Mutation Analysis of Seven Known Glaucoma-Associated Genes in Chinese Patients With Glaucoma

Abstract

Purpose.: To evaluate mutations in the MYOC, WDR36, OPTN, OPA1, NTF4, CYP1B1, and LTBP2 genes in a cohort of Chinese patients with primary glaucoma.

Methods.: Genomic DNA was prepared from 683 unrelated patients, including 50 with primary congenital glaucoma, 104 with juvenile open-angle glaucoma (JOAG), 186 with primary open-angle glaucoma (POAG), and 343 with primary angle-closure glaucoma (PACG). Mutations in the seven genes in 257 patients (36 with JOAG, 89 with POAG, and 132 with PACG) were initially analyzed by exome sequencing and then confirmed by Sanger sequencing. In addition, Sanger sequencing was used to detect MYOC mutations in the remaining 426 patients.

Results.: Exome sequencing identified 19 mutations (6 in MYOC, 9 in WDR36, 3 in OPA1, and 1 in OPTN) in 20 of 257 patients, including 4 patients with JOAG, 8 patients with POAG, and 8 patients with PACG. No mutation was detected in the other three genes. In addition, Sanger sequencing detected additional MYOC mutations in 5 of the remaining 426 patients, including 3 patients with JOAG and 2 patients with POAG.

Conclusions.: Twenty-two mutations in MYOC, WDR36, OPA1, and OPTN were detected in 25 of the 683 patients with primary glaucoma, including nine MYOC mutations in 11 patients, nine WDR36 mutations in 11 patients, three OPA1 mutations in 3 patients, and one OPTN mutation in a patient who also carried a MYOC mutation. Eight mutations in MYOC, WDR36, and OPA1 in 8 of the 343 PACG patients are of uncertain significance and need to be analyzed further.

Introduction

Glaucoma, the second leading cause of irreversible blindness worldwide, ^1,2 is characterized by typical structural damage to the optic nerve, specific visual field defects, ³ and often relatively higher IOP. ⁴ Approximately 20% of glaucoma occurs secondary to other ocular or systemic diseases. ³ Based on anatomical changes in the anterior chamber angle, primary glaucoma can be classified as primary angle closure glaucoma (PACG) or primary open-angle glaucoma (POAG) ³ ; the latter classification can be further subdivided into juvenile open-angle glaucoma (JOAG) and adult onset POAG. ⁵  

Genetic factors have been regarded as the critical risk factor in the development of glaucoma. ^6–8 The following genes have been reported to be associated with glaucoma: myocilin (MYOC), ⁹ optineurin (OPTN), ¹⁰ WD repeat domain 36 (WDR36), ^11,12 neurotrophin 4 (NTF4), ¹³ optic atrophy 1 (OPA1), ^14,15 cytochrome P450, family 1, subfamily B (CYP1B1), ¹⁶ and latent transforming growth factor-beta binding protein 2 (LTBP2). ¹⁷ To date, the mutations of known genes account for only approximately 5% of patients with POAG, ¹⁸ and the influences of these mutations on patients with PACG have been controversial. ¹⁹ It was interesting that different types of glaucoma have overlapping candidate genes. ⁸ MYOC has been confirmed to be associated with JOAG and POAG, with different mutation frequencies in different ethnic groups. ²⁰ Variants of OPA1 are associated with normal tension glaucoma (NTG). ¹⁴ A study of Japanese patients further implied that OPA1 affected the phenotypic features of patients with high-tension glaucoma. ²¹ The repetitive studies in Caucasian, African-American, Ghanaian, and African-Caribbean populations did not find similar relationships. ²² However, no study involving Chinese patients has been reported. CYP1B1, a definite primary congenital glaucoma (PCG) gene,was associated with JOAG as an autosomal recessive trait, ^23,24 and variants of CYP1B1 were also detected in patients with PACG. ²⁵ A systemic analysis of the mutation frequency and mutation spectra in different types of glaucoma might provide a better understanding of the heterogeneity of glaucoma. 

