Genetic Analysis in a Swiss Cohort of Bilateral Congenital Cataract

Key Points

Question

Which genes are associated with bilateral congenital cataract in the Swiss population?

Findings

This cohort study of 37 patients from 25 families with bilateral congenital cataract detected 13 novel and 7 known pathogenic variants in genes associated with congenital cataract. Approximately one-third of these variants were associated with syndromic cataract, with a total detection yield of 80%; a recessive variant in the CRYBB2 gene is described in contrast to the previously reported autosomal dominant trait.

Meaning

These findings demonstrate a very high genetic detection yield and underline the importance of performing genetic testing in congenital cataract for family counseling and excluding a syndromic disease.

This cohort study reports disease-causing variants and their detailed phenotype in patients with bilateral congenital cataract from a single center in Switzerland to draw a genetic map and perform a genotype-phenotype comparison of this cohort.

Abstract

Importance

Identification of geographic population-based differences in genotype and phenotype heterogeneity are important for targeted and patient-specific diagnosis and treatment, counseling, and screening strategies.

Objective

To report disease-causing variants and their detailed phenotype in patients with bilateral congenital cataract from a single center in Switzerland and thereby draw a genetic map and perform a genotype-phenotype comparison of this cohort.

Design, Setting, and Participants

This clinical and molecular-genetic cohort study took place through the collaboration of the Department of Ophthalmology at the University Hospital Zurich and the Institute of Medical Molecular Genetics, University of Zurich, Schlieren, Switzerland. Thirty-seven patients from 25 families with different types of bilateral congenital cataract were included. All participating family members received a comprehensive eye examination. Whole exome sequencing was performed in the index patients, followed by a filtering process to detect possible disease-associated variants in genes previously described in association with congenital cataract. Probable disease-causing variants were confirmed by Sanger sequencing in available family members. All data were collected from January 2018 to June 2020, and the molecular-genetic analyses were performed from January 2019 to July 2020.

Main Outcomes and Measures

Identification of the underlying genetic causes of bilateral congenital cataract, including novel disease-causing variants and phenotype correlation.

Results

Among the 37 patients (18 [49%] male and 19 [51%] female; mean [SD] age, 17.3 [15.9] years) from 25 families, pathogenic variants were detected in 20 families (80% detection rate), which included 13 novel variants in the following genes: BCOR, COL4A1, CRYBA2, CRYBB2, CRYGC, CRYGS, GJA3, MAF, NHS, and WFS1. Putative disease-causing variants were identified in 14 of 20 families (70%) as isolated cases and in 6 of 20 families (30%) with syndromic cases. A recessive variant in the CRYBB2 gene in a consanguineous family with 2 affected siblings showing a nuclear and sutural cataract was reported in contrast to previously published reports. In addition, the effect on splicing in a minigene assay of a novel splice site variant in the NHS gene (c.[719-2A>G]) supported the pathogenicity of this variant.

Conclusions and Relevance

This study emphasizes the importance of genetic testing of congenital cataracts. Known dominant genes need to be considered for recessive inheritance patterns. Syndromic types of cataract may be underdiagnosed in patients with mild systemic features.

Introduction

Pediatric cataract is defined as an opacity of the crystalline lens, which may occur at different ages and can be categorized accordingly.^1,2 Congenital cataracts may be associated with severe amblyopia and secondary vision-threatening glaucoma after successful cataract surgery, or remain asymptomatic and unnoticed, unless a dilated fundus examination is performed. Cataract classification systems are based on the anatomical landmarks of the lens, morphology of the cataract, and etiology.^3,4,5 Inherited cataracts can either be isolated, which is the most common type with a frequency at approximately two-thirds of all inherited cataract cases,^6,7 or occur in association with other ocular developmental anomalies, such as microphthalmia, microcornea, and/or iris anomalies. Syndromic cataracts are reported in association with a number of diseases or chromosomal aberrations, such as Nance-Horan syndrome (NHS), oculofaciocardiodental (OFCD) syndrome, Lowe syndrome, or trisomy 21.^8,9

