Grin2b-mutant mice exhibit heightened remote fear via suppressed extinction and chronic amygdalar synaptic and neuronal dysfunction

Abstract

Individuals with autism spectrum disorders (ASD) frequently show long-lasting traumatic fear memory, but the underlying mechanisms remain unclear. Here, we report that Grin2b-mutant mice carrying a human ASD-risk mutation (Grin2b^C456Y/+ mice) show normal acquisition of contextual fear memory but suppressed fear memory extinction and enhanced remote fear memory responses, along with anxiety- and social-related abnormalities. After footshock and fear extinction, these mutants chronically develop occluded neuronal activation in the basal amygdala (BA) detectable at remote fear memory retrieval, which involves suppressed spontaneous excitatory synaptic transmission and neuronal excitability. Chemogenetic activation of mutant BA neurons during fear extinction improves fear memory extinction and remote fear memory responses without affecting anxiety- or social-related phenotypes. This rescue involves normalized spontaneous excitatory synaptic transmission and neuronal excitability. These results suggest that Grin2b^C456Y/+ mice, through impaired fear memory extinction, chronically develop suppressed spontaneous excitatory synaptic transmission and neuronal excitability in BA neurons that enhances remote fear memory responses.

Amygdalar dysfunctions trigger suppressed fear extinction and enhanced remote fear responses in a Grin2b-mutant ASD mouse model.

INTRODUCTION

Autism spectrum disorders (ASD) are characterized by the core symptoms of social and repetitive behavioral deficits and related comorbidities such as intellectual disability, anxiety, attention deficit hyperactivity disorder–like hyperactivity, epilepsy, sensory abnormalities, and posttraumatic stress disorders (PTSD). PTSD in ASD, which is observed at ~10 times higher rates than in a general population (1–3), is thought to exacerbate ASD symptoms and cause cognitive and behavioral social difficulties, including social anxiety, fear, phobia, and avoidance. Previous studies on mouse models of ASD have provided decent insights into ASD-related mechanisms (4–14). However, PTSD-like phenotypes and related mechanisms in these models have been minimally studied (15), although a medication for PTSD [i.e., (±)3,4-methylenedioxymethamphetamine] (16, 17) has been shown to improve social functions in some mouse models of ASD (18, 19).

N-methyl-d-aspartate receptors (NMDARs) play critical roles in regulating brain development and functions (20–26). The GluN2B subunit of NMDARs is expressed early in development and regulates neuronal and synaptic development, function, and signaling (20–22, 24, 27, 28). Pathogenic variants in GRIN2B, which encodes GluN2B, are linked to a spectrum of neurodevelopmental, neurological, and psychiatric disorders, including developmental delay, ASD, and intellectual disability (28–33). Studies in Grin2b-mutant mice have clarified the in vivo roles of GluN2B-containing NMDARs and illuminated disease mechanisms in GRIN2B-related disorders (34–42). Nevertheless, exactly how GRIN2B mutations disrupt neural development and circuit function remains incompletely understood (42–45).

Grin2b^C456Y/+ mice, which carry the patient-derived GluN2B-C456Y point mutation (31), show NMDAR dysfunctions and behavioral and sensory abnormalities (43, 44). In particular, Grin2b^C456Y/+ mice show suppressed NMDAR-dependent long-term depression (LTD) and related impairments of neural circuit refinements, which are thought to lead to neuronal hyperactivity, cortical/subcortical hyperconnectivity, and sensory hypersensitivity (44). However, it remains unclear how these connectivity changes affect ASD-related brain functions in social, anxiety, and fear domains.

We report here that Grin2b^C456Y/+ mice show normal acquisition of contextual fear memory but suppressed fear memory extinction and enhanced remote fear memory responses, along with anxiety- and social-related phenotypes. A brain-wide c-fos survey after remote fear memory retrieval points to occluded activation of basal amygdalar (BA) neurons in Grin2b^C456Y/+ mice. Mutant BA neurons also exhibit chronic suppression of spontaneous excitatory synaptic transmission and neuronal excitability. Increasing BA activity with designer receptors exclusively activated by designer drugs (DREADD) during fear extinction improves fear memory extinction and remote fear memory responses but not anxiety- and social-related behaviors. This rescue involves normalization of the suppressed spontaneous excitatory synaptic transmission and neuronal excitability. Therefore, Grin2b^C456Y/+ mice seem to show impaired fear extinction that leads to exaggerated remote fear memory responses through chronic maladaptations of synaptic function and neuronal excitability.

RESULTS

Grin2b^C456Y/+ mice show suppressed fear memory extinction and enhanced remote fear memory responses

To test whether Grin2b^C456Y/+ mice show abnormal fear-related responses at the levels of fear memory acquisition, retrieval, extinction, and/or remote retrieval, we subjected mice to a contextual fear conditioning test followed by a fear memory extinction test performed over days 2 to 6 and a remote fear memory retrieval test performed on day 30 (Fig. 1A).

Fig. 1. Grin2b^C456Y/+ mice show suppressed fear memory extinction and enhanced remote fear memory responses.

(A) Experimental scheme for contextual fear memory acquisition (day 1/P35) followed by recent fear memory extinction (days 2 to 6) and a remote fear memory retrieval (day 30) in WT and Grin2b^C456Y/+ mice. (B) Freezing levels during fear memory acquisition (day 1), fear memory extinction (days 2 to 6), and remote fear memory retrieval (day 30) in WT and Grin2b^C456Y/+ male mice. {n = 38 mice (WT) and 42 [heterozygous (HT)], two-way analysis of variance (ANOVA) with Sidak’s test}. (C) Index of fear memory extinction [(day 2 freezing – day 6 freezing)/(day 2 freezing + day 6 freezing)] in WT and Grin2b^C456Y/+ male mice. [n = 38 (WT) and 42 (HT), Mann-Whitney test]. (D) Remote fear memory responses (day 30) in WT and Grin2b^C456Y/+ male mice. [n = 38 (WT) and 42 (HT), Mann-Whitney test]. (E to G) The experiments described above were repeated for female WT and Grin2b^C456Y/+ mice. [n = 10 mice (WT) and 9 (HT), two-way ANOVA with Sidak’s test for freezing levels during fear memory acquisition, fear memory extinction, and remote fear memory retrieval (E), Student’s t test for the index of fear memory extinction (F), and Welch’s t test for remote fear memory responses (G)]. (H to K) Repetition of the experiments presented in (A) to (G) with the fear extinction experiment performed during days 30 to 34 instead of days 2 to 6 (remote fear memory extinction) in WT and Grin2b^C456Y/+ mice (male). {n = 16 (WT) and 14 (HT), two-way ANOVA with Sidak’s test (I), Student’s t test (J), and Welch’s t test (K); index of fear memory extinction [(day 30 freezing – day 34 freezing)/(day 30 freezing + day 34 freezing)]}. Significance is indicated as *P < 0.05, **P < 0.01, ***P < 0.001, or ns (not significant).

Grin2b^C456Y/+ mice at juvenile stages (~postnatal day/P35; males) displayed normal acquisition of contextual fear memory, as shown by electric footshock-induced freezing on day 1 (Fig. 1B). On day 2, upon stimulation of 24-hour fear memory retrieval, Grin2b^C456Y/+ mice showed freezing levels comparable to those of wild-type (WT) mice. On days 3 to 6, Grin2b^C456Y/+ mice showed less fear memory extinction than WT mice (Fig. 1, B and C). On day 30, Grin2b^C456Y/+ mice showed enhanced remote fear memory responses, relative to WT mice (Fig. 1D). Female Grin2b^C456Y/+ mice also showed similarly suppressed fear memory extinction and enhanced remote fear memory responses, as evidenced by a significant genotype effect but no effect of sex or genotype × sex interaction in two-way analysis of variance (ANOVA) (Fig. 1, E to G, and fig. S1). Therefore, Grin2b^C456Y/+ mice show normal acquisition of fear memory but suppressed fear memory extinction and enhanced remote fear memory responses.

