A Novel missense mutation of COL2A1 gene in a large family with stickler syndrome type I

Abstract Stickler syndrome type I (STL1, MIM 108300) is characterized by ocular, auditory, skeletal and orofacial manifestations. Nonsyndromic ocular STL1 (MIM 609508) characterized by predominantly ocular features is a subgroup of STL1, and it is inherited in an autosomal dominant manner. In this study, a novel variant c.T100>C (p.Cys34Arg) in COL2A1 related to a large nonsyndromic ocular STL1 family was identified through Exome sequencing (ES). Bioinformatics analysis indicated that the variant site was highly conserved and the pathogenic mechanism of this variant may involve in affected structure of chordin‐like cysteine‐rich (CR) repeats of ColIIA. Minigene assay indicated that this variant did not change alternative splicing of exon2 of COL2A1. Moreover, the nonsyndromic ocular STL1 family with 16 affected members showed phenotype variability and certain male gender trend. None of the family members had hearing loss. Our findings would expand the knowledge of the COL2A1 mutation spectrum, and phenotype variability associated with nonsyndromic ocular STL1. Search for genetic modifiers and related molecular pathways leading to the phenotype variation warrants further studies.


| INTRODUC TI ON
Stickler syndrome is characterized by ocular, auditory, skeletal and orofacial manifestations. 1 It has been divided into three types: STL1, STL2 (MIM 604841) and STL3 (MIM 184840) with STL1 being the most common form. STL1 is caused by COL2A1 variants and is inherited in an autosomal dominant manner. Except the familiar symptoms in eye such as high myopia, vitreoretinal degeneration, retinal detachment (RD) and cataracts, in some cases STL1 also showed short stature, scoliosis/kyphosis, joint hypermobility /osteoarthritis, cleft palate, midfacial hypoplasia and some degree of deafness. 2,3 Nonsyndromic ocular STL1 is a subgroup of STL1 and is characterized by predominantly ocular features with absent or minimal extraocular abnormalities. 4 However, both STL1 and nonsyndromic ocular STL1 have a wide phenotypic spectrum, with considerable interfamilial and intrafamilial variability in its clinical expression. 5 COL2A1 locates in chromosome 12 and encodes the a1 chain of collagen II (ColIIA), which is the main component of the cartilage extracellular matrix and the vitreous of eye. 6 The pathogenic variants of COL2A1 cause many diseases related to dysosteogenesis, including STL1. 7,8 These variants happen in both introns and exons.
While the former usually affects splice sites, the latter is much more complicated. Pathogenic variants in exons mainly contain premature stop (ie.,nonsense, frameshift) variants and missense variants. These two types of pathogenic variants cause diseases usually because of haploinsufficiency or structure defect of ColIIA, but they sometimes affect cis-elements and alter pre-mRNA splicing process.
In the nucleus, pre-mRNA of pro-ColIIA undergoes alternative splicing at exon2 and produces various isoforms: IIA, a long form that includes exon 2; IIB, a short form that excludes exon 2; IIC, which contains only the first 34 nucleotides of exon 2 and has no protein product; and IID, which contains the same sequence as IIA with an additional three nucleotides. Compared to IIA and IIB, the prevalence of IIC and IID is low, so IIA and IIB are the main isoforms. 9 The alternative splicing event of COL2A1 is developmentally regulated.
Inappropriate splicing often causes chondrogenesis and ocular diseases. Several pathogenic variants of exon 2 related to STL1 have been found. [10][11][12] The present study reports a novel pathogenic variant of COL2A1 resulting in nonsyndromic ocular STL1 and analyses the clinical heterogeneity of this family, which enriches our understanding of nonsyndromic ocular STL1.

| Patient recruitment
Patients and histories were identified in Binzhou Medical University Hospital. The studies were performed with approval of the Ethics Committee of the Binzhou Medical University Hospital (2018-008-01). All persons gave informed consent prior to their inclusion in the study. Venous whole blood samples were collected for molecular genetic testing. IV-2 and IV-4 were missing blood samples and inspection because they were lost to follow-up.

| Clinical examination
Ophthalmic, orofacial, skeletal and auditory features were assessed using previously reported methods. 4