The overall prevalence of congenital and infantile cataract is reported to range from 1 to 15 per 10 000 live births, depending on socioeconomic status and ethnicity.^3,8,10,11 Approximately one-quarter to one-third of congenital cataracts are caused by a genetic alteration,^12,13 with a predominance of autosomal dominant inheritance for the nonsyndromic cases.^2,3,12,14,15

At present, sequence variants in more than 100 genes are associated with congenital cataracts.¹¹ Disease-causing variants in more than 40 genes have been reported in patients with nonsyndromic cataract. These include variants in core genes encoding crystallins (CRYAA, CRYAB, CRYBA1, CRYBA2, CRYBA4, CRYBB1, CRYBB2, CRYBB3, CRYGA, CRYGB, CRYGC, CRYGD, and CRYGS), membrane transport proteins (GJA3, GJA8, and MIP), cytoskeletal proteins (BFSP2 and VIM), developmental regulators and transcription factors (PITX3, MAF, HSF4, FOXE3, PAX6, and EYA1), and transmembrane proteins (LIM2, TMEM114, CHMP-4B, and EPHA2).^11,16,17,18

To date, the reported detection yield of a diagnostic genetic variant in familial cataract cases is approximately 75%, with a higher pick-up rate in nonsyndromic than syndromic cases of 85% vs 63%, respectively, using panel-based next-generation sequencing.¹⁹ The mutation identification yield in sporadic cases ranges from 26% to 68%, suggesting that the causes of sporadic cases may include yet unknown genetic and nongenetic factors.²⁰

The high genetic heterogeneity of congenital cataracts is a challenge for establishing a reliable genotype-phenotype correlation in the clinical setting.²¹ However, knowledge of genotype-phenotype correlation can improve genetic counseling for patients and families.

The aim of this study was to describe novel and recurrent variants and the associated phenotype identified in a cohort of patients with nonsyndromic and syndromic congenital cataracts by whole exome sequencing (WES). We hypothesized that WES would have a higher diagnostic yield and impact compared with panel-based analysis for the diagnosis of congenital cataract.

Methods

Patients

Patients with—or with a history of—bilateral congenital cataracts (including patients previously reviewed²²) were recruited by the senior author (C.G.-K.). All patients and available family members had received a comprehensive eye examination. Data of cataract type³ at initial presentation or before surgery, associated ocular or extraocular anomalies, and time of surgery were extracted from the patient’s medical record and photographic documents when available. Best-corrected visual acuity and presence of aphakic glaucoma at last follow-up were documented. Patients were not offered any compensation or incentives for joining the study. The study was approved by the Cantonal Ethics Committee of Zurich, and written informed consent was obtained by all patients or their legal guardians. The study was conducted in accordance with the principles of the Declaration of Helsinki.²³

WES and Data Analysis

All data were collected from January 2018 to June 2020, and the molecular-genetic analyses were performed from January 2019 to July 2020. Protocols for next-generation sequencing and data analysis (including copy number variation analysis) have been published recently.²¹ A list containing 101 genes previously associated with congenital cataract is provided in eTable 1 in the Supplement. A gene list for syndactyly and/or zygodactyly (available on request) was used for WES data analysis in 1 patient. In case of missense variants, 6 different bioinformatic tools for estimating functional effects were applied: Align GVGD (Grantham Variation/Grantham Deviation),²⁴ SIFT (Sorting Intolerant From Tolerant),²⁵ MAPP (Multivariate Analysis of Protein Polymorphism),²⁶ MutationTaster2,²⁶ PolyPhen2,²⁷ and CADD (Combined Annotation Dependent Depletion) score.²⁶ Variants were considered if estimated to be disease causing by at least 2 of 6 algorithms. The American College of Medical Genetics and Genomics classification²⁸ was assigned to all variants for further interpretation of pathogenicity. The variants were labeled as novel if no previous description was found in the Human Gene Mutation Database at the time of manuscript submission.

