The Effect of ASIC3 Knockout on Corticostriatal Circuit and Mouse Self-grooming Behavior

Abstract

Stereotypic and/or repetitive behavior is one of the major symptoms of autism spectrum disorder (ASD). Increase of self-grooming behavior is a behavioral phenotype commonly observed in the mouse models for ASD. Previously, we have shown that knockout of acid-sensing ion channel 3 (ASIC3) led to the increased self-grooming behavior in resident-intruder test. Given the facts that ASIC3 is mainly expressed in the peripheral dorsal root ganglion (DRG) and conditional knockout of ASIC3 in the proprioceptors induced proprioception deficits. We speculate a hypothesis that stereotypic phenotype related to ASD, pararalled with striatal dysfunction, might be caused by proprioception defect in the peripheral sensory neuron origin. Herein, we investigate in depth whether and how ASIC3 is involved in the regulation of self-grooming behavior. First, we observed that Asic3 null mutant mice exhibited increased self-grooming in social interaction during juvenile stage. Similarly, they displayed increased self-grooming behavior in a novel cage in the absence of cagemate. To further understand the mechanism by which ASIC3 affects grooming behavior, we analyzed neurochemical, neuropathological and electrophysiological features in the dorsal striatum of Asic3 null mutant mice. Knockout of Asic3 increased dopamine (DA) activity and phospho-ERK immunoreactivities in the dorsal striatum. Furthermore, we detected a lower paired-pulse ratio (PPR) and impaired long-term potentiation (LTP) in corticostriatal circuits in Asic3 null mutant mice as compared with wild-type (WT) littermates. Moreover, knockout of Asic3 altered the medial spiny neurons in the striatum with defects in presynaptic function and decrease of dendritic spines. Lastly, genetic ablation of Asic3 specifically in parvalbumin-positive (PV⁺) cells resulted in the increase of self-grooming behavior in mice. These findings suggest knockout of Asic3 in the PV⁺ neurons alters grooming behavior by co-opting corticostriatal circuits.

Keywords: ASIC3; corticostriatal circuit; parvalbumin; proprioception; self-grooming.

Figure 1

Increased self-grooming behavior during social interaction in Asic3^−/− mice. (A) Asic3^−/− mice displayed higher self-grooming and rearing behavior during juvenile social interaction compared to Asic3^+/+ mice. Rearing: t = 4.699, df = 18, p < 0.0002; Self-grooming: t = 3.286, df = 18, p < 0.0041. N = 10 mice per group. **p < 0.01; ***p < 0.001. (B) Cross-fostering the Asic3^−/− mice to wild-type (WT) dam normalized the heightened rearing behavior in Asic3^−/− mice during social interaction, but did not change the level of grooming behavior. bp, biological parents; cf, cross-fostering. Rearing: genotype: F_(1,35) = 5.940, p = 0.020; Cross-fostering: F_(1,35) = 2.065, p = 0.160; interaction: F_(1,35) = 15.444, p < 0.001; Post hoc analysis: **p < 0.01 vs. Asic3^+/+(bp), ^##p < 0.01 vs. Asic3^−/−(bp); Self-grooming behavior: genotype: F_(1,35) = 18.001, p < 0.001; Cross-fostering: F_(1,35) = 3.371, p = 0.075; interaction: F_(1,35) = 0.212, p = 0.648; Post hoc analysis: **p < 0.01 vs. Asic3^+/+(bp), *p < 0.05 vs. Asic3^+/+(cf). N = 9–10 mice per group. Data represent mean ± SEM. Data analyzed by two-tailed unpaired t-test (A) two-way analysis of variance (ANOVA; B).

Figure 2

Increased self-grooming behavior in novel cage in Asic3^−/− mice. (A) The self-grooming behavior was increased in Asic3^−/− mice born to heterozygous dam (Asic3^−/−(he)) and knockout dam (Asic3^−/−(ko)). Data represent mean ± SEM. Genotype: F_(2,39) = 4.109, p = 0.0240; Time: F_(1,39) = 1.958, p = 0.1697; interaction: F_(2,39) = 0.4408, p = 0.6467; Post hoc analysis: *p < 0.05 Asic3^+/+ vs. Asic3^−/−(he), ^#p < 0.001 Asic3^+/+ vs. Asic3^−/−(ko). (B) No change in the rearing behavior in the novel cage in Asic3^−/− mice. (C) No change in the distance traveled in the novel cage in Asic3^−/− mice. N = 13–15 mice per group. Data represent mean ± SEM. Data analyzed by two-way ANOVA.

