Autism spectrum disorders (ASDs) comprise a distinct entity of neurodevelopmental disorders with a strong genetic component. Despite the identification of several candidate genes and causative genomic copy number variations (CNVs), the majority of ASD cases still remain unresolved. We have applied microarray-based comparative genomic hybridization (array-CGH) using Agilent 400K custom array in the first Cyprus population screening for identification of ASD-associated CNVs. A cohort of 50 ASD patients (G1), their parents (G2), 50 ethnically matched normal controls (G3), and 80 normal individuals having children with various developmental and neurological conditions (G4) were tested. As a result, 14 patients were found to carry 20 potentially causative aberrations, two of which were
Autism spectrum disorders (ASD) (OMIM 209850) comprise a group of neuropsychiatric developmental disorders, affecting approximately 1% of the general population [
Genome-wide association studies have linked multiple regions to autism, and the role of genetic alterations within several genes for example,
We have limited our sample heterogeneity by focusing on the small and genetically homogenous population of Cyprus and applied a robust CNV detection technique, namely Agilent 400K custom array-CGH. This platform combines high resolution with relatively low complexity of analysis and can reliably identify deletions and duplications as small as 13 kb. Genetic screening was directed towards four different groups of population: 50 selected patients with ASD or autistic features, their parents, 50 ethnically matched normal controls, and 80 normal individuals having children with syndromic or nonsyndromic mental retardation, developmental delay, or rare neurological syndromes. With the assumption that ASDs are underdiagnosed or misdiagnosed in the Cypriot population, we have also performed a clinical reevaluation of a group of patients who were given a preliminary diagnosis of “ASD”.
We studied four different groups from the population of Cyprus. Group 1 (G1) includes 50 children (45 boys and 5 girls) 3–18 years of age at the time of recruitment, with a preliminary diagnosis of autism spectrum disorders (ASD). There was no gender-based selection of the participants and the prevalence of males in G1 reflects the general male to female ratio among individuals with ASD; group 2 (G2) includes the nonaffected biological parents of the G1 children (with the exception of the father of patient 11, who has mild mental retardation and autistic features); group 3 (G3) includes a control cohort of 50 normal participants (18 males and 32 females), selected to be older than 30 years of age and have at least two biological children with no mental, neurological, or developmental dysfunction; group 4 (G4) includes 80 normal individuals having children with syndromic or nonsyndromic mental retardation, developmental delay, or rare neurological syndromes. The normal individuals of G4 had participated in previous screening studies, focused on array-CGH testing of their affected children (unpublished data). All patients were specifically selected to have normal karyotype and be negative for fragile-X syndrome.
All patients from G1 were reevaluated by a clinical geneticist to rule-out autistic-like syndromes and retested for ASD based on the Diagnostic and Statistical Manual of Mental Disorders for Physician, Text Revision (DSM-IV-TR) and International Classification of Diseases, Tenth Revision (ICD-10), using Gilliam Autism Rating Scale-2 (GARS-2).
DNA was extracted with Qiagen DNA extraction kit (Qiagen Co) according to the manufacturer’s recommendations. Array-based comparative genomic hybridization (array-CGH) was performed using standard protocols with 400K oligonucleotide custom array platform (Agilent Santa Clara, CA). This array includes the entire 4 × 180 K ISCA (International Standard Cytogenomic Array) design which has a strong clinical emphasis and covers 9269 polymorphic regions derived from the Wellcome Trust case control consortium CNV genotyping array, mainly based on the 42 M array data. In addition, the array includes another 209214 probes providing a backbone coverage of 13 kb.
Image analysis, normalization, and annotation were based on Agilent Feature Extraction 9.1, while Nexus Copy Number 5.1 software (BioDiscovery Inc.) was applied for visualization of data, data analysis, and filtering. The main criteria applied to assess the potential causality of detected CNVs were (i) localization within a gene or gene-rich region, (ii) less than 80% overlap with known benign CNVs and with CNVs present in G3, (iii) overlap with syndrome regions described in Decipher database (
All array-CGH findings presented in Table
List of the detected candidate autism susceptibility aberrations.
