Comprehensive Mutation Analysis by Whole-Exome Sequencing in 41 Chinese Families With Leber Congenital Amaurosis

Abstract

Purpose.: Leber congenital amaurosis (LCA) is a genetically heterogeneous disease with, to date, 19 identified causative genes. Our aim was to evaluate the mutations in all 19 genes in Chinese families with LCA.

Methods.: LCA patients from 41 unrelated Chinese families were enrolled, including 25 previously unanalyzed families and 16 families screened previously by Sanger sequencing, but with no identified mutations. Genetic variations were screened by whole-exome sequencing and then validated using Sanger sequencing.

Results.: A total of 41 variants predicted to affect protein coding or splicing was detected by whole-exome sequencing, and 40 were confirmed by Sanger sequencing. Bioinformatic and segregation analyses revealed 22 potentially pathogenic variants (17 novel) in 15 probands, comprised of 3 of 16 previously analyzed families and 12 of 25 (48%) previously unanalyzed families. In the latter 12 families, mutations were found in CEP290 (three probands); GUCY2D (two probands); and CRB1, CRX, RPE65, IQCB1, LCA5, TULP1, and IMPDH1 (one proband each). Based on the results from 87 previously analyzed probands and 25 new cases, GUCY2D, CRB1, RPGRIP1, CEP290, and CRX were the five most frequently mutated genes, which was similar to the results from studies in Caucasian subjects.

Conclusions.: Whole-exome sequencing detected mutations in the 19 known LCA genes in approximately half of Chinese families with LCA. These results, together with our previous results, demonstrate the spectrum and frequency of mutations of the 19 genes responsible for LCA in Han Chinese individuals. Whole-exome sequencing is an efficient method for detecting mutations in highly heterogeneous hereditary diseases.

Chinese Abstract

Introduction

Leber congenital amaurosis (LCA, MIM 204000) is the most severe form of inherited retinal dystrophy, with an estimated prevalence of 1:81,000 ¹ to 1:30,000. ² This condition accounts for more than 5% of all retinal dystrophies and approximately 20% of children attending schools for the blind. ³ LCA is inherited most frequently as an autosomal-recessive trait, but also may be transmitted as an autosomal-dominant trait in rare cases. ⁴ Clinical features of LCA include profound loss of visual function at birth or within the first year of life, nystagmus, oculodigital sign of Franceschetti, sluggish pupillary light reflex, and variable fundus changes ranging from relatively normal appearance to severe pigmentary degeneration. Markedly reduced or no identifiable rod and cone responses on electroretinogram recording in infancy are the hallmark of LCA. ⁵  

To date, mutations in at least 19 genes have been identified as responsible for LCA: guanylate cyclase 2D (GUCY2D), ⁶ retinal pigment epithelium-specific protein 65 kDa (RPE65), ⁷ spermatogenesis-associated protein 7 (SPATA7), ⁸ aryl hydrocarbon-interacting receptor protein-like 1 (AIPL1), ⁹ Leber congenital amaurosis 5 gene (LCA5), ¹⁰ retinitis pigmentosa (RP) GTPase regulator-interacting protein (RPGRIP1), ¹¹ cone-rod homeobox-containing gene (CRX), ¹² crumbs homolog 1 (Drosophila) (CRB1), ¹³ centrosomal protein 290 kDa (CEP290), ¹⁴ inosine 5′-monophosphate dehydrogenase type 1 (IMPDH1), ¹⁵ retinal degeneration 3 (RD3), ¹⁶ retinol dehydrogenase 12 (RDH12), ¹⁷ lecithin retinol acyltransferase (LRAT), ¹⁸ tubby-like protein 1 (TULP1), ¹⁹ inwardly rectifying potassium channel Kir7.1 (KCNJ13), ²⁰ calcium-binding protein 4 (CABP4), ²¹ IQ motif-containing protein B1 (IQCB1), ²² orthodenticle homolog 2 (OTX2), ²³ and nicotinamide nucleotide adenylyltransferase 1 (NMNAT1). ²⁴ Mutations in most of the genes above are associated with autosomal-recessive LCA, whereas mutations in CRX, ⁴ IMPDH1, ¹⁵ and OTX2 ²³ usually are associated with autosomal-dominant LCA. Studies of these genes, based on an individual gene or a subset of genes, have identified numerous mutations. ^25–31 However, to our knowledge no systematic analysis of these 19 genes to evaluate the full spectrum of variations in patients with LCA has been reported. Exon-by-exon analysis of the 19 genes by conventional Sanger sequencing is not suitable for mutation detection in a clinical setting, as it is labor- and time-intensive. ³² Therefore, fast and reliable new techniques are needed to detect mutations in this disease. An LCA mutation chip (Asper Ophthalmics; Asper Biotech Ltd., Tartu, Estonia) was developed that detects all known mutations identified in LCA, but cannot detect novel mutations. ³³ A SNP-chip for LCA diagnosis may be used in patients with a family history, but most LCA cases are isolated patients born to unaffected parents. ³⁴  

