**8. Single Nucleotide Polymorphisms (SNPs), Quantitative Trait Locus (QTL) and risk for ASD**

Of the techniques that supported the rapid advance in molecular genetics, comparative genom‐ ic hybridization (CGH) and single nucleotide polymorphism (SNP) genotyping array have al‐ lowed the qualitatively and quantitatively identification of the genes and genomic rearrangements involved in the autistic phenotype (Alárcon et al., 2008; Whalley, 2011; Mefford et al., 2012). The use of these molecular sequencing tools has also enabled the identification of high frequencies of common variants in some genes, such as SNPs, in individuals with SD. SNPs, causing about 50% of all currently known variations in genetic material, are the most nu‐ merous variants in the genome. Moreover, SNPs are considered genetic markers that can be used to identify genes associated with complex diseases (Malhotra & Sebat, 2012).

important to note that most of these act on neurotransmission in the central nervous system

There are reports of more than two hundred candidate genes in the literature. According to Swanwick et al. (2011), they can be classified into four categories: 1) rare - genes involved in rare monogenic forms of ASD. This type of allelic variant includes rare polymorphisms and mutations directly related to ASD (*NRXN1* and *SHANK3*); 2) syndromic - genes related to syndromes with phenotypic manifestations in a significant subpopulation of carriers that in‐ clude autistic symptoms *(FMR1* and *MECP2*); 3) association - genes with common polymor‐ phisms that confer a small, probably additive risk, for ASD, that were identified from association studies derived from cases of unknown etiology (*MET* and *GABRB1*) and 4) functional - genes with functions related to the biology of ASD that are not included in the other categories (*CASDP2* and *ALOX5AP*). Among these, the ones that belong to the first two categories are the most strongly related to the pathogenesis of ASD (El-fishawy & State, 2010). There are indications that *de novo* point mutations occur in approximately 5 to 20% of

Persico and Bourgeron (2008) proposed that there are three main pathways involved in the pathogenesis of ASD. The first entails genes that affect cell migration, the second disruptions of the glutamate-GABA harmony and the third involves synaptic formation and maintenance and dendritic morphology. All these pathways play a fundamental role in the central nervous sys‐

Recent studies have given more support to evidence that a large subset of genes, involved in the outgrowth and guidance of axons and dendrites, is implicated in the etiology of autism. How‐ ever, many studies are still needed in order to understand the role of isolated genes and gene re‐ gions in ASD and to identify the associations between them and to identify new candidate genes that act within the molecular pathways (Hussman et al., 2011; Griswold et al., 2012).

But despite the large number of previously identified candidate genes, the number of pa‐ tients with changes in these genes does not reach 1% of the total cases, which further high‐ lights the extreme heterogeneity seen in the pathogenesis of ASD. The findings have led to a paradigm shift in the concept of the genetic architecture of common neurodevelopmental diseases, stressing the importance of individual patterns, rare mutations and overlapping in genetic etiology. They have also converged on specific neurodevelopmental pathways, pro‐

**8. Single Nucleotide Polymorphisms (SNPs), Quantitative Trait Locus**

Of the techniques that supported the rapid advance in molecular genetics, comparative genom‐ ic hybridization (CGH) and single nucleotide polymorphism (SNP) genotyping array have al‐ lowed the qualitatively and quantitatively identification of the genes and genomic rearrangements involved in the autistic phenotype (Alárcon et al., 2008; Whalley, 2011; Mefford

tem, particularly in the serotonergic process (Berkel et al., 2010; Durand et al., 2007).

viding insights into pathogenic mechanisms (Mitchell, 2011).

(Sherer & Dawson, 2011).

222 Recent Advances in Autism Spectrum Disorders - Volume I

the cases (O'Roak et al., 2011).