Figure 3

Measurement of dopamine (DA) activity in brain regions in Asic3^−/− mice. (A) The DA activity [(DOPAC+HVA)/DA] was increased in striatum in Asic3^−/− mice as compared with Asic3^+/+ mice: t = 2.532, df = 18, p < 0.0209; *p < 0.05. No difference was detected between Asic3^+/+ and Asic3^−/− mice in (B) frontal cortex, (C) hippocampus, (D) hypothalamus, (E) midbrain, and (F) brainstem. N = 10 mice per group. Data represent mean ± SEM. Data analyzed by two-tailed unpaired t-test.

Figure 4

Phosphorylated extracellular signal-regulated kinase 1/2 (pERK) activity in dorsomedial striatum and bed nucleus of the stria terminalis (BNST) of Asic3^−/− and Asic3^+/+ mice. (A) Representative images of pERK activation in striatal caudate putamen (CPu) in Asic3^+/+ and Asic3^−/− mice. The pERK activity measured by pERK staining of brain sections revealed enhanced activation in striatum in Asic3^−/− mice. (B) Quantification of pERK⁺ cells in CPu. The pERK⁺ cells were significantly increased in CPu in Asic3^−/− mice as compared with Asic3^+/+. N = 3–5 mice per group. (1.1~0.5mm, Asic3^+/+ n = 36 slices, Asic3^−/− n = 36 slices, t = 7.191 df = 70; 0.38~-0.1mm, Asic3^+/+ N = 24 slices, Asic3^−/− N = 18 slices, t = 7.661 df = 40; −0.22~-0.58mm, Asic3^+/+ n = 16 slices, Asic3^−/− n = 14 slices, t = 5.012 df = 28, ***p < 0.001). (C) Representative images of pERK activation in BNST in Asic3^+/+ and Asic3^−/− mice. (D) Quantification of pERK⁺ cells in BNST. The pERK⁺ cells were significantly increased in BNST in Asic3^−/− mice as compared with Asic3^+/+ mice. N = 3–5 mice per group (0.62~0.13mm, Asic3^+/+ n = 22 slices, Asic3^−/− n = 22 slices, t = 3.900 df = 42; 0.02~-0.22mm, Asic3^+/+ n = 14 slices, Asic3^−/− n = 10 slices, t = 3.538 df = 22; -0.34~-0.58mm, Asic3^+/+ n = 12 slices, Asic3^−/− n = 10 slices, **p < 0.01, ***p < 0.001). Data represent mean ± SEM. Data analyzed by two-tailed unpaired t-test.

Figure 5

Electrophysiology properties of corticostriatal circuit in Asic3^−/− and Asic3^+/+ mice. (A) Recording setup of corticostriatal field excitatory post-synaptic potentials (fEPSCs) in sagittal brain slices. The recording electrode was placed in the dorsal striatum, and a bipolar stimulating electrode was placed in the corpus callosum to evoke fEPSPs. (B) fEPSP evoked in cortico-striatal synapses, and fEPSP amplitude plotted as a function of stimulus intensity showing gradually growing responses with increasing intensity. Inset shows a typical field recording. (C) Tracings represent the paired pulse response at 50-ms interval. (D) The averaged paired-pulse ratios (PPRs) of Asic3^+/+ and Asic3^−/− mice in 50- and 100-ms intervals (N = 6–7 mice per group). Paired pulse facilitation was significantly lower in Asic3^−/− mice than Asic3^+/+ mice (50 ms: t = 2.99935, df = 11, p = 0.00605; 100 ms: t = 2.04147, df = 11, p = 0.03297; ^#p < 0.05; ^##p < 0.01). (E) High-frequency stimulation (HFS) of neocortical afferents could induce long-term potentiation (LTP) in the brain slices of Asic3^+/+ animals. The cortico-striatal LTP was impaired in Asic3^−/− slices. Upper panel shows the representative tracings before (in left) and after (in right) HFS of Asic3^+/+ and Asic3^−/− mice. (F) At 30 min after HFS, the mean fEPSP value was 84 ± 10% of pre-HFS values in Asic3^−/− mice and 169 ± 13% of pre-HFS values in Asic3^+/+ mice (Asic3^+/+: pre-HFS vs. HFS: t = 5.30428, df = 5, p = 0.00318; Asic3^+/+ vs. Asic3^−/−: t = 5.08298, df = 10, p = 0.00605, p < 0.001; N = 6–7 mice per group. **p < 0.01 vs. pre-HFS, ^###p < 0.001 vs. Asic3^+/+). Data are mean ± SEM. Data analyzed by two-tailed unpaired t-test.

