Research ArticleMUSCLE PHYSIOLOGY

Autism-associated SHANK3 mutations impair maturation of neuromuscular junctions and striated muscles

See allHide authors and affiliations

Science Translational Medicine  10 Jun 2020:
Vol. 12, Issue 547, eaaz3267
DOI: 10.1126/scitranslmed.aaz3267

Autism and muscles

Mutations in SHANK3 are associated with autism spectrum disorders (ASDs) and Phelan-McDermid syndrome (PMDS). Children with ASD or PMDS present neurodevelopmental abnormalities and skeletal muscle hypotonia; the mechanisms mediating the decreased muscle tone are unclear. Now, Lutz et al. used patient-derived material and a mouse model to show that SHANK3 is expressed in muscle sarcomeres and plays a role in the maturation of the neuromuscular junctions (NMJs). Muscle biopsies from patients with PMDS showed alterations in NMJs and sarcomeres. Similarly, Shank3 deficiency in mice resulted in smaller NMJs, sarcomere abnormalities, and hypotonia. Muscular deficits in mice were rescued with the troponin activator Tirasemtiv, suggesting that the treatment might be effective for treating hypotonia in patients with SHANK3 mutations.

Abstract

Heterozygous mutations of the gene encoding the postsynaptic protein SHANK3 are associated with syndromic forms of autism spectrum disorders (ASDs). One of the earliest clinical symptoms in SHANK3-associated ASD is neonatal skeletal muscle hypotonia. This symptom can be critical for the early diagnosis of affected children; however, the mechanism mediating hypotonia in ASD is not completely understood. Here, we used a combination of patient-derived human induced pluripotent stem cells (hiPSCs), Shank3Δ11(−/−) mice, and Phelan-McDermid syndrome (PMDS) muscle biopsies from patients of different ages to analyze the role of SHANK3 on motor unit development. Our results suggest that the hypotonia in SHANK3 deficiency might be caused by dysfunctions in all elements of the voluntary motor system: motoneurons, neuromuscular junctions (NMJs), and striated muscles. We found that SHANK3 localizes in Z-discs in the skeletal muscle sarcomere and co-immunoprecipitates with α-ACTININ. SHANK3 deficiency lead to shortened Z-discs and severe impairment of acetylcholine receptor clustering in hiPSC-derived myotubes and in muscle from Shank3Δ11(−/−) mice and patients with PMDS, indicating a crucial role for SHANK3 in the maturation of NMJs and striated muscle. Functional motor defects in Shank3Δ11(−/−) mice could be rescued with the troponin activator Tirasemtiv that sensitizes muscle fibers to calcium. Our observations give insight into the function of SHANK3 besides the central nervous system and imply potential treatment strategies for SHANK3-associated ASD.

INTRODUCTION

Autism spectrum disorders (ASDs) are complex neurodevelopmental diseases affecting 1 to 2% of the population (1) and are mainly characterized by deficits in social interaction and repetitive behavior (2). Alterations in synaptic function and development have been found to be a common converging feature in ASD (3). The importance of the synapse is supported by ASD-linked mutations in several genes associated with synaptic structure and function. Syndromic forms of ASD that are caused by synapse-associated single gene mutations include Fragile X syndrome (FMR1), tuberous sclerosis (TSC1 and TSC2), Rett syndrome (MECP2), or Phelan-McDermid syndrome (PMDS; SHANK3) (3).

PMDS results from heterozygous deletions of the long arm of chromosome 22 (22q13.3) (4) including SHANK3 or mutations in SHANK3. SHANK3 is a scaffolding protein known for its postsynaptic localization in excitatory synapses (5, 6). Patients with PMDS not only commonly display typical autism features but also show a global developmental delay, intellectual disability, craniofacial alterations, and skeletal muscle hypotonia (4, 7). Hypotonia is of high relevance for diagnosis because it is one of the earliest clinical hallmarks of PMDS, and affected babies are born as so called “floppy infants” (7, 8). Hypotonia involves the instability of both the trunk and limbs (9). From birth on, patients with PMDS commonly manifest with feeding problems (9) due to compromised sucking and swallowing (4). Later, abnormalities of gait and gross and fine motor coordination are detected in about 90% of PMDS cases (9).

Motor deficits, but especially hypotonia, are commonly observed in children with ASD. Motor disturbance in ASD has been reported in several studies, involving motoric development, motor coordination, motor movement, and stereotypic behaviors [as summarized in (10)]. Low muscle tone in infants has been associated with autistic traits in childhood and was therefore proposed as good marker for early diagnosis (11). Concurrent occurrence of hypotonia and ASD is reported in up to 51% of cases examined (10), and it has already been hypothesized 20 years back that motor deficits can impair the development of appropriate social interaction in children with ASD (12).

SHANK3 deletions and mutations, summarized as SHANK3 deficiencies, have been studied in primary neuronal cultures (13, 14), human induced pluripotent stem cell (hiPSC)–derived neurons (1520), and animal models (2125). Recently, a nonhuman primate model targeting exon 21 of SHANK3 was shown to closely mimic human pathology including decreased muscle strength, increased repetitive behaviors, and social and learning impairments (26). This demonstrates the close association between the wide range of SHANK3-associated clinical phenotypes. However, the pathophysiology of muscular hypotonia associated with SHANK3 mutations has not been examined so far. SHANK3 deficiency in the central synapses, where it is crucial for maturation and maintenance (5, 27, 28), might also contribute to hypotonia. In addition, a recent publication reports that somatosensory neurons in Shank3-mutated mice are dysfunctional and contribute to tactile behavioral phenotypes (29), thus suggesting that impaired proprioception and altered somatosensory feedback to the brain of patients with PMDS may further enhance not only ASD core symptoms but also the delay in motor development.

In this study, we elucidate the effects of SHANK3 deficiency on components of the motor unit, which comprises a motoneuron, the neuromuscular junction (NMJ), and myotube fibers. We chose PMDS hiPSC-derived culture systems, a well-characterized Shank3 isoform–specific knockout mouse model [Shank3Δ11(−/−)], and PMDS muscle biopsies to thoroughly study the development and maturation of NMJs and striated muscles under SHANK3 deficiency. Considering ASD as a neurodevelopmental disorder, we benefit from the advantages of all three systems, because, first, we studied patient-derived neurons in an in vitro coculture system modeling a very early developmental stage. Second, we analyzed Shank3Δ11(−/−) mice at different ages from newborn to adulthood. Last, we confirmed our results using muscle biopsies from a child, a young, and an adult patient with PMDS.

