3D). In addition, the ACTZ-infused rats exhibited a significantly longer freezing time than the vehicle-treated rats in the retention test (Fig. 3D).

The effects of ACTZ infusion into the SNpc are shown in Fig. 3E. Freezing was extinguished over the course of the extinction training session in both experimental groups, with a significant drop in duration in the final (27-30) 3-min interval compared with the initial (0-3) 3-min interval (Fig. 3E). In the retention test, the freezing time did not differ between the vehicle- and ACTZ-treated rats, but it was significantly shorter than that seen during the initial 3 min (0-3) of the extinction training session (Fig. 3E). Taken together, these findings suggest that inactivation of CAs in the vmPFC, BLA, and CA1, but not in the SNpc, impaired the consolidation of CFC extinction.

Effect of Local d-Phen Administration into Selected Brain Regions on CFC Extinction.

To further support the observation that CA inactivation by ACTZ administration in the vmPFC, BLA, and CA1, but not in the SNpc, weakened consolidation of CFC extinction, we tested whether CA activation in the same regions led to the opposite results. The experimental design is shown in Fig. 4A. All animals, distributed into four experimental subsets according to the brain region investigated, were subjected to a CFC training and, 24 h later, a 15-min extinction training session, immediately followed by bilateral infusions of d-phen (50 nmol/side) or vehicle into separate brain regions. To evaluate a potentiating effect, the extinction session was shortened to 15 min to induce weaker extinction, according to the present study and previous studies (38). The effects of drug infusion into the vmPFC are shown in Fig. 4B.

Fig. 4.

CA activation in selected brain regions potentiated the consolidation of extinction memory. (A) Schematic drawing showing the sequence of behavior procedures and treatments. Rats were trained in the CFC and after 24 h subjected to a 15-min extinction training session. Immediately after this session, the rats received bilateral infusions of vehicle or d-phen (50 nmol/side) into the vmPFC (B), BLA (C), CA1 (D), or SNpc (E). A 3-min retention test was performed at 24 h after the extinction training session. (B–E) All rats learned fear extinction memory, as freezing decreased during the extinction session, but this memory was short-lived, as the freezing time of vehicle-treated rats did not differ significantly in the initial 3 min (0-3) of the extinction session and in the retention test. Bilateral infusions of d-phen into the vmPFC (B), BLA (C), and CA1 (D), but not in the SNpc (E), potentiated the consolidation of fear extinction memory. Data are expressed as mean ± SEM of 8 to 13 rats for each treated group. For all regions, ^####P < 0.0001, rat groups during training vs. corresponding groups in the 0 to 3-min interval of the extinction session; *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, rat groups in the 27 to 30-min interval of the extinction session and in the retention test vs. corresponding groups in the 0 to 3-min interval of the extinction session; °°P < 0.01, °°°°P < 0.0001, vehicle vs. ACTZ 10 nmol/side in the retention test. Statistical details are provided in SI Appendix, Table S1.

Comparison of the initial 3-min (0-3) with the final 3-min (12-15) of freezing revealed that freezing extinguished over the course of a 15-min extinction session in both vehicle- and d-phen–treated rats; however, the effect was short-lived, as freezing time of vehicle-treated rats did not differ significantly between the initial 3 min (0-3) of the extinction session and in the retention test (Fig. 4B). In the retention test, rats given d-phen spent significantly less time freezing compared with vehicle-treated rats (Fig. 4B). In addition, the freezing time of d-phen–treated animals was significantly shorter in the retention test than during the initial 3 min (0-3) of the extinction training session (Fig. 4B). The effects of d-phen infusion into the BLA are displayed in Fig. 4C. Freezing extinguished over the course of the extinction session, as its duration in the final 3 min (12-15) was significantly shorter than that of the initial 3-min (0-3) in both vehicle- and d-phen–treated rats (Fig. 4C). However, in the retention test, the freezing behavior of vehicle-treated rats was not significantly different from that exhibited during the first 3 min (0-3) of the extinction session, indicating that extinction learning was short-lived (Fig. 4C). Conversely, rats that received d-phen exhibited a significantly shorter freezing time significantly in the retention test than during the initial 3 min (0-3) of the extinction session (Fig. 4C). Moreover, in the retention test, the d-phen–treated rats spent significantly less time freezing compared with the vehicle-treated rats (Fig. 4C). Fig. 4D shows the effects of d-phen infusion into the hippocampal CA1 region.