Recently, with the rapid development of next-generation sequencing, whole-exome sequencing has arisen as an impressive tool for mutation screening, especially for highly heterogeneous hereditary diseases. ³⁵ In our study, whole-exome sequencing was used to detect variants in 41 unrelated Chinese patients with LCA. Variants found in the 19 LCA-related genes were verified by Sanger sequencing. 

Materials and Methods

Subjects

All probands with LCA were collected from the Pediatric and Genetic Clinic in Zhongshan Ophthalmic Center. In total, 41 probands from unrelated Chinese families were involved in this study, including 25 patients not analyzed before and 16 patients in whom some portion of the 15 LCA-related genes has been screened by Sanger sequencing previously, but without identified mutations. ³⁶ Of the 41 patients, 29 were male and 12 were female; 34 were isolated cases, while six showed autosomal-recessive inheritance, and one showed autosomal-dominant inheritance. Written informed consent conforming to the tenets of the Declaration of Helsinki was obtained from the participants or their guardians before the study. This study was approved by the Institutional Review Board of the Zhongshan Ophthalmic Center. Genomic DNA was extracted from leukocytes of a peripheral blood sample of each participant as described previously. ³⁷  

Whole-Exome Sequencing

Whole-exome sequencing was performed through a commercial service from BGI Shenzhen (Shenzhen, China; available in the public domain at http://www.genomics.cn/index). The methods for exome capture, exon-enriched DNA library construction, sequencing, genotyping, and variant analysis have been reported previously. ³⁸ In brief, a NimbleGen SeqCap EZ Exome (44M; Roche, Basil, Switzerland) array was used to perform the exome capture. Subsequently, exon-enriched DNA fragments were loaded on the Illumina Genome Analyzer II (Illumina, Santiago, CA) platform for sequencing. The mean exome coverage was set as 60-fold. For variant analysis, alignment of the sequencing reads with UCSC hg19 was performed on a SOAPaligner. ^39,40 SOAPsnp was used to calculate the likelihood of possible genotypes in target regions. ⁴¹ Variants in all 19 LCA-related genes detected by whole-exome sequencing were selected for further verification. 

Sanger Sequencing

Sanger sequencing was used to validate the variants found in whole-exome sequencing in the 19 genes. Segregation analyses were performed in patients with available relatives. Genomic information about the 19 genes is listed in Supplementary Table S1. Primers used to amplify fragments harboring individual variants were designed by Primer3 (available in the public domain at http://frodo.wi.mit.edu/primer3/), and the sequences of these primers are listed in Supplementary Table S2. Additionally, a known mutation hot spot outside the capture range of the exome array, c.2991 + 1665A > G in CEP290, was analyzed by direct Sanger sequencing in all probands. Polymerase chain reaction was used to amplify the genomic fragments with variants, and the sequences of the amplicons were determined by Sanger sequencing using a BigDye Terminator cycle sequencing kit v3.1 and an ABI 3130 Genetic Analyzer (both from Applied Biosystems, Foster City, CA). The resultant sequences were compared to consensus sequences using Seqman software (Lasergene 8.0; DNASTAR, Inc., Madison, WI). The possible impact of amino acid substitutions was predicted by SIFT (available in the public domain at http://sift.jcvi.org/) and PolyPhen-2 (available in the public domain at http://genetics.bwh.harvard.edu/pph2/).⁴² Splice site prediction by a neural network was used to predict the effects of variants on splicing sites (available in the public domain at http://www.fruitfly.org/seq_tools/splice.html).⁴³ Base-by-base conservation scores ranging from 0 to 1, with higher scores indicating the higher degrees of conservation, were obtained using PhastCons (available in the public domain at http://varianttools.sourceforge.net/Annotation/PhastCons).⁴⁴ Each novel putative disease-causing variant was evaluated further in 192 normal individuals. 

