Homozygosity analysis in subjects with autistic spectrum disorder

Introduction

Autistic spectrum disorder (ASD) (MIM 209850) is a complex neurodevelopmental disorder that results in social and communication impairment, as well as repetitive and stereotyped patterns (1,2). Other non-psychiatric symptoms of ASD include intellectual disability, motor delay, hypotonia and seizures (2). Individuals affected with ASD have also been reported to exhibit neuroanatomical structure alterations (3). Epidemiologically, the frequency of ASD is estimated to be ~60 cases per 10,000 children, increasing from early epidemiological studies that originally reported 2–5 cases per 10,000 children worldwide (2,4).

Genetically, ASD has been described as a multifactorial genetic disorder, which in certain cases is associated with a specific mutation or syndrome, such as fragile X syndrome (Xq27.3) (5), tuberous sclerosis (9q or 16p) (6) and Angelman/Prader-Willi syndrome (15q11-q13 chromosomal abnormality) (7,8). Mutations in genes that encode proteins associated with synaptic transmission and activity, including NLGN3/NLGN4 and SHANK3, have been reported in patients with ASD (2,9–11). Other reported changes include protein synthesis-associated genes (FMR1 and PTEN), transcription factors (MEF2) (11), and neurotransmitter proteins and receptors [glutamate and gamma-aminobutyric acid (GABA) receptors] (10). Previous reports have linked ASD to dysfunctional energy metabolism in the central nervous system (CNS), which may describe changes in mitochondrial DNA, or nuclear DNA associated with mitochondrial function (10). In addition, there are numerous cases where ASD appears to arise from a de novo genetic alteration (9).

In the effort to characterize novel genetic alterations associated with ASD, the aim of the present study was to investigate potential susceptibility loci of ASD, utilizing the highly consanguineous and inbred nature of numerous families within the population of Saudi Arabia. Families with ASD-affected individuals were recruited in the present study, and loss of heterozygosity analyses were performed. The present study presents possible regions that may be linked to ASD, as obtained from a loss of heterozygosity analysis.

Materials and methods

Patients and samples

The present study was performed under the approval of the tertiary care center King Faisal Specialist Hospital and Research Center (KFSHRC RAC no. 2080001; Riyadh, Saudi Arabia). The present study was a collaboration between the departments of Genetics and Neuroscience, and the Autism Research Center at KFSHRC. Families with >1 affected individual, regarded as multiplex families, were identified and recruited. The majority of the families recruited contained consanguineous marriages. Examples of family pedigrees from a selection of the recruited families are shown in Fig. 1. Diagnosis of ASD was made at KFSHRC, according to the ADI-R and DSM-VI criteria (1,2). Families with individuals diagnosed with ASD secondary to known genetic or metabolic disorders were excluded. Blood samples (3–5 ml) were obtained in EDTA tubes from family members for subsequent genetic analysis. The family members provided written informed consent in adherence to institutional and international guidelines (RAC#2080001).

Figure 1 Examples of family pedigrees. Consanguinity is evident in these families and the autosomal recessive pattern is observed from pedigrees (~60% of cases are consanguineous). (A) Pedigree presents 3 affected female siblings including twins. (B) Pedigree where the parents of one affected individual are related. There are 2 affected males diagnosed with ASD. (C) Pedigree with two male siblings diagnosed with ASD. Double lines indicate consanguinity. ASD, autistic spectrum disorder.

DNA extraction

DNA was extracted from the peripheral blood samples using the Gentra Systems Puregene DNA Isolation kit (Qiagen, Valencia, CA, USA), according to the manufacturer’s instructions. DNA concentration was determined spectrophotometrically using a Nanodrop (Thermo Fisher Scientific, Waltham, MA, USA).