Variant Confirmation and Segregation Analysis

Sanger sequencing was used for verification of variants as well as for segregation analysis according to the previous publication.²¹ For the amplification of GC-rich regions (present in CRYBB3 and MAF genes), high-fidelity DNA polymerase (Phusion; Thermo Fisher Scientific) and DNA polymerase (AmpliTaq Gold 360; Thermo Fisher Scientific) were used. Kinship analysis was performed using a polymerase chain reaction amplification kit (PowerPlex Fusion; Promega), according to the manufacturers’ instructions. The probability of paternity was calculated according to Essen-Möller,²⁹ using the allele frequencies of a Swiss population sample.³⁰ Nonpaternity was excluded in 5 of 8 cases of de novo occurrence of an identified variant. In case of the CRYBB2 gene, we used primers specific for regions only present on the targeted gene in question to avoid sequencing its pseudogene (CRYBB2P1). All variants were assessed by direct and indirect genetic evidence in cosegregation, co-occurrence of other variants, allele frequency, estimations,³ and literature research.

Functional Analysis by Minigene Assay of a Novel NHS Splice Site Variant and NHS Expression Analysis in Peripheral Blood

A minigene construct was cloned according to previous studies.^31,32 In brief, the rhodopsin gene sequence spanning exons 3 to 5 was cloned into pcDNA3.1 (Invitrogen) using EcoRI and XhoI restriction sites (primer sequences).³¹ Rhodopsin exon 4 was excised from the plasmid and replaced by NHS exon 3 (reference and variant c.[719-2A>G]) flanked by intron sequences (NM_001291867.1: c.[719-693] to c.[852 + 684]). Plasmid constructs were transfected into HEK293T cells followed by total cell RNA isolation. NHS RNA in whole blood was isolated using a blood RNA kit (PAXgene; BD Biosciences). RNA was extracted according to manufacturer protocol except for DNase I treatment, followed by a second elution with BR5 buffer and RNA denaturation. Complementary DNA from isolated RNA was generated as described previously.²¹

Results

Our cohort included 37 patients (18 [49%] male and 19 [51%] female; mean [SD] age, 17.3 [15.9] years) from 25 unrelated families, with known consanguinity in 1 family (pedigrees in eFigure 1 in the Supplement). More novel (13 of 20) than recurrent (7 of 20) putative diagnostic sequence variations were identified in 20 of 25 families (80%) (Table).^{11,33,34,35,36,37,38} The inheritance pattern varied among autosomal dominant (11 of 20), autosomal recessive (1 of 20), and X-linked (4 of 20). Cosegregation of the variants remained unknown owing to nonavailability of 1 parent for genetic testing in 4 families. De novo variants were identified in 8 families (40%). The gene spectrum with potentially diagnostic variants is illustrated in Figure 1: variants in crystalline (CRYBA2, CRYBB2, CRYBB3, CRYGC, and CRYGS) and connexin (GJA3 and GJA8) genes represent a larger proportion compared with variants in genes encoding cytoskeletal proteins (MIP), membrane associated signaling and transport proteins (COL4A1 and WFS1), transcription factors (MAF, PAX6 and BCOR), and proteins with an unknown function (NHS) in our cohort.

Table. DNA Variants Identified in Patients With Nonsyndromic and Syndromic Bilateral Congenital Cataract.

