Clinical manifestation and genetic analysis in Chinese early onset X‐linked retinoschisis

Abstract Background X‐linked retinoschisis (XLRS) is one type of retinal dystrophy leading to the schisis of the neural retina and causing reduced visual acuity. The study aimed to investigate the clinical manifestations and retinoschisin 1 (RS1) mutations in Chinese patients with early onset XLRS. Methods Thirty‐eight probands with early onset XLRS were recruited, comprehensive ophthalmic examination was performed. A targeted gene panel was used to test the RS1 mutations. Results All probands had RS1 hemizygous mutations including 16 known and 14 novel mutations. The median onset age was 2 years old (range 0.1–6 years). Probands with onset age ≤1 years. had more complications (retinal detachment and vitreous hemorrhage, p < 0.001), more mutations outside the discoidin domain and more non‐frameshift mutations than probands with onset age >1 years. Macular and peripheral involvement was present in 77.27% of probands, and inner and outer nuclear layer splitting were present in 53.57% of probands. Electroretinography showed an electronegative waveform. The relatively rare phenotypes of lamellar macular hole and macular hole were present in a unilateral eye in three probands. Conclusion In conclusion, the early onset XLRS developed more severe complications which need close monitoring and clinical manifestations illustrated here may facilitate the early diagnosis of retinoschisis.


| INTRODUCTION
X-linked retinoschisis (XLRS; MIM 312700) is one type of retinal dystrophy that leads to the schisis (splitting) of the neural retina, causing reduced visual acuity. It is the leading cause of macular degeneration in male children with an estimated prevalence ranging between 1:5,000 and 1:20,000 (George, Yates, & Moore, 1995;Molday, Kellner, & Weber, 2012). The typical clinical feature of XLRS is a spoke-wheel pattern foveal schisis that is associated with a peripheral schisis in about 50% of the affected patients. Foveal involvement has been reported in all affected patients, with mildly to severely reduced visual acuity (George et al., 1995;Tantri et al., 2004). An absent or severely reduced b-wave found via electroretinography (ERG) is typical for XLRS (Tantri et al., 2004). Spectral-domain optical coherence tomography (OCT) is the major diagnostic technique for identifying XLRS. However, because of the inclusion of the infants and preschool children, it is difficult to perform these examinations on the young children.
The retinoschisin 1 (RS1) gene is located on the X chromosome. It contains six exons that encode retinoschisin, express exclusively in the retina and function as an octamer. The gene is also implicated in cell-cell adhesion and interactions (Sikkink, Biswas, Parry, Stanga, & Trump, 2007). RS1 protein contains three functional domains, including a signal sequence (exons 1 and 2), retinoschisin domain (exon 3), and discoidin domain (exons 4-6) (Sauer et al., 1997), which are important during retinal development (The Retinoschisis Consortium, 1998). To date, more than 240 unique mutations in the RS1 gene have been reported as associated with retinoschisis (RS) phenotypes (Human Gene Mutation Database [HGMD], http://www.hgmd.cf.ac.uk/ac/index.php), including 154 missense and nonsense mutations and 23 splicing mutations, with the remaining as small deletions, small insertions, and gross deletions. A great proportion of the mutations are located in the discoidin domain (HGMD).
The clinical manifestations of XLRS exhibit a large degree of clinical variability (Apaolaza et al., 2016;Xiao et al., 2016). Previous studies have failed to identify a genotype-phenotype correlation because of this variability in clinical features. Clinical variability can also lead to misdiagnosis, especially in patients with early onset like XLRS. In this study, we attempted to illustrate the clinical manifestations of early onset XLRS in a Chinese population.

| Participants
The data from 38 unrelated families with XLRS were collected from the retinal department at Zhongshan Ophthalmic Center, Guangzhou, China, from February 2016 to December 2018. The recruitment criteria included a clinical diagnosis of RS conformed by fundus photography (FP) or OCT and a pathogenic mutation in the RS1 gene as well as a disease onset age of less than 6 years old.