RESULTS

Generation of hiPSCs

To study SHANK3 deficiency in human cells, we generated hiPSC lines of three controls (CTRL1 to CTRL3) and five patients with PMDS (PMDS1 to PMDS5) (Fig. 1A). If not indicated otherwise, CTRL refers to the pool of all CTRL lines and PMDS to the pool of PMDS lines. Clinical data of all patients are summarized in table S1. In addition, the InsG cell line was created artificially via CRISPR-Cas9 as an isogenic SHANK3-deficient cell line from CTRL1 by introducing a heterozygous guanine insertion into SHANK3 at position 3860, as described in a patient previously (30). PMDS and SHANK3 deficiency can be caused by deletion or mutation of SHANK3; therefore, this cell line carrying only a point mutation in SHANK3 can confirm SHANK3 specificity in this study. In silico analysis for potential off-target effects of the respective small guide RNA and sequencing of the first five target sequences revealed no off-target alterations (fig. S1). All generated hiPSC lines were shown to be pluripotent by immunostaining for the transcription factors octamer binding transcription factor 4 (OCT4), sex determining region Y-box 2 (SOX2) and NANOG and the surface markers stage-specific embryonic antigen 4 (SSEA4), TRA-1-60 and TRA-1-81 (fig. S2A). In addition, RNA expression of OCT4, SOX2, NANOG, and KLF4 in comparison to keratinocytes of the respective hiPSC lines was verified (fig. S2B). A decrease of KLF4 RNA expression in some hiPSC lines might be explained by intrinsically high expression of KLF4 in keratinocytes that contributes to efficient reprogramming as reported in literature before (31, 32). The hiPSC lines differentiated spontaneously into all three germ layers with α-ACTININ staining mesoderm, βIII-TUBULIN marking ectoderm, and α-FETOPROTEIN being expressed in endoderm (fig. S3). Karyogram analysis revealed normal chromosomes (fig. S4A), and the respective genotypes were confirmed (fig. S4B).

Fig. 1 Reduced SHANK3 expression in hiPSC-derived PMDS motoneurons and impaired capability to induce AChR clusters in myotubes.

(A) hiPSC lines were pooled as indicated. (B) hiPSCs were differentiated into d42 motoneurons. (C) ICC of neurofilament heavy chain (NFH) and choline acetyltransferase (CHAT) in d42 motoneurons and analysis of differentiation efficiency. n = 3 independent cultures, one-way ANOVA and Tukey’s multiple comparisons test. (D and E) Western blot analysis of SHANK3 in (D) hiPSCs and (E) motoneurons. n = 3 to 5 independent experiments. Means ± SEM, Kruskal-Wallis test and Dunn’s multiple comparisons test [hiPSCs SHANK3a and motoneurons (MN) SHANK3e]; all others one-way ANOVA and Tukey’s multiple comparisons test. (F) ICC analysis of SHANK3 synaptic puncta colocalizing with the presynaptic marker bassoon (BSN). Data collected from two to three independent cultures. Means ± SEM, one-way ANOVA and Tukey’s multiple comparisons test. Puncta: n equals number of neurites analyzed. CTRL, n = 77; PMDS, n = 166; InsG, n = 37. Size and intensity: n equals number of puncta analyzed, CTRL, n = 790; PMDS, n = 1824; InsG, n = 433. (G) hiPSC-derived motoneurons were cocultured with biopsy-derived myotubes from healthy human donors. (H) Bungarotoxin (BT)–positive structures in biopsy-derived human myotubes were categorized into four classes with increasing maturity and percentages calculated. Data collected from three independent experiments. n equals total number of AChR clusters analyzed per cell line. CTRL1, n = 90; CTRL2, n = 137; PMDS2, n = 125; PMDS4, n = 118; InsG, n = 112, χ2 test. (I) BT area was measured in cocultures. Data collected from three independent experiments. n equals number of structures analyzed [see (H)]. Means ± SEM, Kruskal-Wallis test and Dunn’s multiple comparisons test. (J) SHANK3 localization in cocultures. Percentage of BT-positive AChR clusters colocalizing with SHANK3. Data collected from two to three independent experiments. n equals total number of AChR clusters analyzed per cell line. CTRL1, n = 107; CTRL2, n = 149; PMDS2, n = 45; PMDS4, n = 106; InsG, n = 168. Scale bars, 50 μm (C) and 10 μm (others).

PMDS motoneurons express less SHANK3 and have reduced potency to induce AChR clusters in human myotubes

First, we investigated the effect of SHANK3 deficiency on the ability of the hiPSCs to form motoneurons (Fig. 1B). All hiPSC lines showed comparable differentiation efficiency confirmed by the expression of neurofilament heavy chain (NFH) and choline acetyltransferase (CHAT) (Fig. 1C). No substantial differences were found in soma size or number of primary neurites in all CTRL versus all PMDS/InsG lines (fig. S5, A and B). We analyzed expression of the SHANK3 isoforms SHANK3a, SHANK3c/d, and SHANK3e (33) and total SHANK3 as the sum of all isoforms. Total SHANK3 was considerably reduced in the lysates of PMDS/InsG hiPSC (Fig. 1D), but in motoneurons, the SHANK3 reduction did not reach significance (Fig. 1E). To analyze specifically the synaptic compartment, we used immunocytochemistry (ICC) and identified NFH-positive neurites. We found that the size and intensity of SHANK3 synaptic puncta colocalizing with bassoon (BSN) were markedly decreased, whereas the overall number of synaptic puncta was unchanged (Fig. 1F). For analysis of SHANK3 expression, we showed pooled data of all CTRL versus PMDS to highlight the prominent difference in SHANK3 expression between diseased and healthy cells from multiple patients. To analyze the implication of reduced SHANK3 in motoneurons on muscle cells, hiPSC-derived motoneurons were cocultured with healthy human biopsy–derived myotubes (Fig. 1G). PMDS2 + 4 were selected for this experiment because of the severity of their clinical hypotonia phenotype (table S1), along with InsG. Resulting acetylcholine receptor (AChR) clusters labeled with bungarotoxin (BT) were categorized according to their maturity (Fig. 1H). During maturation, AChRs form complex structures containing clusters and invaginations on the myotube surface (34, 35), which can be used to assess NMJ maturation. We categorized AChR clusters according to their morphology (36): “diffuse puncta,” “aligned puncta,” “cluster,” and “complex”. Both PMDS2 + 4 and InsG revealed less mature AChR clusters (Fig. 1H) and reduced cluster area than the controls (Fig. 1I). In this in vitro model, SHANK3 was found localizing throughout the muscle cell but was not clustering at sites of BT staining (Fig. 1J). In summary, SHANK3 deficiency in motoneurons alters the clustering of postsynaptic AChRs in myotubes.