Patient	Gene (RefSeq ID)	Nucleotide change (zygosity)	Predicted amino acid change	Cataract descriptiona	Allele frequency	Predictions (in silico)							Segregation	First publication
						AGVGDb	SIFT	MAPP	MutationTaster2	PolyPhen2	CADD Score	ACMG
Nonsyndromic
Family 1
III:1	CRYBA2 (NM_057093.1)	c.[193G>A]; [193 = ] (heterozygous)	p.[(Gly65Arg)]; [Gly65 = ]	Total	NR	X	X	X	X	X	31	VUS	Paternal (unaffected)	Novel
III:2				Total
Family 2
III:1	CRYBB2 (NM_000496.2)	c.[526G>A]; [526G>A] (homozygous)	p.[(Asp176Asn)]; [Asp176Asn]	Nuclear plus sutural	NR	NA	X	X	X	NA	19.05	VUS	Biparental (unaffected)	Novel
III:2				Nuclear plus sutural
Family 3
II:1	CRYBB2 (NM_000496.2)	c.[458G>A]; [458 = ] (heterozygous)	p.[(Gly153Asp)]; [Gly153 = ]	Total	NR	X	X	X	X	X	25.2	LP	De novo	Novel
Family 4
II:1	CRYBB3 (NM_004076.4)	c.[466G>A]; [466 = ] (heterozygous)	p.[(Gly156Arg)]; [Gly156 = ]	Total (MC, MO)	6.9 × 10−6	X	X	NA	X	X	26.6	VUS	De novo	Li et al11
Family 5
II:1	CRYGC (NM_020989.3)	c.[320_321del]; [320_321 = ] (heterozygous)	p.[(Glu107Glyfs*56)]; [Glu107 = ]	Total (MC, MO)	NR	NA	NA	NA	NA	NA	NA	Pathogenic	De novo	Novel
Family 6
I:2	CRYGS (NM_017541.2)	c.[224G>A]; [224 = ] (heterozygous)	p.[(Gly75Asp)]; [Gly75 = ]	NA	NR	X	X	X	X	X	31	VUS	Paternal (affected)	Novel
II:1				Lamellar/kissing
III:2				Lamellar/kissing
Family 7
II:1	GJA3 (NM_021954.3)	c.[142G>A]; [142 = ] (heterozygous)	p.[(Glu48Lys)]; [Glu48 = ]	Total	NR	X	X	X	X	X	25.6	LP	De novo	Novel
Family 8
II:1	GJA3 (NM_021954.3)	c.[200A>G]; [200 = ] (heterozygous)	p.[(Asp67Gly)]; [Asp67 = ]	Nuclear	NR	X	X	X	X	X	25.4	LP	NA	Novel
				(pulverulent)
Family 9
I:2	GJA8 (NM_005267.4)	c.[773C>T]; [773 = ] (heterozygous)	p.[(Ser258Phe)]; [Ser258 = ]	NA	NR	NA	X	X	X	X	25.7	VUS	Maternal (affected)	Gao et al33
II:1				Pulverulent (restricted to nucleus)
Family 10
II:1	GJA8 (NM_005267.4)	c.[226C>T]; [226 = ] (heterozygous)	p.[(Arg76Cys)]; [Arg76 = ]	NA (MC, LE)	NR	X	X	X	X	X	32	LP	NA	Reis34
Family 11
I:1	MAF (NM_005360.4)	c.[905C>T]; [905 = ] (heterozygous)	p.[(Ala302Val)]; [Ala302 = ]	NA	NR	X	X	NA	X	X	32	LP	NA	Novelc
Family 12
II:2	MIP (NM_012064.3)	c.[97C>T]; [97 = ] (heterozygous)	p.[(Arg33Cys)]; [Arg33 = ]	NA	NR	NA	X	X	X	X	31	VUS	Maternal (affected)	Gu et al35
III:1				Total (MC)
Family 13
II:1	PAX6 (NM_001258462.1)	c.[52G>A]; [52 = ] (heterozygous)	p.[(Gly18Arg)]; [Gly18 = ]	Anterior polar (IA, MH)	NR	NA	X	X	X	X	25.9	Pathogenic	De novo	van Heyningen et al36
Family 14
II:1	WFS1 (NM_006005.3)	c.[1163T>G]; [1163 = ] (heterozygous)	p.[(Leu388Arg)]; [Leu388 = ]	Total	NR	X	X	X	X	X	28.1	VUS	De novo	Novel
Family 15
II:1	BCOR (NM_001123385.1)	c.[1221dup]; [1221 = ] (hemizygous)	p.[(Gly408Argfs*32)]; [Gly408 = ]	Nuclear	NR	NA	NA	NA	NA	NA	26.4	LP	De novo	Novel
Family 16
II:1	BCOR (NM_001123385.1)	c.[4512_4514delinsA]; [4512 = ] (hemizygous)	p.[(Ala1506*)]; [A1506 = ]	Total	NR	NA	NA	NA	NA	NA	NA	Pathogenic	Maternal	Hilton37
Family 17
II:1	COL4A1 (NM_001845.4)	c.[2951G>A]; [2951 = ] (heterozygous)	p.[(Gly984Glu)]; [Gly984 = ]	Total	NR	X	X	X	X	X	23.6	LP	NA	Novel
Family 18
II:1	COL4A1 (NM_001845.4)	c.[2987G>A]; [2987 = ] (heterozygous)	p.[(Gly996Asp)]; [Gly996 = ]	Total	NR	X	X	X	X	X	24.2	LP	De novo	Zagaglia38
Family 19
II:1	NHS (NM_001291867.1)	c.[1108 + 1del] (hemizygous)	p.? hemizygous	Total (MC, MO)	NR	NA	NA	NA	NA	NA	NA	Pathogenic	Maternal (affected)	Novel
Family 20
I:6	NHS (NM_001291867.1)	c.[719-2A>G]; [719-2 = ] (hemizygous)	p.? hemizygous	NA	NR	NA	NA	NA	NA	NA	23.4	Pathogenic	Maternal (either affected or unaffected)	Novel
II:1				Total (MC)
II:2				Total (MC)
II:3				Total

