Genetic testing for autism: recent advances and clinical implications

Keywords::

• autism
• chromosomal disorders
• comparative genomic hybridization
• DNA copy number variations
• genetic testing
• microarray analysis

Understanding the genetics of autism guides the development of clinical genetic testing

Autism is defined and diagnosed as a pattern of behaviors, including difficulties with communication and social interactions, and repetitive or stereotypical behaviors. Autism spectrum disorders (ASDs), the broad category that includes autism and related disorders, are common, with a prevalence of one in 110 or 9.0 per 1000 individuals (95% CI: 8.6–9.3) [1]. Twin studies showing concordance rates approaching 90% among monozygotic twins, and the high recurrence risk for parents of a child along the spectrum, point to a strong genetic component [2,3]. Discovery of genetic variants contributing to ASD is the first step in developing clinical diagnostic genetic tests, but most cases are still unexplained by current genetic testing and will require ongoing discovery to improve the yield from clinical testing.

The ‘common disease–common variant model’ suggests that common diseases such as autism will be caused by common variants. Linkage studies, candidate gene studies and genome-wide association studies [4,5] have identified possible causative genes for autism, but have so far not identified robust associations that replicate across ASD cohorts. Mutations in synaptic genes such as NLGN3 and NLGN4 are rare causes of ASDs that can be tested in a clinical setting, but should not be considered first-tier tests owing to low yield in the ASD population as a whole. No single gene appears to explain a large proportion of cases, as reviewed elsewhere [6,7].

An ideal clinical genetic test for ASDs would identify genetic susceptibility to autism in a majority of patients, but this is an elusive goal for complex disorders, such as ASDs, where genetic heterogeneity has, so far, been the rule. The genetic landscape for ASDs includes a combination of multiple rare point mutations, chromosomal deletions and duplications with variably penetrant Mendelian effects, and also many rare variants that may contribute to ASD susceptibility through oligogenic or polygenic inheritance.

Recent advances in clinical genetic testing for autism

Analysis of genome-wide copy number variants (CNVs) has, so far, been the most successful strategy for identifying both recurrent and unique genetic events affecting patients with ASDs [8–12]. CNV pathogenicity is typically inferred based on features such as gene content and the frequency of de novo events, and comparison with CNV frequency in control populations. Although still rare at the population level, several recurrent CNVs have been identified, including deletions or duplications at chromosomes 1q21.1 [13], 15q11 [14,15], 15q13.3 [16,17], 16p11.2 [9,18–21] and 22q13.3 [22], all of which are associated with neuropsychiatric disorders, including autism.

Published recommendations for the clinical genetic testing of patients with ASD struggle to match the rapid pace of discovery, but typically include testing for chromosomal disorders and fragile X syndrome as a standard minimum [23]. Cytogenetic testing has traditionally involved a G-banded karyotype, with yields of approximately 2–3% for patients with ASDs. Chromosomal microarray (CMA), a form of array comparative genomic hybridization, detects the vast majority of events seen by karyotype and many more events below karyotypic resolution with yields of 7–10% [8,24]. Clinical laboratories have rightfully adopted high-density oligonucleotide arrays with whole-genome coverage as a clinical test for patients with ASDs, a considerable advantage over earlier platforms that provided coverage targeted more to specific recurrent deletions and duplications [25,26]. The increased yield supports the argument that CMA should replace G-banded karyotype as the first-tier cytogenetic test for patients with ASD [27].

Copy number analysis has been useful in ASD, not only by identifying large or de novo events, but also through its ability to generate a list of candidate genes for further analysis. This approach has identified several genes that would be expected to have an impact on brain development and function, such as CNTNAP2[28–30], NRXN1[11,31,32], PCDH10[33], SHANK2[12], SHANK3[34,35] and SYNGAP1[12]. These genes are candidates for clinical genetic testing but, since each gene is a rare cause of ASD, they are not yet clinically available.

Ideally, a clinician would use the clinical evaluation to guide the choice of a specific test and increase the chances of a positive result. Macrocephaly represents an important diagnostic clue among patients with ASD. Mutations in PTEN have been reported in some patients with ASD, but typically in the presence of extreme macrocephaly, with head circumference measurements ranging from +3.7 standard deviation (SD) to +9.6 SD (average: +5.4 SD) [36]. Although macrocephaly is not uncommon among patients with ASD, PTEN mutations are still likely to account for only a small fraction of all ASD cases. For most patients with ASD, a physical examination provides insufficient clues to guide the selection of a single genetic test.