Similar results (suppressed fear memory extinction and enhanced remote fear memory responses) were observed in Grin2b^C456Y/+ male mice when fear memory extinction was performed long after fear memory acquisition (after 30 days instead of 1 day) (Fig. 1, H to K). Similar results were also observed when Grin2b^C456Y/+ male mice were exposed to the initial footshock at an adult (~P65) stage instead of a juvenile (~P35) stage (fig. S2). These results collectively suggest that Grin2b^C456Y/+ mice show normal fear memory acquisition but suppressed fear memory extinction and enhanced remote fear memory responses, and moreover, such changes are independent of the initial age of fear memory acquisition, the interval between fear memory acquisition and extinction, and the sex of the mice.

Grin2b^C456Y/+ mice with shock extinction show long-lasting anxiety-like behavior and social deficits

We hypothesized that the suppression of fear memory extinction in Grin2b^C456Y/+ mice may induce long-lasting anxiety-like behaviors. To this end, we subjected WT and Grin2b^C456Y/+ mice to fear acquisition (day 1), fear memory extinction (days 2 to 6), and remote fear memory retrieval (day 30), followed by locomotion (open-field), anxiety-related (open-field center and elevated plus maze), and social interaction (three-chamber) tests (Fig. 2A). We also analyzed unshocked naïve WT and Grin2b^C456Y/+ mice as controls.

Fig. 2. Grin2b^C456Y/+ mice with shock extinction show long-lasting anxiety-like behavior.

(A) Experimental scheme for tests of anxiety-related behaviors in WT and Grin2b^C456Y/+ mice (naïve/no shock extinction or shock extinction; male). Mice with footshock (day 1/P35) were subjected to fear extinction (days 2 to 6/P36 to P40) and remote fear memory retrieval (day 30/P66), and both naïve and shock extinction WT and Grin2b^C456Y/+ mice were subjected to anxiety/social-related behavioral tests [open-field (OFT), elevated plus maze (EPM), and three-chamber (3CT) tests; >day 32/P68]. (B to I) Locomotor activity and anxiety-like behavior (center time) in the open-field test for naïve (B to E) and shock extinction (F to I) WT and Grin2b^C456Y/+ mice. [n = 31 mice (WT-naïve), 32 (HT-naïve), 31 (WT-shock), and 33 (HT-shock), two-way ANOVA with Sidak’s test (B, D, F, and H), Student’s t test (C and G), and Mann-Whitney test (E and I)]. (J to M) Anxiety-like behavior in the elevated plus maze test for naïve (J and K) and shock extinction (L and M) WT and Grin2b^C456Y/+ mice. [n = 18 (WT-naïve), 16 (HT-naïve), 30 (WT-shock), and 30 (HT-shock), Student’s t test (J and K) and Mann-Whitney test (L and M)]. Significance is indicated as *P < 0.05, **P < 0.01, ***P < 0.001, or ns (not significant).

Hypoactivity in the open-field test was observed in naïve Grin2b^C456Y/+ mice, as compared with naïve WT mice, and Grin2b^C456Y/+ mice with footshock and extinction (shock extinction), as compared with shock extinction WT mice and naïve Grin2b^C456Y/+ mice without shock extinction (albeit to a much lesser extent in the latter comparison) (Fig. 2, B, C, F, and G, and fig. S3, A, B, E, and F). Shock extinction Grin2b^C456Y/+ mice, with attenuated fear extinction, also showed increased anxiety-like behavior (decreased center time) in the open-field test, as compared with shock extinction WT mice and naïve Grin2b^C456Y/+ mice without shock extinction (Fig. 2, D, E, H, and I, and fig. S3, C, D, G, and H). The time-dependent decrease in center time in shock extinction Grin2b^C456Y/+ mice likely reflects normal perception of novelty but diminished willingness, likely involving reduced exploratory drive or increased anxiety, to continue investigating the arena. In the elevated plus maze test, shock extinction Grin2b^C456Y/+ mice, with attenuated fear extinction, showed anxiety-like behaviors, as compared with shock extinction WT mice and naïve Grin2b^C456Y/+ mice (Fig. 2, J to M, and fig. S3, I to Ad). These results suggest that Grin2b^C456Y/+ mice with footshock and fear extinction stimulations show enhanced anxiety-like behaviors.

In the three-chamber test, shock extinction Grin2b^C456Y/+ mice showed no change in the social interaction level but suppression of social novelty recognition, as compared with shock extinction WT mice and naïve Grin2b^C456Y/+ mice (albeit to a lesser extent in the latter comparison) (fig. S4, A to P). Female Grin2b^C456Y/+ mice exposed to shock extinction displayed heightened anxiety-like behavior in the open-field and elevated plus maze tests, although the effect was milder than in males (fig. S5), pointing to the sex-dependent difference reminiscent of the male-female difference in ASD. These results collectively suggest that Grin2b^C456Y/+ mice with footshock and extinction show long-lasting anxiety-like behaviors and suppressed social novelty recognition.

Abnormal activities of BA and infralimbic neurons in Grin2b^C456Y/+ mice with shock extinction

To identify the brain regions involved in the suppressed fear memory extinction and enhanced remote fear memory responses of Grin2b^C456Y/+ mice with shock extinction, we attempted a c-fos labeling experiment. To this end, WT and Grin2b^C456Y/+ mice were subjected to initial footshock (day 1) and fear memory extinction (days 2 to 6), remote fear memory retrieval (day 30) was applied or omitted, and c-fos analyses were performed (Fig. 3A and fig. S6, A to C). All subsequent assays were performed in males, as the fear- and anxiety-related phenotypes are stronger in Grin2b^C456Y/+ males than in females.

Fig. 3. Abnormal activities of BA and IL neurons in Grin2b^C456Y/+ mice with shock extinction.

(A) Experimental scheme for identifying brain regions exhibiting altered neuronal activities in WT and Grin2b^C456Y/+ mice with shock extinction (male) by subjecting mice to footshock (day 1/P35) and fear extinction (days 2 to 6) with/without remote fear memory retrieval (day 30) and then, 90 min later, assessing c-fos. (B to E) Distinct changes in c-fos expression in brain regions of WT and Grin2b^C456Y/+ mice with/without remote fear memory retrieval, as shown by c-fos ratios under retrieval/no-retrieval (Retr/Ctrl) conditions in WT and Grin2b^C456Y/+ mice. Note that c-fos levels are increased in the WT-BA but unaltered in the mutant-BA, while c-fos levels are unaltered in the WT-IL and WT–ventral hippocampus (vHPC) but decreased in the mutant-IL and mutant-vHPC. Scale bar, 1 mm. S1, primary somatosensory cortex; S2, secondary somatosensory cortex; PL, prelimbic cortex; IL, infralimbic cortex; Pir, piriform cortex; NAcC, nucleus accumbens core; NAcSh, nucleus accumbens shell; VP, ventral pallidum; ACC, anterior cingulate cortex; dHPC, dorsal hippocampus; LA, lateral amygdala; CeA, central amygdala; VTA, ventral tegmental area; PAG, periaqueductal gray. (F and G) Example images and quantification of c-fos levels in the BA. The levels of c-fos were normalized by NeuN (neuronal marker) staining [c-fos⁺ and NeuN⁺/NeuN⁺ (%)]. Scale bars, 200 μm. [n = 5 mice (WT-Ctrl/control), 9 (WT-Retr/retrieval), 4 (HT-Ctrl), and 6 (HT-Retr), two-way ANOVA with Sidak’s test]. (H and I) Example images and quantification of c-fos levels in the IL. Scale bars, 200 μm. [n = 5 (WT-Ctrl), 9 (WT-Retr), 4 (HT-Ctrl), and 6 (HT-Retr), two-way ANOVA with Sidak’s test]. Significance is indicated as *P < 0.05 and ns (not significant). DAPI, 4′,6-diamidino-2-phenylindole.