| Editorial policies and ethical considerations
This study was approved by the Institutional Review Board of the Zhongshan Ophthalmic Center and was performed in accordance with the guidelines described in the Declaration of Helsinki. Informed consent was obtained from all participants or their guardians prior to the collection of the clinical data and genomic samples. The genomic DNA was extracted from the leukocytes of venous blood using previously reported methods .

| Sequencing
The genetic testing was performed using a clinical panel targeting genes that are common causes of inherited eye diseases. The targeted resequencing was performed using the Illumina MiSeq platform (Illumina, Madison, WI, USA) with a SeqCap EZ Library SR kit (Roche NimbleGen Roche, Inc., Madison, WI, USA), Illumina MiSeq v2 kit and Illumina PhiX Control v3 kit. The average sequencing depth was set to 100-fold. Strand NGS software version 2.0 (Strand Scientific Intelligence Inc.) was used to set the sequencing reads to University of California Santa Cruz hg19. The variation annotation was conducted using single nucleotide polymorphism (SNP) effects analysis software. SIFT, Polyphen2_HDIV, Polyphen2_HVAR, LRT, MutationTaster, MutationAssessor, FATHMM, MetaSVM, and MetaLR were used for the bioinformatics analysis. Sanger sequencing was used to verify the mutations and perform a segregation analyses on available family members. All the mutation nomenclature followed Human Genome Variation Society (HGVS) guidelines for mutation description (http://www.hgvs.org). Allele frequencies of the variants detected in this study were confirmed by The Exome Aggregation Consortium (ExAC; http://exac. broad insti tute.org).

| Clinical examination
The clinical data collected included the age, gender, onset age, first symptom, and family history of subjects. A complete ophthalmic examination, including a best corrected visual acuity (BCVA) measurement, slit-lamp examination, and fundus biomicroscopy, was performed in each patient and their available family members. FP, ERG, and OCT were done with parental consent as required. FP was performed with an FF450 fundus camera (Zeiss, Oberkochen, Germany), scanning laser ophthalmoscope (SLO, Nidek F-10, Gamagori, Japan), or RetCam (Clarity Medical Systems, Pleasanton, CA, USA). ERG was performed with a RETeval (LKC Technologies, Gaithersburg, MD, USA). OCT was performed with a Spectralis HRA (Heidelberg Engineering, Heidelberg, Germany).

| Statistics
The statistics was performed using the IBM SPSS Statistics 21.

| Clinical data
All of the patients were male, with a median examination age of 4 years old (range 0. of 2 years old (range 0.1-6 years) including 17 patients with the onset age equal to or less than 1 year old and 21 patients with the onset age more than 1 year old (Table 2, Figure 3a). The first symptoms consisted of no following (5/38, 13.15%), poor vision (27/38, 71.05%), and strabismus (6/38, 15.79%). The BCVA examinations were conducted in patients more than 4 years who could cooperate with the examination, and the results ranged from no light perception to 0.60 (Table 2). FP was available for 66 eyes of 33 probands, and a spokewheel pattern combined peripheral RS was detected in 51 (77.27%) eyes (Figure 4a,b). OCT was available for 28 probands, from which inner nuclear layer RS was detected in 13 (46.43%) probands (Figure 4c,d) and inner nuclear layer RS combined with outer nuclear layer RS was detected in 15 (53.57%) probands. ERG was obtained for 18 probands and showed an electronegative waveform pattern with a b-wave amplitude smaller than that of an a-wave. Nine probands had complications, including five patients (seven eyes) with retinal detachments, and four patients (four eyes) with vitreous hemorrhages.

