Autism spectrum disorder (ASD) is a heterogeneous group of neurodevelopmental disorders characterized by impaired social interaction and communication, and stereotyped behaviors and interests. Currently, a number of genetic syndromes are known to be related to ASD. Abnormalities of chromosome 15 are the most frequently reported in ASD, especially 15q11-q13 duplications which have been reported to occur in 0.25–3% of ASD cases depending on sample ascertainment (Cook et al., 1997; Schroer et al., 1998; Depienne et al., 2009; Sanders et al., 2015). Chromosome 15q11-q13, the Prader-Willi/Angelman syndrome (PWS/AS) critical region, contains a number of imprinted and non-imprinted genes including the UBE3A and GABA_A receptor subunit genes that may be associated with the presence of ASD. Several association studies of these genes have found significant associations with autism (Cook et al., 1997; Cook et al., 1998; Buiting et al., 2016; Khatri and Man, 2019).

Angelman syndrome (AS) is a neurogenetic disorder characterized by severe developmental delay usually noticeable at 6–12 months of age. The notable clinical features include developmental delay, severe speech delay with minimal to no use of word, ataxia, hand flapping, happy demeanor and inappropriate laughter or smiling. Other clinical features associated with AS include microcephaly and seizure with a characteristic abnormal electroencephalogram (EEG) pattern with large amplitude slow-spike waves. Irregular sleep-wake cycles, wide-spaced teeth, frequent drooling and hypopigmentation are also frequently observed (Williams et al., 2006). AS is caused by four known genetic mechanisms including deletion of chromosome 15q11-q13, mutations of UBE3A, uniparental disomy (UPD) of chromosome 15 and imprinting center defect (Buiting et al., 2016; Beygo et al., 2019). While the loss of function of the maternally inherited UBE3A gene causes AS, abnormalities of chromosome 15q (deletions/duplications) that contain the UBE3A gene have been reported in association with ASD (Cook et al., 1997; Depienne et al., 2009; Vatsa and Jana, 2018; Khatri and Man, 2019), suggesting that the UBE3A gene may also be implicated in autism. AS has a high comorbidity with autism and shows many overlapping clinical features of autism (Steffenburg et al., 1996; Peters et al., 2004; Trillingsgaard and ØStergaard, 2004; Bonati et al., 2007). AS is also considered as a syndromic form of ASD. However, previous studies have shown a wide range of ASD prevalence in patients with AS (Chan et al., 1993; Steffenburg et al., 1996; Peters et al., 2004; Trillingsgaard and ØStergaard, 2004; Sahoo et al., 2006; Bonati et al., 2007; Peters et al., 2012; Moss et al., 2013; Mertz et al., 2014; Wink et al., 2015) but there are very few studies reporting on the prevalence of AS in patients with ASD (Schroer et al., 1998; Depienne et al., 2009). Several recent studies have suggested that ASD is strongly associated with genetic variations, including copy number variations (CNVs). Chromosomal microarray (CMA) has enabled identification of CNVs with a much higher resolution than standard chromosome analysis. In addition, single nucleotide polymorphism (SNP) arrays based on CMA can also detect copy number neutral abnormalities such as regions of long contiguous stretches of homozygosity (LCSH) and UPD which are associated with an increased risk of genetic diseases including PWS and AS. CMA is now recommended as the first-tier clinical diagnostic test for individuals with ASD, developmental delay/intellectual disability, and multiple congenital anomalies of unknown causes (Miller et al., 2010; Battaglia et al., 2013; Schaefer et al., 2013).

In this study, we aimed to identify novel or pathogenic CNVs in selected Thai children with ASD, and we found a patient with atypical AS who had originally been diagnosed as ASD. SNP arrays identified the absence of heterozygosity (AOH) along the long (q) arm of chromosome 15. Subsequently, methylation-specific PCR (MS-PCR) analysis confirmed the AS diagnosis in this patient. A haplotype analysis of chromosome 15 showed that the patient had inherited both chromosomes 15 from his father, indicating paternal UPD consistent with a diagnosis of AS. In addition, we studied 100 patients with ASD for PWS/AS methylation-specific PCR screening, and no other patients with AS or PWS were detected in our ASD cohort.

