The SHANK gene encodes SHANK1, SHANK2, and SHANK3, multidomain proteins that contain an ankyrin repeat domain (ANK), a Src homology 3 domain (SH3), a postsynaptic density 95 (PSD-95)/Discs large/zona occludens-1 homology domain (PDZ), a proline-rich region and a sterile α motif domain (SAM) [13], another proline-rich region, a Homer-binding site (Hbs), and a cortactin-binding site (Cbs) [14].
PMS has the characteristics of an ASD. Some children first visit pediatric departments for symptoms of autism. In a cohort study of 221 patients (133 from the United States, 88 from Italy), the frequency of SHANK3 mutations was approximately 1% in patients with ASD [15],revealing a role for this gene as a possible major contributor to autism. In a Chinese cohort of 539 patients with ASD, pathogenic and likely pathogenic variants were detected in 9.5% of individuals. Variants of SHANK3 and SHANK2 had the highest frequencies, especially in females, and occurred in 1.2% of patients [16].
In this study, the child visited the outpatient department of neurology due to developmental retardation. The whole-exome sequencing results suggested that the newly identified SHANK3 gene mutation (OMIM: 606230) (NM_033517.1) c.3679dup (p.A1227Gfs*69) is a de novo heterozygous frameshift mutation. Figure 1a shows the family pedigree; Fig. 1b shows verification of the mutation by Sanger sequencing. The mutation site is marked with a red arrow. The child was heterozygous for the mutation, and the parents and older sister of the child harbored wild-type alleles.
We further performed quantitative RT-PCR analysis of the SHANK3 gene (Fig. 2) and found that the gene expression level in the patient was significantly lower than that in the general population. This finding suggests that this mutation can cause a decrease in gene expression. Gene mutation is affected by genetic, environmental, and other factors. Humans have regulatory systems to minimize the damage caused by deleterious mutations. One such surveillance system is the translation-dependent nonsense-mediated mRNA decay (NMD) pathway, which exists in eukaryotic cells[17–19]. NMD recognizes and eliminates mRNAs with suboptimally positioned translation termination codons. NMD-triggering translation termination codons are generically termed premature termination codons (PTCs). PTCs can arise as a result of nonsense or frameshift mutations in the coding DNA; mutations causing errors in splicing or the selection of the translation start site may also introduce PTCs. NMD degrades mRNAs containing PTCs and shortens the half-life of mRNAs [20, 21].
β-Thalassemia is the earliest identified and most classic example of a disease caused by NMD. The disease is caused by mutation of the β-globin gene, which consists of three exons. Nonsense mutations occur in the first two exons, and NMD results in loss or very low levels of the corresponding mRNA transcripts. Individuals homozygous for this mutation cannot synthesize β-globin, resulting in severe erythropoiesis disorders. Individuals heterozygous for this mutation exhibit only mild anemia; thus, the disease has an autosomal recessive (AR) inheritance pattern [22, 23]. However, nonsense mutations in exon 3 do not trigger NMD. The mRNA is stably translated into a truncated protein, resulting in abnormal β-globin. Individuals heterozygous for these mutations have severe anemia; thus, these mutations are associated with an autosomal dominant (AD) inheritance pattern [24].
PTCs at the 3’ end of the gene encoding the muscular dystrophy protein cannot be identified by the NMD pathway in patients with Duchenne muscular dystrophy (DMD), and affected transcripts have been found to be translated into truncated proteins with some functionality. The clinical phenotypes are mild, resulting in Becker muscular dystrophy (BMD). In contrast, when PTCs are located distally upstream, they can be identified and degraded via NMD, resulting in the inability to synthesize partially functional dystrophin, leading to severe phenotypes [25]. Therefore, NMD can either ameliorate or exacerbate diseases. The specific function of NMD in all genetic diseases is unclear.
NMD is a quality control pathway that maintains the normal function of eukaryotic cells. However, not all PTCs trigger NMD. For example, PTCs located less than 50 nucleotides (nt) upstream of the last exon-exon junction typically do not trigger NMD (the 50 nt rule) [26]. n addition, PTCs located in the last exon of a gene do not trigger NMD (the last exon rule) [27, 28]. These two canonical rules of NMD have been widely validated [29].
Figure 5 shows the possible action of NMD in heterozygotes containing PTCs. The child in this case study was heterozygous for a mutation in the SHANK3 gene and exhibited a clinical phenotype. We found by RT-PCR that SHANK3 gene expression was reduced in this patient. The SHANK3 gene contains 22 exons. The mutation site in SHANK3 reported here is located in exon 21, which is more than 50 nt from the last exon. This mutation is speculated to trigger NMD, resulting in degradation of the mutated mRNA transcript and subsequent haploinsufficiency and clinical symptoms.