PMDS myogenic cells show decreased SHANK3 and maturation deficits

To elucidate the impact of SHANK3 on the postsynaptic side of the NMJ, hiPSCs were differentiated into myogenic cells (fig. S5, C to F) (37). As reported previously, during myogenesis, myoblasts expressing the transcription factor paired box protein 7 (PAX7) mature into myotubes characterized by high expression of myogenin (MYOG) (37, 38)). We grew hiPSC-derived myogenic cells as spheres, plated them at 6 weeks for myotube formation, and kept them in culture upon 8 weeks (37). Because of the growth factor–driven protocol, we obtained a mixture of PAX7-expressing myoblasts and MYOG-positive myotubes at week 8 as shown schematically in Fig. 2A. ICC of 8-week-old cultures for PAX7 or MYOG showed no differences in the number of positive cells (Fig. 2, B and C). To follow the maturation of myotubes, RNA samples were collected at weeks 4 to 8 of differentiation. PAX7 increased steadily in CTRL, indicating an active pool of progenitor cells, but remained static in PMDS/InsG (Fig. 2D). PMDS/InsG showed higher PAX7 expression in young cell (4 weeks) but lower PAX7 in mature cells (8 weeks), indicating dysregulation of the PAX7-positive myoblast pool. CTRL cells showed an increase in MYOG expression over time, and in PMDS/InsG, the increase of MYOG was less profound (Fig. 2E). Myoblasts, expressing less PAX7, and myotubes, expressing less MYOG, appeared less mature than control cells. This indicates that SHANK3 mutations might affect the transcription of myogenic factors.

Fig. 2 Impaired maturation of PMDS hiPSC–derived myogenic cells and AChR clustering.

(A) hiPSCs were differentiated into myoblasts (PAX7-positive) followed by myotubes (MYOG-positive). (B) ICC for PAX7 and analysis of the fraction of PAX7-positive cells in 8-week-old cultures. n = 4 to 6 independent cultures per cell line. One-way ANOVA and Tukey’s multiple comparisons test. Scale bars, 50 μm. DAPI, 4′,6-diamidino-2-phenylindole. (C) ICC for MYOG and analysis of the fraction of MYOG-positive cells in 8-week-old cultures. n = 3 to 6 independent cultures per cell line. One-way ANOVA and Tukey’s multiple comparisons test. Scale bars, 50 μm. (D) Quantitative reverse transcription polymerase chain reaction (qRT-PCR) of PAX7 during differentiation of hiPSC-derived myogenic cells. n = 3 to 11 independent cultures per cell line. One-way ANOVA and Tukey’s multiple comparisons test. (E) qRT-PCR of MYOG. n = 3 to 11 independent cultures per cell line. Kruskal-Wallis test and Dunn’s multiple comparisons test. (F) qRT-PCR of SHANK3. n = 3 to 10 independent cultures per cell line. One-way ANOVA and Tukey’s multiple comparisons test. (G) ICC for SHANK3 and BT staining in 8-week-old cultures. Analysis of SHANK3 intensity. Percentage of BT-positive AChR clusters colocalizing with SHANK3. Data collected from two to four independent cultures per cell line. Left graph: n equals number of α-ACTININ–positive cells. CTRL1 + 2, n = 90; PMDS1 + 2, n = 75; InsG, n = 60. Right graph: n equals total number of AChR clusters. CTRL1 + 2, n = 111; PMDS1 + 2, n = 106; InsG, n = 90. Kruskal-Wallis test and Dunn’s multiple comparisons test. Scale bars, 10 μm. (H) ICC and BT staining in 8-week-old myotubes and calculation of difference. n = 3 to 4 independent experiments. One-way ANOVA and Tukey’s multiple comparisons test. Scale bars, 10 μm. All data are displayed as means ± SEM.

We found that SHANK3 expression was lower in PMDS/InsG compared to CTRL during development, but this reduction was not seen in 8-week-old myogenic cells (Fig. 2F). ICC revealed that SHANK3 was highly expressed in myotubes, and its intensity was reduced in both PMDS/InsG myogenic cells, but SHANK3 again did not cluster at AChR sites (Fig. 2G). hiPSC-derived myotubes spontaneously form AChR clusters, which were split into two categories (Fig. 2H) based on the classification used in Fig. 1H. Maturation of BT-positive structures has been reported to start with a punctate pattern that matures into clustered and complex pretzel-like formations (35, 38). In contrasts to the coculture, we only observed two classes of BT structures in the “myotubes-only” culture, most likely due to the lack of motoneurons driving maturation. Analysis revealed a substantial reduction of the more mature “cluster” class in PMDS1 + 2 and InsG compared to CTRL1 + 2 (Fig. 2H). PMDS1 + 2 were selected for the experiments, because PMDS4 had to be excluded due to a negligible number of mature α-ACTININ–positive myotubes and, thus, AChR clusters that could be analyzed (fig. S5, F and G). In summary, not only SHANK3 expressed in the motoneuron but also myogenic SHANK3 deficiency seems to affect the assembly of AChR clusters.

Lithium increases SHANK3 in PMDS and rescues clustering of AChRs

Lithium has not only been reported to increase SHANK3 and restore synaptic deficits in vitro (18) but has also been applied to patients with SHANK3 deficiency (18, 39, 40). Therefore, we tested whether we could restore SHANK3 expression and maturation deficits with lithium. We applied lithium to hiPSC-derived motoneurons, cocultures, and hiPSC-derived myotubes. Lithium or water, serving as a vehicle control, was applied to the cultures in the final 120 hours of differentiation, as published previously (18). Lithium re-established the intensity of SHANK3 synaptic puncta in PMDS4/InsG motoneurons to the amount seen in the control (Fig. 3A) but did not change cluster size or number of clusters (fig. S5H). In addition, lithium was able to completely rescue the maturity of AChR clusters in cocultures of PMDS4/InsG motoneurons and human biopsy–derived myotubes (Fig. 3B). Moreover, BT area increased in cocultures of all cell lines (Fig. 3B). In hiPSC-derived myotubes, lithium not only increased SHANK3 intensity in PMDS2/InsG (Fig. 3C) but also restored the maturity of PMDS2 AChR clusters (Fig. 3D). Likewise, lithium increased BT area in PMDS2 (Fig. 3D). As expected, lithium increased the maturity of AChR clusters not only in PMDS and InsG but also in CTRL1. We show that lithium can influence SHANK3 expression not only in neurons as previously reported (18) but also in other tissues, as, for example, in muscle cells. Moreover, the increase in SHANK3 in the respective cell is in line with the restoration of the phenotype in PMDS/InsG. Thus, the in vitro lithium treatment also supports the SHANK3 dependency of the observed effects.

Fig. 3 Lithium restores SHANK3 expression in iPSC-derived motoneurons and myogenic cells and rescues maturation deficits in muscles and cocultures.

