WHAT IS ALREADY KNOWN ON THIS TOPIC

WHAT THIS STUDY ADDS

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

Participants are enrolled in the Autism Speaks Genomics of Autism (MSSNG) research study across multiple sites (primary site, The Hospital for Sick Children (SickKids) in Toronto) which includes individuals from varying ancestries (see online supplemental table 1). The primary goal of this prospective cohort study is to identify genetic variants contributing to ASD and the cellular processes involved. Individuals with autism, as well as their immediate family members (eg, parents, siblings), are eligible to participate. Voluntary enrolment involves providing a biological sample for genomic analysis (ideally trio analysis, but can involve singleton and multiplex analysis) and completing standardised questionnaires to supply clinical presentation information (ie, phenotype data).12 25 When the genomic analysis is complete, the results are disclosed to the family via the research study team. The consented genomic and phenotypic data are uploaded to the MSSNG database, which is a cloud-based controlled-access data repository that provides a power tool for genomic analysis across the autism study population.12 The cohort described in this paper are those research families who underwent WGS analysis and received a genetic result from genetic counsellors or clinicians. All participants consented to be in the study which allows for publication of results.

The primary research findings generated from the WGS analysis are genetic variants related to autism or NDD. Variants are interpreted using the ACMG variant classification framework with five variant classifications: pathogenic, likely pathogenic, variant of uncertain significance (VUS), likely benign and benign.26 The study reporting protocol focuses on genes associated/potentially associated with autism or NDD.12 Since the genetic findings are not confirmed in a clinically certified laboratory prior to disclosure to participants, the ACMG variant classification terminology was modified to differentiate between research-grade and clinical-grade results: ‘significant’ replacing ‘pathogenic’.26 Research reports are generated if VUS, likely significant variants and significant variants are identified.

A research report is generated for the study participant with autism using a team of genetic counsellors, genomic analysts, investigators and a clinical laboratory geneticist. For the disclosure, the family is contacted by the recruiting site to offer either an in-person or videoconference meeting with the genetic counsellor or study clinician. The families are counselled about their genetic results using a previously outlined approach to communicating complex information.3 It is emphasised that the research results and their interpretation are based on current technology and understanding. As new genomic tools and algorithms are developed and scientific knowledge advances, the family may get updated information about their WGS results, including changes to variant classifications and newly reported variants. During the result disclosure, the research genetic counsellors asked the participant and family how they felt about receiving the research results. Their response was noted in the research file as direct quotes and/or paraphrased and compiled for this manuscript during the chart review process. Following the disclosure of results, the genetic counsellors or study clinician facilitate referrals to genetics centres for clinical confirmation of any significant variants through a Clinical Laboratory Improvement Amendments (CLIA) or Accreditation Canada Diagnostics certified laboratory.

We reviewed all disclosed WGS research reports from January 2012 (when WGS was first implemented) to November 2023; 202 families received results where a genetic variant was disclosed. Here, we describe individuals with autism who received likely significant or significant results relevant to ASD. In addition, we include a few examples of VUS findings that, based on clinical interpretation, are likely to be contributory but require additional supporting evidence.

Using the ACMG description of clinical utility of genetic testing14 and related publications in the NDD/autism population,11 17–19 we designated three outcome categories: (1) genetic diagnosis, (2) counselling benefits and (3) support to family. Descriptions of each category are provided in the Results section. For each family who received genomic sequencing results, we reviewed the research report and any available summary note from the result disclosure to determine if the result impacted each clinical utility category. A single genetic finding can have multiple downstream outcomes; therefore, it can count towards more than one category. Two of the researchers (TS, NH) independently categorised each case to ensure reliability.

Table 1

List of (likely) significant genomic findings in individuals with autism returned to families

Receiving a genetic diagnosis that ends the diagnostic odyssey, informs medical management and prognosis, and treatment is a potential benefit of genetic testing in autism.