A 2-year-old boy initially presented with no speech, poor social interaction skills and repetitive behaviors consistent with ASD based on the DSM-IV criteria, which also fulfilled the DSM-V criteria. The patient was the second child born to healthy and unrelated Thai parents. He had one older sister with normal development. His birth weight was 3,600 g with no perinatal complications. He first stood with support at 10 months, but did not walk without support until the age of 27 months. He was unable to hold a spoon and feed himself until 45 months of age. He never developed verbal language skills. His growth curves were within normal ranges, including head circumference, except he had been overweight since 2 years of age (Supplementary Figure S1). His BMI at 10 years was 29.7 kg/m² (>97th centile, Z score = 3.79). He had no hypopigmentation, and had never shown signs of seizure, ataxia or hypotonia. He usually walked on tip toes. He had poor eye contact and had a repetitive behavior in constantly spinning the wheels of a toy car. He engaged in hand-flapping whenever he was happy. He had no self-injury nor aggressive behaviors. He liked to bite things (i.e., clothes, drinking straws). At the age of 5 years, he was assessed using the Autism Diagnosis Observation Schedule (ADOS)-Module 1 by a developmental pediatrician. His scores for communication were 4 (cutoff = 4), 8 for reciprocal social interaction (cutoff = 7), and 1 for repetitive behaviors (Supplementary Table S1). The scores were consistent with a diagnosis of autistic disorder. At 10 years of age, he had a non-verbal IQ rating of 42 (Stanford Binet Intelligence Scales-Fifth edition) and the Vineland Adaptive Behavior Scales-II showed moderate deficits in all three domains, communications, daily living skills and socialization. We also observed that his autistic features evolved as his social interaction skills improved over time. For example, at 11 years of age, he liked to play with his peers and parents, although he still could not speak understandable words.

Initially, he was tested for common genetic causes of ASD, which revealed normal karyotyping and normal CGG repeats on the FMR1 gene. Electroencephalogram (EEG) was also normal. Following the recommendations from the International Standard Cytogenomic Array (ISCA) and the American College of Medical Genetics (ACMG) for the use of CMA, he was screened with the CMA at 10 years of age to look for pathogenic CNVs. Written informed consent was obtained from the patient’s parents for publication of this case report.

We identified an AS patient with ASD caused by paternal UPD of chromosome 15. The patient was first diagnosed with ASD of unknown cause and a chromosomal microarray was performed following the ACMG and ISCA recommendations. The microarray results revealed UPD of chromosome 15. The MS-PCR analysis of this patient showed abnormal methylation of the maternal SNRPN allele and haplotype analysis showed that the patient had paternal uniparental isodisomy, resulting from meiosis II nondisjunction. Therefore, the results of all molecular genetic testing of this patient were consistent with AS. In general, in approximately 75% of individuals with AS, the AS results from maternal 15q11-q13 deletion, which results in a more severe clinical phenotype than other genetic types of AS. AS is also caused by the UBE3A mutations (5–10%), paternal UPD of chromosome 15 (3–7%), or an imprinting center defect (2–3%). The remaining approximately 10% of patients with AS have no detectable genetic abnormality (Buiting et al., 2016; Beygo et al., 2019). Different genetic types in AS patients may show different phenotypes in performance (Yang et al., 2021).

Many aspects of the case presented here are consistent with previous reports of AS patients with the underlying cause of paternal UPD (Lossie et al., 2001; Poyatos et al., 2002; Thompson and Bolton, 2003; Tsai et al., 2004; Varela et al., 2004; Saitoh et al., 2005; Bonati et al., 2007; Depienne et al., 2009; Horváth et al., 2013; Luk and Lo, 2016), who generally had a lower prevalence of seizure, normal head circumference and absence of hypopigmentation (Lossie et al., 2001; Depienne et al., 2009; Horváth et al., 2013; Samanta, 2021). AS patients with UPD also have a higher risk of obesity than deletion type (Lossie et al., 2001; Poyatos et al., 2002; Varela et al., 2004; Saitoh et al., 2005; Brennan et al., 2015; Luk and Lo, 2016). A previous study found that children with UPD showed hyperphagic behavior and their weight increased significantly after the age of 2 years compared with children in the other genetic groups. These UPD children also had significantly higher birth weight than children with deletions or UBE3A mutations (Mertz et al., 2014). AS patients caused by UPD with no ataxia (Poyatos et al., 2002; Saitoh et al., 2005; Horváth et al., 2013; Luk and Lo, 2016) or no happy demeanor (Fridman et al., 2002; Varela et al., 2004; Bonati et al., 2007; Depienne et al., 2009; Luk and Lo, 2016) have also been reported. Moreover, it is notable that, despite one earlier study suggesting a decreased risk of ASD in AS in the absence of seizure (Thompson and Bolton, 2003), our case presented with autistic features without any history of seizure. This case adds to the pool of milder phenotypes in AS patients with UPD compared to a typical AS patients with deletions. A summary of the clinical features of our case and other AS cases with UPD is shown in Table 1. Of note, ataxic gait and frequent laughter/happy demeanor as consistent (100%) clinical features of patients with AS as claimed in a previous report (Williams et al., 2006) were not present in all AS patients with UPD in our literature review (Fridman et al., 2002; Poyatos et al., 2002; Thompson and Bolton, 2003; Varela et al., 2004; Saitoh et al., 2005; Bonati et al., 2007; Depienne et al., 2009; Horváth et al., 2013; Luk and Lo, 2016), supporting our hypothesis that AS resulting from UPD can manifest as a milder AS phenotype.

