A novel, homozygous nonsense variant of the CDHR1 gene in a Chinese family causes autosomal recessive retinal dystrophy by NGS-based genetic diagnosis

2.1 Ethical statement, proband and clinical assessment

The research was approved by the Ethical Committees of the Southwest Medical University; written informed consent conforming to the tenets of the Declaration of Helsinki (1983 Revision) was obtained from all participants.17 The study consisted of a proband (Figure 1, pedigree II: 4, arrow), and 9-related family members from three-generations, with no consanguineous marriage history, based on their genetic and pedigree analysis (Figure 1). For clinical diagnosis, a detailed clinical history and ophthalmic examinations were performed in proband, including the best-corrected Snellen visual acuity, humphrey visual fields, ‎slit-lamp biomicroscopy, fundoscopy, ‎optical coherence tomography, fundus photographs (FP) and fundus fluorenscent photographs (FFP), and standard electroretinography, which were used in previously studies.17-19

2.2 Blood sampling and DNA extraction

Two millilitres of fresh peripheral bloods were taken, and human genomic DNAs (gDNAs) were extracted using the previously described standard phenol/chloroform method from blood leucocytes of the proband and pedigree members who were accessible.20, 21 Blood samples were also taken from 100 RP-unrelated, ethnically matched and healthy control volunteers no any disease history.

2.3 Capture panel designing, exome sequencing

To access the disease-causing genes and mutations, the panels for TES analyses on the DNA sample from the proband M341 were designed, according to the Illumina paired-end libraries (Illumina, Inc., San Diego, CA, USA).11, 12, 14 The capture Agilent probes were used in previously published studies.11, 12, 14, 18, 22 Two micrograms of extracted proband gDNA was randomly sonicated into 300~500 bp fragments. The 5′ ends of DNA fragments were phosphorylated by polynucleotide kinase, and adenine was added at the 3′ ends. Then hybridization to the pre-capture libraries was quantified (the PicoGreen fluorescence assay kit, Invitrogen, Carlsbad, CA, USA). Each captured DNA libraries were applied for sequencing on Illumina HiSeq 2000 (Illumina, Inc.) at the BCM core facility, following the manufacturer's protocols, after pre-capture libraries pooled, washed and recovered (Agilent Technologies, Santa Clara, CA, USA).18 Then, paired-end sequencing illumina reads were aligned to the human hg19 reference genome using Burrows-Wheeler Aligner version 0.6.1 and available public online UCSC database (http://genome.ucsc.edu/).23 Single nucleotide polymorphisms (SNPs) and Insertions/Deletions (INDELs) variations were refined using a toolkit Atlas-SNP2 and Atlas-Indel2 (GATK version 1.0.5974).24 Variant frequency data were applied to online control databases, CHARGE consortium,25 1000 Genome Project,26 ANNOVAR,27 ESP-650028 and Exome Aggregation Consortium (ExAC) databases, to look for the pathogenic mutations in all candidate genes with a minor allele frequency of more than 5%. Sequencing depth 4, estimated copy number 2, SNP quality 20 (score 20 represents 99% accuracy of a base call), and a distance between two SNPs > 5 are considered the filtration criteria for candidate SNPs, according to previously reported studies.17, 18 Sequence variants should not annotated in any of the above public databases. Finally, we identified the variant in the affected subject (Figure 1, pedigree II: 4).

2.4 Primer design, PCR amplification and Sanger sequencing

For mutation verification and co-segregation analysis, polymerase chain reaction (PCR) amplification and direct Sanger sequencing of variant was applied to gDNA of all the available individuals.18 Locus-specific primer pair (CDHR1-1641) was designed from the online Primer3 program (http://primer3.ut.ee/) by genomic DNA sequences containing identified mutation c.T1641A in CDHR1 (Table 1). A product with 208 bps was amplified using gDNA as the template. Then, the PCR products were sequenced by Sanger method on an ABI-3500DX sequencer (Applied Biosystems Inc., Foster City, CA, USA) through the specific primer CDHR1-1641L in Table 1. All unrelated ethnical-matched controls were sequenced using aforementioned primers.

2.5 Protein structure and bioinformatic analysis

The functional classification of proteins via subfamily domain architectures for CDHR1 was performed through an online NCBI system (https://www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgi).19, 29, 30 Comparison of CDHR1 in different species was also performed by previously online NCBI system.

2.6 RNA extraction and revere transcriptional-polymerase chain reaction (RT-PCR)

RNAs from mice with indicated ages and indicated tissues were extracted using RNAsimple Total RNA kit according to our previously reported standard protocols,31 and the quality was measured by a NanoDrop-2000-spectrophotometer (NanoDrop 2000, Wilmington, DC, USA) and an agarose gel electrophoresis. Whole eye balls from embryos at 12.5 days (12 days) and 20.5 days (20 days) before birth for mice were taken instead of retina because of sampling difficulty.

Then, the total amount of RNA in each reaction system is equal to 500 ng for cDNA synthesis (reverse transcriptase/RT) using random oligomer primers method and RT kit. The primer sequences, product size and PCR conditions (anneal temperature) are listed in Table 1. Semi-quantitative RT-PCR with a 383 bps product was performed by primer pair RT-cdhr1 for mouse cdhr1 gene; ß-actin gene of mouse was served as an internal control by primer pair RT-b-actin-m (Table 1). Each assay was performed in three independent tests.