In an 18-year-old male patient with autism, intellectual disability, seizures and generalised anxiety disorder, clinical CMA and gene panel testing were uninformative. When he was 23 years old, WGS analysis through this study identified a de novo, recurrent frameshift variant in SHANK3 (MIM: 606230) associated with Phelan-McDermid syndrome (PMS (MIM: 606232)). The research result was clinically confirmed and provided a unifying diagnosis for the participant’s clinical presentation. This genetic diagnosis provided guidance on how to monitor and manage associated features. For instance, a consensus guideline recently published advised on the diagnosis and treatment of seizures in patients with PMS.27 This includes recommendations for brain imaging (eg, MRI) for every individual with PMS having neurological symptomology. In this participant, his seizures were already evident, and he is regularly followed up by his neurologist. For some seizure conditions, condition/gene-specific information regarding medication effectiveness may be available. For PMS, the guidelines recommend general seizure treatment protocols as there is currently no evidence of a specific medication that works best for all patients with PMS and no contraindicated drugs. The parents expressed that receiving a genetic diagnosis helped them understand their child better. Prior to the diagnosis, they felt they were ‘floating with nothing to hold on to and with this diagnosis, they now have a starting point.’ They expressed that this knowledge gave them a sense of control prompting them to plan their child’s future differently, especially financially.

After exhausting all clinically available tests (CMA, fragile X testing and MED12 (MIM: 300188) gene sequencing), a participant with high suspicion of an underlying genetic aetiology was enrolled. The 15-year-old male patient presented with a provisional diagnosis of autism with a history of developmental delay, speech delay, febrile seizures, sensory hypersensitivity and strong food preferences. WGS identified a de novo variant in ARID1B (MIM: 614556), associated with Coffin-Siris syndrome (MIM: 135900). Following results disclosure, the participant was reassessed by a clinical geneticist, and the ARID1B variant was confirmed clinically to be pathogenic, providing the family with a new clinical and molecular diagnosis. As a result, the clinical team reviewed the clinical management guidelines for Coffin-Siris syndrome and ordered screening for associated symptoms, including abdominal imaging to rule out associated renal anomalies. For this participant, the ultrasound came back normal, but could have uncovered a hidden feature of the condition. With this molecular diagnosis, the primary care provider can use anticipatory guidelines (eg, vision, dental, scoliosis) to monitor the general health and well-being of this patient. Here, having a genetic diagnosis can inform additional surveillance and assessments, primarily unrelated to ASD but still impactful to the well-being of the participant.

In a 31-year-old male patient with autism, attention deficit hyperactivity disorder (ADHD), obsessive-compulsive disorder with Tourette syndrome and intellectual disability, WGS analysis identified a de novo missense variant in GRIN2B (MIM:138252). Variants in GRIN2B are associated with GRIN2B-related neurodevelopmental disorder (MIM: 613970, 616139), which is characterised by variable intellectual disability, muscle tone and motor differences, and behavioural differences including autism. Approximately 50% of individuals also have seizures.28 Over several years, this participant was noted to have episodes of incontinence, not being able to chew and appearing slow and tired. He was assessed by a neurologist, but the underlying aetiology of these episodes was not identified. After receiving this result, the participant was assessed by a clinical geneticist, and the GRIN2B variant was clinically confirmed. In addition, the participant’s neurologist was notified of the new genetic diagnosis as the additional GRIN2B-related information may help clarify the seizure query. Whether these episodes are confirmed to be seizures or another clinical feature, this participant contributes to the growing knowledge about the presentation of GRIN2B-related disorder. Emerging research on therapies for GRIN-related disorders showed promise with dietary supplementation. For some patients with variants in GRIN2B, taking L-serine supplements improved behaviour, development and/or seizure frequency, but additional clinical trials are needed to fully understand the benefits and safety of this therapy.29