Results

Whole-exome sequencing detected 41 variants that affected either encoded residues or splicing in 15 of the 19 genes. No such variants were detected in RD3, LRAT, KCNJ13, or OTX2 by exome sequencing in this LCA cohort. The 41 variants were present in 28 of the 41 probands. Of the 41 variants, 40 were confirmed by Sanger sequencing, and the remaining variant (c.293G > A in CABP4) was a false-positive with low read depth (G4A2). 

The pathogenicity of the 40 variants was evaluated by bioinformatics and segregation analysis. Homozygous and compound heterozygous variants were initially selected. Single heterozygous variants in CRX or IMPDH1 also were selected for further analysis. Unreported variants were evaluated in 192 normal individuals. By these criteria, 22 of the 40 variants were considered putatively pathogenic for 15 of the 41 patients (Table 1, Fig. 1, Supplementary Fig. S1). Of the 22 variants, 17 were novel and five were known. All of the novel mutations were predicted to be pathogenic in silico and were absent from 384 control chromosomes. The 22 variants found in 15 probands were present in 10 of the 19 genes: CEP290, GUCY2D, CRB1, CRX, RPGRIP1, IQCB1, RPE65, IMPDH1, LCA5, and TULP1. The less likely causal variants are summarized in Supplementary Table S3. 

Figure 1. 

 

Pedigree and segregation analysis in 15 LCA families with mutations identified in this study. +, wild-type allele.

Table 1

 

The 22 Potentially Pathogenic Variants Identified in 15 of 41 Chinese Families With LCA

Gene	Inheritance	Family ID	Variations		State	Bioinformatic Analysis			Reported or Not
			Nucleotide	Amino Acid		SIFT	P/SS	Phastcons
CEP290	AR	Family 1	c.3361G > T	p.E1121*	Het	–	–	1.000	Novel
			c.2817 + 2T > C	SD	Het	–	SSA	0.998	Novel
		Family 2	c.3265C > T	p.Q1089*	Het	–	–	0.999	Novel
			c.4090G > T	p.E1364*	Het	–	–	1.000	Novel
		Family 3	c.2954delT	p.M985fs	Het	–	–	1.000	Novel
			c.7028_7034 + 3 dup	SD	Het	–	SSA	0.298	Novel
GUCY2D	AR	Family 4	c.2015G > A	p.C672Y	Het	D	PrD	1.000	Novel
			c.2476C > T	p.Q826*	Het	–	–	0.997	Novel
		Family 5	c.1956 + 1G > T	SD	Het	–	SSA	1.000	Novel
			c.3034A > C	p.T1012P	Het	D	PrD	1.000	Novel
CRB1	AR	Family 6	c.1841G > T	p.G614V	Hom	D	PrD	1.000	Novel
CRX	AD	Family 7	c.573T > A	p.Y191*	Het	–	–	0.975	Novel
RPGRIP1	AR	Family 8	c.535delG	p.E179fs	Hom	–	–	0.923	Reported36
		Family 9	c.2236G > A	p.G746R	Hom	D	PrD	0.978	Novel
IQCB1	AR	Family 10	c.1090C > T	p.R364*	Hom	–	–	0.993	Reported29
		Family 11	c.994C > T	p.R332*	Hom	–	–	1.000	Novel
RPE65	AR	Family 12	c.200T > G	p.L67R	Het	D	PrD	1.000	Reported28
			c.430T > C	p.Y144H	Het	T	PrD	1.000	Novel
IMPDH1	AD	Family 13	c.592G > T	p.G198C	Het	D	PrD	1.000	Novel
LCA5	AR	Family 14	c.795T > G	p.Y265*	Hom	–	–	0.861	Novel
TULP1	AR	Family 15	c.1198C > T	p.R400W	Het	D	PrD	0.998	Reported19
			c.1444C > T	p.R482W	Het	D	PrD	0.994	Reported31

