A family of fuchs endothelial corneal dystrophy and anterior polar cataract with an analysis of whole exome sequencing

ABSTRACT Purpose Our aim was to introduce a family affected by this rare phenotype, and perform the whole exome sequencing (WES) to explore the potential candidate genes causing the disorders. Methods A five-generation family including five patients affected by FECD with APC, and nine patients suffered from only FECD was recruited from the First Affiliated Hospital of Harbin Medical University. All participants received ophthalmic examinations. Eight family members were selected to perform WES with a bioinformatics analysis and genome-wide linkage analysis. The candidate genes were identified by polymerase chain reaction (PCR) and Sanger sequencing. Results Patients in this family had FECD as the common feature. The proband (a 65-year-old female) was affected by FECD and APC in both eyes, with epithelial bullae in the left eye. Slit-lamp, specular, and confocal microscope and OCT images showed guttae more serious in the central cornea than the peripheral area, confirming the diagnosis of FECD. In this family, most corneal guttae was bilateral with an almost equal degree of progression in the Descemet membrane, APC was found around the age of 10, perhaps even earlier. According to the analysis of bioinformatics analysis, two candidate genes were found and confirmed by PCR and Sanger sequencing, but could not achieve genotype-phenotype co-segregation in the family. Conclusion We introduced a family of FECD with APC, with no known causative gene found by WES, inferring that there may be a novel gene-locus in the non-coding regions of genome, which needs further study by WGS. The contribution of this study was to exclude the possibility of the rare phenotype pathogenic site in exome and narrow the scope of pathogenic genes.

Age and gender are important factors influencing the development of FECD. People over 40 and female have a higher risk, with a female-to-male ratio of 2.5-3:1 (1,5,6). FECD displays in an autosomal dominant inheritance with incomplete penetrance, about 50% of the patients have a positive family history (6,9). Clinically, FECD can be divided into early-onset FECD and late-onset FECD. Early-onset FECD, which began in the first decade of life, shows similar progress to classic phenotypes. The initial clinical manifestation of late-onset form of FECD (corneal guttae) usually occurs in the fourth decade of life (1,6,10). Usually, patients do not need intervention until the sixth or seventh decades (1,6).
Corneal guttea with anterior polar cataract is a rare phenotype which is inherited in an autosomal dominant pattern, initially proposed by Ichikawa and Hiraga in 1951 (12,13). Chen P et al. have reported that mutations in TMCO3 gene were related to this rare phenotype in 2016 (12). In view of FECD may occur independently or in association with other ocular or systemic abnormalities (6,7,12), it remains to be clarified whether the phenotype of cornea guttata with anterior polar cataract is caused by FECD gene alleles or closely linked modifiers. With the breakthrough of next-generation sequencing (NGS) technology, whole exome sequencing (WES) analysis has been applied in the detection of variants in exons (proteincoding regions) and splicing sites in the human genome (14). It is estimated that the exome contains about 85% mutations, which has a great influence on the diseaserelated traits (15,16). In order to find likely causal variant, WES analyses were applied in a family affected by FECD with anterior polar cataract in Hei Longjiang Province, China.

Subjects
All study subjects were recruited from the First Affiliated Hospital of Harbin Medical University. A five-generation family with 33 members (14 affected; Figure 1a) was enrolled. All participants received detailed examinations by ophthalmologists, including vision, slit-lamp microscope, intraocular pressure measurement. Specular microscope, OCT examination, and confocal microscope were performed on selected subjects. There was no other systematic abnormal family history in this family. This study is based on the declaration of Helsinki and approved by the ethics committee of the First Affiliated Hospital of Harbin Medical University. All participants (or their guardians) received written informed consent.

Genomic DNA preparation
Peripheral blood from family members was collected. 5 ml peripheral blood was drawn from the elbow vein of each subjects and was preserved at −80°C prior to use. Genomic DNA was extracted from peripheral leukocytes using QIAamp DNA Blood Mini Kit (Qiagen, Hilden, Germany) according to the manufacturer's protocol. A NanoDrop ND-2000 spectrophotometer was used to DNA quantification (analyzed by agarose gel electrophoresis). DNA was stored at −20°C for subsequent analysis.

Whole exome sequencing and library construction
Exome sequencing was applied on eight patients (II-7, II-9, II-I0, III-7,III-8, III-11 III-12, and IV-4) at CapitalBio Technology Co. Ltd., Beijing, China. Approximate 1 μg of genomic DNA sample was sheared into fragments of 300-500 bp in length. The sheared fragments were bluntend repaired and a single adenine base was added to the 3′ ends using Klenow exonuclease. Illumina adapters were ligated to the repaired ends and DNA fragments were PCR amplified for 8 cycles to each sample. Agilent SureSelect Human All Exon V6 kit (Agilent Technologies, Canada) was used for whole exome capture and library construction. Captured libraries were then sequenced on the Illumina HiSeq X-Ten PE150 (Illumina Inc.,USA).