| Exome sequencing (ES) and bioinformatics analysis
Genomic DNA was extracted from the peripheral leukocytes of all recruited family members using standard protocols for whole blood DNA extraction. The main portion of the ES was provided by Novogene Bioinformatics Institute. The exomes were captured using Agilent SureSelect Human All Exon V6 kits, and highthroughput sequencing was performed in an Illumina HiSeq X-10.
The basic bioinformatics analyses, including reads, mapping, variant detection, filtering and annotation, were also provided by Novogene Bioinformatics Institute. The average coverage for all of the experiments was 70x and was at least 20x for 90% of the targets. Paired sequencing reads were aligned to the reference genome (GRCh37/ hg19) using BWA 26 and sorted with SAMtools27 and Picard (http:// broad insti tute.github.io/picar d/webcite). Post-alignment processing (local realignment around insertions-deletions and base recalibration), single nucleotide variant (SNV) and small nucleotide sequences insertion-deletion (InDel) calling were performed with Genome Analysis Toolkit (GATK) 28, with parameters adapted to the haloplex-generated sequences. The called SNV and InDel variants produced with both platforms were annotated according to the ANNOVAR Web server (in the public domain, http://wanno var. wglab.org/index.php). 15 A tiered filtering strategy was used to prioritize the SNVs and InDels using previously reported methods. 16 The obtained SNVs and InDels were further analysed for conservative and possible deleterious impact by software (dbNSFP version3.0). According to a simple model of the dominant mode of inheritance pattern, we explored the SNVs that were heterozygous for the variant allele in the affected patients but normal in unaffected patients. Then, gene-disease phenotypic correlation analysis for the candidate variants was performed, and the variants were sorted by relevance. Whether the variants involved in the splicing process were predicated in Spidex databases.

| Pathogenic variant validation and co-segregation analysis
Filtered pathogenic variants and co-segregation analysis among all family members were validated by Sanger sequencing. The primer pairs were designed by Primer 5 (Table S1), and sequences of the PCR products were determined using the Eppendorf Mastercycler Genetic Analyzer. Sanger sequencing was performed at Majorbio.

| Conservation analysis of variants in COL2A1
Conservation of mutant cite in COL2A1 orthologs was analysed on UCSC (http://genome.ucsc.edu). Multiple sequence alignment analysis of CR repeats in ColIIA and other proteins was performed using DNAMAN software.  Figure S3).

| Analysis of spliced mRNA isoforms derived from the COL2A1 minigenes
Total RNA was extracted from the cells using Trizol (Invitrogen) according to the manufacturer's instructions. Total RNA was used to synthesize cDNA with a PrimeScript RT reagent kit (Takara, RR037A). Alternative splicing of the minigenes produced either the IIA or IIB isoform. The primer pairs RT-F and RT-R (Table S1) were used to detect the mRNA expression of IIA and IIB isoforms derived from the minigenes by RT-PCR ( Figure 4B). The products of RT-PCR were detected by electrophoresis in 12% polyacrylamide gels and were semiquantitative analysed by ImageJ software.

| Statistical analysis
The data were expressed as the mean ± SD of at least three independent experiments. Statistical analysis of the data was performed using one-way ANOVA followed by Bonferroni's multiple-comparison correction in GraphPad Prism 5.01.

| Clinical presentations of the patients
The family reported in the present study was from Shandong Province of China, and it was a five-generation nonsyndromic ocular STL1 family with 16 affected and 33 unaffected members ( Figure 1).
Individual IV-15 (the propositus) had extreme myopia in both eyes at about age 6. At age 9, this patient suffered a giant-tear retinal detachment in the left eye that was not successfully repaired. Two years later, this patient developed a cataract in this eye. At age 10, he suffered a giant-tear retinal detachment in the right eye and underwent retinal repair surgery. At age 11, a complicated cataract appeared in the right eye, and the retina detached, leaving faint light perception.
At age 36, the left eye was enucleated due to inflammation. Now, at age 45, a slit-lamp biomicroscopy examination for anterior segment F I G U R E 1 Pedigree of the nonsyndromic ocular STL1 family. The propositus was indicated by the black arrow. Normal descendants of the normal individual were not shown due to limited space. Asterisks: individuals which had blood available of the right eye showed that the oval pupil moved up to temporal, and the residual lens cortex and capsule were clouded and subluxated in the anterior vitreous ( Figure 2A). Ultrasound inspection showed vitreous opacities and long-standing retinal detachment ( Figure 2B). On joint examination, this patient could easily touch his fingers to the floor with extended knees despite a herniated disc ( Figure S1A,B). When he was younger, he was able to easily maintain his palms to the floor. Furthermore, on examination he was able to oppose his thumb to the ventral aspect of his forearm, while the same age control could not do this ( Figure  The other affected members of this family showed various degrees of myopia and retinal detachment ( Table 1). All of the patients had normal stature, hearing, speech development and intelligence, and did not have cleft palate or midfacial hypoplasia.

