Late Breaking Chromosomes
SHANK Mutations May Disorder Brain Development
A hemizygous deletion in region 22q13 has frequently been found in children with neonatal hypotonia, developmental delay, absent to severely delayed speech, autistic behavior, and minor dysmorphic features [Phelan and McDermid, 2012]. This syndrome, termed Phelan McDermid syndrome, has been attributed to mutations of the SHANK3 gene, which is part of a family of genes, including SHANK1 and SHANK2, that encode post-synaptic scaffold proteins. They consist of one SPN domain, 3 or 4 ankyrin repeats, a Src homology domain 3, a PSD-95/Dlg/ZO-1 domain, a proline-rich region, and a sterile alpha motif [Sheng and Kim, 2000; Leblond et al., 2014]. The SHANK proteins localize to the post-synaptic density, where they connect the actin cytoskeleton to the glutamatergic receptor region and G protein-coupled signaling pathways, and contribute to synapse formation and dendritic spine maturation. In patients with mutations in one allele or hemizygous deletions of SHANK3, vermis hypoplasia, striatal dysfunction and altered dendritic spine morphology have been found [Peça et al., 2011; Wang et al., 2011, 2014; Aldinger et al., 2013; Durand et al., 2012]. With in vitro RNAi knockdown experiments, altered regulation of metabotropic glutamate receptor 5 (mGluR5) expression and signaling at synapses was demonstrated [Verpelli et al., 2011]. Hemizygous losses and mutations of SHANK3 were detected in patients with developmental delay, autism spectrum disorder (ASD) and schizophrenia [Durand et al.; 2007; Marshall et al., 2008; Gauthier et al., 2010]. Conversely, duplication of SHANK3 induces bipolar disorders with progressive loss of skills, attention deficit hyperactivity disorder (ADHD), or episodes of status epilepticus [Denayer et al., 2012; Han et al., 2013]. So, clinical phenotypes, albeit of strikingly different nature, are caused by either loss or duplication of SHANK3. Thus, mutations of SHANK3, provoking haploinsufficiency or triplosufficiency, are sufficient to elicit a clinical phenotype. Such a dominant effect of mutations suggests that SHANK3 may be a node in a combinatorial genetic network, as for instance CNTNAP2 [Poot et al., 2011; Poot, 2015]. It remains conceivable, however, that some mutations of SHANK3 may alone be sufficient to elicit a clinical disorder. To shed light on these issues and to derive possible phenotype-genotype relationships, a closer examination of a large number of mutations may be a promising approach [Leblond et al., 2014].
Leblond et al. [2014] analyzed copy-number variants (CNVs) in 5,657 ASD patients and 19,163 healthy controls, and ascertained single nucleotide variants (SNVs) in coding sequences of the SHANK genes in 760-2,147 ASD patients and 492-1,090 controls (depending on the gene examined) by performing classical dideoxy sequencing of the 3 SHANK genes. In approximately 1% of patients with ASD, CNVs and gene-truncating mutations were found. SNVs in SHANK1 (0.04%) were only found in male patients with normal IQ and ASD, while SNVs in SHANK2 were detected in 0.17% of patients with ASD and mild intellectual disability. In contrast, SNVs in SHANK3 were found in 0.69% of patients with ASD and in up to 2.12% of the ASD cases, moderate to profound intellectual disability was found. The frequencies of CNVs and SNVs of SHANK2 and SHANK3 were significantly higher than in healthy controls. In previous exome-sequencing efforts, only 1 de novo SNV in SHANK2 and no truncating coding-sequence variants in SHANK1 and SHANK3 have been found [Neale et al., 2012; O'Roak et al., 2012; Sanders et al., 2012]. This may be due to the poor mutation detection rate in stretches of GC base pairs by the current exome-sequencing technology. SNVs were found in all domains of the SHANK proteins, with no difference in distribution between ASD patients and controls. To assess the possible pathogenicity of SNVs, their respective Genomic Evolution Rate Profile and Grantham matrix scores were determined [Grantham, 1974; Cooper et al., 2005]. There was no significant difference between ASD patients and healthy controls regarding these scores for the SNVs examined. However, in ASD patients, but not in healthy controls, in-frame deletions presumed to remove several amino acids in the SHANK2 and SHANK3 proteins were found.
Given the significantly high frequencies of variants in SHANK2 and SHANK3 in ASD patients versus healthy controls and the impact of such variants on neuronal development, a molecular hypothesis to explain these findings is needed. Conceivably, certain SHANK2 and SHANK3 variants may by themselves be sufficient to elicit ASD or ASD with moderate to severe ID, respectively. On the other hand, a polygenic model of gene action should also be considered. Leblond et al. [2012] found 2 patients with SHANK2 mutations showing an additional gain of the nicotinic receptor gene CHRNA7 and in 1 patient a loss of the CYFIP1 gene. This study supports a model of epistatic interactions of genes in distinct biological pathways, which upon disruption may lead to ASD.
In mice, deletion of both Shank2 alleles results in an early, brain-region-specific upregulation of ionotropic glutamate receptors at the synapse and increased levels of ProSAP2/Shank3 [Schmeisser et al., 2012]. Shank2-/- mutant mice have fewer dendritic spines and show reduced basal synaptic transmission, enhanced N-methyl-D-aspartate receptor (NMDAR)-mediated excitatory currents, are extremely hyperactive and display profound autistic-like behavioral alterations including repetitive grooming as well as abnormalities in vocal and social behaviors [Schmeisser et al., 2012]. Comparing ProSAP1/Shank2-/- and ProSAP2/Shank3αβ-/- mice, the authors found that different abnormalities in synaptic glutamate receptor expression can cause distinct alterations in social interactions and communication. Treatment of Shank2-/- mice with a positive allosteric modulator of metabotropic glutamate receptor 5 (mGluR5) normalizes NMDAR function and markedly enhances social interaction [Won et al., 2012]. These results suggest that reduced NMDAR function, resulting from deletion of both alleles of Shank2 in mice, may contribute to the development of ASD-like phenotypes, and that mGluR modulation of NMDARs may offer a potential approach to treat ASD [Won et al., 2012].
Both the data on transgenic mice and the earlier study of Leblond et al. [2012] carry important implications for medical genetic practice. First, screening for SHANK2 and SHANK3 variants by classical dideoxy sequencing should be included in the laboratory workup of all ASD patients. Second, once such a variant has been detected, genome-wide CNV analysis should be done in order to search for other mutated genes so that a complete as possible picture of the mutational landscape in the individual ASD patient can be obtained [Leblond et al., 2012; Poot et al., 2011]. Third, with this information we may be able to label selected patients as ‘SHANK-type' ASD and may then consider therapies that match their underlying synaptopathic phenotype [Schmeisser et al., 2012; Poot, 2013].
Martin Poot
