Behavioral decline in Shank3Δex4-22 mice during early adulthood parallels cerebellar granule cell glutamatergic synaptic changes

Abstract

Background: SHANK3, a gene encoding a synaptic scaffolding protein, is implicated in autism spectrum disorder (ASD) and is disrupted in Phelan-McDermid syndrome (PMS). Despite evidence of regression or worsening of ASD-like symptoms in individuals with PMS, the underlying mechanisms remain unclear. Although Shank3 is highly expressed in the cerebellar cortical granule cells, its role in cerebellar function and contribution to behavioral deficits in ASD models are unknown. This study investigates behavioral changes and cerebellar synaptic alterations in Shank3^Δex4-22 mice at two developmental stages.

Methods: Shank3^Δex4-22 wildtype, heterozygous, and homozygous knockout mice lacking exons 4-22 (all functional isoforms) were subjected to a behavioral battery in both juvenile (5-7 weeks old) and adult (3-5 months old) mouse cohorts of both sexes. Immunostaining was used to show the expression of Shank3 in the cerebellar cortex. Spontaneous excitatory postsynaptic currents (sEPSCs) from cerebellar granule cells (CGCs) were recorded by whole-cell patch-clamp electrophysiology.

Results: Deletion of Shank3 caused deficits in motor function, heightened anxiety, and repetitive behaviors. These genotype-dependent behavioral alterations were more prominent in adult mice than in juveniles. Reduced social preference was only identified in adult Shank3^Δex4-22 knockout male mice, while self-grooming was uniquely elevated in males across both age groups. Heterozygous mice showed little to no changes in behavioral phenotypes in most behavioral tests. Immunofluorescence staining indicated the presence of Shank3 predominantly in the dendrite-containing rosette-like structures in CGCs, colocalizing with presynaptic markers of glutamatergic mossy fiber. Electrophysiological findings identified a parallel relationship between the age-related exacerbation of behavioral impairments and the enhancement of sEPSC amplitude in CGCs.

Limitations: Other behavioral tests of muscle strength (grip strength test), memory (Barnes/water maze), and communication (ultrasonic vocalization), were not performed. Further study is necessary to elucidate how Shank3 modulates synaptic function at the mossy fiber-granule cell synapse in the cerebellum and whether these changes shape the behavioral phenotype.

Conclusions: Our findings reveal an age-related exacerbation of behavioral impairments in Shank3^Δex4-22 mutant mice. These results suggest that Shank3 may alter the function of glutamatergic receptors at the mossy fiber-cerebellar granule cell synapse as a potential mechanism causing cerebellar disruption in ASD.

Keywords: AMPAR; Autism spectrum disorder; Cerebellum; Glutamate receptor; Granule cell; Mouse behavior; Phelan-McDermid syndrome; Shank3.

Fig. 1

Shank3^Δex4–22 KO mice display greater levels of anxiety with age. (A) Representative heatmaps of time spent in each area of the open field arena for one mouse of each genotype, age, and sex. (B-I) Individual (circles) and mean ± SEM (black bars) of the number of center entries (B, C), total open field center time (D, E), total freezing duration during open field exploration (F, G), and total number of fecal boli at the end of open field exploration (H, I) for each genotype at both ages (B, D, F, and H) and further separated by sex (C, E, G, and I). (J) Representative heatmaps of time spent in each area of the zero maze for one mouse of each genotype, age, and sex. (K-N) Individual animal (circles) and group mean ± SEM (black bars) of total open arm time (K, L) and the number of open arm entries (M, N) in the elevated zero maze for each genotype at both ages (K, M) and further separated by sex (L, N). For A and J, the color scale bar at right applies to all heatmaps in the corresponding assay. N = 22–30 mice/group for each genotype at each age and N = 10–16 mice/group for each sex within each genotype at each age. *p < 0.05 for post-hoc test between genotypes with a Bonferroni correction

Fig. 2

Juvenile and adult Shank3^Δex4–22 KO mice display reduced exploratory and locomotion behavior. (A-C) Mean ± SEM of the total distance moved within each 5 min period of open field exploration for each genotype at both ages (A) and further separated into males (B) and females (C) at each age. (D-G) Individual animal (circles) and group mean ± SEM (black bars) of the total cumulative distance moved (D, E) and maximal linear movement velocity detected (F, G) during open field exploration for each genotype at both ages (D, F) and further separated by sex (E, G). N = 22–30 mice/group for each genotype at each age and N = 10–16 mice/group for each sex within each genotype at each age. In open field distance-time plots (A-C), symbols correspond to p < 0.05 in post-hoc comparison of genotypes within age and/or sex: * WT-KO, ‡ Het-KO, and § WT-Het, while *p < 0.05 in scatter dot-mean plots (D-G) in post-hoc comparisons between genotypes, all with a Bonferroni correction

