Whole exome sequencing revealed novel pathogenic variants in Vietnamese patients with FEVR

Subjects

This study recruited pediatric patients diagnosed with FEVR at the National Eye Hospital (VNEH) in Hanoi, Vietnam, and their family members. The patients underwent a clinical ocular examination, including fundoscopy (ophthalmoscopy) or fluorescein angiography. For patients under one year old, only fundoscopy was used to detect retinal vascular anomalies. Patients were diagnosed with FEVR once they met the clinical criteria, which were composed of three standards: the appearance of incomplete vascularisation of the peripheral retina and poor vascular differentiation (marked by peripheral retinal avascularity) in one eye or both eyes in ophthalmoscopy or fluorescein angiography; a full-term birth or a pre-term birth with retinopathy of prematurity (ROP)-unrelated disease progression; and variable levels of vitreoretinal traction, falciform, subretinal exudation, neovascularization, and retinal detachment. This study was approved by the Ethics Committee in Biomedical Research of the Institute of Genome Research, Vietnam Academy of Science and Technology. Written informed consent for the genetic testing was obtained from the parents or guardians of the patients. The peripheral blood samples (2 ml) of patients and their available family members were collected into ethylenediaminetetraacetic acid (EDTA)-containing tubes and stored at –20 °C.

DNA extraction

Genomic DNA was isolated from peripheral blood using the Exgene Blood SV Mini Kit (GeneAll Biotechnology Co. LTD, Seoul, Korea) following the manufacturer’s formal protocol. The quality of the DNA sample was assessed using 0.8% agarose gel electrophoresis and spectrophotometry at 260 and 280 nm (BioSpectrometer basic; Eppendorf, Hamburg, Germany).

Whole exome sequencing

The DNA library was prepared using SureSelect V6-Post (Agilent Technologies, Santa Clara, CA) according to the manufacturer’s guidelines. Quantification of the enriched library was conducted using the Qubit dsDNA HS Assay Kit (Thermo Fisher Scientific, Waltham, MA). A bioanalyzer using a High Sensitivity DNA Chip (Agilent Technologies) was employed to check the library size distribution, with an expected size range from 200 bp to 400 bp. Paired-end sequencing was performed on an Illumina NovaSeq 6000 platform (Illumina, San Diego, CA) with paired reads of 150 bp. The mean exome coverage was more than 100X, and each target base had at least 20X coverage.

Data analysis and pathogenicity evaluation

The sequence reads were mapped to the hg19/GRCh37 human reference genome using the Burrows–Wheeler Aligner tool (BWA.v0.7.12). Picards were used to mark duplicates. GATK and Samtools were used to detect single nucleotide variants (SNVs) and short indels. To exclude false positives, all variants with depth reads lower than 10X were removed. Short indels in the repeat regions and within a 10 bp range from the start and end of the read were also removed. The remaining variants were then filtered from the 1000 Genome database (1000G) and the Exome Aggregation Consortium (ExAC). The variants were annotated using the ANNOVAR program.

To identify the potential causative variants, we first screened the rare and novel variants in six known FEVR genes (FDZ4, LRP5, NDP, TSPAN12, ZNF408, and CTNNB1), followed by 231 retinal dystrophies-associated genes. Novel or rare missense variants that were not reported previously were analyzed using in silico tools, such as Sorting Intolerant From Tolerant (SIFT), PolyPhen-2, and MutationTaster, to anticipate the damaging effect on corresponding proteins. Amino acid conservation was checked using the MULTIZ alignment (UCSC). Furthermore, each pathogenic or likely pathogenic variant was examined for segregation in the family members. The pathogenicity of the variants was evaluated using the guidelines established by the American College of Medical Genetics and Genomics and the Association for Molecular Pathology (ACMG/AMP). The presence of candidate causative variants was further checked in the Human Gene Mutation Database (HGMD), LOVD Database, Genome Aggregation Database (gnomAD), and an in-house WES database of 535 Vietnamese without any FEVR phenotypes (VN WES DB).

Sanger sequencing validation

Specific primers were provided by PHUSA Biochem Company (Can Tho, Vietnam; Appendix 1). Polymerase chain reaction (PCR) was performed using a total reaction volume of 20 µl that included 10 ng genomic DNA, 1X NEB master mix (New England Biolabs, Ipswich, MA), 10 pmol of each primer, 1 µl DMSO (Bioworld, Dublin, Ohio (OH; 43017), and 16.5 µl deionized water. The thermal cycle comprised denaturation at 95°C for 5 min, followed by 40 cycles of 95°C for 15 s, 58°C for 30 s, 68°C for 20 s, and a final extension at 68°C for 5 min. The PCR products were then purified using a MultiScreen-PCR 96 Filter Plate (Merck-Millipore, Burlington, MA) and subsequently bi-direction sequenced using an ABI Prism BigDye Terminator Cycle Sequencing Kit Version 3.1 (Applied BioSystem, Waltham, MA) on an ABI genetic analyzer 3500 (Applied Biosystems, Waltham, MA).

