A family with CNA2 was recruited and identified at the Hunan Jiahui Genetics Hospital (Changsha, China) in October 2017. A pedigree of the family is illustrated in Fig. 1A. The proband, III:2, developed visual disorders shortly following birth, but hadn't been diagnosed or treated. She visited the doctor at the age of 26 years due to a considerable decrease in binocular visual acuity and accompanied by photophobia. Clinical and eye examinations demonstrated temporal iridocorneal synechiae, corneal clouding (Fig. 1B) and an unclear fundus. In addition, she was pregnant at the same time. A similar phenotype was observed in the younger brother of the proband, III:3.

Best-corrected visual acuity (BCVA) was established using the Acta Ophthalmologica 2017 LogMAR charts. Visual acuity was examined using the ETDRS chart. Intraocular pressure was measured by Goldmann applanation tonometry. Anterior chamber depth and axial length were measured with ocular A ultrasound. Motility, pupillary reaction, ocular fundus, slit lamp examinations and keratometry (Zeiss keratometer; Zeiss AG) were performed and documented. The growth and development of III:2 and III:3 were in the normal range. Physical examinations were performed to exclude systemic diseases.

The present study was performed according to the guidelines approved by the Ethics Committee of Hunan Jiahui Genetics Hospital and adhered to the tenets set out in the Helsinki Declaration. Informed consent was obtained from all subjects.

Venous blood samples were collected from the study subjects (III:2 and III:3), their family members (II:1, II:2 and III:1) and 200 randomly selected normal individuals without any history of ophthalmic disease. In addition, amniotic fluid was collected from III:2 to reflect the condition of the fetus (IV:1). Genomic DNA from all subjects was extracted via the phenol/chloroform method. Whole-exome sequencing data were screened for pathological variants in the patients (III:2 and III:3), and sanger sequencing of the candidate variants in KERA was performed on all subjects.

The extracted DNA was fragmented with DNase and purified via the magnetic bead method, followed by ligation of the adaptor sequence and polymerase chain reaction (PCR) amplification. PCR was conducted using a Phusion^® High-Fidelity PCR Master Mix (New England BioLabs, Inc.) and the follows primers: 5′-AATGATACGGCGACCACCGAGATCTACACTCTTTCCCTACACGACGCTCTTCCGATCT-3′ and 5′-CAAGCAGAAGACGGCATACGAGATCGGTCTCGGCATTCCTGCTGAACCGCTCTTCCGATCT-3′. PCR was performed under the following conditions: 98°C for 30 sec; then 12 cycles of 98°C for 10 sec, 65°C for 30 sec and 72°C for 30 sec; and finally 72°C for 5 min. Library preparation was performed as previously described (25). The whole-exome was captured by Nimblegen SeqCap EZ Exome Probes v3.0 (Roche Diagnostics, Basel, Switzerland), extended, purified and finally sequenced on an Illumina HiSeq sequencer (Illumina, Inc.). The coverage of the target sequence was not less than 97% and the average depth of the target area was 130.95.

The original data were obtained following sequencing of all exons. The Burrows Wheeler Aligner algorithm was used to compare the reference sequence (UCSChg19), GATK version 3.7.0 (https://software.broadinstitute.org/gatk/) and VarScan version 2 software (http://dkoboldt.github.io/varscan/) were used to identify mutations, e.g. single nucleotide polymorphisms (SNPs) and InDels, including detection, annotation and statistical analysis. Annovar (http://www.openbioinformatics.org/annovar/) was utilized to catalog the detected variants. The list of mutations was tabulated, with the variants being described according to the HGVS guidelines. Combining this analysis with the public variation database ExAc, mutations with frequencies <0.1% in the population were identified. Mutations were scored for disease effect using a range of predictive tools [e.g. SIFT (http://sift.bii.a-star.edu.sg/), PolyPhen-2 (http://genetics.bwh.harvard.edu/pph2), Provean (http://provean.jcvi.org/) and MutationTaster (http://www.mutationtaster.org/)]. Based on the results of the data analysis, the mutations were identified with reference to the 1000 Genomes Project (http://browser.1000genomes.org), the Exome Variant Server (http://evs.gs.washington.edu/EVS/), dbSNP database (https://www.ncbi.nlm.nih.gov/projects/SNP/index_old.html) and the Human Gene Mutation Database (http://www.hgmd.cf.ac.uk/ac/index.php). HomoloGene (https://www.ncbi.nlm.nih.gov/homologene) and T-Coffee multiple sequence alignment (http://www.tcoffee.org) were utilized to determine the conservation of the mutated amino acid residues between different species. A three-dimensional structure of keratocan was modeled using SWISS-MODEL (https://swissmodel.expasy.org). The analysis was performed in combination with assessment of the clinical manifestations and the mode of inheritance. Finally, the nature of the mutation sites were interpreted according to the guideline by the American College of Medical Genetics and Genomics/the Association for Molecular Pathology (ACMG) (26).

