Autism - Genetics of Autism Spectrum Disorders

ASDs are among the most heritable of neurodevelopmental disorders. Family- and twin- based studies have clearly proved the genetic basis of autism. There is a 36-91% chance of identical twins being autistic. Identical twins share the same DNA since they are derived from the same embryo. Fraternal twins on the other hand, share only 50% of their DNA since they come from two different embryos, and there is only 2-4% chance of both being autistic.

As ASDs are among the most heritable of all psychiatric disorders, much effort has been made to elucidate the underlying genetics, to identify the responsible genetic variants and chromosomal abnormalities. However, unlike some of the monogenic disorders which are caused by defects in a single gene, the genetics of autism is very complicated. It is because autism manifests a wide range of symptoms and probably several genes are involved. Owing to this high degree of heterogeneity, none of the known genetic causes of autism account for more than 1% of the ASD cases; a common susceptibility locus has not been reported until now. Therefore, attention has now been shifted to identify the rare variants associated with the disorder.

During the last decade, scientists from all over the world have worked together to trace autism candidate genes. Data has been collected from thousands of autistic families from different populations across the world. Linkage analyses and family-based association studies have resulted in the identification of hundreds of autism candidate genes. The recent evolution of SNP microarray and high-throughput sequencing technologies accompanied by large well-characterized patient cohorts is now leading to a rapidly accumulating pool of well-established genes and loci. This could lead to the emergence new treatment strategies. Genetic diagnostic tests leading to the early detection of this disorder may come up in the near future.

Classic candidate genes: current status

At present, using modern genetic methods, as many as 235 genes associated with autism, with supportive evidence from ≥ 5 independent studies (Autism database- https://gene.sfari.org/autdb/CNVHome.do) (Table-1) have been identified. These genes were identified through single nucleotide polymorphism (SNP) association studies and/or sequencing and/or gene expression analyses. SNPs, which are single base alterations in a DNA sequence, are responsible for much of the genetic variation in the human genome. Genetic association studies can be used to examine whether a particular genetic variant is associated with the disorder. Since most of the associated SNPs are located in the non-translated portions of the genome, their functional impact cannot be determined. However, several functional SNPs, leading to a defective protein, have also been identified in the exonic regions of autistic individuals.

Despite the diversity, ASD patients exhibit similar behavioral and neuronal conditions, although differing in severities and comorbidities. This commonality of neurological phenotypes suggests that the susceptible genes may act through a limited set of pathways (Lanz et al, 2013). Several studies have suggested a role for the genes involved in synaptic assembly, such as the synaptic cell adhesion molecules (CAMs), ion channels, neurotransmitter receptors, and the scaffolding and cytoskeletal proteins that work harmoniously to provide synaptic structural integrity and functionality (Banerjee et al, 2014). Research in the past few years is being focused on the following most important candidate genes of autism.

• Cell Adhesion Molecules (CAMs)
  1. Neurexins and neuroligins: These are pre- and post-synaptic CAMs respectively. Several studies have reported mutations of these genes in autism (Jamain et al 2003; Yan et al, 2005; Kim et al, 2008; Vaags et al, 2012). Neuroligin-deficient mice display abnormal social and vocal behaviors that resemble ASDs symptoms (Tabuchi et at, 2007). Impairments in social behavior traits have been found in neurexin knockout mice too (Grayton et al, 2013).
  2. Contactins (CNTNs): They have diverse roles in synapse formation and plasticity. Recurrent small deletions and duplications involving CNTN3 and CNTN4 have been reported in several ASD patients (Roohi et al, 2009). Rare mutations have been observed in CNTNAP2, a gene known to be associated with language quantitative traits (Alarcon et al, 2008). Cntnap2 mutant mice revealed deficits in the three core ASD behavioral domains, with hyperactivity and epileptic seizures (Penagarikano et al, 2011). The drug risperidone could reverse some of these behaviors in the mutant mice, opening new avenues for the therapeutic intervention in ASD.
  3. Cadherins and protocadherins (CDH and PCDH): These are a large family of CAMs required for synaptic specificity. CDH18, PCDHA, PCDH10 and PCDH9 have been found to be disrupted in ASD, mental retardation and intellectual disability cases (Marshall et al, 2008; Morrow et al, 2008; Anitha et al, 2013).
• Ion Channels

  Ion channels are essential mediators of neuron excitability and neurotransmitter release within the central nervous system. Recent data has implicated neuronal excitation alterations in ASD, pointing to a potential role for ion channel proteins in disease susceptibility.