Clinical implications of genetic testing for autism

Clinical genetic testing facilitates better patient care through:

• • More accurate genetic counseling regarding recurrence risks. As a population average, the recurrence risk for autism is more than ten-times higher for parents with a child with ASD, but for each family, specific recurrence risks may differ from the average, ranging from a very low risk of 1–2%, to a risk as high as 50%;

• • Early identification of children with a genetic susceptibility to ASD. This can facilitate early behavioral interventions that improve developmental outcomes [37–40]. Even for a child with an affected older sibling, an ASD diagnosis would not be made clinically until after symptoms develop, whereas a child with a genetic diagnosis could qualify for services at an earlier age;

• • Sparing the patient and family a ‘diagnostic odyssey’, owing to a clear genetic diagnosis, involving multiple rounds of diagnostic testing.

Further development of autism genetic testing

Although the use of CMA in clinical practice represents a tremendous improvement over traditional G-banded karyotype in clinical testing, most patients with ASD still do not have a genetic diagnosis. One strategy would be the incremental adoption of more single-gene clinical tests to assess for rare causes of ASDs, but the lack of identifying physical characteristics and the genetic heterogeneity of ASDs implies that each patient would require numerous tests. As clinicians operating in an increasingly economically challenging healthcare environment, however, we will be obligated to make efficient choices about diagnostic testing, guided always by the desire to add value for the patient.

Costs of sequencing and other forms of genetic analysis are decreasing exponentially. The increased diagnostic yield of CMA over karyotype is due primarily to increased resolution. Ongoing revisions of copy number arrays will add information content, such as high-resolution coverage of specific genes, to improve diagnostic yield. One effect of genomic copy number gains or losses could be alterations in gene expression, and gene-expression analysis may play a role in clinical testing for ASDs in the near future. Analysis of other genetic alterations, such as aberrant methylation, will become more routine in clinical practice. Finally, dramatic cost reductions for whole-genome sequencing will usher in a new era for genetic discovery and clinical genetic testing.

The near-term affordability of whole-exome and/or whole-genome sequencing will facilitate gene discovery for ASDs, but this approach is most effective for patients with a distinctive phenotype caused by highly penetrant Mendelian variants [41,42]. The genetic heterogeneity of ASDs represents a significant challenge. Improvements in bioinformatics will help interpret the numerous coding region mutations that will be identified [43], but estimates that every person carries approximately 175 de novo point mutations underscores the challenge [44]. Careful selection of ASD cohorts will also facilitate the discovery process. When applied to patients with ASDs, for example, recessive variants may be easiest to identify, as demonstrated among patients ascertained from consanguineous matings [33].

Next-generation sequencing is just now being adopted by some clinical diagnostic laboratories, primarily as panels of multiple genes related to a single disorder. Advantages of this approach include the massive throughput and the ability to revise and update such panels as new genes and variants are discovered. There are challenges in the implementation in clinical diagnostic settings, both technical [45] and interpretive [43,46], but this approach holds tremendous promise for genetically heterogeneous disorders such as ASDs.

Because ASDs are so common, commercial interest in diagnostic testing will remain strong and will probably increase. Advocacy groups and clinicians who care for patients with ASDs must be aware of the pitfalls of direct-to-consumer testing [47]. Concern is warranted because of the complexity of information involved and the likelihood that some laboratories will promote tests aimed at predictive or presymptomatic testing with the goal of starting early treatment. Patients need to be guided in the use of this clinical information, so genetic testing for ASDs should only be performed by healthcare providers who are capable of explaining the significance of either positive or negative test results [48]. As clinicians, we should balance our enthusiasm for finding a genetic diagnosis with the recognition that autistic traits represent one aspect of a diverse behavioral spectrum, and work to avoid any potential stigmatization of the patient and family through identification of genetic susceptibility. With this in mind, clinical genetic testing for autism can be pursued with the goal of improving patient care.

Financial & competing interests disclosure

The author has no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

No writing assistance was utilized in the production of this manuscript.