Figure 6

Reduced corticostriatal synaptic transmission in Asic3^−/− medium spiny neurons (MSNs). (A) Representative traces of mEPSC from Asic3^+/+ and Asic3^−/− MSNs recorded with whole-cell voltage clamp. Reduced mEPSC: (B) the frequency and (C) the amplitude of mEPSC were reduced in Asic3^−/− MSNs as compared with Asic3^+/+ MSNs (frequency: t = 2.27013, df = 16, p = 0.01869; amplitude t = 2.16085, df = 16, p = 0.02311; N = 8–10 mice per group; ^#p < 0.05 vs. Asic3^+/+). Data are mean ± SEM. Data analyzed by two-tailed unpaired t-test.

Figure 7

The effect of Asic3 knockout on dendritic spines of striatal medium spiny neurons (MSNs). (A) The dendritic spine density of MSNs was lower in Asic3^−/− mice than in Asic3^+/+ mice (Asic3^+/+ n = 108 spines from six mice, Asic3^−/− n = 83 spines from five mice, t = 2.670 df = 189, p = 0.0082 by two-tailed t-test). **p < 0.01 vs. Asic3^+/+. (B) No difference in MSNs cell body size between two genotypes (Asic3^+/+ n = 118 spines from six mice, Asic3^−/− n = 92 from five mice). Data are mean ± SEM. Data analyzed by two-tailed unpaired t-test.

Figure 8

The gene expression of Asic subtypes in the cortex and striatum of Asic3^+/+ and Asic3^−/− mice. The gene expression of Asic1a, 1b, 2a, 2b, 3 and 4 were analyzed by quantitative real-time polymerase chainreaction (qRT-PCR) in each brain region in each genotype. (A) Compared to the Asic3^+/+ control for each gene, Lower expression of Asic1a and Asic3 were detected in the cortex of Asic3^−/− mice. No difference was detected in other Asic subtypes between Asic3^+/+ and Asic3^−/− mice in the cortex. N = 3 mice per group. *p < 0.05 vs. Asic3^+/+. (B) Compared to the WT Asic1a control for each genotype. The distribution of Asic subtypes in the cortex of Asic3^+/+ and Asic3^−/− mice. N = 3 mice per group. (C) Compared to the Asic3^+/+ control for each gene, Lower expression of Asic3 were detected in the striatum of Asic3^−/− mice. No difference was detected in other Asic subtypes between Asic3^+/+ and Asic3^−/− mice in the striatum. N = 3 mice per group. *p < 0.05, **p < 0.01 vs. Asic3^+/+. (D) Compared to the WT Asic1a control for each genotype. The distribution of Asic subtypes in the striatum of Asic3^+/+ and Asic3^−/− mice. N = 3 mice per group. Data are mean ± SEM. Data analyzed by two-tailed unpaired t-test.

Figure 9

Deletion of Asic3 in parvalbumin-positive (Pv⁺) neurons increased grooming behavior. (A) Pv-Cre⁺/Asic3^f/f mice showed more grooming behavior than Pv-Cre⁻/Asic3^f/f mice (genotype: F_(1,14) = 1.738, p = 0.2085; time: F_(1,14) = 7.431, p = 0.0164; interaction: F_(1,14) = 3.517, p = 0.0817; Post hoc analysis: *p < 0.05 0–5 min vs. 6–10 min in Pv-Cre⁺/Asic3^f/f). N = 8 mice per group. No difference was found in (B) rearing and (C) distance traveled behaviors between in Pv-Cre⁺/Asic3^f/f and Pv-Cre⁻/Asic3^f/f mice. No difference was found in (D) grooming behavior, (E) rearing, and (F) distance traveled behaviors between Na_v1.8-Cre⁺/Asic3^f/f and Na_v1.8-Cre⁻/Asic3^f/f mice. N = 7–9 mice per group. Data are mean ± SEM. Data analyzed by two-way ANOVA.