  4. SCN1A and SCN2A are trans-membrane voltage-gated sodium channels. SCN1A has emerged as the most important gene in epilepsy with more than 70% of individuals with epileptic encephalopathy possessing a mutation in this gene (Mulley et al, 2005). Several patients with a mutated gene have been found to exhibit autistic-like behaviours. Loss-of-function mutant mice also exhibit autistic behaviours (Han et al, 2012).
  5. CACNA1C codes for the alpha subunit of voltage-dependent calcium channels. A well known mutation in this gene causes Timothy syndrome; while other genetic variations has been associated with various psychiatric diseases (Lu et al, 2012).
  6. GABRB3 encodes one of the subunit of a chloride channel which serves as the receptor for GABA, a major inhibitory neurotransmitter. Genetic association has been reported in various populations (Menold et al, 2001; Ashley-Koch et al, 2006; Delahanty et al, 2011).
• Scaffolding Proteins

  Shank family of scaffolding proteins are essential for the synaptic architecture.

  7. SHANK2 and SHANK3: The genetic association of ASD with SHANK family genesis well characterized (Berkel et al, 2012; Sato et al, 2012; Uchino and Waga, 2013).Neurobiological studies of Shank mutations in mice support the general hypothesis of synaptic dysfunction in the pathophysiology of ASD (Jiang and Ehlers, 2013).
• Neuropeptide Receptors

  The neuropeptides oxytocin and vasopressin are regulators of social cognitive skills (Israel et al, 2008). Polymorphisms in the genes coding for the receptors of these molecules could be relevant to individual differences in mammalian social behaviours.

  8. AVPR1a- Genetic associations and rare variants in the arginine vasopressin receptor gene were reported in several studies (Yirmiya et al, 2006; Yang et al, 2010). 9. OXTR- The oxytocin receptor gene was found to be associated with ASD in a number of independent studies (Wermter et al, 2010; Campbell et al, 2011; Skuse et al, 2014).
• Signaling Pathways

  Disruption of genes involved in critical signaling cascades often contribute to the development of ASD phenotypes.

  10. The mTOR pathway: The serine/threonine protein kinase, mechanistic target of rapamycin (mTOR), has a pivotal role in synapse formation. The mTOR signaling pathway includes genes such as NF1, PTEN, TSC1, and TSC2; recurrent mutations have been reported in these genes in several cases of autism-like behavioral phenotypes. Inherited or de novo PTEN mutations are one of the most validated causes of autism (O’Roak et al, 2012a). Mutant mice showed deficits in synaptic strength and displayed behavioral deficits common to autism (Xiong et al, 2012).
• Other Important Candidate Genes
  11. RELN- a serine protease that plays a critical role in neuronal migration during brain development. Several studies have shown genetic associations (Serajee et al, 2006; Fu et al, 2013); rare mutations were also reported (Neale et al, 2012). 12. MET- mediated signaling has been implicated in multiple aspects of neocortical and cerebellar neuronal growth and maturation. Number of positive associations have reported (Campbell et al, 2006; Thanseem et al, 2010; Plummer et al, 2013)

 

