Background
Autism spectrum disorder (ASD) is a set of heterogeneous neurodevelopmental alterations characterized by persistent difficulties in verbal and nonverbal communication with restricted and repetitive patterns of behavior [
1]. ASD can be diagnosed during childhood and affects 3 times more males than females [
2]. The basis for the male bias in ASD is unknown with theories including the “extreme male brain”, hormonal differences, and genetic influences [
3]. For instance, increased levels of testosterone have been correlated with autistic traits in toddlers and young children [
4]. Male bias in ASD and intellectual disability have also been associated with rare mutations at chromosome Xp22.11 but that affect less than 1% of patients [
5]. In the absence of biological markers, ASD is currently diagnosed based on clinical scales focused on behavioral symptoms. There is presently no cure for ASD but only symptomatic relief to some of its comorbidities such as anxiety, sleep disorders, or seizures [
6].
ASD etiology remains elusive, and both genetic and environmental components have been identified [
7]. Genetic factors have long been suspected to be involved in ASD, as this is a highly heritable psychiatric disorder [
8]. Hundreds of genetic mutations have been associated with ASD in genes such as neurexin (NRXN), neuroligin (NLGN), SH3 and multiple ankyrin repeat domains proteins (SHANK), tuberous sclerosis (TSC)1/2 and fragile X mental retardation (FMR)1, each one contributing to only a small percentage of the disease [
9]. Although these mutations only account for approximately 20–30% of ASD cases, many of them seem to converge towards synaptic dysfunction as they are implicated in the formation, the function, and maintenance of synapses, mainly glutamatergic [
10,
11]. Based on these findings, the current working hypothesis is that synaptic dysfunction would lead to functional and cognitive impairments observed in ASD [
12].
Several genetic mutations linked to these genes have been modeled in mice and were shown to induce behavioral symptoms reminiscent of ASD: deficits in social interaction, impairments in communication, and dysfunctional behaviors [
12]. Among these mutations,
Shank3 has received a great deal of attention this last decade [
13‐
15]. Shank (also known as ProSAP) proteins consist of three major isoforms (Shank1, Shank2, and Shank3), and are master scaffolding proteins at excitatory synapses [
13,
14]. They interact with more than 30 synaptic proteins through multiple domains and are essential for synaptic formation, glutamate receptor trafficking, and neuronal signaling.
The
Shank3 gene has 22 exons that give rise to numerous mRNA and protein isoforms deriving from multiple intragenic promoters and alternative splicing [
11,
13]. Mouse models bearing mutations or deletions in these isoforms have been generated and have been extensively studied this last decade [
16]. Although many of these animal models exhibit autism-associated behaviors, such as social deficits and repetitive behaviors, they do so to different extents and with a great deal of variability in terms of neuropathology and behavior [
17‐
27]. These myriad outcomes can be linked to (1) the existence of several
Shank3 animal models in relation with the targeted isoform, (2) lack of systematic use of heterozygote mice, whereas several
Shank3 mutations are in fact a haploinsufficiency in clinical settings, (3) lack of systematic use of female mice in some of these studies and even mixing females with males within the same group of mice, (4) and lack of standardized and robust behavioral procedures.
Here, we aimed at investigating the consequences of a
Shank3 mutation on behavioral parameters that we previously investigated in two different environmental animal models of ASD [
28,
29]. These models were obtained following prenatal injection of either valproic acid (Depakine), an anti-epileptic drug targeting the GABA neurotransmission system, or a double stranded RNA (poly IC) mimicking a viral infection and inducing a maternal immune activation. We previously reported that these models recapitulate several ASD symptoms, although at different degrees and magnitude, with a clear sexual dimorphism. Using a full-scale behavioral analysis, we have shown in the VPA model that males and females were differently affected by the treatment with males showing motor, gait and social deficiencies while females showing only motor and gait deficiencies with normal social behavior. At the cellular level, decreased number of cerebellar Purkinje cells (PC) did not affect the same cerebellar sub-regions in males and females. Similar results were obtained with the poly IC injection, although at lower magnitude, where males were also more affected by the treatment than females. These findings recapitulate the spectrum distribution of ASD that affect differently patients in magnitude and in symptomatology with a clear sexual dimorphism.
