Clinical and Genetic Heterogeneity of Autism

Autism (MIM 209850) comprises a heterogeneous group of disorders with a complex genetic etiology, characterized by impairments in reciprocal social communication and presence of restricted, repetitive and stereotyped patterns of behavior [1]. With an early onset prior to age 3 and prevalence as high as 0.9–2.6% [2,3], autism occurs predominantly in males, with a ratio of male: female of 4 to 1. It is one of the leading causes of childhood disability and inflicts serious suffering and burden for the family and society [4].


Introduction
Autism (MIM 209850) comprises a heterogeneous group of disorders with a complex genetic etiology, characterized by impairments in reciprocal social communication and presence of restricted, repetitive and stereotyped patterns of behavior [1]. With an early onset prior to age 3 and prevalence as high as 0.9-2.6% [2,3], autism occurs predominantly in males, with a ratio of male: female of 4 to 1. It is one of the leading causes of childhood disability and inflicts serious suffering and burden for the family and society [4].
Diagnosis of autism is based on expert observation and assessment of behavior and cognition, not etiology or pathogenic mechanism. This is further emphasized by the current trend in the DSM-V, in which the category of Asperger syndrome is removed and the diagnostic criteria for autism are modified under the new heading of autism spectrum disorder (ASD). The change in diagnostic criteria is not based on known similarities or differences in causation between these clinically defined categories, but rather on the consensus of opinions of expert clinicians. For autism, several diagnostic instruments are available. Two are commonly used in autism research: the Autism Diagnostic Interview-Revised (ADI-R) that is a semi-structured parent interview [5], and the Autism Diagnostic Observation Schedule (ADOS) uses observation and interaction with the child(ren) [6]. The Childhood Autism Rating Scale (CARS) is used widely in clinical environments to assess severity of autism based on observation of children [7]. The M-CHAT was developed in the late 1990s as a first-stage screening tool for ASD in toddlers' age 18 to 24 months, with a sensitivity of 0.87 and a specificity of 0.99 in American children [8,9].

Clinical heterogeneity of ASD
Autistic conditions are a spectrum of disorders, rather than a distinct clinical disorder, which means that the symptoms can be present in a variety of combinations with a range of severity. The disease has variable cognitive manifestations, ranging from a non-verbal child with mental retardation to a high-functioning college student with above average IQ with

Autism is a complex genetic disorder
It is widely held that autism is largely genetic in origin; several dozen autism susceptibility genes have been identified in the past decade, collectively accounting for about 20% of autistic cases. There is strong evidence from twin and family studies for the importance of complex genetic factors in the development of autism [12,13]. Family studies have shown that a recurrence rate of autism in siblings of affected proband is as high as 8-10% [12,14]. Thus, the recurrence risk in siblings is roughly 100 times higher than that found in the general population. The substantial degree of familial clustering in ASD could reflect shared environmental factors, but twin studies strongly point to genetics. Several epidemiological studies among sex-matched twins have clearly demonstrated significant differences of concordance rates in the monozygotic (MZ) and dizygotic (DZ) twins. The largest of these studies [15] found that 60% of the MZ pairs were concordant for autism compared with none of the DZ pairs, suggesting a heritability estimate of >90% assuming a multifactorial threshold model. This is what is observed in every twin study in autism, and is overall consistent with heritability estimates of about 70-80% [15,16]. One exception is a very recent study with a large sample of twins, which, despite showing a concordance of about 0.6 for MZ twins and 0.25 for DZ twins, comes to the conclusion that shared environment plays a larger role than genetic factors [17]. However, the question of how a shared environment would have a more major role than genetics is not clear. Moreover, studies in families show that first-degree relatives of an autistic proband have a markedly increased risk for autism relative to the population, consistent with a strong familial or genetic effect observed in twins [18]. This is not to dispute the role of the environment but to emphasize that genes play an important role. Similar to other common diseases with genetic contributions, autism was thought to fit a model in which multiple variants, each with small to moderate effect sizes, interact with each other and perhaps in some cases, environmental factors, to lead to autism; a situation referred to as complex genetics [13].