After shock extinction followed by remote retrieval, Grin2b^C456Y/+ mice showed abnormal c-fos activation in the BA and infralimbic (IL) medial prefrontal cortex (mPFC; Fig. 3, B to E, and fig. S6, D to O). Both BA and mPFC have been reported to be central to fear extinction circuits (46–57). In the BA, the c-fos levels in WT mice were increased upon remote fear memory retrieval, whereas those in mutant BA neurons were not changed (Fig. 3, F and G). In the IL region, the c-fos levels were not changed upon remote retrieval of WT mice, whereas they were decreased in mutant mice (Fig. 3, H and I). Therefore, BA and IL neurons show abnormal changes of c-fos levels in Grin2b^C456Y/+ mice upon remote fear memory retrieval following fear shock extinction.

We next measured c-fos levels immediately after footshock (day 1) or fear memory extinction (days 2 to 6) in addition to immediately after remote fear memory retrieval. Immediately after footshock (day 1), c-fos expression in BA neurons rose in both WT and Grin2b^C456Y/+ mice, but it returned to baseline by the end of extinction training (days 2 to 6) (fig. S7, A to C). IL neurons showed no significant c-fos change at either of these acute time points (fig. S7D). Thus, the genotype-specific BA and IL differences we observe may emerge during remote fear memory retrieval, implicating retrieval-related circuitry rather than initial encoding or extinction learning.

Suppressed spontaneous excitatory synaptic transmission in Grin2b^C456Y/+ BA neurons with shock extinction

To explore the mechanisms underlying the differential responses of BA and IL neurons upon remote fear memory retrieval in WT and Grin2b^C456Y/+ mice with shock extinction, we first examined excitatory and inhibitory synaptic transmissions in BA and IL regions from mice with/without shock extinction. Here, we did not include a remote retrieval session before recording because our goal was to isolate the long-term or chronic synaptic consequences of the shock extinction experience. To this end, WT and Grin2b^C456Y/+ mice with/without shock extinction were subjected to measurements of spontaneous synaptic transmission on ~day 30 without remote fear memory retrieval (Fig. 4, A to C, and fig. S8, A to C).

Fig. 4. Suppressed spontaneous excitatory synaptic transmission in BA neurons of Grin2b^C456Y/+ mice with shock extinction.

(A and B) Experimental scheme for measuring spontaneous excitatory and inhibitory synaptic transmissions in BA neurons from WT and Grin2b^C456Y/+ mice (male; ~P66) with footshock (day 1/P35) and fear extinction (days 2 to 6) without application of remote fear memory retrieval (day 30). Naïve WT and Grin2b^C456Y/+ mice (male; ~P66) without shock extinction were used as controls. (C) Behavioral data validating that fear acquisition and fear memory extinction occurred in WT and Grin2b^C456Y/+ mice with shock extinction (day1/P35 and days 2 to 6, respectively) without remote fear memory retrieval (day 30) before the measurement of spontaneous excitatory and inhibitory synaptic transmissions in the BA region [n = 10 (WT) and 9 (HT), two-way ANOVA with Sidak’s test]. (D to F) mEPSCs in BA neurons from naïve and shock extinction WT and Grin2b^C456Y/+ mice. [n = 22 neurons from seven mice (WT-naïve), 16 from four mice (WT-shock), 19 from four mice (HT-naïve), and 20 from four mice (HT-shock), two-way ANOVA with Sidak’s test]. (G to I) sEPSCs in BA neurons from naïve and shock extinction WT and Grin2b^C456Y/+ mice. [n = 22 neurons from five mice (WT-naïve), 17 from four mice (WT-shock), 21 from five mice (HT-naïve), and 16 from four mice (HT-shock), two-way ANOVA with Sidak’s test]. (J to L) mIPSCs in BA neurons from naïve and shock extinction WT and Grin2b^C456Y/+ mice. [n = 17 neurons from four mice (WT-naïve), 17 from four mice (WT-shock), 14 from three mice (HT-naïve), and 17 from four mice (HT-shock), two-way ANOVA with Sidak’s test]. (M to O) sIPSCs in BA neurons from naïve and shock extinction WT and Grin2b^C456Y/+ mice. [n = 17 neurons from four mice (WT-naïve), 24 from six mice (WT-shock), 15 from three mice (HT-naïve), and 24 from five mice (HT-shock), two-way ANOVA with Sidak’s test]. Significance is indicated as *P < 0.05, **P < 0.01, ***P < 0.001, or ns (not significant).

WT-BA neurons with shock extinction showed an increase in the frequency (not amplitude) of miniature excitatory postsynaptic currents (mEPSCs) compared with those of naïve WT-BA neurons (Fig. 4, D to F). Intriguingly, naïve Grin2b^C456Y/+ BA neurons exhibited a greater mEPSC frequency than naïve WT-BA neurons, indicative of a baseline increase. Shock extinction of mutant mice did not increase the frequency of mEPSCs in BA neurons, indicating that there was no additional increase of mEPSC frequency in shock extinction contexts. Spontaneous excitatory postsynaptic currents (sEPSCs), measured in the presence of network activity, abolished the baseline increase in mEPSC frequency in Grin2b^C456Y/+ BA neurons as well as the shock extinction–induced increase of mEPSC frequency in WT BA neurons (Fig. 4, G to I), reflecting network activity–dependent compensatory changes to suppress the baseline difference as well as shock extinction–induced changes. However, the network activity seems to highlight the significantly lower sEPSC frequency in Grin2b^C456Y/+ BA neurons relative to that in WT BA neurons under shock extinction but not naïve contexts.

In the IL region, we did not observe any baseline difference of mEPSC frequency/amplitude in WT mice versus mutant mice without shock extinction (fig. S8, D to F). Shock extinction did not induce significant changes in mEPSC frequency/amplitude, or sEPSC frequency/amplitude, in WT and mutant neurons (fig. S8, D to I). Unlike excitatory synaptic transmissions, inhibitory synaptic transmissions [miniature and spontaneous inhibitory postsynaptic currents (mIPSCs and sIPSCs, respectively)] did not differ between WT and mutant neurons under baseline and shock extinction conditions in the BA and IL regions (Fig. 4, J to O, and fig. S8, J to O).

When mEPSCs and sEPSCs were measured 24 hours (not 30 days) after the shock phase (before extinction) (fig. S9, A to C), naïve Grin2b^C456Y/+ BA neurons still displayed the higher baseline mEPSC frequency seen in naïve mutants (fig. S9, D to F). Footshock reduced this frequency to WT levels, whereas the smaller reduction in WT did not reach significance. For sEPSCs, the two genotypes were identical at baseline; shock decreased event frequency and increased amplitude in both (fig. S9, G to I). Therefore, footshock alone elicits parallel synaptic adjustments in WT and Grin2b^C456Y/+ BA neurons; divergent changes appear only after the full shock extinction sequence and the ensuing 30-day interval.

Thus, Grin2b^C456Y/+ BA neurons start with higher baseline mEPSC rates but, 30 days after shock extinction, do not gain additional excitatory drive and ultimately fire less than WT neurons under network (sEPSC) conditions. IL neurons, in contrast, show identical baseline excitation in both genotypes and remain unchanged after the same interval.

Suppressed neuronal excitability in the BA of Grin2b^C456Y/+ mice with shock extinction

To further understand the mechanisms underlying the limited activation of BA neurons in Grin2b^C456Y/+ mice with shock extinction upon remote fear memory retrieval, we next measured neuronal excitability. For this, WT and Grin2b^C456Y/+ mice with shock extinction were used for electrophysiological experiments on day 30, without prior fear memory retrieval (Fig. 5, A and B, and fig. S10A).

Fig. 5. Suppressed neuronal excitability in BA neurons of Grin2b^C456Y/+ mice with shock extinction.