In line, we investigate here behavioral, cellular, and molecular consequences of lack of
Shank3 expression [
17]. In this animal model Shank3 is lacking the entire C-terminal region, including the Homer-binding site in the sterile alpha motif (SAM) domain. This results in a partial or total loss of the major naturally occurring isoforms of Shank3 proteins in heterozygotes and homozygotes, respectively. This mutation was chosen as it has strong construct validity given that it mimics a human mutation, which is not the case of several other
Shank3 mutations in mice. Major behavioral deficits were reported in these mice, including social preference deficits and repetitive behaviors, but not always using both males and females or Shank3+/ΔC and Shank3 ΔC/ΔC mice [
22‐
25]. Here we used wild type, Shank3+/ΔC, and Shank3 ΔC/ΔC mice, and we analyzed males and females separately. A special focus was made on cerebellar-related parameters as we have previously shown that motor dysfunction and gait deficits are strongly correlated with the degree of severity of ASD in terms of social symptoms and loss of PC. Standardized behavioral procedures were used and were previously validated in two environmental models of ASD using a prenatal injection of either valproic acid (VPA) or the double-stranded RNA Poly: IC, inducing a maternal immune activation (MIA) [
28,
29].
We report here that loss of major Shank3 isoforms induces several social, motor, and gait behavioral phenotypes reminiscent of ASD symptoms in humans. These are accompanied by a restricted decrease in both cerebellar PC cell number and mGluR5 levels. While social interaction deficits were observed in Shank3 ΔC/ΔC and Shank3+/ΔC, in both males and females, other parameters investigated affected mostly male Shank3 ΔC/ΔC mice.
Methods
Design, and setting
Shank3+/ΔC mice with a C-terminal deletion of
Shank3 at exon 21 (Shank3tm1.1Pfw/J) were obtained from Jackson laboratories and are on a C57BL/6J genetic background (Jackson Laboratories, USA); they were originally donated by Paul Worley from Johns Hopkins University School of Medicine [
17]. Shank3+/ΔC mice were mated to generate all mice of all three genotypes: WT, Shank3+/ΔC, and Shank3 ΔC/ΔC. Sex and age-matched pups were separated from the dams on postnatal day 28 (P28) and were then genotyped and raised by groups of 4 of the same genotype in a randomized fashion to avoid litter effects. All litters of the three genotypes and two sexes were used for this study. Outliers and mice that did not perform a given test appropriately (not finishing the walking beam, jumping out of the apparatus, stalling or reversing during gait…) were taken out from the study. Comprehensive behavioral screening was performed between postnatal days 50 and 60 (P50 and P60) as we previously described in two environmental ASD mouse models [
28,
29]. Mice were tested during their light cycle and the investigators were blind to genotype. At the end of the behavioral experiments, mice were sacrificed and brains were harvested for histological analysis, RT-PCRq, or western blotting experiments.
Animals
Animal housing and experimental procedures were performed under the European Union directive (2010/63/EU) and validated by the regional ethical committee (Approval # 2015020415093780). Mice were housed in ventilated cages with access to food and water ad libitum. Room temperature was maintained at 23 °C on a 12 h light/dark cycle. Offspring were segregated into 6 different groups: wild type males (n = 47), Shank3+/ΔC males (n = 71), Shank3 ΔC/ΔC males (n = 27), wild type females (n = 39), Shank3+/ΔC females (n = 58) and Shank3 ΔC/ΔC females (n = 26). A set of these mice was used for every procedure, as detailed below.
Behavioral analysis
All behavioral tests were performed as we previously described in detail in two different environmental ASD mice models [
28,
29].