Loss of heterozygosity analysis

All DNA samples were genotyped using GeneChip® Mapping 250K and 6.0 arrays, and the Axiom Genome-Wide CEU 1 Human Mapping Array (Affymetrix, Santa Clara, CA, USA), according to the manufacturer’s instructions. Briefly, 25–250 ng of high-quality genomic DNA was digested using the restriction endonuclease StyI for the 250K array, and StyI and NspI restriction endonucleases for the 6.0 array. The products were then ligated with an adaptor. Generic primers that recognize the adaptor sequence were used to amplify the adaptor-ligated DNA fragments in the recommended size range. The amplified polymerase chain reaction products were then fragmented with the Affymetrix fragmentation reagent to 50–200 bp, and end-labeled with biotinylated dideoxyadenosine triphosphate using terminal deoxynucleotidyl transferase. The end-labeled fragments were then hybridized to GeneChip® Human Mapping 250K or 6.0 arrays. For the Axiom Set, the target DNA was amplified using the Axiom 2.0 Amp kit and then fragmented according to the manufacturer’s recommendations (Affymetrix). The pellets were then re-suspended and hybridized, and genotyping was performed using the GeneTitan® Multi-Channel Instrument (Affymetrix).

GeneChip® Operating software (version 1.4) and Genotyping Console™ (GTC; version 3.01) software (Affymetrix) were used for primary data analysis, normalization against internal control features on the chip, genotype calling and quality control (overall single nucleotide polymorphism call rate, 95–99%). Data generated from the arrays were then analyzed using Affymetrix Genotyping Console (version 3.01) and/or a homozygosity mapper program (http://www.homozygositymapper.org/) (12). Regions of shared homozygosity between affected individuals within a family, but not shared with unaffected individuals, were identified. Also, regions of shared homozygosity between affected individuals from different families were identified. The cutoff characteristic was that the region had to be shared between all of the affected individuals from at least four different families. Unaffected individuals from three random families were used as controls. Regions <2 Mb were excluded from the analysis, and gender-specific chromosomes were not analyzed.

Identified regions of homozygosity were entered into the Suspects Candidate Gene Search v.28.3 (http://www.cgem.ed.ac.uk/resources/suspects/) from the University of Edinburgh (Edinburgh, UK), in order to prioritize genes in an order of their reported relevance to the development of ASD (12). The top three candidate genes were reported in the present study.

Results

A total of 13 multiplex families were recruited and analyzed using GeneChip® Mapping 250K and 6.0 arrays, and the Axiom Set. There were >27 affected individuals within the recruited families. A selection of the pedigrees is presented in Fig. 1.

Regions of homozygosity identified from the independent analysis of each family were spread across all chromosomes. The chromosome with the most regions of homozygosity was identified as chromosome 5 (Table I). Shared regions of homozygosity between the affected individuals from all of the families were also determined. There were seven regions that were shared between all individuals, with two regions on chromosome 11 (Table II). The number of genes in the regions identified ranged between 3 genes to several hundreds. Notably, no regions of homozygosity (above 2 Mb in size) were observed on chromosomes 13, 14 and 22 as per the analysis criteria used.

Table I Regions of homozygosity identified in the analysis of multiplex families (n=13).

Table I

Regions of homozygosity identified in the analysis of multiplex families (n=13).