Genetic heterogeneity of autism
Although autism is highly heritable, the identification of candidate genes has been hindered by the heterogeneity of the disease. Autism genetics is highly complex, involving many genes/loci and different genetic variations, including translocation, deletion, single nucleotide polymorphism (SNP) and copy number variation (CNV) [13,19,20]. The most obvious general conclusion from all of the published genetic studies is the extraordinary etiological heterogeneity of autism. No specific gene accounts for the majority of autism; rather, even the most common genetic forms account for not more than 1-2% of cases [21]. Further, these genes, including those mentioned earlier, represent a diversity of molecular mechanisms that include cell adhesion, neurotransmission, synaptic structure, RNA processing/splicing, and activity-dependent protein translation. Genetic heterogeneity of autistic cases has been documented by identification of single gene mutations and genomic variations including CNV.

Genotype/phenotype correlation in ASD
The presence of genetic and phenotypic heterogeneity in autism with a number of underlying pathogenic mechanisms is highlighted in this current review. There are at least three phenotypic presentations with distinct genetic underpinnings: (1) autism with syndromic phenotype characterized by rare, single-gene defects ( Table 2); (2) broad autistic phenotypes caused by genetic variations in single or multiple genes, each of these variations being common and distributed continually in the general population but resulting in variant clinical phenotypes when it reaches a certain threshold through complex gene-gene and gene-environment interactions; and (3) severe and specific phenotype caused by 'de-novo' mutations in the patient or transmitted through asymptomatic carriers of such mutations (Table 3) [48,49]. Understanding the neurobiological processes by which genotypes lead to phenotypes, along with the advances in developmental neuroscience and neuronal networks at the cellular and molecular level, are paving the way for translational research involving targeted interventions of affected molecular pathways and early intervention programs that promote normal brain responses to stimuli and alter the developmental trajectory [50]. Recent genetic results have improved our knowledge of the genetic basis of autism. Nevertheless, identification of phenotypic markers remains challenging due to phenotypic and genotypic heterogeneity.

Gene
Genetic alteration Location Reference

FMR1
The  Abbreviations: ID, intellectual disability; SCZ, schizophrenia; TS, Tourette syndrome; SLI, speech and language impairment; ADHD, attention deficit hyperactivity disorder Table 3. Severe and specific phenotype with rare variants of genes

Copy number variation (CNV): A paradigm shift in autism
The strong genetic contribution shown in family studies and the association of cytogenetic changes, but apparent lack of common risk factors in autism, led to a hypothesis that rare sub-microscopic unbalanced changes in the form of CNVs likely contribute to the autism phenotype. With the development of microarrays capable of scanning the genome at submicroscopic resolution, there is accumulating evidence that multiple CNVs contribute to the genetic vulnerability to autism [80]. de novo CNV has been identified in up to 7-10% of sporadic autism [81,82], but are less frequent in multiplex families, in which CNV accounts only for about 2% of families screened [80,83]. This could possibly suggest different genetic liabilities in simplex and multiplex autism. Recurrent CNVs at 15q11-13 (1-3% of autism patients), 16p11 (1% of autism patients), and 22q11-13 have been confirmed in multiple studies [80,[83][84][85][86]. This hypothesis also has been proven largely successful in identifying autism-susceptibility candidate genes, including gains and losses at SHANK2 [87], SHANK3 [88], NRXN1 [13], NLGN3 and NLGN4 [37], and PTCHD1 [89,90]. Neurexins and neuroligins are synaptic cell-adhesion molecules (CAMs) that connect pre-and postsynaptic neurons at synapses, mediate trans-synaptic signaling, and shape neural network properties by specifying synaptic functions. The Shank family of proteins provides scaffolding for signaling molecules in the postsynaptic density of glutamatergic synapses. Genes encoding CAMs play crucial roles in modulating or fine-tuning synaptic formation and synaptic specification. Localization and interacting proteins at the synapse is shown in Figure 1. It is apparent that many different loci, each with a presumably unique yet subtle contribution to neurodevelopment, underlie the phenotype of autism. These observations have resulted in a paradigm shift away from the previously held "common disease-common variant" hypothesis to a "common disease-rare variant" model for the genetic architecture of autism. The central tenet of this model suggests a role for multiple, rare, highly penetrant, genetic risk factors for ASD, many of which are in the form of CNV. To make sense of the contribution of CNVs to autism, a "threshold" model has been proposed [80]. The model posits that different CNVs exhibit different penetrance depending on the dosage sensitivity and function (relative to autism) of the gene(s) they affect. Some CNVs have a large impact

PSD95
Fyn on autism susceptibility and these are typically de novo in origin, cause more severe autistic symptoms, are more prevalent among sporadic forms of autism, and are less influenced by other factors like gender and parent of origin. Other CNVs have moderate or mild effects that probably require other genetic (or non-genetic) factors to take the phenotype across the autistic threshold.