Patient | Phenotype | Gender | Event | Size | Chr. Position hg18 | Inheritance | Genes | CNVs | Decipher | ASD regions [ |
In G2 |
In G3 |
In G4 |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
1 | Autism | F | Del | 27.16 kb | chr19:59,840,527-59,868,239 | Father |
|
YES | AUTISM | YES | 1 | 0 | 0 |
| |||||||||||||
2 | ASD | M | Dup | 40.27 kb | chr1:229,985,499-230,025,777 |
|
|
NO | MR/DD | YES | 0 | 4 | 10 |
| |||||||||||||
3 | DD, DF, AF | M | Dup | 304.87 kb | chr5:61,885,191-62,190,063 | Father |
|
NO | MR/DD | YES | 1 | 0 | 0 |
| |||||||||||||
4 | MR, epilepsy, AF | M | Dup | 829 kb | chr11:132,447,787-133,276,750 | Father |
|
NO | AUTISM/SPEECH DELAY | YES | 1 | 0 | 0 |
| |||||||||||||
5 | Autism | M | Del | 167.4 kb | chr2:212,521,344-212,688,773 | Mother |
|
NO | SPEECH DEFECTS | YES | 1 | 0 | 0 |
Del | 191.3 kb | chr3:60,047,028-60,238,353 | Mother |
|
YES | SPEECH DELAY/MR/DD | YES | 1 | 0 | 0 | |||
| |||||||||||||
6 | ASD, MR | M | Dup | 101.78 kb | chr17:9,906,791-10,008,572 | Father |
|
NO | NO | YES | 1 | 0 | 0 |
| |||||||||||||
7 | ASD high functioning | M | Dup | 923.6 kb | chr22:47,373,766-48,297,411 | Mother |
|
NO | AUTISM | YES | 1 | 0 | 0 |
Del | 41.87 kb | chr2:178,247,791-178,289,664 | Mother |
|
NO | SPEECH DELAY | YES | 1 | 0 | 0 | |||
| |||||||||||||
8 | ASD high functioning | M | Del | 101.4 kb | chr3:6,981,830-7,103,523 | Mother |
|
NO | SPEECH DELAY | YES | 1 | 0 | 0 |
Dup | 2.24 Mb | chr1:144,513,497-146,753,802 | Mother | 26 GENES | NO | MICRODEL/DUP SYND | YES | 1 | 0 | 0 | |||
| |||||||||||||
9 | ASD | M | Del | 77.79 kb | chr9:9,598,565-9,691,922 | Father |
|
NO | MR/DD, ADHD | YES | 1 | 0 | 0 |
| |||||||||||||
10 | Psychomotor delay, AF | M | Del | 25.45 kb | chr7:146,067,147-146,092,598 | Mother |
|
NO | AUTISM | YES | 1 | 0 | 0 |
Dup | 41.69 kb | chr16:8,715,900-8,757,596 | Father |
|
NO | MR/DD | NO | 1 | 0 | 0 | |||
| |||||||||||||
11 | MR, AF | M | Del | 1.07 Mb | chr3:197216353-198287118 | Affected father | numerous | NO | 3q29 microdeletion syndrome | YES | 1 | 0 | 0 |
Del | 132 kb | chr4:124119791-124251951 | Mother |
|
NO | WITH MR | YES | 1 | 0 | 0 | |||
| |||||||||||||
12 | MR, AF | F | Del | 428.4 kb | chr3:189499640-189928068 |
|
|
NO | MR/DD | YES | 0 | 0 | 0 |
| |||||||||||||
13 | ASD | M | Del | 142 kb | chr9:28719400-28861226 | Father |
|
NO | AUTISM/MR/DD | YES | 1 | 0.5 | 0 |
| |||||||||||||
14 | MR, ASD | M | Dup | 242 kb | chr4:74032829-74275631 | Mother |
|
NO | SPEECH DELAY | YES | 1 | 0 | 0 |
Del | 96.5 kb | chr1:239,357,689-239,454,247 | Father |
|
NO | MR/DD | NO | 1 | 0 | 0 |
ASD: autism spectrum disorder; MR: mental retardation; DD: developmental delay; AF: autistic features; DF: dysmorphic features; CNV: copy number variation.
Clinical reevaluation of the patients has shown that only 23 out of the 50 participants can be accurately characterized as having nonsyndromic ASD, while the remaining 27 patients have autistic features accompanying mental retardation, developmental or psychomotor delay, epilepsy, and dysmorphism.
400K array-CGH resulted in the detection of twenty potentially causative aberrations, in fourteen patients (Table
Array-CGH profiles, highlighting the two
We present the first investigation of the genetic basis of ASD, carried out for the population of Cyprus. Unlike widely performed large-scale screening studies, we opted to focus on this relatively small and genetically homogenous population in order to maximize the probability of association given a smaller cohort size. Moreover, we included three different ethically matched control groups with the aim to gain more insight into variable expressivity and comorbidity aspects, both of which play an important role in ASD. Thus, even though the number of studied individuals is not large enough to draw statistically significant conclusions, the genetic homogeneity of the participants allowed for some cautious assumptions based on general observations.
One of the challenges encountered in this study was the accuracy of ASDs diagnosis, clearly differentiating them from syndromic neurological conditions, developmental delay, or mental retardation with autistic features. The reevaluation of our patients (G1) has demonstrated that clinical genetic assessment is a vital addition to neurological, psychiatric, or psychological evaluation as a tool for ruling-out genetic syndromes that may affect development and mimic the clinical manifestation of ASD.
In Table
The
The
The above data support the multifactorial model of autism, where potentially causative aberrations exhibit variable penetrance depending on the genetic background, often defined as the synergy of other coexisting aberrations [
Finally, we have observed significant differentiation of G2 (parents of ASD patients) from other nonaffected parent groups in terms of the number of CNVs within genes that are associated with ASD. Six out of fifty mothers and 8/50 fathers from a total of 100 parents (14%) appear to carry 16 different rare variants associated with ASD or MD/DD, not found in international CNV databases nor in any of the studied Cypriot groups. In contrast, the difference between G3 and G4 cannot be considered significant as it concerns only one aberration, which is likely to be a population-specific variant. The presence of such “genetic load” within G2 is an addition to the mounting evidence concerning the role of multiple (genetic and nongenetic) factors in the occurrence of autism or autistic features.
In this study, we addressed some important issues relevant to the genetic characterization of ASDs, such as differential diagnosis, cohort homogeneity, variable expressivity, and comorbidity.
400K array-CGH analysis of our patient cohort along with three ethnically matched control groups has provided supporting evidence about the complexity of ASD aetiology in comparison to other developmental disorders involving cognitive impairment.
Our data have demonstrated that a more targeted approach combining accurate clinical description with high-resolution population-oriented genomic screening is a promising strategy for defining the role of CNVs in autism and identifying meaningful associations on the molecular level. Further studies are required in order to reveal the total of genetic and environmental factors that lead to the disease.
The authors are grateful to all patients and volunteers for participating in the study and to the Cyprus Autism Association and the Cyprus Ministry of Education for assistance in patient recruitment. This work was funded by Telethon Fund 2009 and by Grant nos.