Extinction occurred in all rats, since the duration of freezing during the extinction session decreased significantly in the final 3 min (12-15) compared with the initial 3 min (0-3) (Fig. 4D), but this was not long-lasting, as the freezing time of vehicle-treated rats was not significantly different between the initial 3 min (0-3) of the extinction session and in the retention test (Fig. 4D). The d-phen–treated rats exhibited a significantly shorter freezing time in the retention test than during the initial 3 min (0-3) of the extinction session (Fig. 4D). Freezing time in the retention test was significantly shorter in the d-phen–infused rats compared with the vehicle-treated rats (Fig. 4D).

The effects of d-phen infusion into the SNpc are shown in Fig. 4E. Freezing was extinguished over the course of the extinction training session, with a significantly decreased duration in the final 3 min (12-15) compared with the initial 3 min (0-3). The freezing times of the vehicle- and d-phen–treated rats did not differ significantly in the retention test, and did not differ significantly from those observed during the initial 3 min (0-3) of the extinction session (Fig. 4E). Taken together, these findings suggest that the activation of CAs in the vmPFC, BLA, and CA1, but not in the SNpc, potentiated the consolidation of CFC extinction.

ACTZ Systemic Administration Modulated c-Fos Expression in the Ventromedial Prefrontal Cortex.

Rats were trained in CFC and 24 h later exposed to a 30-min extinction training session, followed immediately by i.p. administration of ACTZ (30 mg/kg) or vehicle. Rats were euthanized at 90 min after the retention test (Fig. 5A). In line with the results shown in Fig. 1B, all rats learned CFC extinction, which lasted at least 24 h (SI Appendix, Fig. S1); however, in the rats injected with 30 mg/kg of ACTZ, freezing behavior exhibited in the retention test did not differ significantly from that expressed in the initial 3 min (0-3) of the extinction training session (SI Appendix, Fig. S1), confirming that ACTZ at this dosage impaired the consolidation of CFC extinction. We found that c-Fos expression in the vmPFC was significantly higher in rats given ACTZ compared with vehicle-treated rats (Fig. 5C), whereas no differences in c-Fos expression were found in the BLA, CA1, or SNpc of rats treated with ACTZ and those treated with vehicle (SI Appendix, Fig. S3).

Fig. 5.

Effect of systemic administration of ACTZ immediately after the extinction training session on c-Fos expression in the vmPFC. (A) Schematic drawing showing the sequence of behavior procedures and treatments. Rats were trained in the CFC and after 24 h subjected to a 30-min extinction training session. Immediately after this session, the rats received i.p. injections of vehicle (Veh) or ACTZ (30 mg/kg). A 3-min retention test was performed at 24 h after the extinction training session. Rats were euthanized at 90 min after the retention test. (B) Representative photomicrographs showing the effect of vehicle or ACTZ on c-Fos protein expression in the vmPFC. (C) Quantitative analysis of data shown in B. Data are expressed as mean ± SEM of six rats for each treated group. **P < 0.01, ACTZ vs. Veh. Statistical details are provided in SI Appendix, Table S1.