WES data analysis
Sequencing data were analyzed with NextGene V2.3.4 software (Softgenetics, State College, PA), and the raw reads were then aligned to hg19 genome using Burrows-Wheeler Aligner (BWA) bwa-0.7.15 (17), followed by variant calling using ANNOVAR (18) to annotate the detected SNV and indels. Nonpathogenic polymorphisms were filtered comparing with the 1000 Genomes Project database, ESP-6500, the Exome Aggregation Consortium database (ExAC), and the Single Nucleotide Polymorphism database (dbSNP). Variants were assessed for the potential deleteriousness as determined by insilico analysis included in dbNSFP (19), such as Polyphen2 and SIFT. Human Gene Mutation Database (HGMD) was used to search for known pathogenic mutations. WES data were filtered according to the following strategies: (1) The variation of total reading depth > 5 X and SNP quality score > 50; (2)Considering that the phenotype of corneal guttata with anterior polar cataract was rare and was inherited in an autosomal dominant fashion, we assumed that the variants were also rare. Minor allele frequency <0.1% in all the four variant databases(1000 Genomes Project, ESP-6500, ExAC, and dbSNP) for autosomal genes; (3) variants were SNVs (stoploss, stopgain, non-synonymous, non-frameshift substitution, non-frameshift insertion, frameshift insertion, nonframeshift deletion, frameshift deletion) or splice-site variants (splicing within 7-bp of a splicing joint); (4) variants were consistent with the pattern of inheritance models (autosomal dominant fashion, heterozygous);(5) variants were assessed for the potential deleteriousness as determined by 12 insilico prediction scores included in dbNSFP. If the variant wasn't predicted damaging by 10/12 of the scores, the variation may be considered benign (19).

Linkage analysis to identify causal regions
Linkage analysis is to analyze the SNP and small indel genotyping data of each WES sample. On the basis of WES data, Merlin 1.1.2 was used to perform genome-wide linkage analysis (Multipoint parametric and nonparametric linkage analysis) in this family, exploring genetic variations related to family clustering disease or phenotype. We genotyped eight individuals who went through WES. For the parameter linkage analysis, it was assumed that the disease in this family was an autosomal dominant inheritance and the penetrance was 100%, and the frequency of disease allele was 0.0001. A LOD score more than 3.0 was considered evidence of linkage.

Validation variants
Rare heterozygous variants shared by all six patients were retained and then validated in the whole family members by direct PCR sequencing to determine genotype-phenotype cosegregation. The primer sequences were shown as follows: INTS1 forward primer 5ʹ-GCTTCTGTAACGGGTGCCT-3′ and reverse primer 5ʹ-TCACCAGAAATCTCCCAGCG −3′. SH3GL2 forward primer 5ʹ-CAGCAACTTCCAAAGGTCCG −3′ and reverse primer 5ʹ-TGCAGCATCCCCTCATACCA −3′. PCR was performed under the condition as previously published (20). After purification, amplicons were sequenced on an ABI 3730 XL Genetic Analyzer (ABI, Foster City, California, USA). The sequences were assembled and analyzed using Lasergene SeqMan software (DNASTAR, Madison, WI) and were compared to reference sequences from Ensembl.

Clinical findings
Patients of corneal guttae with anterior polar cataract Proband II-7 is a 65-year-old female who complained of binocular-red eyes with pain, and vision decline in left eye for 2 years, with visual acuity of 0.12 Oculus Dexter (OD) and CF/20 cm Oculus Sinister (OS). She has poor eyesight since childhood. Slit-lamp examination revealed guttae and pigment in both posterior corneas with anterior polar cataract in both eyes. Epithelial bullae in the left eye. The changes in the central cornea is more serious than the peripheral area ( Figure 1b).FECD grade was assessed according to Krachmer et al. (6,9). Specular microscope demonstrated that a large number of vacuolated pathological black areas can be seen in corneal endothelium. The corneal endothelial cells were uneven in morphology, enlarged and pleomorphic, and the density of corneal endothelial cells was significantly reduced (Figure 1c). OCT showed the corneal epithelium of the left eye was edematous, and the corneal endothelium of both eyes was not smooth (Figure 1 D(a-b)). Confocal microscope showed a lot of high reflective, confluent guttae between corneal endothelial cells, and the structure of endothelial cells was unclear. (Figure 1D (c-d)). Her daughter III-8 is a 40-year-old female with guttata and anterior polar cataract phenotype in both eyes who complained of vision decline for 30 years, with visual acuity of 0.25 OD and 0.3 OS. The proband's granddaughter IV-4 is a 16-year-old female who found a little white dot in the center of the pupil in both eyes around the age of 10, with visual acuity of 0.4 OD and 0.3 OS. She was also found corneal guttae in both eyes. The clinical features were similar with the proband's son III-9, and granddaughter IV-5. APC was also found at around 10 years old. The eyesight of proband's mother I-1 has been poor while the proband's father had good eyesight during the lifetime according to the families, indicating that at least FECD may be inherited by I-1.