(A and B) Experimental scheme for measuring neuronal excitability, NMDAR/AMPAR currents, and evoked AMPAR currents in BA neurons from WT and Grin2b^C456Y/+ mice (male; ~P66) with shock extinction (day 1/P35 and days 2 to 6, respectively) but no remote fear memory retrieval (day 30). Naïve WT and Grin2b^C456Y/+ mice (male; ~P66) without shock extinction were used as controls. (C to E) Current-firing curves obtained from BA neurons of naïve and shock extinction WT and Grin2b^C456Y/+ mice. [n = 24 neurons from six mice (WT-naïve), 22 from five mice (WT-shock), 24 from four mice (HT-naïve), and 22 from five mice (HT-shock), two-way ANOVA with Sidak’s test]. (F to H) Input resistance and sag amplitude in BA neurons from naïve and shock extinction WT and Grin2b^C456Y/+ mice. [n = 24 neurons from six mice (WT-naïve), 22 from five mice (WT-shock), 24 from four mice (HT-naïve), and 22 from five mice (HT-shock), two-way ANOVA with Sidak’s test]. (I to M) Rheobase and action potential (AP)–related parameters [rheobase currents, after hyperpolarization (AHP) amplitude, AP amplitude, and full width at half maximum (FWHM) of AP] in BA neurons from naïve and shock extinction WT and Grin2b^C456Y/+ mice. [n = 24 neurons from six mice (WT-naïve), 22 from five mice (WT-shock), 24 from four mice (HT-naïve), and 22 from five mice (HT-shock), two-way ANOVA with Sidak’s test). (N to P) Ratios of NMDAR- and AMPAR-EPSCs in BA neurons from naïve and shock extinction WT and Grin2b^C456Y/+ mice. [n = 9 neurons from six mice (WT-naïve), 15 from eleven mice (WT-shock), 13 from eight mice (HT-naïve), and 12 from nine mice (HT-shock), two-way ANOVA with Sidak’s test]. Significance is indicated as *P < 0.05, **P < 0.01, ***P < 0.001, or ns (not significant).

When BA neuronal excitability was measured in naïve and shock extinction WT mice, we found that shock extinction led to increases in excitability, as supported by our current firing curve, rheobase, and train duration analyses (Fig. 5, C to M, and fig. S10, B to G). In contrast, Grin2b^C456Y/+ BA neurons did not show these shock extinction–induced increases.

When IL neuronal excitability was measured in naïve and shock extinction WT mice, WT-specific increases in excitability were seen in the rheobase and train duration analyses (fig. S11). However, other measures of neuronal excitability, including the current-firing curve, were either unaltered or similarly altered in WT and Grin2b^C456Y/+ IL neurons. Therefore, shock extinction boosts intrinsic excitability robustly in WT BA neurons but only modestly in WT and Grin2b^C456Y/+ IL neurons.

When neuronal excitability parameters were measured 24 hours (not 30 days) after the shock phase (before extinction) (fig. S12, A to C), naïve WT and Grin2b^C456Y/+ BA neurons displayed identical excitability (fig. S12, D to J). Footshock alone did not raise excitability in either genotype, showing only an upward trend in WT neurons. Hence, the excitability divergence observed at ~30 days after shock extinction requires both the shock extinction experience and the 30-day interval.

Grin2b encodes the NMDAR-GluN2B subunit, and NMDARs critically regulate fear memory extinction (46, 58–66). We thus measured NMDAR-EPSCs and α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor (AMPAR)–EPSCs and their ratios in BA neurons from WT and Grin2b^C456Y/+ mice (at P66) without (naïve) and with shock extinction (Fig. 5A and fig. S10H). There was no baseline difference of NMDAR/AMPAR-EPSC ratios in naïve WT and Grin2b^C456Y/+ BA neurons (Fig. 5, N and O). Upon shock extinction, NMDAR/AMPAR-EPSC ratios were similarly decreased to comparable levels in WT and Grin2b^C456Y/+ BA neurons (Fig. 5, N and O). Evoked AMPAR-EPSCs were also comparable in WT and Grin2b^C456Y/+ BA neurons with or without shock extinction (fig. S10, I to K). Therefore, NMDAR currents at ~P66 (30 days after) are indistinguishable between WT and Grin2b^C456Y/+ mice, likely because of the largely normal GluN2A levels in Grin2b^C456Y/+ mice (43). In addition, shock extinction induces comparable decreases in NMDAR currents and NMDAR-dependent synaptic plasticity in WT and Grin2b^C456Y/+ mice.

Last, the baseline decay kinetics of NMDAR currents was decreased in Grin2b^C456Y/+ BA neurons, reflecting decreased levels of GluN2B-containing NMDARs, but shock extinction removed the genotype difference (Fig. 5P). Therefore, NMDAR function and plasticity are unlikely to underlie the limited shock extinction–induced BA neuronal activation in Grin2b^C456Y/+ mice.

These results collectively suggest that Grin2b^C456Y/+ BA neurons fail to show shock extinction–induced increases in excitability exhibited by WT BA neurons. In contrast, WT and Grin2b^C456Y/+ IL neurons show largely similar shock extinction–induced changes in excitability.

Chemogenetic activation of BA neurons improves fear memory extinction and remote fear memory responses in Grin2b^C456Y/+ mice

We hypothesized that the limited excitatory synaptic and neuronal activation of BA neurons may underlie the suppressed fear extinction and enhanced fear memory responses seen in Grin2b^C456Y/+ mice. We thus sought to perform DREADD-mediated chemogenetic activation of Grin2b^C456Y/+ BA neurons during fear memory extinction sessions. To this end, Grin2b^C456Y/+ mice were injected with AAV-CaMKIIα-hM3Dq-mCherry in the BA (~P15) and subjected to fear memory acquisition (day 1/P35) and fear memory extinction in the presence or absence of chemogenetic BA neuronal stimulation [days 2 to 6; clozapine N-oxide (CNO)/dimethyl sulfoxide (DMSO)], followed by remote fear memory retrieval (day 30) and behavioral tests (anxiety and social related; >day 32) (Fig. 6, A to C).

Fig. 6. Chemogenetic activation of BA neurons improves fear memory extinction and remote fear memory responses in Grin2b^C456Y/+ mice.

(A) Experimental scheme for the injection of AAV-CaMKIIα-hM3Dq-mCherry in the BA region of WT and Grin2b^C456Y/+ mice (~P15) followed by intraperitoneal (i.p.) injection of CNO or DMSO for chemogenetic BA neuronal activation. (B) Example images of BA regions with expression of hM3Dq-mCherry in CaMKIIα-positive neurons, as shown by double immunostaining for mCherry and CaMKIIα. Scale bars, 300 μm (left) and 100 μm (right; magnified). (C) Experimental scheme for AAV-CaMKIIα-hM3Dq-mCherry injection (P15), footshock (P35/day 1), and CNO-dependent stimulation of hM3Dq-expressing CaMKIIα-positive BA neurons during fear memory extinction (P36 to P40/days 2 to 6; daily CNO injection), followed by remote fear memory retrieval (P66/day 30) and related behavioral tests (anxiety and social related; >P68/day 32). (D to G) Freezing levels in hM3Dq-expressing WT and Grin2b^C456Y/+ mice during fear memory acquisition, recent fear memory extinction with CNO/DMSO treatment, and a remote fear memory retrieval (D). Freezing levels on individual day 2 (E), day 6 (F), and day 30 (G). [n = 21 mice (WT-DMSO), 22 (WT-CNO), 26 (HT-DMSO), and 26 (HT-CNO), two-way ANOVA with Sidak’s test]. Significance is indicated as *P < 0.05, **P < 0.01, ***P < 0.001, or ns (not significant).

Grin2b^C456Y/+ mice without chemogenetic BA neuronal stimulation (DMSO) showed suppressed fear memory extinction and enhanced remote fear memory responses, compared with WT-DMSO mice (Fig. 6D); these results resembled the findings obtained from naïve Grin2b^C456Y/+ and WT mice.

Upon CNO-dependent BA neuronal stimulation, Grin2b^C456Y/+ mice showed significantly reduced freezing on day 2, compared with DMSO-treated Grin2b^C456Y/+ mice (Fig. 6, D and E). Similar CNO-dependent decreases in freezing were observed in Grin2b^C456Y/+ mice on days 6 and 30 (Fig. 6, D, F, and G). These results suggest that BA neuronal activation improves fear memory extinction and remote fear memory responses. BA neuronal activation in WT-CNO mice enhanced fear extinction on day 2 but not on days 6 and 30, compared with WT-DMSO mice (Fig. 6, D to G), likely because their baseline extinction is already optimal.