(A) ICC for SHANK3/NFH and intensity analysis of SHANK3 synaptic puncta in iPSC-derived motoneurons. Data collected from two to three independent experiments. n equals number of synapses. CTRL1, n = 1847; PMDS4, n = 279; PMDS4 vehicle, n = 189; PMDS4 lithium, n = 124; InsG, n = 989; InsG vehicle, n = 948; InsG lithium, n = 1350. Means ± SEM, Kruskal-Wallis test and Dunn’s multiple comparisons test. Scale bars, 5 μm. (B) ICC for NFH and BT staining and analysis for complexity and BT area of AChR clusters in cocultures of iPSC-derived motoneurons and human biopsy–derived myotubes. Data collected from three independent experiments. n equals total number of AChR clusters analyzed (indicated). Complexity: χ2 test; BT area: means ± SEM, Mann-Whitney test per cell line. Scale bars, 20 μm. (C) ICC for SHANK3/α-ACTININ and intensity analysis of SHANK3 in iPSC-derived myogenic cells. Data collected from three to six independent cultures. n equals number of cells. CTRL1 vehicle, n = 75; CTRL1 lithium, n = 45; PMDS2 vehicle, n = 45; PMDS2 lithium, n = 45; InsG vehicle, n = 75; InsG lithium, n = 72. Means ± SEM, Mann-Whitney test per cell line. Scale bars, 20 μm. (D) ICC for α-ACTININ and BT staining and analysis of AChR cluster complexity and BT area in iPSC-derived myogenic cells. Data collected from three to six independent cultures. n equals total number of AChR clusters analyzed (indicated). Complexity: χ2 test; BT area: means ± SEM, Mann-Whitney test per cell line. Scale bars, 20 μm.

SHANK3 colocalizes with sarcomeric α-ACTININ at Z-discs

Given the impact of SHANK3 on myogenic maturation, we investigated SHANK3 localization in hiPSC-derived myotubes. In the sarcomere, α-ACTININ localizes to the Z-disc, and myosin heavy chain (MHC) is present along the A-band (Fig. 4A). Confocal microscopy of hiPSC-derived myotubes revealed a SHANK3 distribution in a cross-striated manner, comparable to α-ACTININ (Fig. 4B). When analyzing SHANK3 and α-ACTININ intensity profiles according to the Otsu method in R and applying feature computation methods from the EBImage Toolbox (41, 42), we found concordant oscillation of fluorescent intensities (Fig. 4B). A density plot showed a correlation between SHANK3 and α-ACTININ with a PCR (Pearson correlation coefficient) of 0.524 ± 0.006 (±SEM). In contrast, SHANK3 did not colocalize with MHC (fig. S5I) with a PCR of 0.216 ± 0.019, indicating that SHANK3 is a protein of the sarcomere colocalizing with α-ACTININ present in the Z-discs.

Fig. 4 SHANK3 colocalizes with α-ACTININ and sarcomeres are malformed in PMDS myotubes.

(A) Proposed schematic localization of α-ACTININ, MHC, and SHANK3 in the sarcomere (top) and respective ICC fluorescence intensity profile (bottom). (B) ICC for SHANK3 and α-ACTININ. Longitudinal standardized fluorescence intensity profiles (means ± SEM) and density plots of colocalization between SHANK3 and α-ACTININ. n equals number of cross-striated myotubes analyzed per cell line. CTRL1, n = 40; CTRL2, PMDS1 + 2, n = 30; InsG, n = 45. Scale bars, 10 μm. (C) TEM analysis of A-band length. Data collected from two to three independent cultures. n equals number of A-bands. CTRL1 + 2, n = 105; PMDS1 + 2, n = 37; InsG, n = 51. One-way ANOVA and Tukey’s multiple comparisons test. Scale bars, 500 nm. (D and E) ICC analysis of peak width (Z-disc width) for SHANK3 (D) and α-ACTININ (E). Data collected from two to four independent experiments. n equals number of Z-discs. SHANK3: CTRL1 + 2, n = 196; PMDS1 + 2, n = 197; InsG, n = 163. α-ACTININ: CTRL1 + 2, n = 173; PMDS1 + 2, n = 101; InsG, n = 177. Data shown as means ± SEM. Kruskal-Wallis test and Dunn’s multiple comparisons test.

PMDS myotubes show shortened contraction units

We also investigated the sarcomere organization of PMDS/InsG hiPSC–derived myotubes. In both PMDS1 + 2 and InsG, ultrastructural analysis revealed that the A-band length of sarcomeres was substantially reduced (Fig. 4C). In addition, we measured peak width of SHANK3 (Fig. 4D) and α-ACTININ (Fig. 4E) intensities. Reduced peak width was found for both proteins in PMDS1 + 2 and InsG. These data indicate that PMDS/InsG hiPSC–derived myotubes display fundamental structural changes in the sarcomere that might contribute to the muscular hypotonia seen in patients with PMDS.

Shank3Δ11(−/−) mice present smaller soma of motoneurons and synaptic alterations

Because hiPSC-derived cultures generally do not reach the maturity seen in vivo (43, 44), we further analyzed motoneurons in 56-day-old (d56) Shank3Δ11(−/−) mice (Fig. 5A) (19, 45). α-Motoneurons [positive for microtubule-associated protein 2 (MAP2) and vesicular glutamate transporter 1 (vGLUT1) (46)] in the ventral horn of lumbar sections were analyzed (Fig. 5B). Somatic SHANK3 appeared markedly reduced (Fig. 5C), and motoneuronal soma size was decreased in Shank3Δ11(−/−) (Fig. 5D). Besides that, the size of glutamatergic [vGLUT1 (Fig. 5E)] and acetylcholinergic [vACHT (Fig. 5F)] axon-somatic clusters in Shank3Δ11(−/−) was decreased. Moreover, the density of vACHT but not of vGLUT1 clusters was reduced (Fig. 5F), pointing toward specific synaptic alterations affecting α-motoneurons upon SHANK3 deficiency.

Fig. 5 Shank3Δ11(−/−) mice show reduced SHANK3, soma size, and synaptic alterations in ventral horn α-motoneurons.

(A) Lumbar spinal cord sections of adult mice were prepared, and α-motoneurons from ventral horn were evaluated. (B) Lumbar section of the spinal cord. Gray matter and ventral horn are outlined. Scale bar, 300 μm. (C) IHC of spinal cord for SHANK3/MAP2. Soma SHANK3 intensity relative to Shank3(+/+) is displayed. Data collected from three animals per genotype. n equals number of cells. +/+, n = 64; −/−, n = 63. Means ± SEM, Mann-Whitney test. Scale bars, 20 μm. (D) IHC for SHANK3/MAP2 and analysis of soma size. Distribution (left) and soma size (right) are displayed. Data collected from three animals per genotype. n is equal to the number of cells. +/+, n = 248; −/−, n = 257. Means ± SEM, Mann-Whitney test. Scale bars, 20 μm. (E) IHC for SHANK3/MAP2/vGLUT1. The mask displays vGLUT1 clusters. Cluster size and clusters per soma size were analyzed. Data collected from three to four animals per genotype. Size: n equals number of clusters. +/+, n = 402; −/−, n = 286. Number: n equals number of cells. +/+, n = 78; −/−, n = 71. Means ± SEM, Mann-Whitney test. Scale bars, 50 μm. (F) IHC for SHANK3/MAP2/vACHT. The mask displays vACHT clusters. Analysis of vACHT cluster size and clusters per soma size. Data collected from three animals per genotype. Size: n equals number of clusters. +/+, n = 1046; −/−, n = 854. Number: n equals number of cells. +/+, n = 82; −/−, n = 85. Means ± SEM, Mann-Whitney test. Scale bars, 50 μm.