Chromosome	Coordinates	Top 3 genes	Family ID
1	17, 606, 235-20, 525, 209	TAS1R2, HTR6, PLA2G2F	F14
	66, 592, 949-70, 541, 226	NP_065999, NP_060238, SLC35D1	F14
	90, 600, 674-10, 477, 047	ATP6V0B, SLC1A7, ATPIF1	F20
	208, 362, 435-211, 304, 878	NP_061134, ATF3, RPS6KC1	F8
2	16, 280, 027-18, 724, 599	KCNS3, SMC6L1, VSNL1	F14
	111, 883, 724-114, 024, 469	SLC20A1, MERTK, NP_116213	F23
	162, 797, 863-165, 675, 258	KCNH7, FAP, NP_775783	F23
	129, 417, 474-133, 043, 938	NP_079305, PLEKHB2, CFC1	F6
	180, 338, 192-192, 340, 990	ITGAV, NAB1, COL3A1	F7
3	49, 088, 836-51, 462, 363	BSN, DOCK3, LAMB2	F20
	155, 790, 787-160, 134, 049	Q8NGV9, SSR3, GMPS	F14
4	3, 691, 191-5, 842, 856	MSX1, CYTL1, STK32B	F14
	37, 930, 813-44, 149, 564	CHRNA9, APBB2, SLC30A9	F14
	75, 534, 780-81, 987, 898	FRAS1, ANTXR2, GK2	F7
	91, 479, 821-96, 553, 748	PDLIM5, GRID2, BMPR1B	F3
	127, 483, 173-131, 017, 058	SLC25A31, PHF17, PDZK6	F6
5	10, 671, 102-12, 018, 348	CTNND2, DAP	F14
	13, 115, 454-15, 292, 186	DNAH5, TRIO, ANKH	F23
	43, 721, 462-46, 644, 204	HCN1, FGF10, MRPS30	F14
	63, 181, 011-65, 623, 491	TRIM23, ADAMTS6, HTR1A	F14-F7
	70, 580, 012-78, 664, 650	F2RL1, F2RL2, F2R	F14
	91, 871, 294-95, 831, 186	KIAA0372, PCSK1, NR2F1	F14
	105, 398, 456-108, 334, 980	EFNA5, FBXL17	F6
	110, 309, 108-113, 085, 059	SRP19, APC, WDR36	F23
	154, 784, 618-157, 912, 731	ADAM19, CRSP9, SGCD	F23
6	43, 852, 052-46, 694, 741	DSCR1L1, ENPP4, SLC35B2	F20
	51, 472, 664-58, 554, 555	PKHD1, MCM3, ICK	F6-F14
	61, 784, 585-65, 361, 195	PTP4A1, KHDRBS2, GLULD1	F14
	89, 467, 426-91, 642, 040	GABRR2, GABRR1, MDN1	F14
	117, 801, 917-120, 016, 046	MAN1A1, DCBLD1, C6orf60	F14
	128, 741, 476-132, 437, 814	AKAP7, CTGF, LAMA2	F14
7	44, 167-3, 935, 550	GPR146, LFNG, CHST12	F23
	33, 956, 386-38, 816, 634	SEPT7, AMPH, GPR154	F14
	39, 251, 002-42, 234, 071	GLI3, INHBA, CDC2L5	F14
	78, 299, 408-93, 347, 267	GRM3, SRI, CACNA2D1	F14
	88, 881, 477-91, 474, 214	PFTK1, CYP51A1, CLDN12	F14
	118, 385, 860-122, 485, 086	SLC13A1, KCND2, WNT16	F1
	132, 226, 627-137, 446, 264	SLC13A4, PTN, NUP205	F7
	143, 159, 897-147, 967, 548	CNTNAP2, TPK1, O60393	F14
	149, 811, 568-153, 488, 778	MLL3, ACCN3, KCNH2	F14
8	95, 013, 089-99, 313, 599	UQCRB, GDF6, NP_057218	F14
	113, 038, 464-118, 887, 549	CSMD3, THRAP6, NP_115710	F14
	8, 966, 544-10, 388, 436	MSRA, TNKS, PPP1R3B	F20
	29, 493, 059-31, 510, 288	GSR, WRN, LEPROTL1	F23
9	77, 676, 620-85, 812, 95	SLC28A3, NTRK2, TLE1	F14
	125, 138, 695-137, 750, 235	GRIN1, KCNT1, SLC25A25	F23
10	24, 259, 097-26, 357, 328	GPR158, THNSL1, PRTFDC1	F6
	34, 935, 425-36, 373, 753	CREM, NP_699199, CUL2	F23
	56, 919, 560-59, 481, 992	ZWINT, ZWINTAS	F7
	79, 202, 450-85, 447, 212	SFTPD, ANXA11, PPIF	F23
11	43, 773, 141-51, 682, 438	RAPSN, PSMC3, SPI1	F3
	110, 025, 057-112, 216, 763	PTS, TIMM8B, DLAT	F23
12	4, 042, 363-10, 283, 711	A2M, CHD4, TNFRSF1A	F7
	37, 345, 650-39, 627, 393	KIF21A, SLC2A13, ABCD2	F23
15	80, 676, 184-82, 887, 137	AP3B2, BTBD1, RPS17	F23
	95, 798, 343-101, 692, 828	ADAMTS17, CHSY1, IGF1R	F17
16	65, 000, 489-70, 848, 261	SLC12A4, CTCF, ATP6V0D1	F1-F14
	6, 392, 511-8, 255, 216	A2BP_HUMAN, Q8WZ91, Q14229	F14
	63, 133, 419-77, 264, 064	SLC12A4, CTCF, ATP6V0D1	F14
17	32, 587, 614-35, 196, 133	GPR158L1, ERBB2, NEUROD2	F3
	21, 890, 916-27, 461, 374	SLC6A4, SLC13A2, UNC119	F14
	76, 007, 579-78, 771, 650	SLC25A10, FOXK2, SLC16A3	F23
18	36, 589, 650-38, 808, 690	PIK3C3	F23
19	4, 463, 233-7, 013, 199	SLC25A23, NP_775908, PTPRS	F8
20	13, 821, 260-16, 084, 725	FLRT3, C20orf133, Q9NQI9	F14
21	23, 309, 707-39, 155, 954	ATP5J, ATP5O, GRIK1	F23

