Novel compound heterozygous mutations in ABCA4 in a Chinese pedigree with Stargardt disease.

PURPOSE: Stargardt disease (STGD) is a common macular dystrophy in juveniles that is commonly inherited as an autosomal recessive trait. Mutations in five genes (ABCA4, PROM1, ELOVL4, BEST1, and PRPH2) have been reported to be associated with STGD. In the present study, we aimed to identify the pathogenic mutations in affected members in a Chinese STGD pedigree.
METHODS: One patient was selected for whole-exome sequencing. Variants in five candidate genes were identified initially, followed by several filtering steps against public and private variation databases (1000Genomes, ESP6500si, ExAC, and in-house database), as well as bioinformatic analysis of the putative pathogenic roles. Sanger sequencing was used for cosegregation analysis among all members with available DNA.
RESULTS: Two mutations in ABCA4 (NM_000350.2; c.5646G>A; p.Met1882Ile and NM_000350.2; c.3523-2A>G) were found using whole-exome sequencing. Cosegregation analysis confirmed all the affected members carried the compound heterozygous mutations while the other healthy members had at most one. The missense mutation was extremely rare in public databases and predicted to be deleterious. The splice-site mutation was absent from all public and private databases and was predicted to alter the splice pattern, resulting in an exon skip and a frameshift.
CONCLUSIONS: Using whole-exome sequencing, we found novel compound heterozygous mutations in ABCA4 in a Chinese STGD pedigree. These mutations are reported for the first time, therefore widening the mutation spectrum of Stargardt disease. The present study also illustrates the potential of whole-exome sequencing in determining the genetic cause of STGD.

Introduction

Stargardt disease (STGD) is a common juvenile hereditary macular dystrophy that was first proposed by Karl Stargardt in 1909 [1]. It is characterized by progressive loss of central vision, a beaten-bronze appearance, atrophy of the RPE, and flecks around the macula and/or retinal periphery with a prevalence of about 1 in 10,000 [2,3]. Most cases are inherited in an autosomal recessive pattern while in some families the disorder is transmitted in an autosomal dominant form [2,4,5]. This disorder is clinically and genetically heterogeneous. The clinical features, including disease manifestation, progression, prognosis and imaging results, may vary with the disease-causing mutations in various genes. Until now, mutations in five genes (ABCA4-Gene ID: 24, OMIM 601691, PROM1-Gene ID: 8842, OMIM 604365, ELOVL4-Gene ID: 6785, OMIM 605512, BEST1-Gene ID: 7439, OMIM 607854, and PRPH2-Gene ID: 5961, OMIM 179605) have been reported to be associated with STGD [6-10].

STGD1 is the most frequent form of all STGD cases, which is caused by compound heterozygous or homozygous mutations in the ATP binding cassette subfamily A member 4 (ABCA4) gene [11]. The onset of STGD occurs in childhood most commonly, followed by early adulthood and later adulthood. Generally speaking, the earlier the onset of STGD, the worse the prognosis, resulting in more serious vision loss in the children compared with adults [12]. To date, more than 900 sequence variants in ABCA4 have been identified [13]. Increasing evidence suggests that onset is related to the deleterious effect of the identified variants. Early onset STGD1 is more frequently associated with more deleterious variants, including nonsense, splice-site, and frameshift mutations, compared with later onset [12].

Here, we report a Chinese STGD pedigree with three affected members in an autosomal recessive inherited pattern. Whole-exome sequencing (WES) revealed compound heterozygous mutations in ABCA4 completely cosegregated with this disorder. These two mutations were identified for the first time in STGD.

Methods

Ethics statement

This research project was approved by the ethical committee of the National Research Institute of Family Planning. Written informed consent was obtained from all participants. The methods and experiments were performed in accordance with approved guidelines. The study adhered to the tenets of the Declaration of Helsinki and the ARVO statement on human subjects.