Multiplex ligation-dependent probe amplification

MLPA was performed according to the manufacturer’s standard protocol using the SALSA MLPA Probemix P285-LRP5 (MRC-Holland, Amsterdam, the Netherlands). This assay can detect deletions/duplications in the FZD4, LRP5, NDP, and DKK1 genes. The data were analyzed using Coffalyser software (MRC Holland). Genomic DNA samples (50 ng) were denatured at 98°C for 5 min. Hybridization was performed at 60°C for 16 h, followed by ligation at 54°C for 15 min and amplification of the ligated probes. Capillary electrophoresis of the PCR products was performed using a 3500 Genetic Analyzer (Applied Biosystem, Foster City, CA). Coffalyser.Net software was used to analyze the data; relative probe values below 0.7 and above 1.3 indicated the borders for loss and gain, respectively.

Clinical features of patients

A total of 20 pediatric patients from unrelated families were enrolled in this study (Appendix 2). Their median age at diagnosis was 4 months (ranging from 14 days to 11 months), and the ratio of males to females was 15:5. All were born at full term and had no history of oxygen therapy treatment. Fluorescein angiography or the indirect ophthalmoscope were used to detect retinal vascular anomalies in 100% of the patients, while strabismus was found in 95% (19/20). Retinal detachment, exudation, nystagmus, and asymmetry were found in 45%, 35%, 30%, and 20% of the patients, respectively. Other ocular characteristics, such as cataract, microphthalmia, band keratopathy, corneal clouding, posterior synechiae, megalocornea, retinal fibrosis, retinal degeneration, and proliferative vitreoretinopathy, were identified in several patients. Notably, four patients showed signs of microcephaly, and one had hydrocephalus.

Four of the 20 patients had a family history in which at least one family member had visual problems. Patient EVR-06 had an elder brother with highly similar eye conditions. For patient EVR-08, their maternal grandfather and uncle had early-onset glaucoma and primary myopia, respectively. The maternal grandfather of patient EVR-15 had nystagmus. In family F20, 11 males in four generations of his maternal family had severe vision loss. Both eyes of patient EVR-20 presented with a shallow anterior chamber, posterior synechiae, and total cataract. The intraocular pressure (IOP) values were 19 mmHg OD and 13 mmHg OS, which indicated a difference of 6 mmHg in IOP between the two eyes. Moreover, the right eye had megalocornea and corneal clouding. Ultrasound of the right eye showed the density of vitreous opacities distributed almost throughout the entire vitreous body, and it was difficult to distinguish posterior membranes. Meanwhile, the left eye showed a thick membrane attached to the optic disc, and the funnel configuration of retinal detachment in this eye was accompanied by end-stage retinal detachment. Due to his severe ocular phenotypes and family history, patient EVR-20 was suspected of having Norrie disease (Figure 1).

Pathogenic variants identified by whole-exome and Sanger sequencing

Genomic DNA of 20 probands was first used for WES analysis, and then novel and rare variants found in genes responsible for FEVR were evaluated for pathogenicity. Eleven probands were identified as carrying pathogenic or likely pathogenic variants, accounting for 55% of all subjects (Table 1). There were nine heterozygous or hemizygous variants in FZD4 (3 patients), NDP (3 patients), and KIF11 (3 patients), as well as 1 homozygous variant in ATOH7 (2 patients). The mutational types included 2 frameshift, 3 nonsense, 4 missense, and 2 splicing variants. Four of the 11 variants were novel and were not found in the in-house control group of the VN WES DB.

For the FZD4 gene (Appendix 3), two known pathogenic variants c.1282_1285del p.(D428Sfs*2) and c.313A>G, p.(M105V), were found in the probands EVR-05 and EVR-14, respectively. A novel variant c.169G>C, which resulted in an amino acid substitution p.(G57R), was identified in the proband of patient EVR-18. This variant was located in a highly conserved region of the FZD4 protein among species and was predicted to be a damaging variant by the SIFT and PolyPhen-2 tools (Figure 2). It was a de novo variant due to its absence in the proband’s parents. According to ACMG/AMP, it should be a moderate pathogenic variant.

Regarding the NDP gene located in the X chromosome (Appendix 3), three variants were detected in three male probands in the hemizygous form. Two missense variants—c.112C>T, p.(R38C) in the proband of patient EVR-10 and c.131A>G, p.(Y44C) in the proband of patient EVR-15—were previously reported as pathogenic. A novel variant c.175–3A>G in the proband of patient EVR-19 was predicted to cause a splice site alteration. The unaffected mother of EVR-19 carried a heterozygous variant of NDP c.175–3A>G.

For the KIF11 gene (Appendix 3), a previously reported splicing mutation c.388–1G>C was found in patient EVR-08. Two novel variants c.2146C>T, p.(Q716*) and c.2511_2515del, p.(N838Kfs*17) were identified in patients EVR-07 and EVR-11, respectively. Both are de novo and null variants and can be classified as pathogenic variants.