Social interaction was assessed using the three-chambers test (3-CT) in an apparatus that consists of a Plexiglass box (60 × 45 × 22 cm) partitioned into three chambers with retractable doorways. Mice were monitored using an unbiased automated video tracking system (VideoTrack, Viewpoint-France) which calculates several parameters for each observed animal such as position, speed, trajectory, and various other behaviors set up by the experimenter. Data were retrieved using the VideoTrack 3.2 software (Viewpoint-France). Mice were habituated to the 3-CT for 20 min 2 days before phase 1 that consists of two identical non-social stimuli (inverted wire-cups) placed in the opposite chambers. The second phase comprises a non-social stimulus and a social stimulus (a naïve mouse with no previous contact with the tested mouse). The naïve mouse is of the same sex and age and is usually used for two tests per day at the most, with a 2–3 h rest between tests; no mouse is used more than 2 consecutive days. In the third phase, an additional and novel mouse was placed in the cup present in the opposite chamber. After a habituation phase of 10 min in the central chamber, mice underwent the three phases of the experiment, with each phase being of 10 min each during which time spent in each chamber and around each cup was recorded. Mice used for this procedure were Wild type males (n = 13), Shank3+/ΔC males (n = 20), Shank3 ΔC/ΔC males (n = 18), wild type females (n = 14), Shank3+/ΔC females (n = 12), Shank3 ΔC/ΔC females (n = 16).
Spontaneous grooming behavior was scored in mice placed individually in a clean and transparent cylinder with house bedding. Each mouse was videotaped and rated for 10 min. Cumulative time spent grooming and grooming frequency were evaluated. Wild type males (n = 20), Shank3+/ΔC males (n = 12), Shank3 ΔC/ΔC males (n = 7), wild type females (n = 12), Shank3+/ΔC females (n = 12), Shank3 ΔC/ΔC females (n = 13).
Motor coordination was evaluated using the challenging beam test and which consists of four Plexiglass sections (25 cm each) starting with a width of 3.5 cm and gradually narrowed to 0.5 by 1 cm decrements. Mice were first trained for 2 days to traverse the beam, starting at the widest section and ending at the narrowest section that led into the home cage. On the test day, a mesh grid (1 cm squares) was placed over the beam surface. Mice were videotaped while traversing the grid-surfaced beam for five trials. Mice that stalled during the beam walk or tried to reverse course were not included in the final analysis. Time to traverse, errors, number of steps, and errors per step made by each mouse were measured and averaged. Wild type males (n = 10), Shank3+/ΔC males (n = 20), Shank3 ΔC/ΔC males (n = 14), wild type females (n = 14), Shank3+/ΔC females (n = 17), Shank3 ΔC/ΔC females (n = 12).
Gait was analyzed during spontaneous walk using an automated gait analysis system (Gaitlab, Viewpoint, France). The apparatus is made of a 1.5 m long glass corridor with dim green light beamed into the glass walkway. Each mouse was assessed individually for 3 consecutive runs. Mice that stalled during the beam walk or tried to reverse course were not included in the final analysis. All gait parameters were analyzed with a special focus on (1) stride length: distance between two consecutive placements of the same paw, (2) limb base of support: distance between two pair prints at contact during each step cycle, and (3) pair gap: gap between the placement of the two trailing feet, which measures spatial coordination between the two pairs. Wild type males (n = 11), Shank3+/ΔC males (n = 27), Shank3 ΔC/ΔC males (n = 15), wild type females (n = 11), Shank3+/ΔC females (n = 21), Shank3 ΔC/ΔC females (n = 12).
Tissue processing and immunohistochemistry
Males and females from each genotype were randomly selected for either histological, mRNA, or protein analysis. For histological analysis, mice (wild type males (
n = 8), Shank3+/ΔC males (
n = 11), Shank3 ΔC/ΔC males (
n = 8), wild type females (
n = 8), Shank3+/ΔC females (
n = 9), Shank3 ΔC/ΔC females (
n = 6)) were deeply anesthetized with ketamine-xylazine (120–20 mg/kg) and transcardially perfused with 0.9% saline at 37 °C followed by 4% paraformaldehyde (PFA) at 4 °C. Brains were post-fixed in 4% PFA at 4 °C for 24 h before cryoprotection in 30% sucrose for 48 h. Serial 50 µm (cerebellum) free-floating sections were collected and stored at − 20 °C until use in an anti-freeze solution. Every fourth cerebellar section was mounted on gelatin-coated slides for PC quantification. PC were identified based on their morphology on cresyl violet staining. The PC phenotype was confirmed using calbindin immunohistochemistry (1:2500; Swant, Cb-38a) [
30]. Stereological estimates were performed using the optical fractionator method and systematic random sampling to obtain the total number of cerebellar PC. Each region of interest was outlined based on the Franklin and Paxinos’s mouse brain atlas [
31] at 2.5 × objective. Neurons were counted at 40 × objective using the Mercator image analysis system (Explora Nova, France). Upper and lower guard zones of 1 µm were set at the top and bottom of the section to exclude lost profiles and each neuron or visible nucleus was counted as previously described [
28,
29].