The development of whole-exome sequencing makes screening of mutations realistic in a cost-efficient and labor-saving way. ^26,27 In this study, whole-exome sequencing was used to screen variations in the seven genes mentioned above in 257 unrelated Chinese patients with primary glaucoma. Variations detected in the seven genes were confirmed by Sanger sequencing. In addition, the mutation analysis for MYOC was further extended to an additional 426 unrelated patients with primary glaucoma by Sanger sequencing alone. 

Materials and Methods

Subjects

All 683 patients with primary glaucoma from unrelated Chinese families were recruited from the Glaucoma Department of the Eye Hospital of the Zhongshan Ophthalmic Center, including 50 cases with PCG, 104 cases with JOAG, 186 cases with POAG, and 343 cases with PACG. Written informed consent was obtained from the participants or their guardians before the collection of peripheral venous blood. Genomic DNA was extracted from leukocytes of the peripheral blood of each participant or from their guardians in adherence to the Declaration of Helsinki as previously described. ²⁸ This study was approved by the Institutional Review Board of the Zhongshan Ophthalmic Center. 

The diagnosis criteria of POAG were consistent with those of a previous study, as follows, briefly ²⁹ : (a) IOP over 21 mm Hg by Goldmann applanation tonometry (GAT); (b) presence of glaucomatous visual field defect or glaucomatous optic neuropathy; and (c) open angle. Patients whose onset occurred earlier than 35 years of age were further categorized as having JOAG. ³ The diagnosis of PACG was consistent with criteria used in a previous study, as follows, briefly ³⁰ : (a) an IOP of more than 21 mm Hg by GAT; (b) glaucomatous optic neuropathy; (c) glaucomatous visual field defect; and (d) closed angle. Diagnosis of PCG was consistent with criteria used in a previous study, as follows, briefly ³¹ : (a) an IOP of more than 21 mm Hg by Schiotz tonometer or GAT; (b) glaucomatous optic neuropathy; and (c) age of onset at less than 3 years of age or presence of buphthalmos without secondary causes. Patients with secondary glaucoma (resulting from, e.g., trauma, uveitis, steroid-induced or neovascular glaucoma, and other associated ocular or systemic anomalies) were excluded. 

Whole-Exome Sequencing

DNA samples from 257 of the 683 patients were initially analyzed by whole-exome sequencing, including 36 patients with JOAG, 89 patients with POAG, and 132 patients with PACG. Exome sequencing was performed by a commercial service (Macrogen, Rockville, MD, USA; available in the public domain at http://www.macrogen.com/). Exome capture was carried out by using an exome enrichment kit (TruSeq; product no. 62 M; Illumina, San Diego, CA, USA) array. Exome-enriched DNA fragments were sequenced with HiSeq2000 (Illumina), with an average 125-fold sequencing depth. The short reads were mapped to the hg19 reference genome with Burrows-Wheeler Aligner software (Burrows-Wheeler Aligner; available in the public domain at http://bio-bwa.sourceforge.net/). SAMtools software (available in the public domain at http://samtools.sourceforge.net) was used for variant calling based on a Bayesian statistical algorithm. ^32,33 To increase the accuracy of variant calling, default filters were set by discarding: (1) single-nucleotide polymorphisms (SNPs) covered by fewer than three reads; (2) SNPs without reads with a mapping quality ≥ 30, which implies that there is a 1 in 1000 Phred-scaled probability that the read is incorrectly mapped ³³ ; (3) SNPs with quality scores less than 30, which indicates the probability of uncorrected variant calling ³³ ; (4) indel with quality score less than 20; (5) SNPs within a 10-bp-adjacent region around indel alternation; (6) SNPs without reads longer than 10 bp; and (7) indel if there were two gaps within 30 bp. Detected variations affecting coding residues or splicing in the seven genes were selected for further validation. Only homozygous or compound heterozygous variations in autosomal recessive PCG genes were included. Databases of dbSNP138, the 1000 genome project, and the National Heart Lung Blood Institute Grand Opportunity Exome Sequencing Project (ESP) were searched to evaluate variant frequency. ³⁴ Variants with minor allele frequency ≥ 0.01 were excluded. 