 Of the 15 families, Family 1, Family 2, Family 3, and Family 12 had triallelic variants. The remaining variants are listed in Supplementary Table S3. Family 8, Family 9, and Family 11 were screened previously for some of the 15 LCA-related genes by Sanger sequencing, but had no identified mutations. The two mutations of RPGRIP1 in Family 8 and Family 9 most likely were omitted from the previous results, and IQCB1 had not been reported to be associated with LCA at that time, so it was not screened previously. AR, autosomal-recessive; AD, autosomal-dominant; P/SS, Polyphen-2/Splice Site Prediction; SD, splicing defect; Het, heterozygous; Hom, homozygous; D, damaging; T, tolerated; PrD, probably damaging; SSA, splicing site abolished.

The 15 probands with identified mutations included 12 of the 25 previously unanalyzed probands and three of 16 probands who were analyzed previously, but without identified mutations. Among the 25 probands who were not analyzed previously, analysis of the 19 known genes detected 19 mutations in 12 probands, which accounted for 48% (12/25) of the cases. The 12 probands had mutations in individual genes as follows: CEP290 (3 probands), GUCY2D (2 probands), CRB1 (1 proband), CRX (1 proband), RPE65 (1 proband), IQCB1 (1 proband), LCA5 (1 proband), TULP1 (1 proband), and IMPDH1 (1 proband). For the three of 16 probands analyzed previously by Sanger sequencing in whom we found mutations, two homozygous mutations in RPGRIP1 were identified in two probands, respectively, and one homozygous mutation in IQCB1 was identified in one proband by whole-exome sequencing. The intronic mutation c.2991 + 1665A > G in CEP290 was not detected in any proband. 

Clinical features of affected individuals with mutations identified in this study are shown in Table 2. These patients had a variable retinal appearance, ranging from a relatively normal fundus to severe pigmentary degeneration (Fig. 2). 

Figure 2. 

 

Fundus photographs of patients from 10 families with mutations identified in our study. The corresponding patient identification numbers and gene mutations are listed above each photo. Variable fundus changes ranging from a relatively normal appearance to obvious pigmentary degeneration were demonstrated. Further clinical information of these patients is listed in Table 2.

Table 2

 

Clinical Features of Affected Individuals With Mutations Identified in This Study