Abbreviations: ACMG, American College of Medical Genetics and Genomics; AGVGD, Align Grantham Variation/Grantham Deviation; CADD Score, Combined Annotation Dependent Depletion; IA, iris anomaly; LE, left eye; LP, likely pathogenic; MAPP, Multivariate Analysis of Protein Polymorphism; MC, microcornea; MH, macular hypoplasia; MO, microphthalmos; NA, not available; NR, not reported; SIFT, Sorting Intolerant From Tolerant; VUS, variant of uncertain significance; X, pathogenic predictions.

^aData in parentheses indicate ocular anomalies at diagnosis, both eyes.

^bGV measures the degree of biochemical variation among amino acids found at a given position in the multiple sequence alignment; GD reflects the biochemical distance of the mutant amino acid from the observed amino acid at a particular position.

^cVariant published in ClinVar RCV000533204.3.

Figure 1. Spectrum of Pathogenic Variants.

Relative proportion of 14 different genes in which disease-causing variants were identified using next-generation sequencing in our cohort.

^aIndicates syndromic pediatric cataract.

Variants were classified as disease causing (14 of 20 families [70%]; 8 novel and 6 recurrent variants) based on previous reports or fulfillment of different pathogenicity criteria.²⁶ Potentially disease-causing variants (or variants of uncertain clinical significance)^37,39,40 were identified in 6 of 20 families (30%), with 5 of them being novel. We did not classify these variants as disease causing owing to either (1) incomplete penetrance, (2) inconsistent or missing segregation analysis, or (3) allele frequency.

Genotype-Phenotype Correlation

Cataract morphology and severity in the nonsyndromic group varied from anterior polar cataract not requiring surgery (1 of 14 families) to vision-limiting nuclear and sutural or lamellar cataract (6 of 14 families) to total cataract with compulsory early surgery (5 of 14 families), with no information available about cataract morphology in 2 of 14 families (Table). Persistent fetal vasculature was not documented in association with the cataract formation in our patient cohort. Pathogenic variants in crystalline genes were mostly causative for total cataract formation. However, we identified a homozygous variant in the crystalline gene CRYBB2 (NM_000496.2: c.[526G>A]) in 2 affected siblings with peculiar nuclear-sutural cataract receiving cataract surgery at 4 and 6 years of age (family 2). Both parents, second-degree relatives of Indian descent, were heterozygous for this variant. They did not show any lens opacities on results of dilated eye examination. Another striking phenotype was apparent in affected members of 1 family with a lamellar cataract we also termed kissing cataract because the opacification appeared lip-shaped (Figure 2 and eFigure 3 in the Supplement).⁴¹ A putative disease-causing variant in the crystalline gene CRYGS (NM_017541.2: c.[224G>A]) was identified in the index patient, the affected father, and the paternal grandmother. The brother of the index patient also carried this variant but did not show any lens abnormalities on dilated eye examination results at 20 years of age.