Sanger Sequencing

Sanger sequencing was used to confirm the variants in the seven genes detected in the 257 patients by exome sequencing. Sanger sequencing was also used to sequence the coding region of MYOC in 192 normal individuals and the remaining 426 of the 683 patients, including 68 with JOAG, 97 with POAG, 211 with PACG, and 50 with PCG. The primers used for amplification were designed using the Primer3 tool (available in the public domain at http://frodo.wi.mit.edu/primer3/); the sequences of these primers are listed in Supplementary Table S2. PCR was used to amplify the targeted fragments with variants detected by exome sequencing or the coding regions of MYOC. The purified amplicons were then analyzed using an ABI 3130 genetic analyzer (Applied Biosystems, Foster City, CA, USA) with BigDye Terminator cycle sequencing kit version 3.1. The sequence results were compared with consensus sequences to identify variants by using Seqman software (Lasergene version 8.0; DNASTAR, Madison, WI, USA). Variants with minor allele frequency ≥ 0.01 were excluded. Detected variations affecting coding residues or splicing in the seven genes were selected for further validation. The impact of amino acid substitution resulting from the variants was predicted by using the SIFT (available in the public domain at http://sift.jcvi.org/) ³⁵ and PolyPhen-2 tools (available in the public domain at http://genetics.bwh.harvard.edu/pph2/). ³⁶ The influence of the variants on the splice site was predicted using the Berkeley Drosophila Genome Project (available in the public domain at http://www.fruitfly.org/). ³⁷ Novel variants and variants predicted to be pathogenic were evaluated in 192 normal individuals. Variants were considered pathogenic based on three criteria: (a) they were predicted to be damaging in at least one type of bioinformatic analysis listed above; (b) they were not detected in normal individuals; (c) they were novel variants or variants with statistically greater frequency than that in the 1000 genome and ESP databases (P < 0.05). Mutations reported in previous studies were also considered pathogenic if they were not detected in normal individuals. 

Results

Whole-exome sequencing detected 39 heterozygous variants in 4 of the 7 genes in 56 probands. These variants were present in MYOC, WDR36, OPTN, and OPA1, all of which were associated with autosomal dominant glaucoma. No variant was detected in NTF4, and no homozygous or compound heterozygous variants expected to be associated with autosomal recessive PCG were detected in CYP1B1 and LTBP2. Of the 39 variants detected, 36 variants were confirmed by Sanger sequencing, and the others were false positives. 

Nineteen of the 36 variants were considered potential pathogenic mutations, including 5 previously reported and 14 novel ones (Table 1). ^38–44 Mutations were detected in 20 of 257 patients, including 6 in MYOC, 11 in WDR36, and 3 in OPA1. Digenic mutations (c.C244T in MYOC and c.C160G in OPTN) were detected in 1 of the 6 patients harboring a MYOC mutation (patient G542). Two heterozygous mutations in WDR36 were detected in one patient (G438) with phase unknown. Patients with mutations included 4 of 36 with JOAG, 8 of 89 with POAG, and 8 of 343 with PACG (Table 2). The mutation frequencies and spectra in different types of glaucoma are shown in Figure 1. Seventeen less likely pathogenic variants are listed in Supplementary Table S3. ^40,45,46  

Figure 1

 

Classification of glaucoma probands based on the results of whole-exome sequencing. (A) Mutation frequencies and spectra in 257 probands. (B–D) Mutation frequencies and spectra in 36 probands of JOAG, in 89 of POAG, and in 132 of PACG, respectively.

