Novel compound heterozygous EYS variants may be associated with arRP in a large Chinese pedigree

Abstract As a genetically heterogeneous ocular dystrophy, gene mutations with autosomal recessive retinitis pigmentosa (arRP) in patients have not been well described. We aimed to detect the disease-causing genes and variants in a Chinese arRP family. In the present study, a large Chinese pedigree consisting of 31 members including a proband and another two patients was recruited; clinical examinations were conducted; next-generation sequencing using a gene panel was used for identifying pathogenic genes, and Sanger sequencing was performed for verification of mutations. Novel compound heterozygous variants c.G2504A (p.C835Y) and c.G6557A (p.G2186E) for the EYS gene were identified, which co-segregated with the clinical RP phenotypes. Sequencing of 100 ethnically matched normal controls didn’t found these mutations in EYS. Therefore, our study identified pathogenic variants in EYS that may cause arRP in this Chinese family. This is the first study to reveal the novel mutation in the EYS gene (c.G2504A, p.C835Y), extending its mutation spectrum. Thus, the EYS c.G2504A (p.C835Y) and c.G6557A (p.G2186E) variants may be the disease-causing missense mutations for RP in this large arRP family. These findings should be helpful for molecular diagnosis, genetic counseling and clinical management of arRP disease.

Although the EYS gene mutations accounted for ∼5% of arRP in a cohort of RP patients who were mainly of western Europe ancestry [17], the EYS mutations of patients with arRP and genotype-phenotype correlations in the Chinese population have not been well described. Here, next-generation sequencing (NGS) was conducted to identify novel, compound heterozygous variants in EYS from a large Chinese arRP family, thereby extending its mutation spectrum.

Pedigree and clinical assessment
This research study was carried out in accordance with the World Medical Association Declaration of Helsinki. This was approved by the ethical committee at the Southwest Medical University and written informed consent was obtained from all subjects. A Chinese proband ( Figure 1, pedigree II: 1, arrow) was recruited. For clinical assessment, an ophthalmic examination was conducted, including best-corrected Snellen visual acuity, Humphrey visual fields, optical coherence tomography (OCT), slit-lamp biomicroscopy, fundoscopy, fundus photograph (FP) and fundus fluorescein photograph (FFP), as in previous studies [18,19].

Sample collection and gDNA extraction
Fresh peripheral blood samples (2 ml each) from seven people in this family were collected, and gDNA was extracted by using our phenol/chloroform method that has been previously described [20][21][22]. One hundred healthy and ethnically matched individuals were sampled for DNA extraction as controls. A NanoDrop spectrophotometer (NanoDrop 2000, Wilmington, DC, U.S.A.) was used to measure the extracted DNA quality.

Capture panel design, sequencing and data analysis
To characterize the disease-causing genes in the M195 family, TES (targeted next-generation sequencing) analyses were performed on the proband (M195), according to the Illumina paired-end libraries as reported previously [23,24]. The capture Agilent probes were applied in the previous reports [23][24][25][26][27], using a retinal disease capture panel. Then, paired-end sequencing Illumine reads were applied to align to the human hg19 reference genome by BWA (version 0.6.1) and the UCSC database [28]. Variations of SNPs and INDELs were refined to look for the causative mutations in suspected genes. The pathogenicity in each variant was estimated using the following programs: Mutation Taster, PolyPhen-2, SIFT, and I-Mutant2.0. The ExAC and HGMD databases were used to search for novel mutations. 374 bp and 478 bp in length were amplified using gDNA as a template. PCR amplification and DNA sequencing in the EYS variants were used to all the available gDNA for variant verification and segregation analysis [19,30].

Sanger sequencing and co-segregation analysis
All the PCR products were then directly used for Sanger method sequencing using a machine of ABI-3500DX sequencer from Applied Biosystems Inc. in our laboratory through the specific primers EYS-M195L16 or EYS-M195L32 (Table 1). All controls with unrelated ethnical-matched were also used to sequence by aforementioned primers in Table 1. For co-segregation, we conducted an analysis based on our sequenced results and the patient's clinical phenotype.