The two novel candidate variations were validated using direct bidirectional sequencing methods. Based on the relevant gene standard sequences listed at the NCBI, primers were designed using Primer Premier 5 software (Premier Biosoft International). Primers forward (5′-3′): CACCCAGTTTTCCTACTGCTTTATA, primers reverse (5′-3′): GCACTGATTCGGGGAACCTT. PCR amplification was performed on the target DNA fragment (700 bp product size, 20 µl reaction system, annealing temperature 62°C) and the product was verified by 1.0% agarose gel electrophoresis. Direct sequencing was performed on an ABI 3130 DNA Sequencer (Applied Biosystems; Thermo Fisher Scientific, Inc.). The sequenced results were aligned with the reference sequence using SeqMan at the NCBI (no. NM_007035).

III:2 exhibited a BCVA of 0.2 oculus dextrus (OD; right eye) and 0.3 oculus sinister (OS; left eye) with hyperopic refractive values. Additionally, she exhibited esotropia (right eye), microcornea, corneal opacity, short axial length, shallow anterior chamber, iridocorneal adhesion, pupillary irregularities (1×2 mm irregular right pupil shifting to the nasal side; needle-like changes on the left pupil) and slow light reflection. Extraocular motility, confrontation fields and intraocular pressure were normal, and the lens was transparent, although the fundus was difficult to see clearly (Table I). The younger brother of the proband (III:3) presented with similar clinical symptoms, however the parents had not identified any eye problems.

The family of interest, originating from China, displayed an autosomal recessive pattern of inheritance. A compound heterozygous mutation c.242A>G (p.N81S) and c.772_779delGGATTTGA(P.G258Cfs*30) in exon 2 of the KERA gene was identified in III:2 and III:3, respectively derived from the heterozygous father (c.772_779delGGATTTGA) and mother (c.242A>G; Fig. 1C and D), although neither of the two mutations were present in the normal controls. These variants have not been previously reported in the 1000 Genomes Project, the Human Gene Mutation Database, dbSNP or the Exome Variant Server. The first variation, an N81S amino acid substitution in the KERA protein, was predicted to be likely pathogenic by the in-silico prediction programs, Mutation Taster, Provean, n PolyPhen (27) and SIFT (28) (Table II). The second mutation was a 7 base-pair deletion (G258Cfs*30), leading to the loss of LRR8-11 and producing a strongly truncated protein, which may trigger nonsense-mediated mRNA decay (−66 amino acid/>10% missing); the prediction algorithms supported its pathogenic nature. Multiple sequence alignment indicated that the residues at position 81 of KERA are highly conserved across species (Fig. 1E) and the lost LRR8-11 regions due to 258 residual changes are also highly conserved. According to ACMG, p.N81S was judged likely to be damaging (PM1-3 and PP1-4) and P.G258Cfs*30 was also judged as damaging (PVS1, PM2, PP1, PP3 and PP4) (28).

Primary sequence analysis indicated that the N81S substitution occurs at an LRR motif, the first of 11 such motifs present in the sequence (Fig. 2C) (29). G258Cfs*30 indicates a 7 base-pair deletion leading to a frameshift and premature termination codon (PTC), followed by the loss of LRR8-11. The predicted three-dimensional structure of the KERA protein using the SWISS-MODEL online software suggested that the current mutant protein (P.G258Cfs*30) is significantly different from the wild type and these changes may lead to abnormal protein function (Fig. 2D) (30). The architecture of other known LRR proteins reveals that the highly conserved LXXLXLXXNXL repetitive motifs form a series of parallel β-strands and these β-strands form a horseshoe-like concave shape. The asparagines (Asn, N) develop a stable Asn-ladder structure. The present study identified that the p.(Asn81Ser) mutation affects the conserved Asn81 residue of LRR1. This substitution to an aspartic acid residue will alter a hydrogen bond donor to make it either an acceptor or neutral, possibly resulting in the instability of the Asn-ladder structure or abnormal interactions between the mutant KERA and other proteins, including collagen. The second mutation, G258Cfs*30, that is reported in the present study results in the loss of LRR8-11. This deletion may form an abnormal truncated protein, thereby causing disease.