Table 1

 

The 22 Potential Pathogenic Variants Identified in 25 of 683 Chinese Glaucoma Families

Gene	Patient ID	Variations		State	Sequencing Technique	Bioinformatic Analysis			Reported or Not	Ref	Frequency in NC
		Nucleotide	Amino Acid			SIFT	PolyPhen2	BDGP
MYOC	G485	c.158T>C	p.V53A	Het	WES	D	PrD	N	Reported	39	0/384
MYOC	G542†	c.244C>T	p.R82C	Het	WES	D	PrD	SD	Reported	41	0/384
MYOC	G473	c.271C>T	p.R91*	Het	SS	NA	NA	N	Reported	38	0/384
MYOC	G571	c.878C>A	p.T293K	Het	WES	T	B	N	Reported	41	0/384
MYOC	G360	c.1109C>T	p.P370L	Het	WES	D	PrD	N	Reported	44	0/384
MYOC	G127, G366	c.1021T>C	p.S341P	Het	SS,WES	D	PrD	N	Reported	42	0/384
MYOC	G336	c.1133A>G	p.D378G	Het	SS	D	PrD	N	Novel		0/384
MYOC	G719, G346	c.1160G>A	p.G387D	Het	SS	D	PrD	N	Novel		0/384
MYOC	G60	c.1456C>T	p.L486F	Het	WES	D	PrD	N	Novel		0/384
WDR36	G666	c.283C>G	p.R95G	Het	WES	T	PrD	N	Novel		0/384
WDR36	G674	c.451A>G	p.N151D	Het	WES	D	PrD	N	Novel		0/384
WDR36	G438	c.[577+37_577+38dupGGGT(;)2635A>G]	p.[SD(;)M879V]	Het	WES	NA;T	NA;PrD	SD;N	Novel		0/384
WDR36	G387	c.598T>G	p.F200V	Het	WES	D	PrD	N	Novel		0/384
WDR36	G445	c.898+23A>G	SD	Het	WES	NA	NA	ASS;N	Novel		0/384
WDR36	G411, G501, G626	c.1645G>A	p.G549R	Het	WES	T	PrD	N	Novel		0/384
WDR36	G392	c.1646G>A	p.G549E	Het	WES	T	PrD	N	Novel		0/384
WDR36	G229	c.2635A>G	p.M879V	Het	WES	T	PrD	N	Novel		0/384
WDR36	G166	c.2803C>T	p.L935F	Het	WES	T	PrD	N	Novel		0/384
OPA1	G493	c.1240G>A	p.V414I	Het	WES	T	PrD	N	Novel		0/384
OPA1	G566	c.658_660del	p.Q219_I221del	Het	WES	NA	NA	N	Novel		0/384
OPA1	G577	c.1781G>A	p.R594Q	Het	WES	D	PrD	N	Novel		0/384
OPTN	G542†	c.160C>G	p.L54V	Het	WES	D	PrD	N	Novel		0/384

 AD, autosomal dominant; ASS, additional splicing site; B, benign; BDGP, Berkeley Drosophila Genome Project; D, damaging; Het, heterozygous; Hom, homozygous; N, no change; NA, not applicable for variants other than amino acid substitution; NC, normal control; PrD, probably damaging; SD, splicing defect; SS, Sanger sequencing; T, tolerated; WES, whole-exome sequencing. Of the 25 families, patient G542 had digenic mutations in MYOC and OPTN, and patient G438 had two mutations in WDR36 with phase unknown. Mutations with reference citations were reported to be pathogenic. No mutations were detected in the NTF4, CYP1B1, and LTBP2 genes.

†  Patient with digenic mutation.