Quantitative immunoblot analysis
Mice from each group [wild type males (
n = 4), Shank3+/ΔC males (
n = 8), Shank3 ΔC/ΔC males (
n = 8), wild type females (
n = 5), Shank3+/ΔC females (
n = 8), Shank3 ΔC/ΔC females (
n = 9)] were sacrificed, brains were retrieved, and cerebellum region was dissected out and frozen in – 80 C until use. For control purposes, we also used cerebella obtained from offspring of pregnant mice that received VPA injections at E12.5, as previously described [
28]. Proteins were extracted using a 1% sodium dodecyl sulfate (SDS) solution in Tris Hcl 0.1 M with EDTA 0.01 M and PMSF, protease inhibitor, and phosphatase inhibitor at 1%. Equal amounts of proteins from cerebellum lysates were separated by SDS-PAGE and migrated proteins were transferred to nitrocellulose membranes (Bio-Rad). After blocking the membranes at room temperature for 1 h in Tris Buffer Solution with Tween-20 0,1 M (TBST) and (pH 7.4, TBS) 5% non-fat milk, the blots were incubated with corresponding primary antibodies at room temperature for 3 h. The blots were then washed three times in TBST and incubated with HRP-conjugated secondary antibodies overnight at 4 °C. Following 3 washes in TBST, the blots were incubated with ECL reagent (GE Healthcare Life Sciences, NJ, USA). For quantification, the films were scanned by a PXi image system and gray signals were analyzed by GeneTools software (Syngene, Cambridge, UK), and normalized to that of corresponding internal controls (actin or β-tubulin III). The primary antibodies mTOR 7C10, P-mTOR ser 2448, ERK1/2, and P-ERK1/2 (Cell Signaling Technology, Leiden, The Netherlands) were used at the dilution 1/1000 to 1/1500. NR2A, NR2B, GluR1, and GAPDH antibodies (Millipore, Paris, France) were used at the dilution of 1/500 to 1/1500. The following dilutions were used for antibodies against CAMKII (1/1000), α-tubulin (1/10,000), and β-actin (1/10,000) (Sigma-Aldrich, Lyon, France). Shank3 and mGluR5 antibodies (Abcam, Paris, France) were used at the dilution of 1/500 and 1/5000, respectively. NR1 antibodies (Antibodies Incorporated, California, USA) were used at the dilution of 1/500. GluR2 antibodies (Neuromab, California, USA) were used at the dilution of 1/1000.
RT-PCRq analysis of mGluR5 mRNA levels
Mice from each group (wild type males (n = 7), Shank3+/ΔC males (n = 8), Shank3 ΔC/ΔC males (n = 8), wild type females (n = 7), Shank3+/ΔC females (n = 8), Shank3 ΔC/ΔC females (n = 8)) were sacrificed, brains were retrieved, and cerebellum region was dissected out and frozen in -80C until use. Total RNA was isolated using TRIzol Reagent/chloroform, then purified using a NucleoSpin RNA kit (Macherey-Nagel) and quantified using NanoDrop ND-1000 spectrophotometer (Thermo Fisher Scientific, USA). Reverse transcription was performed on 1 μg of total RNA for each sample using Verso cDNA Synthesis kit (Thermo Fischer Scientific). qPCR was performed on LightCycler 480 system (Roche Diagnostics, France). Ct values were averaged from triplicates. Results were subtracted to the mean Ct of the housekeeping genes Gapdh, Beta actin and cyclophilin A (ΔCt) and log transformed (2−ΔCt). mGluR5 primers sequence was designed using Primer-Blast (NCBI); primers were synthesized by IDT: 5′-aggacagataaaggtgatcc-3′ and5′-agatactggactgggatcaa-3′.