In control experiments, treatment with CNO alone in the absence of AAV-CaMKIIα-hM3Dq-mCherry injection had no effect on fear memory extinction and remote fear memory responses in both WT and Grin2b^C456Y/+ mice (fig. S13, A to E). Expression of mCherry alone without that of hM3Dq also did not affect fear memory extinction and remote retrieval in WT and Grin2b^C456Y/+ mice (fig. S13, F to J). In an additional control experiment, CNO treatment was confirmed to increase measures of neuronal excitability in hM3Dq-mCherry–expressing WT and Grin2b^C456Y/+ mice (fig. S14). Last, chemogenetic inhibition of BA neurons during fear memory acquisition failed to alter fear extinction or remote fear memory responses (fig. S15), suggesting that BA neuronal activation is required to rescue the mutant phenotype. These results suggest that activation of BA neurons improves the enhanced fear responses seen during fear extinction and remote retrieval phases in Grin2b^C456Y/+ mice while having little effect in WT mice.

Chemogenetic BA neuronal activation does not improve anxiety/social-related behaviors in Grin2b^C456Y/+ mice

We next tested whether chemogenetic activation of BA neurons could improve anxiety-like behaviors in addition to fear memory extinction and remote fear memory responses in Grin2b^C456Y/+ mice. CNO-dependent BA neuronal activation in Grin2b^C456Y/+ mice during fear memory extinction did not affect locomotor activity in the open-field test performed after remote fear memory retrieval (day 30), although the baseline hypoactivity of Grin2b^C456Y/+ mice was reproduced (fig. S16, A and B).

The following were also unaffected by BA neuronal activation in Grin2b^C456Y/+ mice: center time in the open-field test (a measure of anxiety-like behavior) (fig. S16, C and D); anxiety-like behaviors in elevated plus maze test (fig. S16, E to H); and social interaction and social novelty recognition (fig. S16, I to M). Baseline anxiety- and social-related measures in the virus-injected, DMSO-treated cohorts differed from those in our earlier shock extinction animals that received no surgery or injections (e.g., open-field center time). We attribute this shift to the cumulative impact of stereotaxic surgery, viral expression, and repeated systemic DMSO injections rather than to genotype or shock history. These results suggest that chemogenetic activation of BA neurons in Grin2b^C456Y/+ mice does not improve anxiety- and social-related behaviors, in contrast to the improvements seen in fear extinction and remote fear memory.

Chemogenetic activation of Grin2b^C456Y/+ BA neurons normalizes spontaneous excitatory synaptic transmission and neuronal excitability

To better understand how chemogenetic activation of Grin2b^C456Y/+ BA neurons improves fear memory extinction and remote fear memory responses, we attempted to measure baseline/spontaneous excitatory synaptic transmission (mEPSCs and sEPSCs) and neuronal excitability in BA neurons from Grin2b^C456Y/+ mice with shock extinction and chemogenetic activation during fear extinction (Fig. 7, A to D).

Fig. 7. Chemogenetic activation of BA neurons normalizes spontaneous excitatory synaptic transmission and neuronal excitability in Grin2b^C456Y/+ mice with shock extinction.

(A and B) Experimental scheme for AAV-CaMKIIα-hM3Dq-mCherry injection in the WT and Grin2b^C456Y/+ BA (~P15) followed by i.p. injection of CNO or DMSO during fear memory extinction for chemogenetic BA neuronal activation and measurements of spontaneous excitatory synaptic transmissions and neuronal excitability in WT and Grin2b^C456Y/+ BA neurons (male; ~P66) with shock extinction (day 1/P35 and days 2 to 6, respectively) without remote fear memory retrieval (day 30). (C) hM3Dq-mCherry expression in the BA. Scale bar, 200 μm. (D) Validation of fear acquisition and extinction in WT and Grin2b^C456Y/+ mice with shock extinction and CNO/DMSO (day1/P35 and days 2 to 6, respectively) without remote fear memory retrieval (day 30). [n = 10 (WT-DMSO), 8 (WT-CNO), 9 (HT-DMSO), and 8 (HT-CNO), two-way ANOVA with Sidak’s test]. (E to G) mEPSCs in WT and Grin2b^C456Y/+ BA neurons. [n = 16 neurons from three mice (WT-DMSO), 16 from four mice (WT-CNO), 13 from three mice (HT-DMSO), and 17 from three mice (HT-CNO), two-way ANOVA with Sidak’s test]. (H to J) sEPSCs in WT and Grin2b^C456Y/+ BA neurons. [n = 18 neurons from four mice (WT-DMSO), 14 from four mice (WT-CNO), 17 from four mice (HT-DMSO), and 21 from four mice (HT-CNO), two-way ANOVA with Sidak’s test]. (K to N) Currentspike number, currentmean firing rate, and currenttrain duration in WT and Grin2b^C456Y/+ BA neurons. [n = 16 neurons from four mice (WT-DMSO), 10 from four mice (WT-CNO), 14 from three mice (HT-DMSO), and 15 from four mice (HT-CNO), two-way ANOVA with Sidak’s test]. (O and P) Rheobase current in WT and Grin2b^C456Y/+ BA neurons. [n = 16 neurons from four mice (WT-DMSO), 10 from four mice (WT-CNO), 14 from three mice (HT-DMSO), and 15 from four mice (HT-CNO), two-way ANOVA with Sidak’s test]. (Q) Resting membrane potential (RMP) in WT and Grin2b^C456Y/+ BA neurons. [n = 16 neurons from four mice (WT-DMSO), 10 from four mice (WT-CNO), 14 from three mice (HT-DMSO), and 15 from four mice (HT-CNO), two-way ANOVA with Sidak’s test]. Significance is indicated as *P < 0.05, **P < 0.01, ***P < 0.001, or ns (not significant).

Chemogenetic activation had minimal effects on mEPSCs in BA neurons from WT or Grin2b^C456Y/+ mice with shock extinction (Fig. 7, E to G). In contrast, sEPSC frequency, which is decreased in BA neurons from Grin2b^C456Y/+ mice with shock extinction compared to those of WT mice with shock extinction (Fig. 4, G to I), was normalized by chemogenetic activation (Fig. 7, H to J). In addition, measures of neuronal excitability, which are decreased in BA neurons from Grin2b^C456Y/+ mice with shock extinction compared to those of WT mice with shock extinction (Fig. 5, A to M), were normalized by chemogenetic activation (Fig. 7, K to Q). These results suggest that chemogenetic activation of BA neurons from Grin2b^C456Y/+ mice with shock extinction normalizes sEPSC frequency and neuronal excitability at the time of remote fear memory retrieval, which may contribute the improved remote-fear memory responses.

DISCUSSION

Our study reveals that a mouse model of ASD carrying a patient-derived GluN2B-C456Y mutation (Grin2b^C456Y/+ mice) shows normal contextual fear memory acquisition but suppressed contextual fear memory extinction and enhanced remote fear memory responses along with anxiety- and social-related abnormal behaviors. The amygdalar BA neurons of Grin2b^C456Y/+ mice with shock extinction show occluded activation upon remote fear memory retrieval, which involves chronic development of suppressed spontaneous excitatory synaptic transmission and neuronal excitability. Chemogenetic activation of mutant BA neurons during fear memory extinction improves fear memory extinction and remote fear memory responses without affecting anxiety- and social-related behaviors. This rescue involves normalized spontaneous excitatory synaptic transmission and neuronal excitability in mutant BA neurons. Therefore, Grin2b^C456Y/+ mice display suppressed fear memory extinction that leads to chronic suppression of spontaneous excitatory synaptic transmission and neuronal excitability in BA neurons, which, upon remote fear memory retrieval, enhances fear responses.