Table 2

 

Clinical Data From 25 Probands With Potential Pathologic Mutations Identified in this Study

Family ID	Diagnosis	Inheritance	Gene	Mutation	Sequencing Technique	Sex	Age at Diagnosis, y	BCVA		Max IOP, mm Hg		VCDR		Visual Field MD
								OD	OS	OD	OS	OD	OS	OD	OS
G485	JOAG OU	AD	MYOC	c.[158T>C];[=]	WES	M	31	0.7	0.2	NA	NA	0.7	0.8	NA	NA
G542	JOAG OU	AD	MYOC (digenic)	MYOC: c.[244C>T]; OPTN: c.[160C>G]	WES	M	15	1	1	NA	NA	0.9	1	−13.97 dB	−23.60 dB
G473	JOAG OU	AD	MYOC	c.[271C>T];[=]	SS	F	21	FC	1	51	41	0.95	0.6	NA	VFD
G360	JOAG OU	AD	MYOC	c.[1109C>T];[=]	WES	F	32	0.5	1	39.3	37.7	0.8	0.7	NA	−16.12 dB
G127	JOAG OU	Sporadic	MYOC	c.[1021T>C];[=]	SS	F	14	0.08	0.3	52.7	NA	1	1	NA	VFD
G336	JOAG OU	Sporadic	MYOC	c.[1133A>G];[=]	SS	M	23	FC	1.5	47.5	46	1	0.8	NA	VFD
G366	POAG OU	AD	MYOC	c.[1021T>C];[=]	WES	F	41	0.8	0.7	24.3	40	0.9	1	VFD	VFD
G346	POAG OU	Sporadic	MYOC	c.[1160G>A];[=]	SS	M	59	1	0.9	NA	21	0.6	0.9	VFD	VFD
G719	POAG OU	Sporadic	MYOC	c.[1160G>A];[=]	SS	M	39	1	0.04	38	27	0.7	0.6	VFD	VFD
G60	POAG OU	Sporadic	MYOC	c.[1456C>T];[=]	WES	M	36	0.2	0.6	22.4	33	1	0.8	VFD	VFD
G571	PACG OU	Sporadic	MYOC	c.[878C>A];[=]	WES	F	68	0.6	0.1	10	60	0.5	0.3	NA	NA
G438	JOAG OU	Sporadic	WDR36	c.[577+37_577+38dup (;)2635A>G]	WES	F	33	0.7	0.8	29	65.3	0.4	0.7	−18.1 dB	−22.54 dB
G387	POAG OU	AD	WDR36	c.[598T>G];[=]	WES	M	54	1	0.8	23	26	0.3	0.4	−13.36 dB	−7.85 dB
G445	POAG OS	Sporadic	WDR36	c.[898+23A>G];[=]	WES	M	42	1.5	0.9	14.3	44.7	0.2	0.9	NA	NA
G501	POAG OU	Sporadic	WDR36	c.[1645G>A];[=]	WES	M	47	1.2	1	NA	NA	0.9	0.8	−10.78 dB	−3.94 dB
G411	POAG OU	Sporadic	WDR36	c.[1645G>A];[=]	WES	M	60	1.2	1.2	23	33	0.6	0.7	−4.06 dB	−12.5 dB
G392	POAG OU	Sporadic	WDR36	c.[1646G>A];[=]	WES	M	69	1	0.9	25.8	25.8	0.7	0.5	−3.68 dB	−2.22 dB
G229	PACG OU	AD	WDR36	c.[2635A>G];[=]	WES	F	55	1	0.9	17	32	0.4	1	NA	NA
G166	PACG OU	AD	WDR36	c.[2803C>T];[=]	WES	F	47	0.2	0.7	37	21.9	0.9	0.5	−31.44 dB	−7.94 dB
G666	PACG OU	Sporadic	WDR36	c.[283C>G];[=]	WES	M	75	0.3	0.1	35.3	55.7	0.9	0.9	−30.59 dB	NA
G674	PACG OU	Sporadic	WDR36	c.[451A>G];[=]	WES	F	60	0.4	0.4	43.2	35.1	0.5	0.3	−22.81 dB	−13.27 dB
G626	PACG OU	Sporadic	WDR36	c.[1645G>A];[=]	WES	F	88	HM	0.2	47	10	0.8	NA	NA	NA
G493	NTG OU	Sporadic	OPA1	c.[1129G>A];[=]	WES	F	75	1	0.9	20	20	0.8	0.7	−2.80 dB	−0.01 dB
G566	PACG OU	Sporadic	OPA1	c.[658_660del];[=]	WES	F	40	1.2	0.6	17	41	0.3	1	−6.13 dB	−27.60 dB
G577	PACG OU	Sporadic	OPA1	c.[1670G>A];[=]	WES	F	58	1	1	11.4	34.7	0.3	0.3	−1.63 dB	−1.64 dB