Figure 2. Phenotype Spectrum.

A, Slitlamp photography demonstrating the cataract phenotype with pathogenic variants in the CRYBB2 gene in patient III:1 of family 2 (left eye: nuclear, partially sutural cataract). B, A pathogenic variant in the CRYGS gene is identified in patient III:2 of family 6 with a lamellar cataract and a distinct shape resembling a lip (left eye). C, The different cataract phenotype with a pathogenic variant in the GJA8 gene is shown in patient II:1 of family 9 (right eye with a pulverulent cataract restricted to the nucleus). Detailed phenotype images are available in eFigure 3 in the Supplement.

The 2 index patients in 2 families with disease-causing variants in the COL4A1 gene phenotypically showed total cataracts and intracerebral and/or intraventricular hemorrhages. Structural intracerebral changes in both patients are shown in eFigure 3 in the Supplement.

Syndromic cataracts were present in 6 of 25 families, 4 of them in association with NHS or OFCD syndrome, whereby all affected family members presented with dense nuclear or total cataract requiring early cataract surgery. The only exception was the index patient’s uncle with NHS from 1 family, who received cataract surgery at 14 years of age with resulting bilateral mild amblyopia (detailed data available in eTable 2 in the Supplement). The cataract diagnosis led to the diagnosis of the syndrome in all but the familial case (previously described³⁷). The 2 patients carrying disease-causing BCOR variants showed ocular, facial, cardiac, and/or skeletal features as summarized in eTable 3 in the Supplement. The index patient in 1 family additionally presented with a patent foramen ovale and a simian crease. Her maternal aunt (I:3) showed bilateral second-third toe syndactyly similar to the index patient, but no cataract or any other ocular or extraocular signs of OFCD syndrome. Neither the parents of the index patient nor the aunt harbored the identified BCOR variant. The analysis of the genes for syndactyly and zygodactyly did not reveal a potentially associated sequence variant in the index patient nor in the aunt (I:3). Ocular phenotype variability in OFCD was apparent in 1 family, with total congenital cataract without other ocular anomalies in the daughter and unilateral microphthalmia in association with bilateral congenital cataract in the mother.³⁷

All available male individuals in 2 families carrying the NHS variant were affected. One female carrier was clinically affected and presented with a bilateral sutural cataract (see the promotional Figure), whereas her sister, also a carrier of the variant, did not show any lens opacities on results of the dilated eye examination. This prompted us to perform functional analyses to study this variant in more detail. The putative disease-causing variant is located on the X chromosome, in the splice acceptor site of exon 3 of the NHS gene. We observed an increase in transcript size in both minigene assay and patient blood RNA if the variant was present (Figure 3). Transcript sequencing analysis revealed a retention of 83 bases from intron 2 of the NHS gene within the transcript, due to an alternative splice acceptor site that was inactive in the control.

Figure 3. Functional Analysis by Minigene Assay of a Novel NHS (Nance-Horan Syndrome) Splice Site Variant.

The NHS variant c.[719-2A>G] detected in family 20 shows a different splice product in a minigene assay. A, Schematic diagram of the transcript derived from the minigene plasmid construct. Major splice products obtained by polymerase chain reaction (PCR) (primers indicated) and analyzed by Sanger sequencing (eFigure 2 in the Supplement) are shown for the reference sequence and the NHS variant c.[719-2A>G]. B, Gel electrophoresis of amplified complementary DNA derived from minigene construct transfected HEK293T cells using primers indicated in part A. Two independent clone constructs were used for each separate transfection. C, Reverse transcriptase (RT)–PCR on RNA from whole blood obtained from family 20 (patient I:2) carrying the NHS variant c.[719-2A>G] in heterozygous state and a control sample (CS). PCR was performed using primers in exon 2 and 3 of the NHS gene. Expected PCR product size of the reference transcript is 122 base pairs (bp). D, Alamut Visual splicing window view of NHS variant c.[719-2A>G]. Different indicated estimation algorithms show a cryptic splicing acceptor site 83 nucleotides (83nt) upstream of NHS exon 3 (ex3). The red box indicates premature termination codon (PTC) of the NHS variant splice product if the reading frame from NHS exon 2 is continued. NTC indicates no template control; part, partial; and RHO, rhodopsin.