Shank3Δ11(−/−) mice demonstrate smaller NMJs and ultrastructural changes of the sarcomere along with signs of hypotonia

Similarly, skeletal muscle tissue from Shank3Δ11(−/−) mice was analyzed to complete our developmental data with a more mature system (Fig. 6A). In contrast to hiPSC cultures, SHANK3 colocalized with BT in the postsynaptic NMJ compartment in adult mice (Fig. 6B). In the muscle, SHANK3 strongly colocalized with α-ACTININ (PCR = 0.687 ± 0.005) (Fig. 6C) but not with MHC (PCR = −0.032 ± 0.027) (fig. S6A). SHANK3 was reduced in newborn (d0 to d1) and adult (d56) Shank3Δ11(−/−) but not in heterozygous Shank3Δ11(+/−) muscle (fig. S6, B and C). In d56 mice, all SHANK3 isoforms (SHANK3a, SHANK3c/d, SHANK3e, and the SHANK3 isoform at 120 kDa) could only be detected by Western blot after immunoprecipitation (IP) (Fig. 6D): Nonprecipitated lysate did not show all SHANK3 isoforms (see Input in Fig. 6E). In Shank3Δ11(−/−) mice, expression of SHANK3 isoforms a, b/c, and e was almost absent, whereas the isoform at 120 kDa was not reduced (Fig. 6D). We co-immunoprecipitated SHANK3 and α-ACTININ and found that the two form a complex (Fig. 6E). Especially the SHANK3 isoform at 120 kDa was detected in the co-immunoprecipitate with α-ACTININ. However, we cannot exclude that also larger isoforms interact with α-ACTININ, since our method might be not sensitive enough.

Fig. 6 Shank3Δ11(−/−) mice show reduced muscular SHANK3, altered NMJs and contraction units, and respond to CK2017357 (Tirasemtiv).

(A) Gastrocnemius muscle was prepared from mice aged 0 to 1, 35, and 56 days. (B) SHANK3/NFH/IHC and BT staining in sagittal skeletal muscle sections. Percentage of BT-positive AChR clusters colocalizing with SHANK3 is indicated. Data collected from three animals per genotype. n equals number of AChR clusters. +/+, n = 26; −/−, n = 38. Scale bars, 50 μm. (C) SHANK3 colocalized with α-ACTININ in longitudinal skeletal muscle sections. Longitudinal standardized fluorescence intensity profiles (means ± SEM) and density plots of colocalization between SHANK3 and α-ACTININ. Data collected from three animals per genotype. n equals number of cross-striated myotubes analyzed. +/+, n = 36; −/−, n = 39. Scale bars, 10 μm. (D) Western blot of immunoprecipitated SHANK3 in muscle tissue. Isoforms are indicated. n = 3 animals were analyzed per genotype. Means ± SEM, Student’s t test. (E) Co-immunoprecipitations of muscle lysate with antibodies against SHANK3 or ɑ-ACTININ. (F) Whole mount BT labeling of muscle in mice day 0 (d0) to d1. BT labeling is categorized according to shape. BT area was analyzed. Data collected from three animals per group. n equals total number of AChR clusters (indicated). Categories: χ2 test; BT area, means ± SEM, Student’s t test. Scale bars 5 μm. (G) Whole mount BT staining of muscle. BT area was analyzed. Data collected from four to five (d35) and three (d56) animals. n equals number of BT structures. d35: +/+, n = 415; −/−, n = 293; d56: +/+, n = 245; −/−, n = 252. Means ± SEM, Mann-Whitney test. Scale bars, 20 μm. (H) Electron microscopy of NMJs at d56. Data collected from three animals. n equals number of NMJs. +/+ and −/−, n = 14. Means + SEM, Student’s t test. Scale bars, 1 μm. Synaptic vesicle, dashed squares; basal lamina, arrows; muscle, M. (I) Electron microscopical analysis of Z-disc width/A-band length. Data collected from three (d0 to d1), three to four (d35), and three (d56) animals. n equals number of Z-discs or A-bands. Z-disc: d0 to d1 +/+, n = 193; −/−, n = 188; d35 +/+, n = 189; −/−, n = 188; d56 +/+, n = 190; −/−, n = 139. A-band: n = 90 for all. Means + SEM, Mann-Whitney test. Scale bars, 1 μm. (J) Muscular strength (latency to fall/wire hanging test) of d28 mice (n = 4 animals, means + SEM, Student’s t test). (K) Muscular strength of d30 mice before/after treatment with CK2017357 or vehicle [vehicle, n = 7; Shank3Δ11(−/−), n = 6 animals; left: means + SEM; right: single mice, Wilcoxon matched-pairs signed-rank test].

To closely analyze NMJ structures in mouse skeletal muscle tissue, whole muscle mounts were labeled with BT. BT-positive areas of d0 to d1 mice were classified as complex, crescent, oval, or wave with declining maturation (movies S1 to S4). In Shank3Δ11(−/−), the categorization pattern was considerably altered (Fig. 6F). Shank3Δ11(−/−) mice of all ages demonstrated a smaller BT area, for d0 to d1, especially the mature (complex) form (Fig. 6, F and G, and fig. S6D). Ultrastructural analysis of NMJs from Shank3(+/+) revealed a normal axon terminal, complex postsynaptic membrane folding with properly formed secondary synaptic clefts, and a well-developed postsynaptic membrane (Fig. 6H). In contrast, Shank3Δ11(−/−) mice displayed an undefined postsynaptic cleft with less complex postsynaptic foldings, as shown by fewer secondary branches (Fig. 6H) in adult mice. No differences were found in young animals (fig. S6E). However, it was difficult to find a considerable number of NMJs in the muscle tissue, and the complex three-dimensional (3D) structure of an NMJ cannot be analyzed thoroughly in a 2D picture. Ultrastructural analysis of skeletal muscle sarcomeres demonstrated considerably thinner Z-disc width in Shank3Δ11(−/−) mice of all ages (Fig. 6I) in accordance with our findings in iPSC-derived myogenic cells. In contrast, A-band length was increased in newborns but unchanged in young and adult Shank3Δ11(−/−) mice (Fig. 6I), suggesting differential developmental regulation of sarcomere organization under SHANK3 deficiency. Muscular grip strength as measured in a wire hanging test was markedly reduced in young Shank3Δ11(−/−) mice (Fig. 6J), reflecting hypotonia in PMDS. This could be increased with the fast skeletal muscle troponin activator Tirasemtiv (CK2017357) (47) in both Shank3(+/+) and Shank3Δ11(−/−) (Fig. 6K).