Patients of corneal guttae with no anterior polar cataract
The proband's sister II-9 and her daughter III-11 and son III-12 had corneal guttae phenotype with no anterior polar cataract (Figure 1e,f), the same as II-11, II-12, II-13, and III-5. (See Table 1). Patients in this family had FECD as the common feature. Most corneal guttae was bilateral with an almost equal degree of progression in Descemet membrane, and the central corneal injury was more serious than the peripheral endothelial injury. The degree of injury varies from person to person. Only the sub-pedigree of the proband had anterior polar cataract at the same time. Anterior polar cataract was found around the age of 10, perhaps even earlier in this family. There was no family history of other systemic abnormalities. No obvious abnormality was found in an ophthalmic examination of other members. Fundus examination showed no abnormalities in all subjects.

Heredity findings
To identify the causative mutation underlying this phenotype, we initially performed whole exome sequencing on eight subjects (II-7, II-9, II-I0, III-7,III-8, III-11 III-12, and IV-4) in this family with high quality (mean coverage: 148.56×; regions with >30× coverage: 97.84%). According to the genetic pattern map of the family, it can be inferred that the phenotype may be an autosomal dominant inheritance. The family can be further divided into two sub-pedigrees, subpe-digree1, and subpediree2 ( Figure 1g). The samples actually sequenced in the family subpedigree1 are II-7, III-8, IV-4 with both FECD and APC, and their normal control III-7. The samples actually sequenced in the family subpedigree 2 are II-9, III-11, III-12 with only FECD and their normal control II-10.
Since FECD may occur independently or in association with other ocular or systemic abnormalities (6,7,12), and only sub-pedigree 1 had two abnormalities, while FECD only occurred in the rest of their whole family. Patients in this family had FECD as the common feature. Besides, there was no APC with normal cornea subject in this family. It is speculated that the mutation of genes sharing by both sub-pedigree 1 and 2 may cause FECD and the formation of APC at the same time. According to disease type and sub-pedigree classification, the genetic model analysis was carried out according to the Table 2. Statistical results of autosomal dominant genetic pattern variations were shown in Table 3. The selected SNVs and small indels were used to extract the genes of the mutation that affected the amino acid sequence, including stoploss, stopgain, nonsynonymous, non-frameshift substitution, non-frameshift insertion, non-frameshift deletion, frameshift insertion, frameshift deletion) or splice site variants, which were shown in Table 4.
Based on the above analysis, it was speculated that this rare phenotype was caused by the common variation shared by two sub-pedigrees. The intersection of 01, 02, and 03 was the key genes, which had 83 genes in total in this family. WES data were shown on Supplementary data. No interval with LOD greater than 2 was found in genome-wide linkage analysis (Multipoint parametric or nonparametric linkage analysis) (Supplementary data).After further data filtering, 2 rare genes (2 variants) were shared in six patients ( Table 5). The remaining members of the family (II-1, II-11, II-12, II-13, III-1, III-2, III-3, III-5, III-6, IV-1, IV-2, IV-3, IV-5, V-1) went through Sanger sequencing after PCR. However, it was not found that the variations were co-segregated with the rare clinical phenotype. The sequence chromatograms were shown in Figure 1h. According to the results, no TMCO3 (Chen P et al.) or other known genes leading to FECD were found, indicating that there may be a new gene or locus leading to this rare phenotype in the non-coding sequence of the genome.