For the ATOH7 gene (Appendix 3), the pathogenic variant c.145G>T, p.(E49*) was detected in the two probands patients EVR-06 and EVR-16 in the homozygous state. According to the OMIM database, the ATOH7 gene is responsible for the persistent hyperplastic primary vitreous condition that follows the autosomal recessive model. Sanger sequencing identified a heterozygous variant c.145G>T, p.(E49*) in both unaffected parents of EVR-06. Unfortunately, the samples of EVR-16’s parents were not available for genetic testing.

Pathogenic deletion variant identified by multiplex ligation-dependent probe amplification

Since nine of the 20 patients had not been identified as having causative variants by WES, their DNA samples were used for MLPA analysis to determine whether they had a gross deletion/insertion in the FZD4, LRP5, and NDP genes. The results showed that only patient EVR-20 had a hemizygous deletion of exon 2 of the NDP (Figure 3B). The genetic examination of his family members showed that his affected brother also had a hemizygous NDP exon 2 deletion, while his unaffected mother and sister carried a heterozygous deletion (Figure 3B). The validation of this variant by the PCR method indicated a complete absence of NDP exon 2 in the proband EVR-20 and his affected brother (Figure 3C).

Spectrum of pathogenic variants and their correlation with clinical phenotypes

In the current study, 12 pathogenic variants were identified in four genes, including FZD4, NDP, KIF11, and ATOH7. Among them, NDP variants made up the largest proportion (20%), followed closely by FZD4 (15%), KIF11 (15%), and ATOH7 (10%; Figure 4). No significant variants were detected in the remaining FEVR-causing genes (LRP5, TSPAN12, ZNF408, and CTNNB1). Unfortunately, there were still 8 probands in whom genetic causes were not detected.

All 12 patients positive for variants had retinal vascular anomalies, and 11 presented with strabismus. Approximately 33% (1/3) of patients with a FZD4 variant /a NDP variant and 50% (1/2) with an ATOH7 variant had exudation. Retinal detachment was detected in about 33% (1/3) of patients with FZD4 variants, 50% (2/4) with NDP, 33% (1/3) with KIF11, and 50% (1/2) with ATOH7. Nystagmus and cataracts were not present in the patients with ATOH7 variants. Asymmetry was found only in patients with the NDP variant, with 50% of them showing this feature. All three patients carrying the KIF11 variants presented with signs of microcephaly, and one patient even had microphthalmia in the left eye (Table 2, Figure 5).

Variable expressivity of ocular and extraocular phenotypes

Apart from eye abnormalities, Hu et al. highlighted a specific correlation between genotype and systemic phenotype for KIF11, in which patients with KIF11 variants had a considerably higher possibility of microcephaly, lymphedema, and intellectual disability [22]. Likewise, we found that all three patients harboring the KIF11 variants in this study displayed microcephaly, suggesting that this feature might be particular to patients with the KIF11 variant (Figure 5).

Regarding the patient with Norrie disease, it is noteworthy that 11 out of 19 males in his family presented with ocular manifestations. This is the first time we have found such a large number of patients in a Vietnamese family with Norrie disease. It is apparent that copy number analysis and direct sequencing of the NDP gene may prove highly helpful in quickly identifying the genetic causes of NDP-related retinopathies, particularly Norrie disease.

Even though genetic tests indicated the same NDP exon 2 deletion as in previous studies, the clinical phenotypes were widely variable [23,24]. The proband in our study exhibited multiple ophthalmic abnormalities, including megalocornea, corneal clouding in the right eye, total cataract, posterior synechiae, and total tractional retinal detachment when he was just 23 days old, which was the time of detection (Figure 1). Notably, patients with Norrie disease in previous studies often showed microphthalmos [25,26], whereas our patient’s right eye exhibited megalocornea. We assumed that his right eye had possibly developed secondary glaucoma due to pupillary block and a close anterior chamber angle. In addition, he was likely to have intellectual disabilities. At present, he is nearly 20 months old and still shows no symptoms of hearing loss; however, it remains unknown whether a future event will occur.

Intrafamilial variable expressivity

Clear evidence of intrafamilial variability and incomplete penetrance was observed in this study. A good illustration of this evidence is the intrafamilial occurrence of non-syndromic to syndromic manifestation in family F8 with the KIF11 variant. Notwithstanding being heterozygous for the same KIF11 c.388–1G>C variant, patient EVR-08 exhibited a phenotype with retinal detachment, while his mother did not show any symptoms. Notably, his uncle and grandfather also had several subtle ocular conditions; however, their DNA was not available for genetic testing. Another remarkable case was family F14 with the pathogenic variant FZD4 c.313A>G, p.(M105V), in which the proband showed a range of eye conditions (retinal vascular anomalies, strabismus, nystagmus, and cataract), while the carrier’s father’s eyes were completely normal. The pattern was entirely different from previous observations, with all family members positive for such variants being affected [27,28].

Conclusion

In summary, this study was successful in using comprehensive WES analysis together with other methods to identify variants in the FZD4, NDP, KIF11, and ATOH7 genes that caused FEVR phenotypes in Vietnamese patients. Our findings provide useful information for the clinical management and genetic counseling of this disease, especially in asymptomatic cases.