Muscle biopsies from patients with PMDS show less complex NMJs and alterations in sarcomeres and sarcoplasmic reticulum

To further investigate the potential clinical importance of our in vitro and murine findings, we obtained skeletal muscle biopsies (Fig. 7A) from three patients with PMDS between 2 and 58 years old that share autistic clinical manifestations and SHANK3 deficiency (table S2). Regarding protein, SHANK3 was reduced in the 2 years old (2y) patient with PMDS but not in the young (22y) and old (58y) patient (Fig. 7B). In ultrastructure, NMJ complexity was decreased in PMDS 22y and 58y but not in 2y, which presented a broader range in the number of foldings (Fig. 7C). Similar to mice and hiPSC muscle, Z-discs were smaller in all patients with PMDS (Fig. 7D). In the biopsies, A-band alterations demonstrated interesting dynamics as they were thinner in the child, as seen in hiPSC muscle, but wider in the young and adult patients, as observed in young mice. In addition, in ultrastructure, we could detect very wide sarcoplasmic reticulum (SR) (Fig. 7E) in all biopsies of patients with PMDS, which correlated with an increase in the area positive for the SR Ca2+ binding protein calsequestrin (CSQ) (Fig. 7F).

Fig. 7 Skeletal muscle biopsies of patients with PMDS show ultrastructural changes of NMJs, contraction units, and SR.

(A) Skeletal muscle biopsies of three patients with PMDS and four age-matched controls were collected. (B) SHANK3/α-ACTININ IHC in sagittal human skeletal muscle sections. n = 3 to 9 pictures per biopsy. means + SEM, Student’s t test. Scale bars, 10 μm. (C) Electron microscopical analysis of NMJ in human muscle biopsies. n equals number of NMJ. CTRL 3y, n = 15; PMDS 2y, n = 18; CTRL 22y, n = 16; PMDS 20y, n = 15; CTRL 40y, n = 33; CTRL 32y, n = 34, PMDS 58y, n = 33. Means + SEM, Mann-Whitney test for child and young and Kruskal-Wallis test and Dunn’s multiple comparisons test for adult. Scale bars, 1 μm. (D) Electron microscopical analysis of Z-disc width/A-band length. n equals number of Z-discs or A-bands. Z-discs: CTRL 3y, n = 63; PMDS 2y, n = 62; CTRL 22y, n = 64; PMDS 20y, n = 63; CTRL 40y, n = 100; CTRL 32y, n = 100; PMDS 58y, n = 100. A-bands: CTRL 3y, n = 60; PMDS 2y, n = 69; CTRL 22y, n = 60; PMDS 20y, n = 60; CTRL 40y, n = 60; CTRL 32y, n = 60; PMDS 58y, n = 60. Means + SEM, Mann-Whitney test for child and young and Kruskal-Wallis test and Dunn’s multiple comparisons test for adult. Scale bars, 1 μm. (E) Electron microscopical pictures of skeletal muscle. Analysis of SR area per picture in percentage. n equals pictures per biopsy. CTRL 3y, n = 9; PMDS 2y, n = 9; CTRL 22y, n = 7; PMDS 20y, n = 7; CTRL 40y, n = 5; CTRL 32y, n = 5; PMDS 58y, n = 5. Means + SEM, Student’s t test for child and young and one-way ANOVA and Tukey’s multiple comparisons test for adult. SR, sarcoplasmic reticulum. Scale bars, 2 μm. (F) SHANK3/calsequestrin (CSQ) IHC in sagittal human skeletal muscle sections. Data collected from three to seven pictures per biopsy. n equals muscle bundles. CTRL 3y, n = 8; PMDS 2y, n = 8; CTRL 22y, n = 5; PMDS 20y, n = 8; CTRL 40y, n = 10; CTRL 32y, n = 10; PMDS 58y, n = 10. Means + SEM, Student’s t test for child and young and one-way ANOVA and Tukey’s multiple comparisons test for adult. Scale bars, 20 μm.

DISCUSSION

Two important aspects arise from our study: First, SHANK3 localizes in the NMJ and in the Z-discs of striated muscles, and second, SHANK3 is a fundamental protein driving maturation of these structures. PMDS, as a SHANK3-associated ASD (7, 30, 48), is considered a neurodevelopmental disorder that has been modeled in rodents and primates (22, 26) and in in vitro systems including hiPSCs (1520). Muscle impairment has been observed in distinct Shank3 mutations in mice (25, 49, 50), and reduced muscle strength has been demonstrated in a recently published macaque model of SHANK3 deficiency (26). However, to date, it is not known how hypotonia is associated with SHANK3 deficiency.

To determine the role of SHANK3 in the development of muscular hypotonia and muscle weakness, a translational approach was used, including advanced in vitro techniques, such as hiPSC-derived cells from patients with PMDS together with an engineered CRISPR SHANK3 mutation cell line, a Shank3 transgenic mouse model, and human biopsy material. This combination allowed us, first, to effectively follow development and maturation of skeletal muscle and, second, to confirm the SHANK3 dependency of the defects. hiPSC-derived cells reflect an early developmental stage with fetal (43) or embryonic identity (44). SHANK3 deficiency has extensively been studied in adult mice; however, there was no focus on the interaction of the nervous system with skeletal muscles (22). In addition, newborn or adolescent offspring can provide fundamental insight into the progression of motor function. Furthermore, the clinical relevance of SHANK3 deficiency was additionally examined in direct muscle biopsies from patients with PMDS ranging from childhood to adulthood, a tissue that has never been closely analyzed before.

SHANK3 dependency of the alterations we found was confirmed by the combination of the three different model systems used in this study. hiPSC lines and muscle biopsies from patients with PMDS have SHANK3 deletions that might also comprise further genes. The Shank3Δ11 mice harbor a specific deletion that results in reduced expression of the main SHANK3 isoforms. However, distinct SHANK3 isoforms are still expressed, making the homozygous Shank3Δ11(−/−) mice a perfect system to model SHANK3 deficiency in heterozygously affected patients with PMDS. In the muscle biopsies, SHANK3 was reduced in the 2 years old (2y) patient with PMDS but not in young (20y) and old (58y), indicative of age-related SHANK3 compensation, as shown for other genes (51, 52). The CRISPR InsG hiPSC line is a “SHANK3-only” mutation. Results obtained from PMDS and the InsG cell line are highly similar, and differences between them might be explained by the above-mentioned additional genes deleted in the PMDS lines.