Previous studies have reported a wide range (1%–100%) of prevalence of ASD in AS caused by deletion, imprinting center defect, UPD or UBE3A mutations. The wide variety of percentages of ASD incidence in AS patients may be due to the small cohorts in such studies, which ranged from 4 to 93 AS cases (Chan et al., 1993; Steffenburg et al., 1996; Peters et al., 2004; Trillingsgaard and ØStergaard, 2004; Sahoo et al., 2006; Bonati et al., 2007; Mertz et al., 2014; Wink et al., 2015). The findings of the studies of ASD in AS are summarized in Table 2. However, many of the AS cases in these studies did not meet the full diagnostic criteria of ASD and lacked genetic verification for the diagnoses of AS. These studies mostly used standardized observation schedules such as ADOS for ASD diagnosis, however the ADOS is not a valid measure for children with mental ages below 18–24 months. Thus social and communication deficits in individuals with AS may be qualitatively different from these phenotypes in individuals with ASD. Among the studies of ASD in AS, Peters et al. and Bonati et al. reported two cases of AS with UPD who met the autism criteria from a total of 4 cases of AS with UPD in their studies. Furthermore, there is little information about the prevalence of AS in patients with ASD. Schroer et al. reported one AS case (a maternal 15q11-q13 deletion) among 100 patients with ASD (1%) and Depienne et al. reported two AS cases in a study of 552 cases with ASD (0.36%), one with a deletion and the other with UPD. Another previous study reported a prevalence of ASD in PWS individuals of 26.7% (Bennett et al., 2015). A systematic review found that ASD occurred more frequently in PWS patients with maternal UPD (35.3%) than those with deletion (18.5%) (Bennett et al., 2015). Although we found no cases of PWS or AS in our 100 patients with ASD, there is sufficient evidence for the occurrence of ASD in AS (Table 2), indicating that MS-PCR screening should be considered for AS and PWS in ASD patients, especially in patients with severe speech impairment and obesity (Lossie et al., 2001; Poyatos et al., 2002; Mertz et al., 2014; Luk and Lo, 2016). MS-PCR analysis is an easy and rapid method to screen for these syndromes simultaneously. MS-PCR analysis can detect more than 99% of affected individuals with PWS and approximately 80% of affected individuals with AS, including those with deletion, imprinting center defect or UPD (Beygo et al., 2019). However, further investigations such as haplotype analysis are needed to elucidate the underlying genetic mechanisms of disease for genetic counseling. The recurrence risk of having AS to the siblings of an affected child depends on the pathogenic mechanism. The risk to siblings is less than 1% if the affected child has a deletion or UPD, but up to 50% if the affected child has an imprinting center defect or a mutation of the UBE3A gene (Beygo et al., 2019).

In conclusion, we report an atypical AS child with autistic features and obesity caused by UPD of chromosome 15. Microarray analysis first identified the absence of heterozygosity of the q arm of chromosome 15. Subsequently, MS-PCR confirmed AS in this patient and microsatellite haplotype analysis revealed that he had received both chromosomes 15 from his father, which confirmed paternal UPD, leading to a final diagnosis of AS with UPD, with clinical features milder than a typical AS patient. Our findings suggest that a PWS/AS investigation may be useful in ASD cases with unknown etiology, as finding the underlying cause of ASD can lead to better genetic counseling for the family of the patient.