| Difference between patients with onset age equal to or less than 1 year old and more than 1 year old
The patients with the onset age equal to or less than 1 year were assigned to group A, and the patients with the onset age more than 1 year old were assigned to group B. Retinal detachment and vitreous hemorrhage were considered to be severe complications of RS which were presented in more patients of group A than patients of group B (p < 0.001, Figure 3b) between groups. Significant difference was also detected in mutation locations (inside the discoidin domain and outside the discoidin domain) and mutation types (nonframeshift mutations and frameshift mutation) (p < 0.001,

| Clinical data of proband with c.540_675+682del mutation
The patient (QT2390) with the gross deletion mutation (c.540_675+682del) was noticed to have strabismus at the age of 4 years. No treatment was applied to him over the next 1 year. At the age of 5, his BCVAs were 0.15 in the right eye and finger counting at 30 centimeters in the left eye. Macular ectopia in the left eye was showed in the fundus images, which was the reason for poor vision. And the poor vision contributed to the development of strabismus. Splitting of the inner nuclear layer was found in both eyes via OCT (Figure 2c).

| Clinical data of probands with the same mutations
Similarity and differences were present in probands with the same mutations. RS was present in the macular area and peripheral retina in all the patients with the p.Arg102Gln mutation; however, retinal detachment also developed in two patients. Among the three patients with the c.579dup mutation, two patients had foveal and peripheral RS and one patient had foveal RS only. Two patients had inner nuclear layer RS combined outer nuclear layer RS, and one patient presented with inner nuclear layer RS.

| Rare XLRS phenotypes
Clinical features varied even within the same patient. A rare phenotype was present in this study. A lamellar macular hole was present in one eye but absent in the contralateral eye in patient F1. A macular hole was present in the left eye but absent in the right eye in patient F4 as well as in patient QT2462 ( Figure 5).

| DISCUSSION
In this study, 38 early onset XLRS probands with a pathogenic RS1 hemizygous mutation were included. Thirty different mutations were detected, including 16 known and 14 novel mutations. Mutation c.214G>A (p.Glu72Lys), c.305G>A (p.Arg102Gln) and c.579dup (p.Ile194fs) were considered to be the most common mutations in this cohort. This cohort emphasized the clinical manifestations with early onset XLRS. Foveal and peripheral RS was present in 77.27% of patients. Inner nuclear layer RS only was detected in 46.43% of patients, and inner nuclear layer RS combined with outer nuclear layer RS was detected in 53.57% of patients. Significant difference was detected in complications, mutation types, and mutation locations between patients less than or equal to 1 year old and old than 1 year old. Phenotypes of this cohort showed differences from a previous study. Molday et al. mentioned that peripheral RS was present in fewer than 50% of the affected XLRS patients (Molday et al., 2012); however, peripheral RS was present in 77.27% of patients in the current study. Inner nuclear layer RS was detected in 46.43% of patients, and inner nuclear layer RS combined with outer nuclear layer RS was detected in 53.57% of patients. Retinal detachment and vitreous hemorrhages was present in approximately 5% of all affected males (Molday et al., 2012) in the general population. Another cohort demonstrated retinal detachment or severe macular lesion happened in 21.4% (3/14) before the age of 4 years, which was consistent with our study. The severe complication including retinal detachment and vitreous hemorrhages presented in 21.1% (8/38), especially in patients with the onset age less than 1 year old.
XLRS exhibits a high degree of phenotypic variability (Apaolaza et al., 2016). Several studies have illustrated that even among patients with the same mutations, there is still a great variation in disease severity (Eksandh et al., 2000;Pimenides et al., 2005;Xiao et al., 2016). Our study also demonstrated the clinical variability in XLRS patients but failed to find a phenotype-genotype correlation. We summarized the clinical features of patients with the same mutation, differences in the severity of clinical phenotypes, extent of the retinal schisis, and complications. Additionally, even in the same patient, the ocular manifestation varied between the eyes. Some of the complications, such as a vitreous hemorrhage or retinal detachment, were detected in only one eye but absent in the contralateral eye. Overall, phenotypic variability makes it difficult to diagnose XLRS, especially in early