 Of the 25 probands, anterior chamber angle closure was recorded in all nine probands with PACG. BCVA, best corrected visual acuity; F, female; FC, figure count; HM, hand movement; M, male; Max IOP, maximum intraocular pressure; MD, mean deviation; NA, not available; NLP, no light perception; VCDR, vertical cup-to-disc ratio; VFD, visual field defect.

In the mutation screening of MYOC by Sanger sequencing, 11 variants were detected in 32 probands. Four of the 11 variants were considered potentially pathogenic mutations, including 2 previously reported variants and 2 novel ones. One of the four mutations was also detected by whole-exome sequencing in another patient. The mutations were detected in 3 of 38 patients with JOAG and 2 of 97 patients with POAG (Table 1). However, no mutations were identified in 343 patients with PACG or in 50 patients with PCG. When all 683 unrelated families were included, the MYOC mutation frequency was 5.8% (6 of 104 patients) in JOAG, 2.1% (4 of 186 patients) in POAG and 0.3% (1 of 343 patients) in PACG. There was one rare nonsynonymous variant predicted to be damaging in the group of 192 normal individuals when the polymorphisms were excluded. Compared to the normal controls, the mutation frequency was higher in JOAG (P < 0.01), but the frequencies in POAG and PACG were not statistically different from those in normal controls (P > 0.01). Given that the mutation detected in the patient with PACG was also reported to be pathogenic in a previous study, ⁴¹ its influence on glaucoma is still uncertain based on the current evidence. 

Segregation analysis was possible in four families (Fig. 2). Mutation c.1021T>C and c.271C>T in MYOC were cosegregated with POAG. All siblings of the proband in family G366 had a diagnosis of POAG and harbored the MYOC mutation c.1021T>C. As for the G473 family, parents and all siblings of the proband underwent segregation analysis. Only the father harbored the same MYOC mutation, c.271C>T, and POAG was diagnosed based on the arcuate visual defect and nighttime elevated IOP. The proband of the G378 family, POAG was diagnosed at the age of 54; the c.598T>G mutation in WDR36 was detected in two of her sisters: the 51-year-old sister had a typical glaucoma visual function defect, but the 48-year-old sister was normal. Because glaucoma is a complex disease with diverse ages of onset, intensive follow-up was advised. The c.658_660del mutation in OPA1 was not transmitted to two children of proband G566, and neither the evidence of PACG nor the risk factors in proband G566 (i.e., shorter axis length, shallow anterior chamber, and thicker lens) were observed in her children. Neither verbal symptoms of glaucoma nor severe visual defect were evident in the proband's mother, who carried the same mutation, but exact clinic data were not available. 

Figure 2

 

Cosegregation analysis in glaucoma pedigree with potential pathogenic mutations. +, wild-type allele.

Clinical data for probands with potentially pathogenic mutations are shown in Table 2. The average age of onset in JOAG, POAG, and PACG was 24, 52, and 61 years old, respectively. Elevated IOP, enlarged cup-to-disc ratio, and narrow visual field were the common manifestations, whereas anterior chamber angle closure was observed in all patients in whom PACG was diagnosed (Fig. 3). 