Clinical examinations

Eight members of a STGD pedigree were enrolled for clinical and genetic analysis (Figure 1A). The medical history of each participant was recorded, and ophthalmic examinations, including measurement of visual acuity, fundus examination, optical coherence tomography (OCT), full field electroretinography (ERG), and electrooculogram (EOG), were performed at the Maternal and Children Healthcare Hospital, Liuzhou, Guangxi, China for the diagnosis of STGD disease.

Figure 1

Compound heterozygous mutations in ABCA4 identified in the STGD family. A: The Chinese Stargardt disease pedigree and the complete cosegregation pattern of the compound heterozygous mutations in ABCA4. B: The fundus images of both of the proband’s eyes showed yellow flecks on the retina and macular atrophy. C: Sanger sequencing results for the two mutations. D: Multiple alignment of ABCA4 indicated p.M1882 is highly conserved. m1: c.3523–2A>G; m2: c.5646G>A, p.M1882I; square: male; circle: female; filled symbol: patient; unfilled symbol: unaffected member; black arrow: the proband; R: right eye; L: left eye.

WES and validation

Peripheral venous blood was collected from all participants, and genomic DNA was extracted using the QIAamp DNA Blood Mini Kit (QIAGEN, Hilden, Germany). All exons were captured with the SureSelect Human All Exon V5 (Agilent, Santa Clara, CA) and then sequenced on the Illumina Hiseq4000 sequencer (Illumina, Saint Diego, CA). The reads were aligned against the human reference genome (hg19) by using BWA −0.7.10 (the Burrows-Wheeler Alignment Tool). The GATK 3.v4 (Genome Analysis Toolkit) was used to call variants that include single nucleotide polymorphisms (SNPs) and indels. Then the variants were annotated with SnpEff _v4.1.

The impact of missense mutations on the proteins was predicted with online programs, including MutationTaster, SIFT, PolyPhen2, and SNP&GO. Conservation analysis was predicted with CLC Workbench software. Whether the splice site mutation altered the splice pattern was assessed with MaxEntScan (MaxEnt), Human Splicing Finder 3 (HSF3), and NetGene2. Cosegregation analysis of the two mutations in ABCA4 was performed with Sanger sequencing among pedigree members with available DNA. The primers used for Sanger sequencing were as follows: forward, 5′-GGC TCT TGC TCA GTT CCC-3′; reverse, 5′-TTT AAC CCG ATC CTC TAC TTG-3′ (ABCA4 c.5646G>A); and forward, 5′-GCT GCG AAG TGA ACA AAC-3′; reverse, 5′-CCA GAC GGA ACC CAA GTA-3′ (ABCA4 c.3523–2A>G).

Results

Clinical manifestations and examination results

The proband (K206–2) was 25-years-old and admitted to the hospital for evaluation of his eyesight. The uncorrected visual acuity of each eye was 20/400, which has progressively decreased for 13 years. Ophthalmoscopy showed atrophic and pigmentary changes in the macular area and some evident yellow flecks (Figure 1B). The retinal vessels were normal. OCT displayed thinning of the nerve fiber layer the outer nuclear layer (ONL), and the inner segment–outer segment junction (IS-OS), as well as various degrees of atrophy of the RPE (Appendix 1). Therefore, he was diagnosed with Stargardt disease 1. His two siblings, K206–1 and K206–3, had had progressive vision loss for 14 and 17 years, respectively. Fundus examinations showed the typical yellow flecks and macular atrophy while OCT indicated results similar to the proband’s (Appendix 1 and Appendix 2). The other family members did not have obvious ocular abnormalities after undergoing medical examinations. The clinical characteristics and examination results are shown in Table 1, Appendix 3, and Appendix 4.

Table 1

Clinical features of a Chinese pedigree with STGD.