Statistical analyses
Data are expressed as mean ± Standard Error of the Mean (SEM) and analyzed using GraphPad Prism-7 software (La Jolla, California, USA). Data followed normal distribution as evidenced by the Shapiro–Wilks normality test with the hypothesis for normality rejected at a p value less or equal to 0.05. Data were analyzed using Student’s t test, one-way or two-way analysis of variance (ANOVA) when relevant and as stated for each experiment. Tukey’s or Fisher’s LSD post-hoc multiple comparisons were applied. Outliers were identified using the Grubbs statistical test with alpha at 0.05. For all analyses, a p value < 0.05 was considered significant.
Discussion
This study aimed at investigating several behavioral, cellular, and molecular outcomes related to ASD in Shank3+/ΔC and Shank3 ΔC/ΔC male and female mice with a mutation that is also observed in human conditions.
The
Shank3 gene encodes a master scaffolding protein in the postsynaptic density where it interacts with multiple key synaptic components, mainly implicating the glutamate receptor clusters, along with the cytoskeleton and signal transduction cascades [
10,
26,
43,
44,
46]. The gene is located on chromosome 22q13.3 in humans and was first implicated in ASD in the 22q13.3 microdeletion syndrome, also known as Phelan-McDermid Syndrome [
13]. Haploinsufficiency of the
Shank3 gene accounts for 0.5–2.0% of ASD and intellectual disability cases [
16,
47]. Human genetic studies of ASD have found that the C-terminal region (exon 21) of
Shank3 harbors several mutations [
16]. This region encodes binding sites for actin/Cortactin and mGluR5/Homer and plays a crucial role in the synaptic targeting and postsynaptic assembly of Shank3 scaffolding proteins [
48]. A dozen different genetically alerted mice with various
Shank3 point mutations or deletions of a given exon have been constructed [
17‐
19,
24,
26,
42,
49]. These mice show variable outcomes at the behavioral, cellular, and molecular levels.
We have set up our behavioral, cellular, and molecular analysis in a
Shank3 mouse model [
17] based on our previous and recently reported findings in two different environmental mouse models of ASD [
28,
29]. These models were obtained following prenatal injection of either VPA or poly IC and recapitulate several ASD symptoms, although at different degrees and magnitude, with a clear sexual dimorphism.
Here, we show that a
Shank3 mutation affects mostly male homozygote mice in several social and motor behavioral parameters. This is accompanied by a significant reduction in cerebellar PC and specific decreased levels of mGluR5 proteins. We first investigated social preference parameters, as they constitute one of the hallmarks of ASD [
50]. We report that Shank3+/ΔC and Shank3 ΔC/ΔC males and females both have normal social preference and motivation but show impairments in social novelty preference. Of interest is the fact that ASD patients are reported to avoid unfamiliar social partners and display diminished interest in novelty [
50]. Indeed, ASD patients tend to avoid social contacts with a new individual compared to someone familiar [
51].
Self-grooming is a core behavior for mice that spend about 40% of their waking time performing it. It is aimed to maintain physiological stasis and comfort and thus tends to increase under stressful situations and in pathological conditions implicating motor brain centers as reported in clinical and animal models of ASD [
33]. Here, we replicated findings in these animal models showing increased grooming, reflective of stereotyped repeated behavior. We further show that this increased grooming affects both Shank3 ΔC/ΔC males and females.