We propose that SHANK3 loss is detrimental for the maturation of complex neuronal and muscular structures, including human NMJs, which was reflected in the immature morphology of NMJs from patients with PMDS. SHANK3 is implicated in the maturation of central nervous system (CNS) synapses (27, 53, 54), suggesting that it might play a comparable role in the peripheral nervous system and in skeletal muscle. To date, localization of SHANK3 in peripheral tissue has only been studied marginally (29, 5558). We found SHANK3 in the pre- and postsynaptic specialization in adult mice NMJs of skeletal muscle sections. Presynaptic localization of SHANKs has been described in primary hippocampal neurons, and SHANK3 was found to modulate N-methyl-d-aspartate receptor expression at axon terminals (59). In somatosensory neurons, SHANK3 has been shown to be present in presynaptic terminals, and peripheral SHANK3 loss induced region-specific alterations in the CNS (29). Likewise, SHANK3 might influence presynaptic receptor function in peripheral motoneurons and thereby affect the assembly of the postsynaptic part of the NMJ.

With high consistency, SHANK3 deficiency leads to shortened Z-discs in all three systems, hiPSCs, Shank3Δ11(−/−) mice, and in PMDS muscle. At central synapses, SHANK3 functions as a scaffolding platform anchoring receptor molecules and connecting to the actin cytoskeleton (53, 60, 61), We show that SHANK3 forms a complex with α-ACTININ localizing to sarcomeric Z-discs in muscle cells, and thus, we propose SHANK3 as an important scaffolding protein that defines and stabilizes sarcomeres. Beyond scaffolding, the complex hierarchical expression of myogenic factors including PAX7 and MYOG, although only RNA and not protein expression, was disturbed under SHANK3 deficiency, indicating a potential role for SHANK3 in the transcriptional program driving muscle maturation. SHANK3 is known to translocate from the synapse to the nucleus in an activity-dependent manner in primary hippocampal neurons (62). Within the nucleus, SHANK3 regulates transcription of several genes. SHANK3 does not act as a transcription factor per se, because it cannot bind DNA but forms a complex with heterogeneous nuclear ribonucleoproteins, which are RNA binding proteins that form complexes with RNA polymerase II transcripts (62). Translated to our model system of myotubes, PAX7 and MYOG transcription might be similarly controlled by nuclear SHANK3 (as we observed nuclear SHANK3 localization). SHANK3 might control the transcription of further myogenic genes or even AChRs, a hypothesis that should be tested in further experiments.

SHANK3 deficiency can be rescued in vitro by applying lithium (18), a substance already known for its positive effects on mood stabilization and bipolar disorders (63). Even adult patients with PMDS have been treated with lithium for psychiatric symptoms (39, 40), but its effect in young patients or on muscle and NMJ maturation have never been studied. In this study, lithium rescued decreased maturation of AChR clusters in vitro, broadening the potential benefit of lithium in PMDS treatment. Because not only neurons (18) but also cocultures and hiPSC-derived muscle cells increased SHANK3 upon treatment, lithium seems to activate a general SHANK3-dependent mechanism independent of the cell type. However, lithium has manifold molecular targets and potential interaction partners in various biological pathways (64), with glycogen synthase kinase 3β (GSK3β) being the most studied. GSK3β has multiple interaction partners and targets itself (65) and is connected to extracellular signal–regulated kinases, protein kinase B, and mammalian target of rapamycin signaling (66). Mitogen-activated protein kinases are essential players not only in the regulation of neurogenesis, synaptic function, synaptic plasticity but also in cell survival, and they substantially influence gene transcription (67). It is not known whether and how transcription of SHANK3 is influenced by downstream effectors of lithium or whether lithium might have a direct effect on SHANK3 that then induces downstream effects itself.

In addition to structural defects, we aimed at ameliorating functional motor defects and treated Shank3Δ11(−/−) mice with the troponin activator Tirasemtiv, thus directly targeting muscle contraction (47). Grip strength in Shank3Δ11(−/−) mice was improved to the same extent as in wild-type mice, suggesting that the cause for hypotonia and reduced muscle force was due to a less mature state. In PMDS muscles, the SR was enlarged, and the calcium binding protein CSQ accumulated, hinting at molecular changes in the calcium storage and release system in muscle cells. We propose that the hypotonia in SHANK3 deficiency such as PMDS might be caused by dysfunctions in all elements of the voluntary motor system: motoneurons, NMJs, and sarcomeres. We hypothesize that these functional defects might add up to the clinical appearance we observe in patients with PMDS.

There are some limitations of our study. Considering the burden to obtain a muscle biopsy from patients with PMDS (especially for children), we concentrated as “proof of principle” on the temporal aspect of muscular hypotonia and analyzed only one patient per age period. Therefore, these data are based on technical replicates and should therefore not be interpreted as statistically significant effects but rather as a presentation of the most obvious results on a more descriptive basis. The data obtained were nearly identical to the results from the other model systems; however, to draw a general conclusion, and relate muscle morphology to a specific phenotypical characteristic, more patient material would be required. Ideally, biopsies from the same patient at different ages should be examined to closely follow the development of NMJs and the striated muscle morphology. In our study, we focused on the description of pathogenic SHANK3-associated changes in the voluntary motor system. The respective function of SHANK3 and underlying mechanisms at all the different sites of the motor unit are, however, far from being completely elucidated. For example, it is not known whether SHANK3 loss itself is triggering all structural and functional changes we observed or if some alterations are secondary to this. We added the hiPSC line InsG to the study to control the dependency on SHANK3 of our results. Overall, the PMDS and the InsG line show highly similar phenotypes; however, for example, in hiPSC-derived myotubes, the InsG phenotype is not as strong as in the PMDS lines. We cannot exclude that SHANK3 mutations behave different than SHANK3 deletions, and therefore, future experiments should include further controls to describe the precise role of SHANK3 in striated muscle and at NMJs. Last, we tested two compounds in this study, namely, Tirasemtiv and lithium. In mice, Tirasemtiv was given for 5 days; however, its ameliorating effects on hypotonia after chronic application need to be studied in detail. In bipolar disorders, the treatment with lithium is well established. However, patients with PMDS would benefit most from lithium treatment at a much younger age than established in current protocols. This safety issue needs to be clarified, especially because lithium is such a promising approach in adjusting SHANK3 expression, the primary defect in PMDS.