Discussion

Spontaneous recovery of fear memories may occur at any time and triggers much distress for people affected by fear-based disorders such as PTSD, generalized anxiety, and phobias. The recommended treatment for these disorders is exposure therapy (39), during which extinction memories override the original fear memory. However, exposure therapy is not always effective, and the original fear often relapses spontaneously after extinction, suggesting that extinction forms a new memory that inhibits or competes with the original fear but does not erase it (34, 40). Thus, novel interventions that may augment extinction and inhibitory learning are needed. In addition to behavioral methods implemented during psychotherapy and device-based stimulation techniques that enhance or reduce activity in different brain regions, there is also increasing support for novel drugs that may augment extinction and inhibitory learning, specifically when combined with exposure-based psychotherapy (41–43). Further investigation of extinction processes is needed to identify targets for these novel drugs.

Here we examined the role of CAs in fear extinction memory in rats by assessing the systemic effects of two CAs inhibitors, ACTZ (31) and C18 (37), and of a CA activator, d-phen (38). We found that CAs in the brain are implicated in fear extinction mechanisms. Administration of ACTZ (30 mg/kg i.p.) but not of C18, a compound that does not penetrate the brain, immediately after the extinction session clearly impaired consolidation of fear extinction memory. It is widely accepted that extinction of a behavioral response requires new inhibitory learning (5, 34, 44). Our results are in agreement with earlier evidence suggesting that brain CA activity is necessary for new memory formation, as inactivation of brain CAs impairs the formation of spatial and fear memory (28, 33). In line with these findings, mice genetically deficient of the CA IX isoform exposed to the Morris water maze test had more difficulty than their wild-type littermates in learning to find the hidden platform (30). More recently, our research group reported that systemic administration of ACTZ to male CD1 mice caused amnesia in the object recognition test (32). A detrimental effect on cognition by ACTZ has been reported in humans as well, as a randomized, double-blind, placebo-controlled study showed that this CA inhibitor impaired cognitive performance, executive function, short-term memory, and sustained attention during acute high-altitude exposure (45).

We previously reported that systemic administration of ACTZ (30 mg/kg i.p) significantly decreased CA activity in the mouse brain (32). In the present study, ACTZ caused amnesia at the same dose used in our previous work; therefore, it is conceivable that extinction memory impairment occurred as a consequence of CA inhibition. Thus, increasing CA activity would be expected to improve extinction memory. Our results support this idea; administration of d-phen, a CA activator, produced long-lasting fear extinction memory when given immediately after the extinction session. Accordingly, administration of d-phen to CD1 mice enhanced memory in the OR test and increased brain CA activity (32). Coadministration of ACTZ fully blocked d-phen–elicited memory improvement in both the present study and the previous study (32), clearly indicating a critical role of CA activation as an underlying mechanism and ruling out possible contributions of other effects of d-phen, such as facilitation of aminergic neurotransmitter synthesis and/or transmission (32).

Using local administration of ACTZ or d-phen into discrete brain regions, the present study demonstrates that CAs have a critical role in the consolidation of extinction memory in the vmPFC, BLA, and CA1. All these regions are involved in memory extinction, as shown by studies using techniques ranging from tissue lesions to pharmacologic, optogenetic, and chemogenetic approaches (22, 46). Conversely, CAs in the SNpc are not involved in this process, at least not under the experimental conditions explored in the present study. Our study unequivocally demonstrates that CA activation in these brain regions is necessary to ensure extinction memory consolidation. Neuronal circuits engaging the vmPFC, BLA, and CA1 also have been implicated in fear memory acquisition, consolidation, and retrieval (34). Although mnemonic processes for fear memory consolidation and extinction share some similar molecular mechanisms and pathways (34), the patterns of neuronal activation and neuronal signal transduction mechanisms within fear consolidation and fear extinction circuits show marked differences (13, 22, 47). These differences might be conferred by procedural dissimilarities regarding the presence or absence of the unconditioned stimulus.