In the initial paper using this line of mice [
17], authors have investigated a wide range of behavioral parameters including motor-coordination deficits, grooming and novelty avoidance, but only in homozygote mice. Of interest is the observation that homozygote mice showed social preference in the phase 2 of the 3-CT but no preference for social novelty (phase 3), a finding that we replicated here both in males and females. In the report by Duffney et al., (2015), Shank3+/ΔC did show social preference in phase 2 trial but that was of a lesser magnitude than in wild type animals [
18]. Qin et al. [
19] used the 3-CT with only two phases and showed no social preference in Shank3+/ΔC mice, mixing measurements of ‘investigating behavior” and time spent in the area surrounding the cup with the social cue. Using a similar mutation leading to a premature stop codon in exon 21 of
Shank3, Speed et al. [
24] found no social interaction deficits in both heterozygote or homozygote mouse mutants. Also in this paper, homozygote mice, but not heterozygotes, showed deficits in hippocampus-dependent spatial learning, impaired motor coordination, altered response to novelty, and sensory processing deficits. Jaramillo et al. [
22] used a different approach to generate Shank3 mutant mice with a loss of the two higher molecular weight isoforms by disrupting the PDZ domain with a transcriptional stop cassette prior to exon 13. Homozygote mice with this mutation, but not heterozygotes, displayed a preference for the social target over the inanimate object (phase 2), a finding that we replicated here. In the phase 3 of the 3-CT, neither heterozygote nor homozygote mice displayed preference to social novelty as also we show here for homozygote mice. In addition, interaction time and approach were altered only in homozygote animals, in line with our findings of altered social behavior only in homozygotes. In the grooming behavior, both heterozygotes and homozygotes displayed increased time spent grooming compared to wild type, but only homozygotes showed increased grooming bouts, a finding that we also replicate in the
Shank3 mouse models that we used. Although apparently conflicting, collectively these results show alterations in grooming and social behavior in
Shank3 mutant mice, which are often present only in homozygotes or at least with a much higher degree than in heterozygote mice.
Several reports in ASD described impairments in visio-motor and manual dexterity tasks, limb coordination during tasks requiring balance, agility, and speed, as well as in gait and ataxia [
52]. Furthermore, motor impairments may be amongst the earliest signs of some forms of ASD and their assessment might help the early and quantitative diagnosis of the pathology and the identification of dysfunctional brain regions and circuits in ASD [
53]. Here, we found only mild motor coordination impairments in this mouse model. These impairments were revealed only when the task required significant challenge (a very narrow beam with a grid mech above it). This is in line with our findings with a Poly IC environmental model where no deficiencies were observed in this task [
29].
One of the most robust and consistent findings in our previous studies and here are anomalies in gait. Early studies in ASD patients have revealed irregular gait [
54]. Quantitative measures indicated increased variability and irregularity in gait suggestive of cerebellar, rather than basal ganglia, involvement. Later studies reported increased missteps, increased step width, and higher ataxia ratios similar to what is found in patients with cerebellar lesions [
55]. In fact, several spatiotemporal gait parameters of children with ASD were similar to those of patients with cerebellar ataxia [
56]. Of interest is the fact that abnormal gait patterns seem to be correlated with the severity of social impairments in both humans [
57] and ASD animal models [
28]. Here, we have used an automated rodent quantitative gait analysis system with optical touch sensors, which records and tracks rodents’ footprints as they move freely in a walkway. This allowed qualitative and quantitative readout of footprint pattern and contacts, highlighting gait abnormalities. As nearly all gait parameters are correlated, a collection of these parameters is needed to determine the potential alterations’ nature. Gait is made up of strides that comprise a stance phase with the foot in contact with the ground and a swing phase with the foot off the ground. Fundamental descriptors of gait can be divided into two main sections: (1) spatial patterns that include stride length, step length, and step width and (2) temporal parameters including duty factor (limb stance time divided by stride time) and limb phase (the time between forelimb and hindlimb foot-strike on the same side divided by the stride time). We report here that several temporal parameters were altered and that these alterations were mainly found in Shank3 ΔC/ΔC mice and to a wider extend in males. Interestingly, spatial patterns did not seem to be altered by the mutation (stride length did not differ between groups), indicating no major anatomical alterations in these mice. However, all temporal parameters measured were reduced in Shank3 ΔC/ΔC mice (stance time, swing time, stride frequency, and duty factor). Along with these temporal deficiencies, there were also alterations of the way paws are applied to the surface in terms of surface contact and strength, and that were also reduced mainly in both males and females Shank3 ΔC/ΔC. Also of interest is the fact that more gait parameters were affected in males than in females. The temporal and surface alterations observed are probably not due to change in velocity as only females, but not males, Shank3 ΔC/ΔC mice showed a mild increase (less than 20%) in velocity; in addition, there was no change in the regularity of gait nor the stride length. Altogether, these gait elements argue against an anatomical alteration and a primary motor disorder implicating the basal ganglia. In contrast, they seem to point to a cerebellar dysfunction as they mainly affect movement coordination, a cerebellar function, which is in line with clinical findings (see for instance reference [
56]).