Figure 3

 

Typical glaucoma damage (i.e., enlarged cup-to-disc ratio and arcuate visual field defect) was observed, and the shallow anterior chamber and anterior chamber angle closures were recorded in probands of patients with PACG. There was no special link between mutation and phenotype in this study.

Discussion

In this study, 22 potential pathogenic mutations in 4 of the 7 genes tested were identified in 25 of the 683 unrelated Chinese patients with primary glaucoma. Mutations in WDR36 were detected in patients with JOAG, POAG, and PACG, as were the mutations in MYOC. In addition, mutations in OPA1 were found in POAG and PACG. 

It is interesting that 1 MYOC mutation, 5 WDR36 mutations, and 2 OPA1 mutations were identified in 8 of the 343 patients with PACG (2.3%). Seven of these mutations were predicted to be pathogenic, and 1 in MYOC was reported to be associated with POAG. ⁴¹ None of the mutations was present in the 192 normal individuals. Analysis of the clinical records confirmed the diagnosis of PACG in the patients with these mutations. Of these mutations, c.2635A>G/ p.M879V was identified in 1 patients with JOAG and in 1 with PACG, and the c.1645G>A/p.G549R mutation in WDR36 was identified in 2 patients with POAG and in 1 patient with PACG. Previously, mutation analysis in patients with PACG was performed for MYOC, WDR36, OPTN, NTF4, and CYP1B1 genes but not the LTBP2 gene, but no potential pathogenic mutation was identified. In a cohort of 106 Chinese patients with PACG, eight variants were detected in MYOC, all of which were present in normal individuals. ⁴⁷ In an investigation of NTF4, the A88V variant was detected in 3 of 111 patients with PACG and in normal individuals. ⁴⁸ In a study of CYP1B1 mutations, screening in Indian glaucoma patients, including 90 with PACG, only heterozygous mutations were identified in PACG. ²⁵ In another systemic mutation analysis in 29 Middle Eastern patients with PACG, no mutations were detected in MYOC, OPTN, CYP1B1, WDR36, or OPA1. ⁴⁹ To date, we cannot determine the pathogenicity of the mutations identified; functional investigation might lead to valuable insights in the future. 

Of the JOAG group, 6 patients were identified with MYOC mutations (including 1 patient with digenic mutations in MYOC and OPTN) and 1 patient with 2 heterozygous mutations in WDR36. MYOC was the first gene to be associated with JOAG, and the mutation frequency ranged from 10% to 30% in previous studies. ²⁰ In our study, the overall mutation frequency of MYOC in JOAG was 5.8% (6 of 104 patients), whereas a mutation frequency of 12.5% was reported in a Taiwanese study in which the mutations were detected in 6 of 48 Taiwanese patients with JOAG. ⁵⁰ However, MYOC was still the major genetic reason for JOAG in the Chinese mainland population. The age of onset for the patient with digenic mutations in both MYOC and OPTN was 15 years old, earlier than the average age in this study. Results concerning the effect of OPTN on glaucoma were controversial, ^10,51–53 but the expression of MYOC might be regulated by OPTN. ⁵⁴ The frequency of mutations in WDR36 varies in previous studies. In a study that aimed to screen glaucoma patient for mutations in WDR36, a single mutation was detected in 1 of 6 probands with JOAG. ⁵⁵ In another study, 4 WDR36 mutations were identified in 5 of 47 German patients with JOAG. ⁵⁴ However, in a study involving 135 Chinese patients, no WDR36 mutations were detected in 11 patients with JOAG. ⁵⁶ In our study, patient G348, who harbored two heterozygous mutations in WDR36, suffered from a more severe phenotype (i.e., more severely elevated IOP, optic nerve defect, and progressive visual field defect in the condition of normal IOP under eye drops). One of the mutations might act as a modifying variant. ⁵⁵ Investigation of a larger sample with JOAG is needed to estimate the WDR36 mutation frequency in a Chinese population. 