Epigenetics plays an important role in autism
In addition to structural genetic factors that play causative roles for autism, environmental factors also play an important role in autism by influencing fetal or early postnatal brain development, directly or via epigenetic modifications. Epigenetic modifications include cytosine methylation, post-translational modification of histones, small interfering RNA and genomic imprinting. Involvement of epigenetic factors in autism is demonstrated by the central role of epigenetic regulatory mechanisms in the pathogenesis of Rett syndrome and fragile X syndrome (FXS), both are the monogenic disorders resulted from single gene defects and commonly associated with autism [38][39][40]. FXS is a result of a triplet expansion of CGG repeats at the 5' untranslated region of FMR1 gene, which encodes the FMRP (fragile X mental retardation protein). FMRP is proposed to act as a translation regulator of specific mRNAs in the brain and involved in synaptic development and maturation, through its nucleo-cytoplasmic shuttle activity as an RNA-binding protein. It has been shown that FMRP uses its arginine-glycine-glycine (RGG) box domain to bind a subset of mRNA targets that form a G-quadruplex structure. FMRP has also been shown to undergo the post-translational modifications of arginine methylation and phosphorylation [91,92]. Our recent study demonstrated that alteration of methylation patterns at loci of Neurex1 and ENO2 are associated with autism [Wang and Zhong, manuscript in preparation].
Research has recently focused on the connections between the immune system and the early development of brain, including its possible role in the development of autism [106]. Immune aberrations consistent with a deregulated immune response may target neuronal development and differentiation [107,108]. Our study has suggested that a close contact with natural rubber latex (NRL) could trigger an immunoreaction to Hevea brasiliensis (Hev-b) proteins in NRL and resulted in autism [109]. This led us to a hypothesis that immune reactions triggered by environmental factors could damage synapse formation and neuronal connections, which would result in missing normal structure or function of synaptic proteins that are encoded by genes NLGNs, NRXN1, CNTNAPs, SHANKs, or in deregulation of gene expression of FMR1, PTEN, FOXPs, and GRIK2.

Converging molecular pathways of autism
Autism is a heterogeneous disorder with a fundamental question of whether autism represents an etiologically heterogeneous disorder in which a myriad of genetic or environmental risk factors perturb common underlying molecular pathways in the brain [110]. Two recent studies have suggested there could be convergence at the level of molecular mechanisms in autism. The first study on molecular convergence in autism identified protein interactors of known autism or autism-associated genes [111]. This interactome revealed several novel interactions, including between two autism candidate genes, SHANK3 and TSC1. The biological pathways identified in this study include synapse, cytoskeleton and GTPase signaling, demonstrating a remarkable overlap with those identified by the gene expression. The second, an analysis of gene expression in postmortem autism brain, provides strong evidence for a shared set of molecular alterations in a majority of cases of autism. This included disruption of the normal gene expression pattern that differentiates frontal and temporal lobes and two groups of genes deregulated in autistic brains: one related to neuronal function, and the other related to immune/inflammatory responses [111]. Genes associated with neuronal function were enriched in metabolic signal pathways, providing evidence that these changes were causal, rather than the consequence of the disease [112]. In contrast, the immune/inflammatory changes did not show a strong genetic signal, indicating a non-genetic etiology for this process and implicating environmental or epigenetic factors instead. These results provide strong evidence for converging molecular abnormalities in autism, and implicating transcriptional and splicing deregulation as underlying mechanisms of neuronal dysfunction in this disorder.

In summary
Autism is a heterogeneous set of brain developmental disorders with complex genetics, involving interactions between genetic, epigenetic and environmental factors. The heterogenerous genetics involves many genes/loci and different genetic variations in autism, such as deletion, translocation, SNP and CNV. Recent studies have also suggested there could be convergence at the level of molecular mechanisms in autism. Although the genetic basis is well documented, considering phenotypic and genotypic heterogeneity, correspondences between genotype and phenotype have yet to be well established.