Perirhinal cortex abnormalities impair hippocampal plasticity and learning in Scn2a, Fmr1, and Cdkl5 autism mouse models

Learning and memory deficits, including spatial navigation difficulties, are common in Autism Spectrum Disorder (ASD), yet the neurobiological underpinning of this phenotype remains unknown. Several ASD mouse models (Scn2a+/- , Fmr1-/-, Cdkl5-/-) exhibit impaired spatial learning, with these deficits often attributed to hippocampal dysfunction. However, we identify the perirhinal cortex (PRC) as a critical driver of these deficits. Cortical-wide Scn2a reduction in excitatory neurons replicated the spatial learning and long-term potentiation (LTP) impairments— a cellular correlate of learning—seen in Scn2a+/- mice, while hippocampal-wide reduction did not. PRC-specific viral-mediated Scn2a reduction in excitatory neurons decreased release probability, which consequently disrupted synaptic transmission and LTP in the hippocampus, as well as spatial learning. As PRC activity was reduced, chemogenetic activation of the PRC reversed these deficits in Scn2a+/- mice and rescued spatial learning and LTP impairments in Fmr1 and Cdkl5 knockout mice. Thus, in several genetic models of ASD, PRC abnormalities may disrupt hippocampal function to impair learning and memory.

Methods

Slice Preparation

 Using standard methods, horizontal brain sections (350 µM) containing hippocampus and perirhinal cortex were collected from mice aged postnatal day 21 (surgically naïve mice) to postnatal day 70 (mice which underwent viral injection and behavior testing). For slice preparation from surgically naïve mice, the cutting solution contained (in mM): 85 NaCl, 2.5 KCl, 1.25 NaH2PO4-H2O, 3 MgSO4·7H2O, 0.5 L-ascorbic acid, 0.5 CaCl2, 75 sucrose, 25 D-(+)-glucose, and 25 NaHCO3. For slice preparation from mice which underwent viral injection, the cutting solution contained (in mM): 2.5 KCl, 1.25 NaH2PO4-H2O, 0.4 L-ascorbic acid, 3 MgSO4·7H2O, 200 sucrose, 20 D-(+)-glucose, and 26 NaHCO3. Both solutions were neutral pH (i.e., 7.3), 310 mOsm, and saturated with 95% O2 and 5% CO2. Once cut, slices were transferred to an incubation chamber (Campden Instruments, Model 7470) filled with oxygenated artificial cerebrospinal fluid (ACSF) (Standard Ringer’s Solution, 300 mOsm, in mM: 119 NaCl, 1.3 MgSO4-7H2O, 2.5 KCl, 1 NaH2PO4-H2O, 2.5 CaCl2, 11 D-(+)-glucose, and 26.2 NaHCO3, pH 7.4) and then incubated at 32°C for 30 minutes. Slices were kept at room temperature for 30 minutes before transfer to the recording chamber (30°C oxygenated ACSF, 2 – 3 mL/min). Heating was maintained with LinLab 2 (Scientifica, Version 1.0.19.50). 

Field Potential Recording Electrophysiology

Recordings utilized bipolar electrodes for stimulation and 2-6 megaohm micropipettes containing ACSF for recording. For perirhinal cortex, the stimulating electrode was positioned into layer II/III on the temporal side of the rhinal sulcus, while the recording electrode was placed in the same layer 400 µm away in the rostral direction (54). For hippocampal recordings, stimulation was applied to Schaffer collaterals from area CA3, and stratum radiatum in area CA1 was recorded. A stimulus generator (Model DS3, Digitimer Ltd.) delivered constant-current square pulses (0.1 ms, 25 - 300 µA), and signals were lowpass filtered at 1 kHz. Parameters were recorded and analyzed utilizing the Clampex suite (Molecular Devices, Version 11.1) Field excitatory postsynaptic potentials (fEPSPs) were recorded every 10 seconds. Input-output curves were determined by measuring the fiber volley (FV) and fEPSP slope at 25 µA, 50 µA, 100 µA, 200 µA and 300 µA for 10 sweeps. Stimulus intensity was adjusted to induce a half-maximal fEPSP amplitude. Paired pulse ratios were determined by calculating the ratio of the slope of the test EPSP (second response) to the conditioning EPSP (first response) at interstimulus intervals of 15, 25, 50, 100, and 200 milliseconds (ms). For long-term potentiation, a baseline was established by recording responses every 30 seconds for fifteen minutes before applying theta-burst stimulation (TBS: four trains every 15 seconds, each train comprising 10 bursts of 5 pulses at 100 Hz, inter-burst interval 150 ms). After TBS, responses were recorded every 30 seconds for 60 min. For studies involving hM3Dq, 1 µM CNO was bath applied to activate this designer receptor. FV amplitude and fEPSP slope were compared using a two-way repeated measures ANOVA followed by a Tukey post-hoc test across genotypes and stimulus intensities to derive input-output curves. A two-way repeated measures ANOVA with a Fisher’s LSD post-hoc test was used to assess the effect of genotype and interstimulus interval on the paired-pulse ratio. Baseline and post-LTP induction responses across genotypes and time were also compared by repeated measures ANOVA. As reported in figure captions, n represents the number of slices recorded, followed by the number of animals utilized. Significance was defined as p-values < 0.05, with α = 0.05.