Patient ID	Gene	Mutations	Sex	Age, y, at		Inheritance	Symptoms	Visual Acuity, Right; Left	Fundus Change	ERG
				Exam	Onset					Rod	Cone
Family 1-II:1	CEP290	c.[3361G > T];[2817 + 2T > C]	M	1.2	FMB	Isolated	PV, RN	NPL*	RNA	NDR	NDR
Family 2-II:1	CEP290	c.[3265C > T];[4090G > T]	M	6.8	FMB	Isolated	PV, NYS	FC; FC	CRD, PWF	NDR	NDR
Family 3-II:1	CEP290	c.[2954delT];[7028_7034 + 3dup]	F	1.9	FMB	Isolated	PV, NYS	NPO	AV, CRD	NDR	NDR
Family 4-II:1	GUCY2D	c.[2015G > A];[2476C > T]	M	0.4	0.2	Isolated	PV, NYS, ODS	NPO	RNA	NDR	NDR
Family 5-II:1	GUCY2D	c.[1956 + 1G > T];[3034A > C]	F	0.5	FMB	Isolated	PV, PA, ODS	NPO	AV, MD	NDR	NDR
Family 6-II:2	CRB1	c.[1841G > T];[1841G > T]	M	23	FMB	AR	PV, NYS	FC; FC	AV, PNP	NDR	NDR
Family 6-II:3	CRB1	c.[1841G > T];[1841G > T]	F	20	FMB	AR	PV	0.1; 0.1	AV, PNP	NA	NA
Family 7-II:1	CRX	c.[573T > A];[ = ]	M	0.5	0.5	Isolated	PV, RN	NPL	AV, PD	NDR	NDR
Family 8-II:1	RPGRIP1	c.[535delG];[535delG]	F	0.5	0.3	Isolated	PV, ODS	NPO	AV, CRD	NDR	NDR
Family 9-II:2	RPGRIP1	c.[2236G > A];[2236G > A]	M	7.5	FMB	AR/Cons	PV, RN	HM; HM	AV, CRD	NA	NA
Family 10-II:1	IQCB1	c.[1090C > T];[1090C > T]	M	28	FMB	Isolated	PV, NYS	HM; HM	AV, CRD	NDR	NDR
Family 11-II:1	IQCB1	c.[994C > T];[994C > T]	F	0.3	FMB	Isolated	PV, RN	NPL	AV, CRD	NDR	NDR
Family 12-II:1	RPE65	c.[200T > G];[430T > C]	M	2.0	FMB	Isolated	PV, NYS	PL	AV, CRD	NDR	NDR
Family 13-III:2	IMPDH1	c.[592G > T];[ = ]	M	23	FMB	AD	PV, NYS	LP; LP	AV, MD, PD	NDR	NDR
Family 14-II:1	LCA5	c.[795T > G];[795T > G]	M	20	FMB	AR	PV, NYS	HM; HM	AV, MD, CRD	NDR	NDR
Family 14-II:3	LCA5	c.[795T > G];[795T > G]	M	17	FMB	AR	PV, NYS	FC; FC	AV, MD, CRD	NA	NA
Family 15-II:1	TULP1	c.[1198C > T];[1444C > T]	M	11	FMB	Isolated	PV, NYS	0.09; 0.05	AV, CRD	NDR	NDR

 M, male; F, female; FMB, first few months after birth; Cons, consanguineous marriage of parents; PV, poor vision; RN, roving nystagmus; NYS, nystagmus; ODS, oculodigital sign; PA, photoaversion; NPL, no pursuit of light; FC, finger counting; NPO, no pursuit of objects; HM, hand movement; PL, pursuit of light; LP, light perception; RNA, relatively normal appearance; AV, attenuated vessels; CRD, carpet-like retinal degeneration; PWF, peripheral white flecks; MD, macular dystrophy; PNP, posterior nummular pigmentation; PD, pigmentary deposit; NDR, no detectable responses.

*  Biocular pursuit of light or objects (such as a pen or pencil) was tested in those young children who were unable to have visual acuity measured.

Discussion

In our study, we tested whole-exome sequencing followed by confirmatory Sanger sequencing as a possible diagnostic approach for LCA. We identified 22 potentially pathogenic mutations in 15 of 41 unrelated Chinese families with LCA. Of the 15 families with potentially pathogenic mutations, 13 had homozygous or compound heterozygous mutations in genes associated with recessive LCA, and two had heterozygous mutations in genes associated with dominant LCA (CRX and IMPDH1). 

Additionally, triallelic variations were identified in Family 1, Family 2, Family 3, and Family 12. Digenic and trigenic variations also were identified in our study (Supplementary Table S3). Segregation analyses were performed on Family 2 and Family 12 because of the availability of samples from family members (Fig. 1). The four probands with triallelic variants did not manifest a more severe phenotype than other patients, which suggests that the third allele does not always aggravate retinal dysfunction (Table 2). On the other hand, the parents who carried digenic variants in Family 2 and Family 12 had normal visual acuity, without any detectable abnormalities on fundus examination. A similar phenomenon also was described in a previous LCA case report. ⁴⁵ Although digenic and triallelic inheritance patterns have been established in probands with RP ⁴⁶ and BBS, ⁴⁷ respectively, further studies are needed to determine whether digenic variations are causative and whether triallelic mutations result in more severe retinal dysfunction. Because these types of variants would be expected to be detected with increasing frequency with the rapid development of high-throughput sequencing, rigorous investigation should be undertaken before reaching a final conclusion regarding pathogenicity. 