Gait disturbances, and more generally behavioral abnormalities in neurodevelopmental disorders including ASD, are a consequence of underlying alterations in circuit maturation and function. Post-mortem and brain imaging studies have consistently identified the cerebellum as one of the most abnormal brain regions associated with ASD with a specific reduction of cerebellar PC [
40,
41]. Here we show a decreased number of PC within the hemispheric part of lobule VII, within the in Crus I and Crus II subregions and only in Shank3 ΔC/ΔC males. Females showed only a mild decrease in PC number in Shank3+/ΔC. This is in line with VPA and Poly IC models showing also decreases of PC in these brain areas and that were more pronounced in males than in females. Cerebellar defects seem to be widespread in ASD mouse models. Using five different ASD mouse models, including the one used in this study, Kloth et al. [
49] have shown major cerebellar associative sensory learning deficits with the eyeblink conditioning task, a task that relies on cerebellar plasticity. Further on, mice with a Tsc1 mutation specifically in cerebellar cells PC showed ASD-related altered behavior that was rescued with rapamycin treatment targeting mTOR signaling further pointing to these cells as key players in ASD [
42].
In this line, we investigated the levels of dozens of cerebellar proteins implicated in the cerebellar glutamatergic transmission and signal transduction. In accordance with previous findings in cortical or hippocampal areas, we did not detect any difference in the cerebellar expression level of several glutamatergic related proteins [
18,
19,
24]. We further extend these findings and show here a dramatic decrease in the levels of cerebellar mGluR5 protein, and only in male Shank3 ΔC/ΔC, pointing to that receptor as the major target of Shank3 and perhaps even the major component underlying the observed behavioral deficits linked to this mutation. These decreased protein levels were probably due to post-translational regulation as we did not detect any alterations in the corresponding mRNA levels. Of interest is also the fact that such a decrease was not observed in the VPA environmental ASD mouse model. Shank3 is essential to mediating mGluR5 signaling by recruiting Homer1b/c to the postsynaptic density [
17,
18,
44]. Pharmacological enhancement of mGluR5 activation in
Shank3 knock-out mice, through the administering 3-Cyano-
N-(1,3-diphenyl-1H-pyrazol-5-yl)benzamide, ameliorated functional and behavioral defects [
58], suggesting that pharmaceutical treatments that increase mGluR5 activity may represent an interesting approach for treating ASD patients with a
Shank3 mutations [
18,
59,
60]. In this line, restoration of Shank3 expression in adult using a
Shank3 conditional knock-in procedure in mice, rescued several ASD phenotypes through improvements in synaptic protein composition, spine density, and neural function [
27]. This is of interest as it further emphasis the major role of Shank3 in postsynaptic function and indicates that this neurodevelopmental pathology may in fact be manageable at adult age in patients with
Shank3 mutations.
Shank3 mutations in ASD patients are either point mutations in one copy of the
Shank3 gene or haploinsufficiencies where a single functional
Shank3 allele is insufficient to maintain a normal behavior [
14,
61]. Thus, one would have expected that Shank3+/ΔC mice would show clear behavioral alterations relevant to ASD. In our hands however, Shank3+/ΔC mice showed deficiencies in the social novelty recognition task and did not show other behavioral or cellular dysfunction, despite over a 50% decrease in mGluR5 expression. This is in line with previous publications with this or similar mouse model [
17,
24] but at odds with other previous reports exploring behavioral and cellular consequences only in males [
18,
19].
Several studies utilizing
Shank3 mutations or deletions do not report measurable changes in animal behavior. For instance, exons 4–9 of
Shank3 (JAX 017890) show only mild ASD-related phenotype and have normal sociability and social novelty in the 3-chamber test [
61]. Mutations targeting exons 4–7 only yielded mild behavioral phenotypes, while targeting exons 13–16 (JAX 017688) showed profound phenotypes such as impaired sociability and preference for social novelty, as well as reduced pair interaction, and profound self-grooming leading to skin lesions [
62]. Targeting exon 11 induced prominent self-grooming and skin lesions [
63], while targeting exon 21 of
Shank3 (JAX 018398) results in normal sociability but impaired social novelty in the 3-chamber test [
17], as observed in our study here.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.