Discussion
The phenotype of corneal guttae with anterior polar cataract (OMIM: 121390) is rare. Dohlman described a Swedish family in which 15 members had this rare phenotype in 1951 (13). He demonstrated that corneal changes were limited to the posterior cornea, i.e. endothelium and Descemet membrane, and no change in other layers of the cornea. In his study, the central cornea was more affected than the periphery in all patients, which was consistent with our patient's phenotype since in II-7 (OS), III-11(OS), guttae concentrated in the central area, which was too serious to be captured by specular microscope. Dohlman also reported that at least one symptom was observed in the other three cases. It appeared that some relatives of the deceased may had at least polar cataracts. Guttae cornea did not seem to exist at birth, but later appeared and progressed slowly, resulting in a "beaten metal" appearance in the reillumination, and the changes of cornea were mostly bilateral with almost the same progress in each eye, but the severity of changes varied with patients. Polar cataracts were not always evident at birth. It occured most often between the ages of 3 and 10 becoming motionless after puberty (13), which was matching with the features of II-7, III-8, IV-4, III-9,IV-5 in our study, who found a little white dot in both eyes since childhood, at the age around 10. Besides, with the increase of age, the more serious the corneal injury, the more likely it is to cause corneal edema, leading to vision loss, which was similar with the proband II-7 in our study. However, there was no patient with a single anterior polar cataract in our study. In young people, polar cataract is the main cause of visual impairment, but with the increase of age, corneal exacerbation is the main cause of visual impairment (13). This was also identified with our family characteristics. Dohlaman inferred that the two ocular abnormalities were caused by the interference of the posterior surface of the cornea and the anterior part of the lens due to the formation  Sub-Pedigree with FECD and sub-Pedigree with both FECD and APC were analyzed together (but only FECD was considered) (1)The first column is the coding of three cases of genetic pattern analysis (2)The second column is the subpedigree involved in the corresponding genetic model analysis   InDel  Total  Gene   01  1,287  71  1358  1001  02  932  43  975  705  03  97  4  101  85 (1)Type:No. 01, 02, 03 (2)SNV: The number of SNVs that conform to the autosomal dominant genetic pattern and have influence on amino acid sequence (3)InDel: Indel number in accordance with autosomal dominant inheritance pattern and having an influence on amino acid sequence (4)Total: The total number of mutation that conforms to the autosomal dominant genetic pattern and have an influence on amino acid sequence of the anterior chamber during the eighth week of embryonic development (13). Traboulsi and Weinberg observed 12 members who were affected by this rare phenotype in an American family. The eyesight of all affected subjects were excellent. They were descended from a family who immigrated from Ireland to the United States in the seventeenth century. They settled in Ireland from Scandinavia in the thirteenth century (21). Chen et al. performed Genome-wide linkage and exome sequencing analysis showed a possible association between variation in the TMCO3 gene and the rare phenotype of cornea guttata with anterior polar cataracts (12).
For many years, Sanger sequencing has been the mainstream method to identify the pathogenic mutations of single-gene diseases. However, this method needs to know the pathogenesis of the disease in advance in order to make an appropriate gene selection. Due to low throughput and high cost, this technology is not suitable for conventional large-scale sequencing projects. Whole exome sequencing (WES), making sequencing of all protein-coding regions(exome) in the human genome rapidly became the most widely used targeted enrichment method, especially for Mendelian diseases, compared with the selection of genes followed by Sanger sequencing in the past decades (12,14). This approach allowed the detection of both exons (coding) and splice-site variants, while requiring about 2% of the sequencing "load" compared to whole genome sequencing (WGS). Unbiased analysis of all genes removes the need for time-consuming selection of candidate genes before sequencing. It has been estimated that the exome contains about 85% of mutations with a great influence on the disease-related traits (14). In view of the obvious ophthalmic changes in our family, we performed WES in this family to explore the pathogenic genes. Traditional linkage analysis, which could narrow down the candidate regions with the method of microsatellites, was used to successfully identify a small enough candidate region from the large pedigree samples of several affected individuals for further analysis. By combining linkage analysis and WES, we can maximize the opportunity to identify pathogenic mutations (12). However, linkage analysis also had limitations: it usually requires multiple generations of families, and it is difficult to analyze small families and sporadic cases. Sometimes even multiple generations of families can not determine the pathogenic locus, and it is more difficult to select the appropriate gene in the locked region (22). Our linkage analysis did not lock into a candidate region (LOD score > 2), which may be due to insufficient normal members in this family for linkage analysis. Therefore, the candidate regions cannot be effectively linked.