During fear conditioning, brain circuits are activated by animal's exposure to both an unconditioned stimulus (shock) and a conditioned stimulus (in our case, the context), whereas fear extinction is induced presenting only the conditioned stimulus. Several protein kinases (Fyn, CDK5, PKA, and PKC), protein phosphatases (calcineurin and SHP1/2), transcription factors, and immediate early genes (CARP, CREB, c-Fos, c-Jun, JunB, and JunD) come into play during fear conditioning or extinction, being either increased or decreased in various ways (reviewed in ref. 47). Impairment of both fear memory consolidation (33) and extinction (the present study) appears to be associated with reduced brain CA activity. The brain is rich in various CA isoforms, differentially distributed in different brain regions (26). Since ACTZ is a nonselective inhibitor of several of these isoforms, its effect may mask distinct contributions by different isoforms (48) and specific involvement of certain brain areas (see below).

Findings from several studies focusing on the role of the vmPFC in extinction indicate that neurons of this region are necessary for extinction learning (12). In particular, brief stimulation of the infralimbic cortex (a subregion of the vmPFC) was found to reduce fear memory and strengthened extinction (12, 49). More recently, systemic administration of ACTZ was found to inhibit prefrontal cortical single-unit firing in vivo, leading to reduced basal neuronal activity of the prefrontal cortex (50). Taken together, these findings suggest that ACTZ impairs fear extinction memory through inhibition of vmPFC firing.

At 90 min after the extinction retention test, c-Fos–immunopositive nuclei were numerous in the vmPFC of ACTZ-treated rats, significantly more numerous than in vehicle treated rats. The most parsimonious interpretation of our results suggests that ACTZ-treated rats exhibited high levels of freezing behavior and an increased number of c-Fos–positive nuclei in the vmPFC at retention, as they did not consolidate extinction but rather remembered the CFC training administered 48 h earlier. Indeed, much experimental evidence suggests a key role of the PFC in processing and recalling remote contextual fear memories (51, 52) and during extinction consolidation, rather than after a retention test (52–54). c-Fos immunostaining in the BLA and hippocampal CA1 region after extinction retrieval was low in brains of vehicle-treated rats and not significantly different from that in brains of ACTZ-treated rats. A time-limited role of BLA and CA1 neurons in fear memory retrieval has been described (55) that may explain why we detected very low c-Fos immunostaining in rats that did not learn extinction (ACTZ-treated) and supposedly remembered the training experience. Furthermore, a bidirectional regulation of c-Fos induction in the mPFC (high) and the BLA (low) has been reported during extinction training and spontaneous recovery (56, 57); however, other studies have found high c-Fos expression in the BLA and CA1 after extinction retrieval (54). The literature covering the role of the hippocampus in fear memory recall and extinction is vast and quite controversial (reviewed in ref. 52). Orsini et al. (58) found no significant increase of c-Fos expression in the ventral hippocampus of extinguished rats, and another study found no significant cFos increase in the dorsal hippocampus on remote contextual fear recall or extinction (59). Furthermore, early work suggested that the involvement of the dorsal hippocampus in contextual consolidation is limited to the early period of the memorization phase (60). The aforementioned studies used different protocols (e.g., auditory extinction, repeated extinction sessions over several days), which may be partially responsible for these discrepancies. Clearly, further work is necessary to explore these controversial results.