Copy Number Variation (CNV) Studies-Recent Progress

CNVs are the most common structural variations found in the human genome and represent a major part of genetic variability. It can be vary in size from a few kilo bases to several mega bases and may affect the information coded for the relevant genomic region. CNVs could affect the phenotype though the mechanism of “dosage sensitivity”, though the idea still remains far from established. Copy number gains or losses can alter the expression of critical genes which may leads to partial dysfunction (Iossifov et al, 2012). The genes that are required for the recently acquired human traits of speech and complex social behaviors may be particularly vulnerable. The concept of CNVs as a possible genetic contributor to the development of ASD was emerged, ever since the publication of first CNV based study on autism families (Sebat et al, 2007).Since then, multiple large studies have been performed using different platforms to examine different cohorts. Several novel or rare CNVs, both de novo and inherited, were identified, especially affecting regions like 1q21, 5p15.2, 7q11-13, 15q13.3, 16p11.2, 17p11.2 and 22q11.2 (Malhotra and Sebat, 2012; Prasad et al, 2012). A list of recurrent CNVs associated with the disorder identified in multiple studies across various populations is shown in Table -2 (extracted from SFARI gene database). De novo events were often more deleterious than inherited variations (Veltman and Brunner, 2012). The reported CNVs in autism are often enriched in genes with important functions in the nervous system like post synaptic translational regulation, neuronal cell adhesion, neuronal activity modulation and excitatory and inhibitory function imbalance (Devlin and Scherer, 2012). Functional impact of CNVs were recently been investigated by studying the expression profile of genes within the rare de novo CNVs (Luo et al, 2012). Genes which are involved in neural-related pathways were found to be mis-expressed.

Most recent data supports an oligogenic model in which severity of the neurodevelopmental disease increases with the increasing CNV burden (Coe et al, 2012). Smaller CNVs with a median size of 18 kb which are often not detected by SNP microarrays were recently analyzed using exome sequence data (Krumm et al, 2013). It was found that probands inherited more CNVs than their siblings and that their CNVs affected more number of genes. A novel study focused on genomic ‘hotspot’ areas which are prone to recurrent rearrangements provided evidence for a phenotypic dependence on the size of the CNV (Girirajan et al, 2013a). As the size of deletions increases, nonverbal IQ significantly decreases, but there was no impact on autism severity; and as the size of the duplication increases, autism severity significantly increases but nonverbal IQ was not affected. It was proposed that this increased duplication load in genomic hotspot areas predisposes to autism (Girirajan et al, 2013b).

Exome Sequencing-Unmasking Rare Variants

Though human genome consists of nearly 3 billion nucleotides, only a small portion (around 1.5%) of it actually constitutes the exons which are translated into proteins. Instead of sequencing the whole genome which still requires considerable cost and time, it would be an easier and faster strategy to sequence all the protein coding regions of the genome. The “exome” consists of all the exons of the genome and has a total size of only around 38Mb. Exome sequencing involves the targeted capture of exons followed by the high-throughput sequencing of exon-enriched samples using next generation sequencers. Genome-wide interrogation of the protein coding regions is a powerful approach to identify causal variants and candidate genes in individuals and families with a complex genetic disorder.

Over the past three years, several independent groups have conducted whole exome sequencing using non-overlapping ASD samples (Iossifov et al, 2012; Neale et al, 2012; O’Roak et al, 2012b; Sanders et al, 2012; Nava et al, 2012; Michaelson et al, 2012; Yu et al, 2013; Jiang et al, 2013). O’Roak et al, (2012a), using 189 trio families, showed that de novo point mutations are overwhelmingly paternal in origin. Such a paternal bias in de novo point mutations has also been reported by Iossifov et al (2012), during the same period. Several novel genes were identified through these studies as playing a role in liability for ASD. These results further highlight the extreme genetic heterogeneity of ASD, while pointing towards a relatively small number of implicated biological pathways. Many of these de novo mutations identified disrupted genes in neural development and plasticity, chromatin related proteins involved in transcriptional regulation, especially during prenatal brain development (Ben-David and Shifman, 2013).

Genetic Tests for Autism

The high heterogeneity and the absence of common variants make the genetic testing of non-syndromic forms of ASD, which account for around 90% of the cases, very difficult. Still some laboratories and companies are now offering panels or arrays which target several dozens of autism candidate genes.

Future directions

Estimation of the functional consequences of CNVs and rare mutations, and the translation of this knowledge to support clinical decisions towards personalized pharmacological interventions is a real future challenge. Much greater wealth of available data concerning CNVs, rare variants, common variants and gene pathways involved in autism is indeed spearheading this ultimate aim of developing novel pharmaceuticals and tools for early genetic diagnosis. Further, in exploring the genetic landscape of ASDs, we need to broaden our perspective of the etiology to examine the extensive gene-regulatory elements, including epigenetic modifications and non-coding RNAs, which, in turn, are both influenced by the environment.

References