Suppressed fear memory extinction and the amplified remote fear response in Grin2b^C456Y/+ mice are likely causally linked: Chemogenetic activation of BA neurons during fear extinction improves the remote fear memory phenotype. Mechanistically, shock extinction mutants chronically acquire deficits in BA spontaneous excitatory transmission (lower sEPSC frequency) and neuronal excitability that remain evident at remote retrieval and coincide with blunted c-fos activation. These deficits are also improved by chemogenetic activation of BA neurons during fear extinction. Because NMDAR function is unchanged, these findings implicate diminished basal synaptic activity and excitability, rather than NMDAR-dependent plasticity, as the principal drivers of the heightened remote fear response.

What initiates the extinction deficit? WT and Grin2b^C456Y/+ mice exhibit similar fear acquisition and comparable c-fos, synaptic, and excitability profiles in BA neurons shortly after fear acquisition, arguing against immediate activity differences. A more plausible cause could be impaired NMDAR-dependent LTD. Fear extinction relies on NMDAR-mediated plasticity in the amygdala (46, 58, 59, 63, 67), and Grin2b^C456Y/+ mice exhibit pronounced deficits in GluN2B-dependent LTD in hippocampus (43), consistent with the established role of GluN2B-containing NMDARs in LTD (20–22, 24, 27, 28) and synaptic pruning (68). If a similar LTD impairment occurs in BA neurons, then it would hinder the synaptic weakening normally required for extinction and leads to increased excitatory synaptic density, as supported by the increased mEPSC frequency in naïve mutant BA neurons. In further support of this view, Grin2b^C456Y/+ mice show sensory hypersensitivity combined with neuronal hyperactivity and hyperconnectivity in the anterior cingulate cortex (44). Thus, suppressed GluN2B-dependent LTD within BA circuits likely underlies the extinction failure.

Our work advances the field in three key ways. First, while prior studies showed that BA circuits gate fear memory extinction in WT mice (46, 48–51, 54, 56, 57, 69–72), we extend this principle to Grin2b^C456Y/+ mutants that model ASD- and PTSD-linked pathology. Second, we establish a causal chain, where suppressed extinction causes exaggerated remote fear, by chemogenetically restoring BA activity and simultaneously uncover underlying deficits in spontaneous excitatory transmission and neuronal excitability. Third, we identify previously unreported adaptations in WT mice: Shock extinction experience elevates BA mEPSC frequency and intrinsic excitability at remote retrieval. However, whether these synaptic changes shape long-term fear remains unclear; artificially boosting BA activity during extinction in WT mice improves extinction but leaves remote fear and the synaptic/neuronal parameters unchanged.

The IL region is an established hub for fear extinction (52, 53, 55, 56, 61, 65, 66, 73–80) and also shows altered activity in Grin2b^C456Y/+ mice. IL neurons in mutants with shock and extinction chronically develop reduced c-fos activation, whereas WT neurons do not. Yet, WT and mutant IL neurons display comparable shock extinction–induced changes in synaptic transmission and intrinsic excitability. These observations suggest that IL alterations are unlikely to drive the extinction deficit or the exaggerated remote fear memory in Grin2b^C456Y/+ mice.

Notably, enhancing BA activity during extinction improves the electrophysiological defects and fear-related phenotypes, yet it fails to correct the elevated anxiety-like and impaired social behaviors. This dissociation indicates that, while the BA is a common hub, fear extinction versus anxiety/social outcomes are likely mediated by separable, projection-specific BA microcircuits. Future studies that combine cell type– and projection-specific manipulations with circuit-level recording will be essential to pinpoint how discrete BA outputs differentially govern these distinct behavioral domains in ASD.

PTSD-related comorbidity has been increasingly reported in clinical studies (1–3), pointing to its importance for ASD-related social anxiety, fear, phobia, and avoidance. However, relatively few studies have addressed PTSD-like phenotypes and related neural mechanisms in animal models of ASD. Our Grin2b^C456Y/+ mice, which show clear fear extinction phenotypes (suppressed fear memory extinction and enhanced remote fear memory responses) reminiscent of PTSD symptoms (81) through specific mechanisms clarified at brain regional, synaptic, and neuronal levels, could serve as a useful model for the study of PTSD-related mechanisms in ASD. In summary, we demonstrate that Grin2b^C456Y/+ mice exhibit PTSD-like deficits in fear memory extinction and heightened remote fear responses, driven by chronically suppressed excitatory synaptic transmission and reduced neuronal excitability in basolateral amygdala neurons.

MATERIALS AND METHODS

Animals

Grin2b^C456Y/+ mice, carrying the patient-derived C456Y variant (31), were maintained on a congenic C57BL/6J background for >10 generations, as described previously (43, 44). WT C57BL/6J littermates were used as controls. For polymerase chain reaction genotyping of Grin2b^C456Y/+ mice, the following primers were used: forward: ACGACTCTTTGTGGAGGAGGG and reverse: CCATATCACAGCTTACTTCAATGT. All mice were bred and housed in ventilated cages where temperature and humidity were carefully controlled. Neonatal mice underwent toe clipping for labeling and genotyping at postnatal day 7 (P7), separated from their mothers, and weaned between P21 and P28. Following weaning, a maximum of six littermates of the same sex with mixed genotypes were housed together in groups before experiments. The animals were maintained under a 12-hour light-dark cycle (13:00 to 01:00) and fed ad libitum according to the Requirements of Animal Research at Korea Advanced Institute of Science and Technology (KAIST). All experimental procedures were approved by KAIST Institutional Animal Care and Use Committee (KA2020-90).

Behavioral assays

All behavioral assays were conducted using age-matched animals or littermates during the light phase of the circadian cycle, with experimenters blinded to the mouse genotype. Before behavioral assays, mice were handled for 3 consecutive days (10 min/day) and accustomed to the dark behavior room with white noise for at least 30 min before the test on the experimental day. Experimental conditions included a temperature range of 19° to 24°C, a humidity range of 40 to 55%, and the presence of white noise.

Contextual fear conditioning test

All contextual fear conditioning sessions were combined with a fear memory extinction session over 5 days following fear memory acquisition and a remote fear memory retrieval session 30 days after fear memory extinction. These experiments were conducted in a fear conditioning system with a Plexiglas chamber (20 cm by 20 cm by 30 cm) and a stainless steel grid floor (Coulbourn Instruments). The luminous intensity in the chamber was 0 lux with dim lights on the upper and middle parts of the wall. In the fear memory acquisition session, mice were allowed to freely explore the chamber for 30 s before being exposed to eight electric footshocks of 1.0-mA intensity, each lasting 2 s and spaced 28 s apart. In fear memory extinction and remote fear memory retrieval sessions, mice were placed in the same chamber without electric footshock for 5 min. Freezing levels were monitored by a video camera and analyzed using FreezeFrame 3 software with a threshold set at 10 (Actimetrics).

Recent and remote fear memory extinction

For recent fear memory extinction, mice were subjected to the fear memory acquisition (day 1) followed by fear memory extinction sessions 24 hours after fear memory acquisition (days 2 to 6) and remote fear memory retrieval session (day 30). For remote fear memory extinction, mice were subjected to the fear memory acquisition (day 1) followed by fear memory extinction sessions 30 days after fear memory acquisition (days 30 to 34) and remote fear memory retrieval session (day 60).

Open-field test

The open-field test was conducted to assess locomotor activity and anxiety-like behavior as reported previously (43). Mice were placed in the corner of the open-field box surrounded by white opaque acryl walls (40 cm by 40 cm by 40 cm) where mouse movements were recorded for 60 min. The center zone was defined as a square of 20 cm by 20 cm in the center of the open-field apparatus, with the light intensity set at 200 lux to induce anxiety. Mice were allowed to freely move and explore the open-field apparatus during recordings. A 60-min session was chosen because it encompasses both the initial novelty-driven exploration and the later habituation phase. The total distance moved and time spent in the center region of the open-field arena were analyzed using Ethovision XT 17 software (Noldus).