Discussion

Results from this genetic congenital cataract study from Switzerland emphasize the need for accessible genetic testing in this vulnerable patient cohort. Herein, we were able to show an overall detection yield for pathogenic variants of 80% for all families and 70% for families with nonsyndromic variants in a congenital cataract population with low consanguinity (consanguineous in 1 of 25 families; 4% in our study compared with the higher rate of ≤28% in a recent study from the UK).¹⁹ Variants in crystalline and connexin genes in nonsyndromic cases account for most of our recorded congenital cataracts, in agreement with previously reported results.^15,42,43 Most of our detected variants resulted in an amino acid substitution that does not permit functional predictions easily. Total cataracts were often associated with variants affecting the crystalline proteins in our cohort; however, nuclear sutural (family 2, individuals III:1 and III:2) or lamellar (family 6, individuals II:1 and III:2) cataract phenotypes were also associated with variants in crystalline genes as reported previously.¹⁷ Similarly, almost all syndromic cataract cases in our cohort showed total cataracts, which appears as the most common phenotypic description.^44,45 In contrast, the cataract phenotype among patients in our unresolved patient group consisted of primarily nuclear or pulverulent cataracts. Some genes causing nontotal cataract may be unknown, variants in known genes may not be detectable, or noncoding variants affecting expression of known genes may not be detected by exome sequencing. Incomplete penetrance was suggestive in 2 families with identified variants in CRYBA2 (family 1) and CRYGS (family 6), as already reported, for congenital cataract as a general observation, and specifically for variants in CRYBA2.³⁴

New Genotype-Phenotype Correlation

Nuclear and sutural cataract, as seen in the consanguineous family 2, have already been reported with CRYBB2 variants.^{11,17,18,19,46} However, the variant we identified clearly exhibited an autosomal recessive inheritance pattern, because only homozygous individuals were affected. Until now, 32 variants in the CRYBB2 gene have been described in the literature, all of them causing autosomal dominant cataracts (Human Gene Mutation Database; October 20, 2020). Our results suggest that variants in CRYBB2 can have an autosomal dominant or autosomal recessive inheritance pattern, similar to other crystalline cataract genes.^47,48 We now identified the genetic cause of the eye-catching lamellar cataract in family 6, the kissing cataract⁴¹ (or “Valentin’s cataract”; F. Munier, MD, oral communication, September 8, 2018), in the CRYGS gene (NM_017541.2:[c224G>A]). The 9 previously published CRYGS variants associated with congenital cataract showed a different lens phenotype, with pulverulent, nuclear lens opacities,⁴⁹ or lamellar cataract.⁵⁰

The identified variant in the PAX6 gene (NM_001258462.1: c.[52G>A]; family 13) showed anterior polar cataract associated with iris anomaly and macular hypoplasia, a similar but less severe phenotype than previously described.^36,51,52 This confirms the high phenotype variability associated with variants in the PAX6 gene.^53,54