In conclusion, the present study provides several important insights into the involvement of CAs in fear memory. We have demonstrated that (i) the selective inhibition of CAs in the brain correlates with impairments of extinction; (ii) the selective activation of CAs in the brain has a beneficial effect on extinction; and (iii) CA activity is involved in extinction modulation only in specific brain regions. The mechanisms underlying the effects of CAs on extinction remain mostly unknown. By inhibiting CAs, ACTZ diminishes the buffering capacity, thereby influencing intracellular and extracellular pH and affecting proteins, NMDA, and γ-aminobutyric acid (GABA) receptor function (51). CA inhibitors also increase intracellular CO₂ accumulation, block anion transport, and increase GABA levels, leading to modulation of the firing rate (61). Early studies demonstrated that in the CA1 hippocampal region, the associated activation of multisynaptic inputs on pyramidal neurons transiently transform GABAergic inhibitory postsynaptic potentials (IPSPs) to excitatory postsynaptic potentials (62–64). The transformed synaptic inputs from the GABAergic interneurons provides a mechanism for modulating signal flow through the hippocampal network, enhancing the signal-to-noise ratio and selectively amplifying synaptic weights relevant to a particular memory (64). This synaptic transformation depends on a depolarizing transmembrane HCO₃⁻ flux that is reduced or eliminated by ACTZ (64). Furthermore, in the presence of d-phen, subthreshold inputs to pyramidal neurons switched the GABA-mediated IPSPs to depolarizing responses (65). The modification of information flow through the hippocampal network may explain the memory-impairing effects of ACTZ. Hippocampal output to the mPFC also would be affected by CA inhibitors, as ACTZ was found to increase the afferent drive from the hippocampus while reducing the basal neuronal activity of prefrontal cortex neurons (50).

It was recently reported that ACTZ inhibits fear conditioning-induced ERK phosphorylation in the amygdala (33), consistent with the reported impairment of extinction (66). In this regard, it is important to note that d-phen administration has been found to rapidly activate ERK pathways in the cortex and the hippocampus and to enhance OR memory (32) and water maze performance (29).

The extinction of unwanted responses when exposed to reminders of a previous trauma is a core process underlying exposure therapy. Systemic drugs that facilitate extinction, such as cannabinoids, noradrenergic, histaminergic drugs, and neurotrophic factors, might be useful in improving the clinical response of exposure-based therapies (3, 5, 12). Based on results from this study, brain CA activators (67), as molecules able to improve extinction, may improve the exposure-based treatment of such disorders as phobias, anxiety, and PTSD. The clinical potentials of these compounds are not diminished by the lack of knowledge about the underlying mechanism, although attempts to optimize their use will have a much greater likelihood of success when their mechanism of eliciting fear attenuation is understood.

Materials and Methods

Three-month-old male Wistar rats were purchased from Charles River Laboratories Italia and group-housed in the Center of Services for Laboratory Animal Housing (CeSAL), University of Florence. Fear conditioning was conducted with three electrical foot shocks (0.5 mA, 2 s) at 30-s intervals. The extinction of contextual fear conditioning was performed as described previously (38, 68) with little modification. Then, 24 h later, the animals were placed in the same chamber for a 15- or 30-min extinction training session (depending on the experimental set) in the absence of the foot shocks. Immediately after these sessions, vehicle, CA activator (d-phen), or CA inhibitor (ACTZ or C18) was administered either systemically (i.p.) or infused locally into selected brain regions through cannulae stereotaxically implanted bilaterally. The time the animal spent freezing was manually recorded by a trained observer unaware of the treatments, and statistically significant differences between the experimental groups were determined with repeated-measures two-way ANOVA. The source of the detected significances was determined using Bonferroni’s multiple-comparison post hoc test. P values < 0.05 were considered statistically significant. Details of the statistical analysis are provided in SI Appendix, Table S1.

For c-Fos measurements, rats were trained in the CFC task and after 24 h were submitted to a 30-min extinction training. Immediately after this session, they received an i.p. injection of vehicle or ACTZ 30 mg/kg. A 3-min retention test was performed at 24 h after the extinction training session. Rats were euthanized at 90 min after the retention test and then perfused transcardially with cold physiological saline, followed by 4% (vol/vol) paraformaldehyde in 0.1 M phosphate buffer (pH 7.4). Tissue preparations, immunostaining, and analysis were performed as described previously (69). Student’s t test was used to compare the number of c-Fos–positive nuclei among different experimental groups.

Details of the study protocols and a list of materials are available in the SI Appendix. The raw data supporting the findings of this study are provided in Dataset S1. Requests for further information should be directed to the corresponding authors.