Patients	Age	Gender	Vision acuity	Fundus manifestations	OCT	ERG	EOG
K206–2	26	F	20/400 20/400	macular atrophy, yellow flecks on RPE	thin nerve epithelium; ONL,IS/OS become thin; RPE atrophy	a, b-wave amplitude declined moderately	light peak/dark trough ratio (Lp/Dt) declined heavily
K206–1	28	M	20/400 20/400	macular atrophy, yellow flecks on RPE	thin nerve epithelium; ONL,IS/OS become thin; RPE atrophy	a, b-wave amplitude declined moderately	N.D.
K206–3	24	M	20/500 20/500	macular atrophy, yellow flecks on RPE	thin nerve epithelium; ONL,IS/OS become thin; RPE atrophy	a, b-wave amplitude declined moderately	light peak/dark trough ratio (Lp/Dt) declined moderately

F: female; M: male; RPE: retinal pigment epithelium; OCT: optical coherence tomography; ONL: outer nuclear layer; IS/OS: inner segment/outer segment; F-ERG: flash electroretinogram; EOG: electro-oculogram; N.D.: not detected

Whole-exome sequencing in one patient in the STGD pedigree

The proband was selected for WES. Using WES, we identified 78,715 SNPs and 9,428 indels. Initially, we focused on the non-synonymous mutations, including the missense, nonsense, splice-site, and frameshift mutations of five known candidate genes (ABCA4, PROM1, ELOVL4, BEST1, and PRPH2) and found seven missense mutations and one splice-site mutation in ABCA4, ELOVL4, and PRPH2 (Table 2). Subsequent searching for the minor allele frequency of each variant in public databases (1000Genomes, ESP6500si, and ExAC) excluded five variants as they are common in various populations. In addition, another variant (ABCA4 p.G1183C) was predicted to be benign by all four online programs and therefore was discarded. Finally, one rare mutation (c.5646G>A; p.M1882I in exon 40) and a novel splice-site mutation (c.3523–2A>G at the 3′ splice site in intron 24 in ABCA4) remained. Both variations were highly conserved according to online predictions.

Table 2

Mutations in known candidate genes of Stargardt disease identified by whole-exome sequencing.

Gene	Chromosome Position	Mutation	1000G (%)	ESP6500si (%)	ExAC# (%)	Online Prediction
						PolyPhen2	SIFT	MutationTaster	SNPs&GO
ABCA4	chr1:94,476,424	NM_000350.2:c.G5646A:p.M1882I	0	0	0.035	+	+	+	+
	chr1:94,505,685	NM_000350.2:c.3523–2A>G:	0	0	0	NA	NA	NA	NA
	chr1:94,505,659	NM_000350.2:c.G3547T:p.G1183C	0.34	0	1.34	-	-	-	-
	chr1:94,544,234	NM_000350.2:c.A1268G:p.H423R	30.26	30.94	24.77	-	-	-	-
ELOVL4	chr6:80,626,375	NM_022726.3:c.A895G:p.M299V	24.16	10.92	21.16	-	-	-	-
PRPH2	chr6:42,666,061	NM_000322.4:c.A1013G:p.D338G	75.74	19.88	86.02	Unknown	-	-	-
	chr6:42,666,145	NM_000322.4:c.G929A:p.R310L	94.13	9.46	99.82	-	-	-	-
	chr6:42,666,164	NM_000322.4:c.C910G:p.Q304E	75.66	19.90	85.85	+	-	-	-

refer to the frequency in ExAC East Asian population comprising over 8600 chromosomes. “+”: predicted to be deleterious; “-“: predicted to be benign; NA: not applicable

Cosegregation analysis

Cosegregation analysis was performed among the family members who had available DNA. All affected members carried these two mutations. The proband’s unaffected father and healthy sister both harbored the splice-site mutation only, while the unaffected daughter of K206–1 had only the missense mutation, demonstrating that these two mutations were compound heterozygous and completely segregated with the disorder (Figure 1A,C).

Online predictions

The missense mutation (c.5646G>A), resulting in an amino acid substitution from Met to Ile at the 1882th residue, was not found in 1000 Genomes, ESP6500si, or an in-house database (about 400 exomes), and was rare in ExAC (carried by three individuals in an East Asian population comprising more than 4,300 sequenced samples). The mutation was predicted to damage the protein function by online programs including MutationTaster, SIFT, PolyPhen2, and SNP&GO, and was highly conserved among various species (Figure 1D, Table 2).