[i] Samples were genotyped using GeneChip^® Mapping 250K and 6.0 arrays and the Axiom Set. Samples from affected and unaffected individuals from all 13 families were genotyped using GeneChip® Mapping 250K, 6.0 arrays and Axiom. (~54 individuals, 27 affected subjects).

Table II Regions of shared homozygosity obtained from analyzing all affected individuals from multiplex families using GeneChip® Mapping 250k and 6.0 arrays.

Table II

Regions of shared homozygosity obtained from analyzing all affected individuals from multiplex families using GeneChip® Mapping 250k and 6.0 arrays.

Chromosome	Coordinates and Axiom	Top 3 genes
1	102, 688, 721-105, 073, 165	COL11A1, AMY1A, AMY1A
2	134, 545, 725-136, 584, 137	ACMSD, LCT, MGAT5
3	49, 856, 601-51, 968, 474	GRM2, DOCK3, SLC38A3
7	77, 339, 194-80, 531, 718	SEMA3C, CD36, GNAI1
11	47, 240, 293-49, 279, 489	PSMC3, RAPSN, SPI1
	48, 549, 401-50, 779, 621	FOLH1, TYRL, OR4C13
15	25, 581, 119-27, 659, 946	APBA2, HERC2, OCA2

Discussion

The regions of homozygosity identified in the present study contained numerous genes that have previously been reported as possible causes of ASD (13). The results included the following genes: GRIN1, GRIK1, GRID2, GPR158L1, GLULD1 and GRM3, which are associated with glutamate receptor and proteins, RAPSN, which is associated with acetylcholine receptors and transmission, and GABRR1 and GABRR2, which are associated with GABA receptors (14). The results of the present study support the numerous reports that suggest ASD may be a synaptopathic disease (15). Furthermore, numerous genes in these regions are associated with mitochondria and are energy-associated, including ATP5J, SLC25A25, SLC25A31 and MRPS30, which have also been implicated in previous studies (16,17). A third group of genes identified in these regions were consistent with embryonic/development-associated genes, including, WNT 16 and NEUROD2, suggesting a possible early insult to the central nervous system in the developing embryo (18).