RNA expression profiles
The EYS mRNA expression profiles in 27 normal human tissues were obtained by RNA-sequencing to determine tissue-specificity through an online NCBI database (https://www.ncbi.nlm.nih.gov/gene/346007#gene-expression) [33]. Since no retinal tissue was obtained from RNA sequencing, the EYS mRNA expression profiles in different tissues, including the human retina were also obtained from the three transcriptomic datasets (https://www.proteinatlas. org/ENSG00000188107-EYS/summary/rna).

Pedigree and clinical characteristics
A large Chinese pedigree, consisting of 31 members including a proband and two additional patients was recruited ( Figure 1, pedigree II: 1 with arrow indicated). The proband, who was from a non-consanguineous family, is a 56-year-old male, who first showed night blindness symptom at 15 years of age. The FPs and FFPs of the proband (II:1) in both eyes and control images are shown in Figure 2. The proband clearly showed clear severe RPE atrophic changes, pigmentation with bone spicules in the retina of the peripheral-mid and transparent macula. The vessels were extremely small, and the optic disc was pale or waxen in both eyes (Figure 2A-D). Electroretinography (ERG) showed reduced cone and core responses or low amplitude ERG in the patient (data not shown). Two younger brothers of the proband showed similar RP feature. The proband's parents and other family members didn't exhibit any RP features, suggesting an autosomal recessive inheritance pattern. Based on pedigree analysis, the cases in this family are likely considered arRP.  Table 2.

Variant confirmation and analysis of co-segregation results
Although deficient, variant confirmation and analysis of co-segregation for EYS were conducted by Sanger sequencing (Figure 3). The c.G2504A and c.G6557A variants of EYS were confirmed in the mutant compound heterozygous types in the proband (pedigree II: 1; Figure 3A,F); and we revealed wild-type without RP in the proband's wife (pedigree II: 2; Figure 3B,G), mutant heterozygous types without RP symptoms in the proband's mother (pedigree II: 2; Figure 3C,H) and mutant heterozygous types without RP symptoms in his younger sister (pedigree II: 8; Figure 3D,I), wild-type without RP symptoms in the proband's daughter (pedigree III: 2; Figure 3E,J), and mutant compound heterozygous types in the proband's two younger brothers (II: 3 and II:5 with RP symptoms, data not shown). Thus, these findings show co-segregation with disease in this family we tested, and suggest its role in pathogenesis. Furthermore, 100 ethnically matched normal controls were sequenced for both variants of EYS; no variants were revealed (data not shown).

Effects of function in EYS variants c.G2504A (p.C835Y) and c.G6557A (p.G2186E)
Searching of the Conserved Domain Database (CDD) from NCBI was performed. Comparing human EYS protein to other species indicated that EYS is conserved among chicken, dog, rhesus monkey, and zebrafish ( Figure 4A). EYS   has two repeated conserved domains: calcium-binding EGF-like domain (EGF CA, cl09941) and laminin G domain (LamG, cl17353). EYS variant p.C835Y, located in the EGF CA domain (amino acid 810∼846), and p.G2186E, located in the lamG domain (amino acid 2149∼2315), were also highly conserved ( Figure 4B,C), which may affect protein function. However, no EYS ortholog in mice was found. MutationTaster revealed disease-causing, SIFT revealed tolerated, and I-Mutant2.0 showed decrease stability for both p.C835Y) and p.G2186E variants; Polyphen 2 for p.G2186E showed probably damaging, whereas Polyphen 2 for p.C835Y showed benign (Table 2). Thus, these compound and heterozygous variants c.G2504A (p.C835Y) and c.G6557A (p.G2186E) in EYS might damage protein function in this arRP family. The variant c.G2504A (p.C835Y) revealed a novel mutation even though c.G6557A (p.G2186E) was not found [16] in a search of in the HGMD and ExAC databases (Table 2).
Comprehensively, the present study shows that compound heterozygous, pathogenic missense variants of EYS c.G2504A (p.C835Y) and c.G6557A (p.G2186E) may cause arRP in this large Chinese pedigree.