**(QTL) and risk for ASD**

Unlike CNVs, SNPs seem to be more penetrating in ASD. *De novo* SNPs, although less fre‐ quent, seem to have more deleterious effects and confer higher risk for autistic behavior (Chahrour et al., 2012). Associations between some SNPs of mitochondrial and nuclear genes and predisposition to ASD have been reported by several studies (Ramoz et al., 2004; Silverman et al., 2008; Smith et al., 2009). Studies have shown SNPs in the *GABRA2, GA‐ BRA3, GABRA4, SLC25A12, FOXP2, CNTNAP2, CNTNAP2* and *BDNF* genes (Li et al., 2005; Segurado et al., 2005; Alárcon et al., 2008; Cheng et al., 2009; Scott-Van Zeeland et al., 2010; Jiao et al.,2011). In the Genome-wide Association Studies (GWAS) involving 4305 individu‐ als with ASD and 6941 controls, strong association signals were revealed in six SNPs of two genes encoding neuronal cell-adhesion molecules, cadherin 10 (*CDH10*) and cadherin 9 (*CDH9*). These findings were replicated in two independent cohorts and implicated neuro‐ nal cell-adhesion molecules in the pathogenesis of ASD (Wang et al., 2009).

It is possible that screening for SNPs may identify new biological mechanisms that are in‐ volved in predisposition to ASD. However, GWAS, although promising, has revealed few common alleles and many results have still to be replicated (Ma et al., 2009; Manolio et al., 2009; Klein et al., 2010). It is clear that SNPs may have variable expressions or reduced pene‐ trance. But, while it is apparent that rare variations can play an important role in the genetic architecture of these diseases, the contribution of common variations to risk for ASD is less clear (Jiao et al., 2011; Sherer & Dawson, 2011). Additionally, a strong Association between SNPs in the 5p14.1 and 5p15.2 regions and ASD has also been reported (Wang et al., 2009).

However, the results of Stage 2 of the Autism Genome Project Genome-Wide Association Study, which incremented 1301 ASD families to the investigation bringing the total to 2705 families analyzed (Stages 1 and 2), showed that no single SNP has a significant association with ASD or selected phenotypes at the genome-wide level and concluded that common variants affect the risk for ASD but their individual effects are modest (Anney et al., 2012).

These controversial results about the role of SNPs in the predisposition for ASD do not rule out participation in the phenotype, but motivate the investigation of biological phenomenon that would explain their participation. Probably a single SNP does not affect the risk, but perhaps the additive effect of several SNPs, in specific combinations, with the participation of environmental determinants cannot be discarded in etiologically complex diseases.

From another standpoint, the Quantitative Trait Locus (QTL) approach is one of the most suitable methods to find susceptibility of loci. This approach follows the assumption that ASD occur as a continuum of severity, a position supported by findings of elevated levels of ASD symptoms in parents and siblings of cases compared to controls, and variations in ASD traits that have been found in the general population. One study to identify the loci that un‐ derlie ASD symptoms in children with attention-deficit/hyperactivity disorder (ADHD) in‐ vestigated both the total level of ASD symptoms as well as scores of three ASD symptom domains, thus taking into account potential differential genetic origins of different ASD symptom domains. QTL linkage analyses for the different ASD domains were carried out using 5407 SNPs spanning the entire genome. Findings suggest that some QTLs are ASD specific, although the 15q QTL potentially has pleiotropic effects for ADHD and ASD (Nij‐ meijer et al., 2011). The genetic analysis of quantitative traits that are phenotypically linked, such as in ASD and ADHD, can reduce the heterogeneity of diagnosis and indicate loci re‐ lated to susceptibility (Lu et al., 2011).

genomic variation (as measured in nucleotides) are similar. Thus, in addition to 0.1% of ge‐ netic difference at the nucleotide sequence level, another 0.1% of genetic difference is appa‐

Genetic Etiology of Autism http://dx.doi.org/10.5772/53106 225

The rate of genome-wide nucleotide substitutions is estimated at 30–100 per generation and ∼1 per exome. In contrast, the global rate of structural mutation events is lower: CNVs > 10 Kb in size occur at a rate of ∼0.01–0.02 per generation (Marshall et al., 2008; Conrad et al., 2011; Levy et al., 2011; Sanders et al., 2011). Nucleotide substitutions probably account for the majority of disease risk alleles, but based on sheer size and potential to impact genes (or multiple genes), structural mutations are, on average, more pathogenic. Thus, CNVs, *de novo* CNVs in particular, seem to be a class of variants that have large effect on disease risk (Mal‐