Elevated plus maze test

The elevated plus maze test was conducted to assess anxiety-like behavior. The apparatus was composed of two open arms (30 cm by 5 cm), a center zone (5 cm by 5 cm), and two closed arms which were surrounded by 30-cm-high walls (30 cm by 5 cm by 30 cm). The height of the apparatus was 75 cm from the floor, and the luminous intensity of open arms was 200 lux to induce anxiety. Mice were placed in the center zone and recorded for 10 min by a video camera. Time spent and entry number in each arm were analyzed using Ethovision XT 17 software (Noldus).

Three-chamber test

The three-chamber test (82) was conducted to assess social-related behaviors. It involved three sessions: habituation, social interaction, and social novelty recognition session. The chamber, surrounded by white opaque aryl walls, was composed of a central area (20 cm by 40 cm by 20 cm) and two side zones (20 cm by 40 cm by 20 cm), each separated from a central area by two transparent triangular walls featuring a smoothly curved hypotenuse. The luminous intensity of the chamber was 100 lux. Mice were placed in a central area, allowed to freely move and explore the chamber for 10 min during each session, and recorded by a video camera. Before the social interaction session, mice were shortly confined to a central area by opaque partitions following the habituation session. Meanwhile, a novel “object” and the first unfamiliar mouse “stranger 1” were placed in containers within each side zone, which were separated by stainless steel mesh. In this session, each test mouse was expected to distinguish between the inanimate object and the social target and to spend more time interacting with the social target. Before the social novelty recognition session, mice were again shortly confined to a central area by opaque partitions following the social interaction session. Meanwhile, the object was replaced with a second unfamiliar mouse, “stranger 2.” In this phase, the former stranger 1 now served as the familiar mouse, and each subject had to discriminate between this familiar mouse and a new stranger, thereby testing social novelty recognition. The used stranger mice were age-matched 129S4/SvJae (129Sv strain) mice. The social interaction index was calculated as (S1 − OB)/(S1 + OB), where S1 is the time spent with stranger 1 and OB is the time spent with the inanimate object. The social novelty recognition index was computed as (S2 − S1)/(S2 + S1), with S2 representing the time spent with the novel stranger. Time spent in each chamber and exploration time to object or “stranger” were analyzed using Ethovision XT 17 software (Noldus).

Immunohistochemistry

Immunohistochemistry was conducted to analyze c-fos expression and virus-mediated marker protein expression. For the analysis of c-fos expression immediately after shock or extinction, mice were subjected to either contextual fear conditioning (day 1) (after shock) or contextual fear conditioning (day 1) followed by fear memory extinction (days 2 to 6) (after extinction). Naïve juvenile mice were used as the control group. For the analysis of c-fos expression during remote fear memory retrieval, mice were subjected to contextual fear conditioning (day 1), followed by fear memory extinction (days 2 to 6) with remote fear memory retrieval (experimental group) or without retrieval (control group) (day 30). After resting in their home cages for 90 min on day 1, day 6, or day 30, respectively, mice were anesthetized with isoflurane (Terrell) and underwent intracardial perfusion with heparinized phosphate-buffered saline, followed by 4% paraformaldehyde (PFA). Brains were then extracted from perfused mice and immersed in 4% PFA overnight at 4°C for postfixation. After postfixation, brains were sliced into 50-μm-thick sections using a vibratome (VT1200s, Leica). Brain sections were incubated overnight with primary antibodies against c-fos [rabbit monoclonal antibody (mAb), 2250S, Cell Signaling, 1:1000] and Neuronal Nuclei (NeuN; mouse mAb, MAB377, Millipore, 1:500) at 4°C and with secondary antibodies [anti-rabbit Alexa 594 (1:500) for c-fos and anti-mouse Alexa 488 (1:500) for NeuN] for 3 hours at room temperature. Slice images were captured using a slide scanner (Axio Scan.Z1) and analyzed using the Automated Mapping of Single Neurons (AMaSiNe) program (83). To visualize virus-mediated protein expression, brains were extracted from the mice that completed behavioral experiments. Brain slices were incubated with primary antibodies for mCherry (rabbit polyclonal antibody, ab183628, Abcam, 1:500) and CaMKIIα (mouse mAb, SC-13141, Santa Cruz Biotech, 1:500) and secondary antibodies [anti-rabbit Alexa 594 (1:500) for mCherry and anti-mouse Alexa 488 (1:500) for CaMKIIα].

Ex vivo electrophysiology

The ex vivo electrophysiology was conducted to measure the spontaneous synaptic transmissions, neuronal intrinsic excitability, NMDAR/AMPAR-mediated excitatory postsynaptic currents (EPSCs), and evoked AMPAR-mediated EPSCs in male mice aged postnatal weeks 8 to 11 (adult) or postnatal weeks 5 to 6 (juvenile).

Brain slices

Brains were extracted from anesthetized mice by isoflurane and coronally sliced into 300-μm-thick sections using a vibratome (VT1200s, Leica) in a sucrose-based dissection buffer maintained at 0°C. The dissection buffer comprised 75 mM sucrose, 76 mM NaCl, 25 mM NaHCO₃, 2.5 mM KCl, 1.25 mM NaH₂PO₄, 25 mM d-glucose, 7 mM MgSO₄, and 0.5 mM CaCl₂, continuously bubbled with 95% O₂ and 5% CO₂. Subsequently, the sliced brain sections were incubated in a prewarmed 32°C recovery chamber containing the dissection buffer for 20 min to facilitate recovery. Following the recovery period, the brain sections were transferred to a holding chamber filled with artificial cerebrospinal fluid (aCSF) at room temperature and allowed to be further recovered for at least 30 min. The aCSF composition included 124 mM NaCl, 26.2 mM NaHCO₃, 2.5 mM KCl, 1 mM NaH₂PO₄, 20 mM d-glucose, 1.3 mM MgCl₂, and 2.5 mM CaCl₂, also bubbled with 95% O₂ and 5% CO₂.

Whole-cell patch-clamp recording

For whole-cell patch-clamp recordings on principal neurons of the BA region and pyramidal neurons in the IL layer 2/3 region, sliced brain sections were positioned on a recording stage with a circulating aCSF solution maintained at 26° to 28°C. The recording and stimulation glass pipettes used in the measurement were crafted from thin-walled borosilicate capillaries (30-0065, Harvard Apparatus) with a resistance of 2.0 to 4.0 megohm using a two-step vertical micropipette electrode puller (PC-10, Narishige): Temperatures initially set at 73.6°C followed by set at 45° to 60°C. The target cells were visualized under a differential interference contrast illumination in an upright microscope (B50WI, Olympus) and were contacted with a pipette filled with an internal solution. To facilitate cell membrane rupture, a gigaohm seal was established between the cell and the pipette. After cell rupture, the electrophysiological signals of the cells were filtered at 2 kHz and digitized at 10 kHz using the MultiClamp 700B Amplifier (Molecular Devices) and the Digidata 1550 Digitizer (Molecular Devices). It was monitored whether the access resistance of target cells exceeded 20 megohm during whole-cell patch clamping procedures; if it exceeds the threshold, the corresponding data were not included in the analysis.