Syndromic cataracts were identified in 4 of 25 of our families (16%) with either NHS or OFCD syndrome. Pathogenic variants in the COL4A1 gene are not yet clearly considered as a syndromic type of cataract, although there is an unquestionable association with intracerebral complications.³⁸ Thus, when we grouped patients with pathogenic variants in the COL4A1 gene into the syndromic category, more than 24% of our patients (6 of 25 families) showed a syndromic association. Late diagnoses, such as occurred in patient II:1 of family 15 at 7 years of age, are alarming because OFCD syndrome can be associated with potentially life-threatening conditions, such as intestinal malrotation. This possibility reinforces an urgent call for availability of genetic testing in children with congenital cataracts, especially to exclude syndromic associations. Two disease-causing variants in the BCL6 corepressor, encoding the key transcriptional regulator in early embryogenesis,⁵⁵ were identified in families 15 and 16 with manifestation of the X-linked OFCD syndrome. Most features we observed in our patients are already reported to be associated with the disease,⁵⁶ whereas patent foramen ovale and simian crease have not yet been described. Because patent foramen ovale is present in 20% to 25% of adults,⁵⁷ it does not have to be associated with the BCOR variant in question. However, any cardiac sign in association with bilateral congenital total cataract should raise the likelihood of an underlying syndromic type of cataract. The phenotype in patient II:1 of family 15 with central hypothyroidism supports the suggestion by Ragge et al⁵⁸ of pituitary underdevelopment being an additional sign of OFCD syndrome. Furthermore, in family 15, syndactyly was observed in a family member (I:3) not carrying the BCOR variant. No variants (or copy number variations) in syndactyly genes were identified in the index patient (II:1) with the BCOR variant nor in the patient’s aunt (I:3). Thus, patient II:1 could have the foot anomaly as part of OFCD syndrome that shows syndactyly or camptodactyly as a common feature.^59,60,61 However, the genetic cause for the syndactyly in patient I:3 remains unknown.

Nance-Horan syndrome is the second common syndromic association identified in our cohort. In family 20, 1 female carrier of the NHS (NM_001291867.1: [c719-2A>G]) variant was affected, whereas her sister did not show any lens opacities. Variable expressivity by skewed X-chromosome inactivation, which coincides with the known phenotype variability in X-linked diseases, may explain this intrafamilial variability. The estimations for the variant suggested loss of the acceptor site of exon 3, which could lead to activation of an alternative acceptor site upstream of exon 3 and a transcript containing 83 additional nucleotides. This consequently leads to an estimated frameshift in the open reading frame of the gene and a premature stop codon upstream of exon 3 (Figure 3A). The functional experiment we performed confirmed an effect on splicing and indicated partial NHS intron 2 retention. We were able to show the effect of this variant in vitro (minigene assay) and in vivo (peripheral blood). Several splice site variants in the NHS gene have previously been reported,^61,62,63,64 including 1 assessment using reverse transcriptase polymerase chain reaction.⁶⁴ Our functional splicing analysis and the analysis performed in a previous report³⁵ were both performed on non–eye-derived cells. However, if normal splicing is affected in vivo and in vitro owing to this single nucleotide variant, the probability is very high that this would also be the case in the lenticular cells.

The potentially pathogenic variants found in our cohort are present in a total of 14 genes consisting of 123 exons and spanning 31.7 kb of DNA. In addition, more than 100 genes have been screened in this cohort for pathogenic variants. These numbers justify WES as the appropriate and first-line method for genetic testing in patients with congenital cataracts. Once new genes are identified, repeated analyses would be possible using the available WES data. In contrast, whole genome sequencing may not be advantageous as a first-line analysis as suggested by recent data from the UK Genomics England WGS (Whole Genome Sequencing) study.⁶⁵

Limitations

This study has some limitations. Specifically, no functional analysis was performed of the identified variants in this study.

Conclusions

We suggest WES as a preferred first-line molecular genetic approach to analyze congenital cataracts and detect novel, recurrent, and de novo variants. This cohort study demonstrated a very high detection yield, with a high number of de novo variants. Therefore, we recommend WES trio analysis in isolated cases of nonconsanguineous families to limit the volume of data and to further improve efficiency of genetic testing. Confirmed genetic diagnoses will permit counseling of families with regard to recurrence risks and, in cases of additional extraocular signs, consideration of syndromic types of congenital cataract. Both inheritance patterns (autosomal dominant and autosomal recessive) need to be considered for variants in the same gene.

Supplement.

eTable 1. Gene List for Congenital Cataract

eTable 2. Phenotype Description in Patients With Nonsyndromic and Syndromic Bilateral Congenital Cataract

eTable 3. Phenotypic Features in Patients With Variants in the BCOR Gene

eFigure 1. Pedigrees and Sequencing Chromatograms

eFigure 2. NHS cDNA Sequencing Chromatograms

eFigure 3. Phenotype Spectrum

eReferences.