The splice-site mutation (c.3523–2A>G) was absent from all public databases and our internal database. We applied MaxEnt and HSF to predict whether the mutation could alter the splice pattern. When it was mutated, the score given by MaxEnt decreased remarkably from 8.69 to 0.73 with a reduction of 91.6% exceeding the threshold of 30% [14]. Meanwhile, the HSF score decreased from 88.54 to 59.59 with an alteration over the threshold of 10% [15]. Both programs predicted this splice site might be abandoned in the mutant. Additionally, the NetGene2 program predicted the acceptor site mutation that led to the skipping of exon 25, a frameshift change, and a premature termination.

Discussion

The ABCA4 gene, located on human chromosome 1p, is a member of the subfamily A of ATP-binding cassette (ABC) transporter family [16]. The gene is expressed in the outer segments of the photoreceptor cells of the vertebrate retina and encodes a retina-specific ABC transporter protein called ABCR [4,17]. Previous studies have shown that ABCR is involved in the transport of N-retinylidene-PE (a precursor of potentially toxic substances) and then prevents the accumulation of precursor inside the disks. The mutation in ABCA4 can ultimately lead to the accumulation of lipofuscin content in the RPE cells and influence retinoid cycle kinetics [4,18].

ABCA4 is a large gene comprising 50 exons and spanning 130 kbp with more than 900 mutations identified in patients with STGD or cone-rod dystrophy. The numerous sequence variants make the establishment of a genotype–phenotype relation problematic and genetic testing of ABCA4-related disorders challenging. Direct sequencing of all exons or screening with microarrays for all known disease-causing mutations in ABCA4 can detect the majority of pathogenic variants [19,20]. However, they are either low cost-effective or unable to detect novel variants. With the rapid development of sequencing technology, massively parallel sequencing or next-generation sequencing (NGS) has been frequently used for genome-wide or region-wide identification of disease-related variants, which includes whole-genome, whole-exome, and target-region sequencing. Exome sequencing projects in Stargardt disease cohorts have shown the enormous potential and good performance of NGS in STGD genetic research and diagnosis [21-24].

Here, we employed WES to identify the genetic pathogenic cause of a Chinese STGD family. We identified compound heterozygous mutations in ABCA4 that was cosegregated with the phenotype among all members. In general, STGD disease is caused by mutations in the ABCA4 gene in an autosomal recessive pattern. That is to say, only when the ABCA4 gene is abnormal on both chromosomes and the structure or function of the ABCR protein is almost completely disrupted is the mutation pathogenic because the protein cannot participate in the transport of N-retinylidene-PE. In the present study, the family members who carry only one mutation exhibit no symptom of vision loss, because nearly half of the ABCR protein can work normally, which is sufficient for the normal transportation of N-retinylidene-PE. Members who carry only one mutation are healthy while the members who carry the two mutations are affected, demonstrating that these two mutations are pathogenic in a recessive pattern.

The impact of the missense mutation (c.5646G>A; p.M1882I) on protein function was predicted to be deleterious by four online prediction programs. This missense mutation has been previously reported in a patient with STGD with screening three exons (17, 40, and 47) of the ABCA4 gene [25]. A splice-site mutation is another common type of disease-causing mutation that disrupts the protein function through altering mRNA splice patterns. Several splice-site mutations (c.5461–10T>C and c.5312+3A>T, for instance) have been reported to cause exon skipping, intron retaining, and frame shifting that result in premature protein products [26,27]. In the present study, a novel mutation (c.3523–2A>G) in a canonical splice acceptor site was identified for the first time. The acceptor site was abandoned, and the 25th exon was skipped, as predicted by HSF3, MaxEnt, and NetGene2. These caused the frameshift of the open reading frame (ORF) and led to premature termination of protein translation.

In summary, we identified novel compound heterozygous mutations in ABCA4, by using WES, in a Chinese STGD pedigree. These results make the genetic studies of STGD more comprehensive and may help in genetic counseling for patients with STGD. In addition, the results indicate that WES is a useful tool for genomic analysis that can help us better understand the genetic mechanism of STGD.