CNVs are gaining importance in the scenario of ASD. They represent a significant source of genetic diversity and seem to significantly contribute to changed behavioral phenotypes (Se‐ bat et al., 2007; Rees et al., 2011). To have an idea, *de novo* CNVs have already been reported to be three to five times more common in families of individuals with ASD than in controls, and more often presenting the syndromic form of autism, that is, with the most severe phe‐ notypes (Miller et al., 2010; Pinto et al., 2010; Shen et al., 2010; Sanders et al., 2011). In fact, CNVs, in particular *de novo* CNVs involving many genes, confer risk for ASD. However, al‐ though they are important in this respect, they rarely interrupt a single gene or have com‐ plete penetrance and many give a wide-ranging risk including risk for other problems such as intellectual deficiency, epilepsy and schizophrenia (Geschwind, 2011; O'Roak et al., 2011). Up to 40% of CNVs in autism are inherited from apparently normal parents, consistent with the suggestion of incomplete penetrance. Both *de novo* (non- inherited) or inherited CNVs occur at the same locus in unrelated individuals, and some of them coincide with those seen in other gene-related diseases associated with ASD, including developmental delay and in‐ tellectual deficiency (Cook & Scherer, 2008; Lee & Scherer, 2010). Thus, some apparently

*De novo* CNVs have been observed in from 7-10% of cases in simplex families, in 2-3% in multiplex families and approximately 1% of normal controls. Rare *de novo* CNVs have al‐ ready been observed in 5.8-7.9% of carriers and in 1.7-1.9% of unaffected siblings in simplex families (Levy et al., 2011; Sanders et al., 2011), while *de novo* mutations in coding regions participate in < 20% of cases of ASD (Malhotra & Sebat, 2012). In addition, about 10% of ASD cases with *de novo* CNVs have two or more CNVs (Sebat et al., 2007; Christian et al.,

Many of the variations occur in gene regions that contain synaptic genes, and it seems that some involve haploinsufficient regions or dominant inheritance. Others seem to express re‐ cessive forms as in the cases of the *NHE926*, *PCDH10* and *DLA1* genes identified in studies of individuals with consanguineous parents. Other rare variations were found deleted in ho‐

There are descriptions of *de novo* and inherited CNVs, sometimes in combination in a given family, implicating many novel ASD genes such as *SHANK2, SYNGAP1, DLGAP2* and the

mozygous (Bourgeron, 2009; Ramocki & Zoghbi, 2008; Morrow et al., 2010).

rent at the structural level (Malhotra & Sebat, 2012).

hotra & Sebat).

have a pleiotropic effect.

2008; Marshall et al., 2008).

There is a QTL related to language delay located close to the 7q35-36 region. Interestingly this region is mapped in the *CNTNAP2* gene, a strong candidate for predisposition to autis‐ tic behavior; it is well known that communication abilities are qualitatively impaired in au‐ tistic individuals. The significant delay in language ("age of first word") is observed in about half of affected children (Alárcon et al., 2008). The relationship of SNPs in the *CNTNAP2* gene has already been described, as has the association of the gene and its SNPs to language development delay in autistic and non-autistic individuals (Alárcon et al., 2008; Arking et al., 2008; Tan et al., 2010; Stein et al., 2011; Whalley et al., 2011).

Additionally, the transcription factor encoded by the *FOXP2* gene has already been linked to the development of language. This factor binds to the promoter of the *CNTNAP2* gene regu‐ lating its expression during development. There are reports of changes in the *FOXP2* bind‐ ing site in patients with ASD, which suggests that a reduced expression of the *CNTNAP2* gene may be the underlying etiology of one of the phenotypic characteristics of ASD (Vernes et al., 2008; Poot, et al., 2009). In addition to these, it has been suggested that *WNT2* and *EN2* are related to language development in autism (Lin et al., 2012).

Promising results on the influence of genetic bases in neurobehavioral disorders have also been obtained through studies on CNVs. All these emerging genetic technologies have brought more valuable approaches to improve the understanding of the etiology of ASD. Advances in the use of molecular biology tools have provided a promising manner to study gene-gene and gene-environment relationships in disorders (Gurrieri, 2012; Li et al., 2012). This combination of tools in the search for the etiology will reflect in the possibility of target‐ ing the diagnosis, prognosis, early interventions and genetic counseling. However, more da‐ ta and the reproducibility of findings are necessary to establish the genetic components of these diseases.