In the current study, we described a Chinese family with this rare phenotype. The common characteristic of patients was FECD in this family. Due to FECD may occur independently or in association with other ocular or systemic abnormalities (6,7,12), and there was no single phenotype of anterior polar cataract, the main consideration should be that the variations in candidate genes shared by all patients can cause both FECD and APC at the same time. There were five individuals who had both abnormalities, while only corneal changes in the other seven patients(another two FECD patients were deceased). Since IV-4 and IV-5 have been diagnosed with FECD at 16 and 18 years old, or maybe the age of the disease was even earlier, indicating that corneal changes were more likely to be early-onset FECD (3 to 40 years) (7), which was influenced by genetic factors, and was usually a familial autosomal dominant disease, first reported by Magovern et al. in 1979 (23). Affected children developed corneal guttae were as young as 3 years old. In contrast to the rough and clear guttae of late-onset FECD, the early-onset FECD was characterized by small and patchy guttae in the reillumination. Guttae in early-onset FECD appears in the center of endothelial cells whereas big guttae in late-onset FECD positioned at edges of endothelial cells by a specular microscope, which was corresponding to the slit lamp and specular images of our patients. Gottsch et al. suggested that the average onset age of familial FECD patients without COL8A2 mutation was 50 years old. The disease develops from early to late stage in 25 years, and its incidence is similar to that of the more common late-onset FECD, except that it appeared earlier and had obvious clinical symptoms at the age of 30-40 (23-25). The pathogenic mutation of COL8A2(MIM 120252) gene which encodes for the a-2 chain of collagen VIII at chromosome 1p34.3 -p32.3 (FECD 1) is related to the early form of FECD pathology, since this mutation affects the structure of the Descemet membrane (1,4,6,7). Mutations (L450 W, Q455 K) positioned in the triple-helical domain of α2 alter the structure and composition of Descemet's membrane, leading to the early onset type of FECD (24,26,27). To date, there has been no further report on the relationship between genotype and phenotype in early-onset FECD except for mutations in COL8A2.
For mutations in late-onset FECD, TCF4 (MIM 602272) gene is the most common cause of FECD (4). (CTG repeats) in intron 3 of the TCF4 gene encoding E2-2 protein are associated with FECD in many different populations, which is the most commonly identified genetic contributor to FECD. Baratz et al. first found there was a strong correlation between TCF4 intron polymorphic marker rs613872 and late-onset FECD after processing a genome-wide association study (GWAS) (5,8). Severity of disease seems to be correlated with repeat length in Caucasian populations, whereas no link could be detected in a Japanese cohort (28)(29)(30). Males were also found to have a higher risk of developing FECD based on the presence of CTG repeats, suggesting that the interaction of this locus with gender could be important (5). It can be seen from our family that the onset age of FECD is early (16 years old or maybe earlier), with more female patients than male patients, which were not very similar to the characteristics caused by TCF4 gene. In addition, we cannot detect GTC duplication using WES. There were other causative genes related to late-onset FECD. SLC4A11 (MIM 610206) gene encodes an ion channel that promotes the absorption of water in the endothelial layer, and is an important medium for corneal deturgescence. Mutations in this gene can cause corneal edema and are associated with FECD (6,10,(31)(32)(33). Similarly, mutations in the ZEB1 (TCF8, MIM 189909) gene encoding the transcription factor zinc finger E-box binding domain 1 are correlated with late-onset FECD (4,6,34,35). AGBL1 (MIM 615496) gene encodes deglutamylase enzyme ATP/GTP binding protein like 1, and gene mutation is related to FECD. The missense mutation of LOXHD1 gene is related to progressive hearing loss and corneal endothelial cell dysfunction in FECD (4-7,10,36). However, the candidate genes didn't cosegregate with the disease phenotype in this family followed by variant exclusion and prioritization. In addition, we haven't found TMCO3 gene, or any known causative gene in FECD as a candidate gene in our analysis at present, indicating there may be a novel variant or gene-locus associated with this rare phenotype in the non-coding region of genome. This disorder may also have genetic heterogeneity like FECD which needs to be explored via whole genome sequencing (WGS). In conclusion, we described a family of FECD with APC, which was autosomal dominant. The on-set age of both FECD and APC was about 10. FECD was more prone to be early onset. We also performed a WES analysis in this family to explore the cause and mechanism of the rare phenotype. No gene was found to cause the rare phenotype in our analysis via WES and Sanger sequencing, inferring there may be a novel variant or gene-locus in the non-coding sequence of the genome, which needs further study by WGS. The contribution of this study was to exclude the possibility of the presence of pathogenic sites in exome and narrow the scope of pathogenic genes.

Declaration of interest
The authors report no conflicts of interest. The authors alone are responsible for the content and writing of this article.

Funding
This work was supported by the Foundation of Heilongjiang Province (NO. CR201809) and National Natural Science Foundation of China (NO. 81970776 and 81671844).

Ethical approval and informed consent
This study has been approved by the Ethics Committee of the First Affiliated Hospital of Harbin Medical University. All participants provided informed written consent that was endorsed by the First Affiliated Hospital of Harbin Medical University.

Availability of data and materials
The datasets used and/or analyzed during the current study are available from the corresponding author up on reasonable request.