In our study, only mutations in MYOC, WDR36, and OPA1 were detected in patients with POAG. According to the results of whole-exome sequencing, WDR36 was the gene most frequently affected with a mutation frequency of 5.6% (5 of 89 patients), more than twice that of MYOC. The mutation frequencies of MYOC in whole-exome sequencing and Sanger sequencing were 2.2% and 2.1%, respectively. This result was consistent with previous reports in Dutch, ⁵⁷ North American, ⁵⁸ African American, Canadian, Australian, Japanese, ⁴⁶ and Indian populations, ⁵⁹ in all of whom the MYOC mutation frequency ranged from 2% to 4%. The mutation frequency of WDR36 in POAG varies in different ethnic groups; for example, no mutations have been found in Russian ⁶⁰ and Australian ⁶¹ subjects, but frequencies range from 2.9% to 3.7% in American, ⁵⁵ Italian, ⁶² Spanish, ⁶³ Chinese, ⁵⁶ and German ⁵⁴ population groups. In our study, the mutation frequency (5.6%) was slightly higher. It was noteworthy that two nearby mutations (c.1645A>G and c.1646G>A) in WDR36 were found in 4 patients (including 3 POAG and 1 PACG patient), which indicates a mutation “hotspot” in the Chinese glaucoma cohort. OPA1 was associated with NTG in patients from Caucasia, England, and Japan. ^14,22 In our study, the c.G1240A variant in OPA1 was detected in one patient with NTG, but replication with different ethnic groups is necessary. 

No mutations were detected in NTF4, CYP1B1, and LTBP2. NTF4 was initially reported as a rare cause of POAG. ¹³ In consideration of its low mutation frequency in the Chinese population, ^64,65 research with a larger POAG cohort was encouraged. There was no evidence to support the contributions of CYP1B1 and LTBP2, two genes both associated with PCG, to the causes of POAG and PACG in the Chinese population. 

Whole-exome sequencing has been considered an effective method for Mendelian diseases gene discovery because most mutations are located in protein-encoding regions. ^26,66 In our study, 36 of 39 variants (92.3%) detected via whole-exome sequencing were confirmed, including 22 novel and 14 reported variants. The overall mutation frequency in glaucoma was 7.8% (20 of 257 patients) via whole-exome sequencing. For most glaucoma patients, the genetic risk factor is still unknown. It is currently impractical to design a chip for mutation screening in glaucoma patients. Compared with the traditional technique of Sanger sequencing, whole-exome sequencing can cover a wide range of genes in a cost-efficient and labor-saving way. For probands with biallelic or triallelic variants such as the G366II7 family, who harbor digenic variants predicted to be pathogenic, although only one variant is cosegregated with POAG, it is apt to bias genetic diagnosis if only a single gene analysis is undertaken. Therefore, whole-exome sequencing is more reliable in detection of pathogenic mutations in multifactor diseases. 

This study provides an overview of the mutation spectra and frequencies in PACG, JOAG, and POAG patients of Chinese ancestry. This systematic analysis involved both a POAG and a PACG cohort. Whole-exome sequencing was a promising gene discovery tool, and it might provide crucial clues to the research of novel candidate genes in glaucoma patients without mutations in MYOC, WDR36, and OPTN. 

Supplementary Materials

pdf S4 

xlsx S1 

xlsx S2 

xlsx S3 

Acknowledgments

The authors thank all patients for their participation. 

Supported by the National Natural Science Foundation of China (81170881, U1201221) and the Fundamental Research Funds of the State Key Laboratory of Ophthalmology. 

Disclosure: X. Huang, None; M. Li, None; X. Guo, None; S. Li, None; X. Xiao, None; X. Jia, None; X. Liu, None; Q. Zhang, None 

Footnotes

 XH, ML, XG, XL, and QZ contributed equally to the work presented here and should therefore be regarded as equivalent authors.