After being part of the study for 13 years, a male participant with autism, developmental delay/intellectual disability, limited verbal communication, obesity, macrocephaly, distinctive facial features and hypothyroidism received a genetic diagnosis at age 18. The WGS analysis identified a novel de novo frameshift variant in CNOT3 (MIM: 604910), a gene implicated in a newly recognised condition called CNOT3-associated neurodevelopmental syndrome (MIM: 618672). The condition was first characterised in 2019 where 16 patients with a similar presentation were described with de novo variants in CNOT3.30 This participant’s neurodevelopmental phenotype is explained by this finding. However, it is not currently clear whether this finding fully explains his other features such as macrocephaly and hypothyroidism. As more individuals with CNOT3 variants are described, we may learn of more associated features. This story highlights that even with newer technology like WGS to uncover previously unidentifiable variants, the immediate medical impact for this family is limited. However, families like this one, where the participant is now an adult, are valuable contributors to the growing understanding of gene-specific conditions as autistic children age. The parents expressed that this result provided closure, ending their diagnostic odyssey.

Given that the aetiology of autism is multifactorial, with many genes implicated, and varying levels of contribution, knowing the individual variant involved is helpful in all aspects of genetic counselling.

A family with a 10-year-old male child diagnosed with autism, global developmental delay, hypotonia and photorefractive keratectomy enrolled looking for an underlying aetiology for their son’s constellation of symptoms. Clinical genetic examination including CMA and fragile X testing was negative. WGS identified a nonsense variant in ASXL3 (MIM: 615115), a gene where loss-of-function variants cause Bainbridge-Ropers syndrome (MIM: 615485). The variant was present in a mosaic state in the mother, who had no reported clinical diagnoses, but did have a history of speech delay. Given the inheritance and reportedly discordant clinical presentation, the variant was reported out by the research team as a VUS, but a referral for clinical follow-up was suggested. On assessment by a clinical geneticist, the participant’s clinical features aligned with Bainbridge-Ropers syndrome, and the ASXL3 variant was clinically confirmed and classified as pathogenic. Knowing the Bainbridge-Ropers syndrome diagnosis allowed the clinical team to counsel the family on the likelihood of recurrence based on a mosaic inheritance model. Since the proportion of gametes in the mother that carry this variant is unknown, the family was counselled that the chance of having another child with Bainbridge-Ropers syndrome is up to 50%.

When a child is diagnosed with autism in our study, some parents pursue an autism assessment for themselves. Here, the mother received an autism diagnosis following her daughter’s diagnosis. WGS analysis of the 19-year-old daughter with autism, ADHD, mild Tourette syndrome and chronic ankle weakness identified a 15 kb deletion impacting multiple exons of EP300 (MIM: 602700). The deletion in the daughter was maternally inherited; however, this same deletion was found to be a de novo event in the mother (ie, not inherited). Variants in EP300 are associated with Rubinstein-Taybi syndrome (MIM: 613684). Individuals with this condition have a range of clinical presentations including short stature, distinctive facial features and autism. This research finding prompted assessments to look for features associated with Rubinstein-Taybi syndrome in both the mother and daughter. Follow-up clinical testing confirmed this deletion as pathogenic in both the proband and mother, and a geneticist assessment found that they both do not have the typical features associated with Rubinstein-Taybi syndrome. The knowledge that this deletion is pathogenic and likely contributed to the autism in the family provides the daughter with refined genetic counselling information. The de novo status in the mother provides insights for other family members as well.

Having a genetic diagnosis allows families to connect with one another and build a social support network to share experiences. Identifying families with rare variants in shared genes also allows researchers to uncover biologically relevant pathways and develop molecular intervention targets.