With this study, we provide evidence that SHANK3 is involved in the regulation of growth and development of the NMJ and striated muscles. To date, SHANK3 is primarily regarded as a protein of the CNS; however, our study emphasizes that it might also play crucial roles in peripheral tissue. We hypothesize that the peripheral structural defects we observed in skeletal muscle and NMJs upon SHANK3 deficiency might lead to dysfunctional somatosensory feedback and even influence the proper development and connectivity of the CNS in a retrograde manner. The observations of our study might be used for developing potential treatment strategies for patients with PMDS. Clinical studies examining the effect of lithium or Tirasemtiv not only on motor behavior but also on cognitive abilities of patients with PMDS of different age can help translating our findings into clinics.

MATERIALS AND METHODS

Study design

The aim of this study was to investigate the SHANK3 dependency of the maturation of the NMJ and striated muscles that might underlie the muscular hypotonia seen in patients with PMDS. To address this question, we analyzed (i) motoneurons and myotubes derived from hiPSCs of healthy controls and SHANK3 patients, (ii) a Shank3 knockout mouse model, and (iii) skeletal muscle biopsies of controls and patients with PMDS of different ages. hiPSC lines of PMDS1, PMDS2, PMDS3, and PMDS5 were chosen randomly from our cohort of patients with PMDS, and PMDS4 was added intentionally due to the severe and persistent muscular hypotonia. The donors of the skeletal muscle biopsies with PMDS were chosen on the basis of their age and independent of their symptom severity. Lithium was applied in cell cultures, and Shank3 mice were treated with Tirasemtiv to rescue the observed phenotypes. To reveal changes in protein localization and protein/RNA expressions, cells and tissue sections were immunostained, and protein and/or RNA were extracted. For ultrastructural changes, samples were analyzed by electron microscopy. Examiners were blinded for the analyses. Different numbers of replicates were used for the experiments as specified in the “Data analysis” section and in the figure legends. Original data of all figures containing less than 20 values in one group are given in data file S1.

Statistical analysis

Statistical analysis was performed using GraphPad Prism 6, R Studio 3, and SPSS version 24 (IBM). Data were tested for normality using Shapiro-Wilk test. For parametric data, Student’s t test and one-way analysis of variance (ANOVA) followed by post hoc analysis (Tukey’s multiple comparisons test) were used. Nonparametric data were analyzed using Mann-Whitney test or Kruskal-Wallis test followed by post hoc analysis (Dunn’s multiple comparisons test). Chi-square (expected frequency, ≥5) or Fisher’s exact (expected frequency, <5) test was used to test categorical data (classified objects). Performance of mice before and after treatment was tested using Wilcoxon matched-pairs signed-rank test. Significance level was set to 0.05 (*P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001) with a corresponding 95% confidence interval. No outliers were removed.

SUPPLEMENTARY MATERIALS

stm.sciencemag.org/cgi/content/full/12/547/eaaz3267/DC1

Material and Methods

Fig. S1. Identification of CRISPR-Cas9 potential off-target effects.

Fig. S2. Proof of pluripotency.

Fig. S3. Differentiation into all three germ layers.

Fig. S4. Karyogram analysis.

Fig. S5. Alterations in PMDS/InsG hiPSC–derived motoneurons and myogenic cells.

Fig. S6. Analysis of skeletal muscle features in Shank3Δ11(−/−) mice.

Fig. S7. Graphical illustration of the findings of this study.

Table S1. Genotype, gender, diagnosis, and clinical phenotype of donors for hiPSC lines.

Table S2. Muscle biopsy donors.

Data file S1. Raw data.

Movie S1. Mice p0 to p1 BT whole mount category wave.

Movie S2. Mice p0 to p1 BT whole mount category crescent.

Movie S3. Mice p0 to p1 BT whole mount category oval.

Movie S4. Mice p0 to p1 BT whole mount category complex.

References (6875)

REFERENCES AND NOTES

Acknowledgments: We are grateful to all patients and families participating in the study and providing the basis for our work with their constant support. We acknowledge the excellent technical support of S. Seltenheim, R. Zienecker, and L. Dietz. We acknowledge the expertise and advice on electron microscopy of P. Walther and thank the core facility for electron microscopy at Ulm University. We thank C. Jacob and J. Philip Delling for providing the colocalization tool for ICC. The MyoG antibody developed by W. E. Wright, University of Texas Southwestern Medical Center, was obtained from the Developmental Studies Hybridoma Bank (DSHB), created by the NICHD of the NIH and maintained at The University of Iowa, Department of Biology, Iowa City, IA 52242, USA. Funding: T.M.B. is supported by the DFG (SFB1149), BIU2, the Else Kröner Foundation, the Helmholtz Gesellschaft (DZNE Ulm), and the Innovative Medicines Initiative (IMI) Joint Undertaking under grant agreement nos. 115300 and 777394, resources of which are composed of financial contribution from the European Union’s Seventh Framework Program and EFPIA companies’ in kind contribution. T.M.B. and M.D. are supported by Bundesministerium für Bildung und Forschung (project no. 01EK1611C). A.M.G. is supported by Baustein 3.2 (L.SBN.0083) and by the Else Kröner-Fresenius Stiftung. C.V. was supported by the Comitato Telethon Fondazione Onlus (grant no. GGP16131) and French Foundation Jérôme Lejeune. FR is supported by the Thierry Latran Foundation (projects “Trials” and “Hypothals”), by the Radala Foundation, by the Deutsche Forschungsgemeinschaft (DFG) as part of the SFB1149 and with the individual grant no. 431995586 (RO-5004/8-1) and no. 443642953 (RO5004/9-1), by the Cellular and Molecular Mechanisms in Aging (CEMMA) Research Training Group and by BMBF (FKZ 01EW1705A, as member of the ERANET-NEURON consortium “MICRONET”). Author contributions: A.-K.L., S.P., V.I., K.J.F., J.C., J.H., M. Stetter, and N.S. performed hiPSC culture, differentiation, and analysis. B.I., I.O., N.O.A., and M.D. stained and analyzed mouse tissue sections. M.Z. performed IPs. M.F. analyzed colocalization of SHANK3 with α-ACTININ/MHC. M.O. provided human muscle biopsies. S.L. generated hiPSC lines CTRL1 and PMDS1 to PMDS4. M.S., S.J., and A.C.L. provided patient and clinical material. G.B. performed karyotyping of hiPSC lines. A.M.G. and F.R. critically revised experiments. R.D. and T.B. provided patient/control hair roots for hiPSC lines PMDS5 and CTRL3. C.V. and F.G. performed mouse hanging wire test and mouse functional rescue experiments. B.M. advised on and performed statistical analyses. A.-K.L., S.P., M.D., and T.M.B. designed the experiments and wrote the manuscript. Competing interests: The authors declare that they have no competing interests. Data and materials availability: All data associated with this study are present in the paper or the Supplementary Materials.
View Abstract

Stay Connected to Science Translational Medicine

Navigate This Article