ColIIA orthologs of various species and CR repeats of different proteins
We compared the ColIIA orthologs and found that C34 was highly conserved in various species ( Figure 3B). CR repeats also known as Von Willebrand factor type C (VWC) domains 17 are ~60-80 amino acids in length and are mainly defined by a consensus sequence of 10 cysteines, which have been identified in ~200 extracellular matrix proteins. As matrix protein, ColIIA contains CR repeat which is encoded by exon2 of COL2A1, and might be involved in bone and eye development through binding transcription factors by CR repeats. Therefore, variants in exon2 often are pathogenic. 18,19 We exerted multiple sequence alignment of CR repeat of ColIIA and other proteins containing CR repeats, such as CHRD, CHRDL1, CHRDL2 and VWCE. The results showed that several cysteines including C34 and C57 were strongly conserved ( Figure 3C), which suggests that they are very important to maintain the structure of this domain.
We also performed deleterious prediction and conservation annotation of the variant using several general international software. The results showed that the c.T100>C(p. Cys34Arg) variant of COL2A1 changed the protein's structure or function with high probability and the mutant site was highly conserved (Table S2). These results suggested that the pathogenic mechanism of this novel variant might be involved in changes in protein structure. and c.G170>A (p. Cys57Tyr) variants in COL2A1 involved in the splicing process using Spidex databases. In general, it is thought that the variant affects splicing when the score greater than 4 or less than −4; however, the score of the two variants could not meet this standard (Table 2), which indicated that these variants affected splicing with low probability.
Furthermore, we constructed COL2A1 minigenes, and they represented wild type, c.T100>C (p. Cys34Arg) and c.G170>A (p.  IIB ( Figure 4B). The expected sizes of the spliced isoforms IIA and IIB derived from RT-PCR were shown in (Figure 4C), and expression level of isoforms IIA was higher than isoforms IIB in all minigenes.
The total expression quantities of IIA and IIB and the ratios of IIA/IIB of the three minigenes showed no difference ( Figure 4D,E). These results suggested that the c.T100>C (p. Cys34Arg) and c.G170>A (p. Cys57Tyr) variants in COL2A1 did not alter the alternative splicing of exon2, which was different from the previous reports.

| Phenotypes of this nonsyndromic ocular STL1
family showed significant heterogeneity and more serious in male gender STL1 has obvious heterogeneity in clinical manifestations. 5 In our study, the nonsyndromic ocular STL1 family showed a wide phenotypic spectrum ( that noted by black arrow were highly conserved extreme myopia, retinal detachment in both eyes before 15 years old and the joints were lax occasionally ( Figure 2, Table 1, Figures S1,S2).
We then analysed the heterogeneity in clinical manifestations in different genders of this nonsyndromic ocular STL1 family. The results showed that all of the affected male members of the STL1 family had high myopia in both eyes; however, a few female members only showed mild myopia ( Figure 5A). Furthermore, the severity of retinal detachment between male and female patients in the left and right eyes showed a similar trend. Males had more retinal detachments and had them earlier in life than females ( Figure 5B).

| DISCUSS ION
In this report, we identify a novel pathogenic variant (c.T100>C, p. The extracellular matrix provides a structural scaffold that imparts physical properties to connective tissue, and they also act as an instructive platform for soluble modulators of cell behaviour. 25 TGFβ (transforming growth factor-beta) binds to extracellular matrix macromolecules and plays an important role in embryonic tissues, and its dysregulation involved in the pathogenesis of retinal detachment. 26 BMP4 (bone morphogenetic protein 4) is a secreted ligand of the TGFβ superfamily and plays critical role in ocular development. 27,28 In recent report, a novel heterozygous BMP4 variant was first identified to cause STL1. 29 The CR repeats can bind members of TGFβ superfamily and are proposed to regulate growth factor signalling. ColIIA contains a 69-amino acid chordin-like CR repeat, which encoded by exon2 that exists in type IIA exclusively. 22 ColIIA interacts with proteins of the TGFβ superfamily by CR repeat and therefore potentially regulates growth factor signalling during development. 18 The isolated CR ColIIA repeat binds BMP-2 and TGFβ in solid-phase binding assays, 19 and binding to BMP-4 is competed by BMP-2 and TGFβ. 30  but several studies suggest that RD might have a higher incidence in males. We report that many major characters of STL1 including myopia and RD seem to have an association with gender. In our study, the affected male members displayed more serious myopia than females, which is different from the trend of nonpathological myopia. Perhaps with this particular genetic background, the progression of myopia is more sensitive to the sex hormones such as testosterone, which was found at a higher level in males.
Furthermore, RD in males was more serious than in female in our study, possibly due to more serious high myopia, previously published male trend in RD, genes modification related to sex that have not been yet described in literature. It was reported that 32 STL1 patients complained of hearing loss (37%, 95% CI 27-48) of whom 17 required hearing aids. 45 In contrast, our current report did not have any patients complaining of hearing loss, neither required hearing aids. Our previous report shows that a mouse model with a missense mutation in the mouse Col2a1 gene resulted in a mouse phenotype similar to human STL1, including F I G U R E 5 Clinical phenotypes showed more serious in male than that in female in both eyes. (A) Compare the serious degree of myopia between male and female in the left and right eyes. (B) Compare of serious degree of retinal detachment between male and female in left and right eyes hearing impairment ranged from 35 to 50 dB hearing loss to completely deaf in a mixed genetic background. 8 All of these reports suggest that genetic background plays important roles in the phenotypic variations. Search for genetic modifiers and related molecular pathways leading to the phenotype variation warrants further study.
In summary, in this study, we identified a novel missense pathogenic variant in COL2A1 in a nonsyndromic ocular STL1 family. The pathogenic mechanism of this variant may be involved in protein structure changing. The nonsyndromic ocular STL1 family showed clinical variability and a male gender trend with unknown reason.
Our findings expand the knowledge of the COL2A1 mutational spectrum and clinical heterogeneity associated with STL1.