A number of the regions reported to be associated with the development of ASD include: i) Duplication or deletion of 16p11.2, which has been implicated in patients with ASD as well as in animal models, which may cause alterations in brain anatomy and behavior (19,20); ii) a 1.5-Mb microdeletion in region 14q23.2–23.3, which is associated with ASD, as well as spherocytosis (21); iii) a deletion at 2p15 (22); iv) an alteration at 8p23.2 (23); and v) a 535-Kb deletion at 3p26.3 (24). However, the results of the present study did not indicate any alterations to these regions, suggesting that ASD-associated gene alterations differ across various populations.

As shown in Table II, the number of regions identified when analyzing all of the affected individuals from the different families was lower, as compared with the regions obtained from analyzing the affected individuals within their respective families. In addition, there were no regions that were universally present in all of the affected individuals from the different families. However, 7 shared regions were observed between certain affected individuals. This difference may be attributed to the fact that ASD is a genetically complex and multifactorial disorder, which concurs with numerous previous studies and reports regarding the genetics of ASD (3,4,9). There may be more than one genetic change per family, and different genetic changes between families. Biallelic mutations have been reported in other disorders with complex phenotypes (25–27).

One of the challenges in studying the genetics of ASD in Saudi Arabia is that the Saudi population, as evident in the families that were recruited in the present study, is highly consanguineous and inbred. This represents a challenge in identifying genetic regions that are associated with ASD, as opposed to regions that may be normally shared as a result of identity by descent. In order to overcome this challenge, it has been proposed for future studies that appropriate tests and statistics are conducted prior to genetic analysis that allow for discrimination of these two types of regions, such as the coefficient of inbreeding (28).

Another challenge facing the genetic study of ASD is that ASD represents a variable spectrum, in which the milder forms may not be detected with conventional clinical methods. This is relevant to the results of the present study in that when the results of affected individuals are compared with normal controls, there may be missing/excluded regions. This may possibly be due to the controls having a mild form of ASD that the clinic was unable to detect. Further study of ASD genetics is therefore dependent on the development of more specific clinical tools. Another way to overcome this issue in future studies is to include information on clinically normal individuals, including their intelligence quotient, social interaction status and their school performance. As has been reported previously, the possibility of the presence of multiple phenotypes within the same family suggests involvement of multiple interacting loci, which adds to the complexity of the inheritance of ASD (28).

As mentioned previously, all of the regions identified in the present study were >2 Mb. It is possible that more regions and possible genes may be identified for analysis by examining smaller regions. This is the aim for future studies by our group, as well as the inclusion of gender-associated chromosomes in the analyses.

Candidate autism genes usually have distinct biological roles and interactions. In order to attain an improved understanding regarding the pathogenesis of ASD, it is important to discover the location and functional role of autism-associated proteins in biological pathways, and hence in neuronal function and resulting behavior (29). An ASD interactome was generated to meet the obstacles caused by this hereditary neurological disorder, and to identify proteins that interact with ASD-associated proteins (30). Of note, when the ASD interactome was created, a high connectivity was identified between two ASD-associated proteins that initially appeared functionally unrelated, SHANK3 and TSC1. SHANK3 is an adapter protein in the post-synaptic regions that may have a role in the organization of the dendritic spine and synaptic junction, and TSC1 is a tuberous sclerosis 1 protein that regulates mechanistic (serine/threonine kinase) target of rapamycin, a promoter of protein synthesis. However, these two proteins share ≥21 protein associates and were found to interact in a complex scaffold at the post-synaptic region (30), thus suggesting that various ASD-associated proteins may share a common key pathway associated with disease development. Similarly, the genes identified in the present study involve various categories of proteins that may also share a common molecular pathway, and may be further investigated through a protein-protein interaction assay. Furthermore, since ASD genes differ between populations, creating an interactome for these Saudi ASD proteins would be beneficial not only in the development of treatment, but also in developing powerful therapies for a specific ethnic group.

Acknowledgments

The present study was supported by the King Faisal Specialist Hospital and Research Center (KFSHRC) (RAC# 2080001) and the Center for Autism Research. The authors of the present study would also like to acknowledge the genotyping core facility at KFSHRC.

References