Expression profiles of EYS mRNA
RNA-seq data showed that EYS has low expression in 27 tested different human tissues but fat has highest expression (RPKM value: 0.397 + − 0.042), followed by the testis (RPKM value: 0.258 + − 0.112); the pancreas has the lowest expression (RPKM value: 0.008 + − 0.002) ( Figure 5A and Table 3). The protein is expressed in the photoreceptor layer of the retina [34,35]. However, no eye tissue data were shown by RNA-seq; no mouse EYS ortholog existed. Then the EYS mRNA expression profiles in more different tissues, including the retina and cells (55 tissue types and 6 blood cell types) were also obtained, and the results showed that EYS is most highly expressed in the retina of humans, with an NX value of 40.7 in the retina, but an NX value of only 2.9 in the testis ( Figure 5B). Thus, this demonstrated that EYS is only highly expressed in retinal tissue and probably only plays a vital role in the retina of the eye.

Discussion
The EYS protein (NP 001136272.1) has different isoforms. Consistent with RNA expression data ( Figure 5), expression analysis by PCR of cDNA from tissues and cell lines showed that the EYS gene is only expressed in the retina and in a retinoblastoma cell line in humans [6]. Immunohistochemistry (IHC) in mature pig retinas showed Eys expression at the photoreceptor layer. In a zebrafish model, Eys absence led to degeneration in the photoreceptor outer segments, photoreceptor death, disorganized retinal architecture, decreased ERG responses and caused visual dysfunction [7,36,37], whereas in humans, by novel EYS mutations, IHC revealed advanced retinal degenerative changes of in all eyes with rod photoreceptor absence [38]. Collin et al. in 2008 independently identified a large transcript, encoding 3165 amino acids (isoforms 4, NP 001278938) with a signal peptide, and domains with EGF-like and laminin A G-like. All four EYS transcripts were expressed in the retina and the Y79 cell line of humans, whereas isoforms 2 and 3 of EYS are also expressed in the testis [35]. BLAST analyses found that this EYS gene was the true ortholog of the Drosophila 'eyes shut' (eys) gene. Surprisingly this EYS gene is abundantly expressed in the retina in humans, but Eys is completely absent in mouse. Expression profiles of EYS mRNA showed that it is only highly expressed in the retina tissue of humans but not in other tissues ( Figure 5), demonstrating that EYS probably only plays a vital role in the retina of the eye. The targeted NGS approach has been proven to be efficient in the genetic diagnosis of RP [9,10,30]. In this paper, we identified that compound heterozygous, missense variants (c.G2504A, p.C835Y; c.G6557A, p.G2186E) of EYS may cause arRP in the large Chinese family by targeted NGS. Analyses by MutationTaster revealed disease causing, SIFT revealed tolerated, and I-Mutant2.0 showed decrease stability for both variants; Polyphen 2 for p.G2186E was probably damaging, whereas Polyphen 2 for p.C835Y was benign. Thus, combined all information, these compound and heterozygous variants in EYS might be pathogenic in this large Chinese pedigree by damaging and reducing the stability of the EYS protein. Variant p.C835Y for EYS is located in a conserved EGF CA domain that may be crucial for protein-protein interactions, whereas variant p.G2186E for EYS is located at a conserved LamG domain that has the role in signal transduction, adhesion, migration and differentiation by mediating cell adhesion molecule through binding. In this regard, the EYS variants p.C835Y and p.G2186E may interrupt protein-protein interactions and signal transduction, leading to the pathogenesis of arRP. To the best of our knowledge, the EYS mutation of c.G2504A (p.C835Y) is novel and enriches the EYS mutation spectrum, and is linked with arRP phenotypes. Therefore, these findings will help in understanding the molecular pathogenesis of RP for diagnosis, prevention as well as genetic counseling.

Conclusion
In summary, our study identified the compound heterozygous variants c.G2504A (p.C835Y) and c.G6557A (p.G2186E) of EYS, which may cause arRP in this large Chinese family. It is the first study to reveal the novel EYS mutation (c.G2504A, p.C835Y) for RP disorder in our Chinese patients, expanding the EYS mutation spectrum. These findings should help in understanding the molecular pathogenesis of arRP disease for diagnosis, clinical management and genetic counseling.