Figure 3 shows the proportions of known pathogenic and likely pathogenic mutations. Frameshift mutations are the most common mutations, accounting for 56% of all known mutations, and other missense mutations and nonsense mutations accounted for 22%. Figure 4 summarizes the specific locations of the known pathogenic and likely pathogenic mutation sites identified in ClinVar. Most mutations occurred at sites upstream of the SAM. If NMD is not triggered, mutations cause changes in protein structure and function. The red arrow in Fig. 3 indicates the mutation site of the patient in this paper. A frameshift mutation at this site in exon 21 leads to premature termination of protein transcription, resulting in a truncated protein with deletion of the Homer-binding site, cortactin binding site, and SAM. Deletion of the C-terminal SAM causes the protein to be unable to interact with zinc ions, which may cause structural and functional changes in the protein. PMS has an autosomal dominant inheritance pattern, which is speculated to be related to haploid insufficiency in heterozygous individuals.
The mutation site identified in this study was previously reported in 3 children with ASD or language delay, including 1 patient harboring a de novo mutation [16]. The other two patients are brothers; both have severe mental retardation, and their mother has germline mosaicism [30]. This variation has not been included in the Human Gene Mutation Database (HGMD) and is reported as a pathogenic variant in ClinVar. The allele frequency of this mutation was less than 0.01% (16/160994) in the total population among the normal population in the Genome Aggregation Database (gnomAD) and 0% among the East Asian population. According to the variant classification guidelines of the American organization ACMG, this variant is classified as a pathogenic variant.
SHANK acts as the main scaffolding protein at the PSD, interacting with various synaptic molecules [31]. For example, Shank proteins are associated with the N-methyl-D-aspartate receptor (NMDA-R) via the postsynaptic density-95/guanylate kinase-associated protein (GKAP) complex, with the metabotropic glutamate receptor (mGluR) via Homer, and the GluR1 α-amino-3-hydroxy-5-methyl-1-4-isoxazole propionic acid receptor (AMPA-R) [32]. Some reports have indicated that Shank1 targeting of synapses is dependent on the PDZ, but the targeting of Shank2 and Shank3 depends on their C-terminal domains, including the SAM [33, 34], and the PDZ of Shank3 also plays a role [35]. Shank2 and Shank3 multimerize and form a platform or framework in the PSD that depends on Zn2+ binding to the SAM [36, 37]. In contrast, Shank1 does not bind Zn2+ but forms a large framework complex with Homer in the PSD [38]. In the PSD complex, the Shank3 protein binds to neuroligin [39] and forms a complex with axon proteins in glutamate synapses. The current data show that the Shank protein can bind to glutamate receptors via cytoskeletal proteins, thereby regulating the size and volume of excitatory synapses and dendritic spines [40].
The SHANK3 gene is rich in GC motifs (with a GC content of up to 69%), and the encoded protein is highly expressed in the human striatum, especially in neuronal synapses. In synapses, neurexin in the presynaptic membrane binds to neuroligin in the postsynaptic membrane and then binds to the Shank3 protein [4], forming a complex in the synapses of glutamatergic neurons. This complex plays a key role in maintaining normal synaptic function and dendritic spine morphology, as well as in regulating the balance between neuronal excitability and inhibition [41].
Phenotypic variability may be related to the type of SHANK3 mutation. Studies have confirmed that large-scale deletions are generally associated with obvious structural abnormalities and social communication disorders, while small-scale deletions or point mutations are often associated with haploinsufficiency of SHANK3 and with abnormalities including ASD, convulsions, abnormal electroencephalogram findings, hypotonia, sleep disorders, abnormal brain MRI findings and gastroesophageal reflux [7]. SHANK3 gene deletion was found in ASD patients, suggesting that haploinsufficiency is a main cause of ASD [6].
Chiara Verpelli et al found that reduced metabotropic glutamate receptor subtype 5 (mGluR5) activity in Shank3-knockdown neurons can impair mGluR5-dependent signaling pathways and plasticity [42]. An allosteric agonist of mGluR5, 3-cyano-N-(1,3-diphenyl-1H-pyrazole-5-yl) benzamide (CDPPB), was proven to be brain penetrant, to reverse mGluR5 activity [43, 44] and to rescue ERK1/2 phosphorylation and miniature excitatory postsynaptic current (mEPSC) frequency. These results demonstrate that positive allosteric modulation of mGluR5 produces behavioral effects and suggest that mGluR5 activity could be a therapeutic target.
In addition, based on the locations of pathogenic and likely pathogenic mutations in the SHANK3 gene (Fig. 4), we consider that the NMD mechanism may be involved in PMS. Further study is needed to determine whether the truncated protein product of the SHANK3 gene has partial functionality. Then, we could study whether inhibition of the NMD process has a therapeutic effect.