The detection rate of 48% (12 of 25 families) in newly recruited probands was higher than that in our prior study, which had a detection rate of 35.6% when screening exons reported previously to contain mutations by Sanger sequencing. ³⁶ Given that the number of newly recruited probands was limited and the other 16 families had been included in a previous study, ³⁶ the analyses of frequency and spectrum of mutated genes were based on 112 families, combining the previous 87 probands with 25 new cases. The most frequently mutated genes in our LCA cohorts were GUCY2D (10%), CRB1 (7%), RPGRIP1 (5%), CEP290 (4%), and CRX (3%). The five most frequently mutated genes are similar to those in other studies conducted in Caucasian populations. ⁴ Potential genotype-phenotype correlations could not be extracted for mutations in most of these patients because of variable manifestation and limited number of cases, except for the association of nummular pigmentation with CRB1 mutations (Table 2, Fig. 2). CRB1 mutations were identified in two patients in Family 6 in this study and five probands out of 87 LCA families previously, ³⁶ all were associated with similar fundus change of nummular pigmentation. 

Compared to conventional sequencing methods, whole-exome sequencing is a rapid, reliable, and less labor-intensive way to establish a precise molecular diagnosis for genetically heterogeneous diseases. Most mutations identified in the 41 LCA probands were novel (77.3%, 17/22); therefore, they could not be detected by an LCA mutation chip. ³³ Considering the heterogeneous nature of LCA, whole-exome sequencing also reveals a broader picture of the molecular background of LCA and provides an opportunity to identify new genes responsible for the disease. Given that the numerous variants identified by exome sequencing may easily lead researchers astray, it would be reasonable to extend this study to include other RP-related genes because it can be difficult to distinguish LCA from early-onset severe RP by clinical examination, and because some genes have been implicated in LCA and RP (available in the public domain at https://sph.uth.edu/retnet/). Among the 26 LCA families without identified mutations, 18 variants that affected either encoded residues or splicing in known RP genes were detected (Supplementary Table S4). However, none of these variants could be confirmed to be pathogenic mutations after bioinformatic and segregation analyses. Extensive additional analysis may be required to identify new genes responsible for LCA in patients without mutations identified in the known genes. 

In conclusion, our study suggests that it should be possible to identify mutations in approximately half of Han Chinese families with LCA by screening the 19 known LCA genes. Together with our previous studies, these results delineate further the mutation spectrum and mutation frequencies of LCA in Han Chinese individuals. Our study also shows that whole-exome sequencing is an efficient method to establish a precise molecular diagnosis for highly heterogeneous hereditary diseases, such as LCA. As additional genes are implicated in the pathogenesis of LCA and more patients become available over time, it should be possible to explore further the underlying pathogenesis of the remaining probands. 

Supplementary Materials

Chinese Abstract 

pdf S1 

xls S2 

xls S3 

xls S4 

xls S5 

Acknowledgments

The authors thank all patients and their family members for their participation. 

Supported by the National Natural Science Foundation of China (81170881, U1201221), Guangdong Translational Medicine Public Platform (4202037), the “985 project” of Sun Yat-sen University, and the Fundamental Research Funds of the State Key Laboratory of Ophthalmology. 

Disclosure: Y. Chen, None; Q. Zhang, None; T. Shen, None; X. Xiao, None; S. Li, None; L. Guan, None; J. Zhang, None; Z. Zhu, None; Y. Yin, None; P. Wang, None; X. Guo, None; J. Wang, None; Q. Zhang, None 

Footnotes

 YC, QZ, JW, and QZ contributed equally to the work presented here and should therefore be regarded as equivalent authors.