For mEPSCs, sEPSCs, NMDAR/AMPAR-mediated EPSCs, and evoked AMPAR- mediated EPSCs, recording pipettes were filled with an internal solution containing 130 mM CsMeSO₄, 10 mM TEA-Cl, 10 mM Hepes, 10 mM EGTA, 4 mM Mg–adenosine triphosphate (ATP), 0.3 mM Na–guanosine triphosphate (GTP), and 0.5 mM QX-314 with pH 7.35, 295 mOsm. In addition, for NMDAR/AMPAR-mediated EPSC and evoked AMPAR-mediated EPSC recording, the stimulation pipettes were filled with aCSF. To measure the mEPSCs, 1 μM tetrodotoxin (TTX) (Tocris) and 100 μM picrotoxin (PTX) (Sigma-Aldrich) were added to the aCSF to prevent spontaneous action potential–mediated synaptic currents and inhibitory postsynaptic currents (IPSCs), respectively. To measure sEPSCs, only 100 μM PTX was added to the aCSF. The holding potentials of target cells for mEPSCs and sEPSCs were set at −70 mV. To measure the NMDAR/AMPAR-mediated EPSCs, 10 μM SR-95531 (Sigma-Aldrich) was added to the aCSF to prevent IPSCs. AMPAR-EPSCs were measured at the holding potential of −70 mV, while the holding potential was shifted to +40 mV to measure NMDAR-EPSCs. EPSCs were elicited at 15-s intervals, and 20 consecutive responses were recorded. The NMDAR component was assessed at 50 ms after the onset of stimulation, and the NMDAR/AMPAR ratio was calculated by dividing the mean value of NMDAR-EPSCs by the mean value of AMPAR-EPSC peak amplitudes. To determine the decay kinetics (tau constant, τ) of NMDAR-EPSCs, each trace of NMDAR-EPSCs was fitted with the following exponential curve formula: F(t) = Ae − t/τ + C (t is the time, C is an arbitrary constant, and A is the amplitude of NMDAR-mediated current). To measure the evoked AMPAR-mediated EPSCs, 100 μM PTX and 50 μM D(-)-2-Amino-5-phosphonopentanoic acid (D-AP5; Tocris) were added to aCSF to prevent IPSCs and NMDAR-EPSCs, respectively. The evoked AMPAR-EPSCs were measured at the holding potential of −70 mV. EPSCs were elicited at 10-s intervals, and 15 consecutive responses were recorded. The stimulus intensity was increased stepwise from 5 to 20 μA in intervals of 5 μA.

To record mIPSCs and sIPSCs, recording pipettes were filled with an internal solution containing 115 mM CsCl, 10 mM TEA-Cl, 8 mM NaCl, 10 mM Hepes, 10 mM EGTA, 4 mM Mg-ATP, 0.3 mM Na-GTP, and 5 mM QX-314-Cl with pH 7.32, 285 mOsm. To measure mIPSCs, 1 μM TTX, 10 μM 2,3-dihydroxy-6-nitro-7-sulfamoyl-benzo[f]quinoxaline (NBQX; Tocris), and 50 μM D-AP5 were added to aCSF to prevent spontaneous action potential–mediated synaptic currents, AMPAR-mediated EPSCs, and NMDAR-mediated EPSCs, respectively. To measure sIPSCs, 10 μM NBQX and 50 μM D-AP5 were added to aCSF. The holding potentials of target cells for mIPSCs and sIPSCs were set at −70 mV.

To record the intrinsic excitability of neurons, recording pipettes were filled with an internal solution containing 135 mM K-gluconate, 7 mM NaCl, 10 mM Hepes, 0.5 mM EGTA, 2 mM Mg-ATP, 0.3 mM Na-GTP, and 10 phosphocreatine with pH 7.35, 295 mOsm. To prevent postsynaptic currents, 100 μM PTX, 10 μM NBQX, and 50 μM D-AP5 were added to aCSF. Target cells were initially maintained at a voltage-clamped state with a holding potential of −70 mV. Subsequently, currents were clamped, and resting membrane potentials were measured. To stabilize the membrane potential around −70 mV, minimal currents were applied. Afterward, to induce neuronal excitation, a series of currents ranging from −300 to 600 pA were incrementally injected (50 pA per sweep) with 10-s intervals. To measure rheobase current, a series of currents ranging from 50 to 440 pA were incrementally injected (10 pA per sweep) with 2-s intervals.

To validate the effects of CNO-dependent activation in the principal neurons of the BA region by the measurement of neuronal intrinsic excitability, virus-transfected BA neurons were initially identified by expressed mCherry signal (red-fluorescence protein). Following the initial measurement of excitability in target cells using the same method described above, the buffer (containing 100 μM PTX, 10 μM NBQX, and 50 μM D-AP5 in aCSF) was supplemented with 10 nM CNO (C0832, Sigma-Aldrich) and circulated for at least 15 min to activate virus-transfected BA neurons. Subsequently, neuronal intrinsic excitability was measured in the same target cells. The data were acquired using Clampex 11.0.3 (Molecular Devices) and analyzed using Clampfit 11.0.3 (Molecular Devices) and Minhee Analysis Package.

Virus reagents

All viruses were obtained from Addgene. The pAAV-CaMKIIα-hM3D(Gq)-mCherry virus and pAAV-CaMKIIα-hM4D(Gi)-mCherry virus were provided by B. Roth (Addgene viral prep #50476-AAV9 and #50477-AAV9), and the pAAV-CaMKIIα-mCherry virus was provided by K. Deisseroth (Addgene viral prep #114469-AAV9).

Stereotactic surgeries and virus injection

The stereotactic surgeries and virus injections were conducted to activate the principal neurons of the BA region through the DREADD system. The virus was injected into mice aged between P14 and P16 under sterilized conditions. Initially, mice were anesthetized with a mixture of isoflurane and O₂ (4.5% v/v; total flow 44.6 mmol per min, corresponding to 1 liter per min at 0°C and 1 atm; O₂ flow 2.0 mmol per min) for induction. Subsequently, anesthetized mice were placed onto a stereotactic stage (model 940, Kopf Instruments) and head-fixed by maintaining under respiratory anesthesia (2.5%, v/v). After exposing the scalp, the bregma-lambda axis was aligned, and the virus was bilaterally injected into the BA regions using a 33-gauge NanoFil needle (NF33BL-2, World Precision Instruments). The exact injection coordinates were set at anterior-posterior (AP) = −1.3, medial-lateral (ML) = ±2.75, and dorsal-ventral (DV) = −4.2, with a bregma-lambda length of 3.8 mm. The injection rate was 50 nl/min, administered four times at 15-s intervals, totaling a volume of 200 nl. Upon completion of the injection, the needle was carefully removed, and mice were allowed to recover under an infrared irradiator.

Chemogenetic rescue

Mice were bilaterally injected with AAV-CaMKIIα-hM3D(Gq)-mCherry virus or AAV-CaMKIIα-hM4D(Gi)-mCherry virus into the BA region. Following 3 weeks to facilitate virus activation and mouse recovery, a contextual fear conditioning test was conducted. To induce DREADD-mediated chemogenetic neuronal activation during the fear memory extinction session, mice were administered CNO (0.5 mg/kg) via intraperitoneal injection at a volume of 5 ml/kg, 30 min before the start of each fear memory extinction session. To induce DREADD-mediated chemogenetic neuronal inhibition during the fear memory acquisition session, mice were administered CNO (0.5 mg/kg) via intraperitoneal injection at a volume of 5 ml/kg, 30 min before the start of fear memory acquisition session. For the control groups, DMSO (D4540, Sigma-Aldrich) was administered instead of CNO.

Statistical analysis

All statistical analyses were performed using GraphPad Prism 7 and 9 software (RRID: SCR_002798). For the comparison of two samples, both D’Agostino and Pearson and Shapiro-Wilk normality tests were used to determine the normality of data distribution. If the data show a Gaussian distribution, then a parametric unpaired Student’s t test was applied for the analysis. If the data show a Gaussian distribution but have significantly different variances based on the F test, then a parametric Welch’s t test was used. For data with non-Gaussian distribution, a nonparametric Mann-Whitney U test was used. For the analysis of data with two independent variables, two-way ANOVA with Sidak’s multiple comparisons test was used. When comparing two matched groups within the same sample, a parametric paired t test was used for data with Gaussian distribution, while a nonparametric Wilcoxon matched-pairs signed rank test was used for data without Gaussian distribution. Data are shown as mean ± SEM. Outliers of the data were identified and removed by Grubbs’ test (α = 0.05). The numbers of samples (n) are indicated in figure legends. Statistical significance is denoted at each figure as follows: *P < 0.05, **P < 0.01, ***P < 0.001, or ns (not significant).

Supplementary Materials

The PDF file includes:

Figs. S1 to S16

Legend for table S1

Other Supplementary Material for this manuscript includes the following:

Table S1