Fig. 3

Shank3^Δex4–22 KO mice develop motor function deficits with age. (A-F) Mean ± SEM of the time until the mouse rotates completely around the rotarod (A-C) or falls to the landing platform (D-F) for three subsequent accelerating rotarod tests (4–40 RPM, 5 min) repeated over two days total for each genotype at both ages (A, D) and further separated into males (B, E) and females (C, F) at each age. (G-N) Individual (circles) and mean ± SEM (black bars) of the time to cross (G, H, K, and L) and the number of left and right total foot slips (I, J, M, and N) on a 6 mm wide (G-J) and 12 mm wide (K-N) beam for each genotype at both ages (G, I, K, and M) and further separated by sex (H, J, L, and N). N = 20–35 mice/group for each genotype at each age and N = 10–18 mice/group for each sex within each genotype at each age. In rotarod time plots (A-F), symbols correspond to p < 0.05 in post-hoc comparison of genotypes within age and/or sex: * WT-KO, ‡ Het-KO, and § WT-Het, while *p < 0.05 in scatter dot-mean plots (G-N) in post-hoc comparisons between genotypes, all with a Bonferroni correction

Fig. 4

Shank3^Δex4–22 KO mice develop an elongated stride length as juveniles. (A) Sample gait analysis raw data with location, stride length, and width of the forelimb identified in blue and hindlimb in red. (B-E) Individual (circles) and mean ± SEM (black bars) of the forelimb stride length and width (B, C) and the hindlimb stride length and width (D, E) for each genotype at both ages (B, D) and further separated by sex (C, E). N = 22–25 mice/group for each genotype at each age and N = 10–14 mice/group for each sex within each genotype at each age. *p < 0.05 for post-hoc test between genotypes with a Bonferroni correction

Fig. 5

Loss of Shank3 increases repetitive behavior in male mice and decreases exploratory behavior in juvenile mice of both sexes. (A, B) Individual (circles) and mean ± SEM (black bars) of the total duration of grooming time during open field exploration for each genotype at both ages (A) and further separated by sex (B). (C) Representative images of marble location after 30 min in the marble burying arena for one mouse of each genotype, age, and sex. (D-F) Mean ± SEM of the number of marbles buried after each 5 min period during the marble burying assay for each genotype at both ages (D) and further separated into males (E) and females (F) at each age. (G, H) Individual (circles) and mean ± SEM (black bars) number of marbles buried after 30 min for each genotype at both ages (G) and further separated by sex (H). (I-L) Individual animal (circles) and group mean ± SEM (black bars) of the total number of arm entries (I, J) and percent of alternations (K, L) in the Y-maze for each genotype at both ages (I, K) and further separated by sex (J, L). N = 20–31 mice/group for each genotype at each age and N = 10–18 mice/group for each sex within each genotype at each age*p < 0.05 for post-hoc test between genotypes with a Bonferroni correction

Fig. 6

Shank3^Δex4–22 KO mice develop reduced social preference with age in the three-chamber sociability assay. (A) Representative heatmaps of time spent in each area of the three-chamber arena for one mouse of each genotype, age, and sex. The color scale bar at the center applies to all heatmaps in the A. (B-E) Individual (circles) and mean ± SEM (black bars) of the social preference index (B, C) and social novelty index (D, E) for each genotype at both ages (B, D) and further separated by sex (C, E). N = 23–33 mice/group for each genotype at each age and N = 9–17 mice/group for each sex within each genotype at each age. *p < 0.05 for post-hoc test between genotypes with a Bonferroni correction

Fig. 7

Shank3 is expressed at CGC dendrites around VGlut1- and VGlut2-positive mossy fibers (MF) terminals. (A, B) Confocal fluorescence images at 4x (A) and 20x (B) demonstrating expression of VGlut1 (cyan) and VGlut2 (magenta) throughout the internal granule cell layer at mossy fiber terminals. (C-H) Grayscale (C-E) and pseudocolor (F-H) 60x confocal fluorescence single-plane images of the same image location in the internal granule cell layer in parasagittal sections labeled with VGlut1, VGlut2, and Shank3. Fluorescence color is assigned to enhance contrast in comparing magenta and green. (F-H) Shank3 is expressed around VGlut1- and VGlut2-expressing terminals. (I) Example of how VGlut1-positive (white, top) terminals were used to define ROIs (yellow) for Shank3 colocalization analysis. (J) Individual (circles) and mean ± SEM (bars) Mander’s coefficient for each analyzed image reflect similar colocalization of Shank3 at VGlut1- and VGlut2-expressing mossy fibers. VGlut1-expressing (281 terminals) and VGlut2-expressing (285 terminals) were evaluated from n = 10 images per mossy fiber marker from N = 5 C57BL/6J mice

Fig. 8

MF–CGC sEPSC amplitude is augmented in adult mice lacking Shank3. (A, B) Representative (20 s) traces of CGC sEPSCs (in 10µM gabazine) recorded from two genotypes each juvenile (A, black, gray) and adult (B, red, light red) wildtype (+/+) and knockout (-/-) Shank3^Δex4–22 mice. Each trace is from a different CGC. (C-F) Cumulative distribution histograms for all events for each group with corresponding inset individual (circles) and mean ± SEM (bars) for sEPSC amplitudes (C, E) and interevent intervals (D, F) from juvenile (C, D) and adult (E, F) wildtype (+/+) and knockout (-/-) Shank3^Δex4–22 mice. (G) Average normalized (to total event number) distribution histogram for sEPSC values from each CGC with the Gaussian fit of the averaged distribution provided in the inset (H). n = 19–22 cells/genotype from N = 11–14 adult mice and n = 11–14 cells/genotype from N = 6–9 juvenile mice. For comparison of mean group sEPSC values (C-F insets) or averaged sEPSC histogram bin percentages (G) *p < 0.05 for t-test between genotypes