WGS of a 7-year-old male child with autism, oromotor apraxia and history of hypotonia, delayed fine and gross motor skills identified a de novo variant in ASH1L (MIM: 607999). The reported associated features included intellectual disability, autism and a broad range of additional manifestations. The limited resources prompted the mother to seek out other families with this ASH1L-associated condition. She wanted to learn what this would mean for her young son and meet other individuals with ASH1L variants. The family formed a non-profit organisation connecting ASH1L families from around the world, as well as researchers studying the gene (https://www.care4ash1l.com/). The organisation’s goal is to advance the understanding of the molecular mechanism of ASH1L to develop tailored therapeutics. Beyond creating an online community to share personal experiences and resources, the organisation has raised funds to support additional research, including stem cell studies, clinical phenotype and natural history studies, and functional work.

In an 18-year-old male patient with autism, WGS identified a de novo nonsense variant in SHANK2 (MIM: 607999). This gene is a member of the SHANK family, which encodes synaptic proteins with important roles in the signalling pathway in the brain. Following the finding of SHANK3 in idiopathic autism,31 our research group followed to implicate SHANK genes in the susceptibility to autism.32 33 For this family, the variant was clinically confirmed and provided an answer for why their son had autism. The question of how SHANK2 variants impact neuron biology remained unclear, which our team investigated using induced pluripotent stem cells (iPSCs). The family consented to participate in this iPSC study and provided blood samples, which were used to generate nerve cells with and without the SHANK2 variant. The functional impact of the variant on neuronal structure, function and signalling was measured. Along with a different SHANK2 variant contributed by another research family, the study found that neurons from participants with SHANK2 variants were overconnected (increased dendrite length and complexity, and synapse number) and overactive (increased frequency of spontaneous excitatory postsynaptic currents) compared with control neurons.34 This study provides insight into the pathophysiology of SHANK2 variants in neurodevelopmental conditions, which may inform research into targeted supports for these families.

For some individuals with ASD and a known genetic syndrome, undergoing genomic analysis can uncover additional genetic contributions and prompt research. This was the case for a 7-year-old female child with Beckwith-Wiedemann syndrome (BWS (MIM: 130650)) due to hypomethylation on chromosome 11p15.5 at IC2. Her clinical team felt that her complex medical history and uncharacteristic clinical presentation was not fully explained by the BWS diagnosis. Research WGS analysis identified a de novo missense variant in AMOTL1 (MIM:614657), which was of uncertain significance at the time. There were three reports in the literature of patients with missense variants in AMOTL1, including the specific variant identified in our participant.35–37 The phenotype described in the three patients overlapped with our participant, prompting us to reach out to the researchers and contributing to a case series characterising a new Mendelian condition in patients with variants in AMOTL1.38 These cohort studies also inform the development of management guidelines. For individuals with AMOTL1 variants, several screening recommendations were suggested: examination for evidence of orofacial clefting and velopharyngeal insufficiency, regular eye examinations to evaluate for myopia, audiology examinations for hearing loss, echocardiograms and ECG for congenital heart disease and arrhythmia, screening for and aggressively treating constipation, evaluation of liver function, scoliosis and obtaining a developmental assessment.

We evaluated genetic results for clinical utility and benefit to an individual family. We did not assess the potential harms of receiving these findings (noting, however, no overt harm was observed). Most individuals who received results were of European ancestry (expected proportion of individuals of this background in Ontario is 66%), although other ancestry groups were represented. We did not explore potential cultural differences in response to genetic findings. While we share some family perspectives on receiving results, we did not systematically collect these experiences from all families that received results.

We present examples of families who have received a genetic diagnosis and discuss our experiences of genetic testing in the autism population. As our understanding of the genetic contributions to autism grows, there is potential to aid in early detection, inform the development of targeted supports and therapies, and provide improved care and outcomes for individuals with autism and their families.

Consent obtained directly from patient(s) or parent(s)/guardian(s).

This study involves human participants and was approved by The Hospital for Sick Children Research Ethics Board (1000080561). All recruiting sites also obtained ethics approval from their respective institutions. Participants provided informed consent to participate in the study before taking part.