3.1 Pedigree and clinical characteristics

The proband (Figure 1, II: 4) from non-consanguineous RP family is a 45-year-old Chinese female with a clinical signs of progression of blindness characteristic of retinitis pigmentosa; she who noticed the simultaneous onset of dark adaptation difficulties, trouble with colour vision and light sensitivity by age of 24. She did not complain of nyctalopia. She was diagnosed with “optic atrophy” at local hospital at age of 27. In the fourth decade of life, visual acuity was markedly decreased, and colour vision was severely impaired. The family included 10 members and 3 generations, and all others were normal. The proband had no known family history of retinal disease through three generations, making her an isolated case and suggesting an autosomal recessive inheritance manner. The fundus photographs (FP) and fundus fluorescent photographs (FFP) of the proband II: 4 in both eyes are shown in the Figure 2. The image of FP in proband displayed attenuated vessels, absence of the foveal reflex, punctated salt- and pepper-like appearance, circular patches of RPE atrophy both at the macula and in the periphery with associated peripheral pigment migration (Figure 2A~B); further FFP results showed a hyperautofluorescent ring surrounding a central area of hypoautofluorescence and an atrophic macular region (Figure 2C~D). For comparing, Figure 2E~F shows age-matched normal control in Chinese for FP and FFP. Her electroretinogram (ERG) in different waves showed markedly reduced rod-and-cone responses (Figure 3A~F). As a result, the proband in our study was presented with typical retinal dystrophy or macular/cone–rod dystrophy (MD/CRD). Invariably, this proband patient was developmentally normal upon review at the age of 45 years old.

3.2 Next-generation sequencing analysis and putative pathogenic mutation screening

To access RP disease-causing gene mutation, targeted capture high-throughput ‎sequencing of 195 RP-related genes was performed successfully using a capture panel on the gDNA sample of proband (Figure 1, pedigree II: 4). Targeted regions with evenness scores more than 0.8 across of sample were converged. Commonly, 96.0% of the targeted regions have coverage of >20× and 91.1% of the targeted regions have coverage of >40×.18 After quality assessment, more than 97% of the billions of bases were aligned to the human reference sequences and, among those, billions of bases were recovered with a 10-fold coverage target region. Then, causative mutations were identified by automatic variant calling, filtering and annotation pipeline in the capture sequencing data.12, 14, 32 SIFT, Polyphen 2, LRT, MutationTaster, MutationAssessor and dbNSFP were used to filter out non-pathogenic population variations, which were not annotated in any of the above public databases and were prioritized for further confirmation and characterization. Surprisingly, a single nucleotide homozygous, nonsense variant (c.T1641A) of exon 15 in the CDHR1 gene (NM_033100.3) in this patient was identified, leading to an amino acid change from Tyrosine (Tyr, Y) to stop codon at position 547 of the CDHR1 protein (p.Y547*), and caused a loss of more than one-thirds of its C-terminus (CDHR1: NP_149091.1) (Figure 1 II: 4). The deleterious and pathogenic aspect of c.T1641A (p.Y547*) mutation in the CDHR1 gene is presented in Table 2. This nonsense variant c.T1641A (p.Y547*) in CDHR1 gene most likely damaged protein function in the analysis of this Chinese non-consanguineous RP family. This variant was searched in the ExAC and HGMD databases and found as a novel mutation (Table 2). Other variants in genes for CROCCP2, SLC6A6, RP1, MYO7A, RDH5, FBLN5, RLTPR and GPR179 by NGS were excluded as deleterious and pathogenic mutations.

3.3 Mutation verification and segregation analysis

Albeit deficient, the Sanger sequencing was used for confirmation and segregation analysis (Figure 4). The c. T1641A variant of CDHR1 was confirmed in the mutant homozygous type patient (pedigree II: 4; Figure 4A), and identified mutant heterozygous types in proband mother as a carrier (pedigree I: 2; Figure 4B), wild types with normal phenotype of proband's elder sister and younger brother (pedigree II: 1, II: 5; Figure 4C,D,), mutant heterozygous type with normal phenotype of proband's younger sister (pedigree II: 6; Figure 4E), and mutant heterozygous types with normal phenotype of proband's two sons (pedigree III: 1, III: 2; Figure 4F&G) in the family; the husband of proband with wild type showed no mutation (pedigree II: 3; Figure 4H). Thus, the c. T1641A variant of CDHR1 was co-segregated with the disease phenotype in all the family's members we tested. This homozygous mutant was absent in 100 unrelated, normal, ethnically matched controls (data not shown). Notably, proband's father I: 1 with normal phenotype was not available because of death, and another proband's elder sister II: 2 was not available either. But this father may carry the same variant c. T1641A by pedigree analysis (pedigree I: 1; Figure 1).

All together, these findings show complete co-segregation in the pedigree for the retinal dystrophy family and pinpoint its role in MD/CRD pathogenesis.

3.4 Functional effects of pathogenic variant c.T1641A (p.Y547*) for CDHR1 and cdhr1 mRNA expression profiles

CDHR1 structure and position for variant p.Y547* are shown in Figure 5. Searching through the Conserved Domain Database (CDD) in NCBI revealed that CDHR1 has six cadherin repeat domains (Figure 5A), which are calcium-dependent cell adhesion proteins that preferentially interact with themselves in connecting cells, and calsyntenins, which modulate calcium-mediated postsynaptic signals. Mutation of p.Y547*, located within fifth cadherin repeat domain, leads to a loss of more than one-thirds of CDHR1 at the C-terminus, including almost one and half of cadherin repeat domains (Figure 5A), possibly changing its function. Comprehensively, this study shows that recessive CDHR1 homozygous mutations, c.T1641A (p. Y547*), which most likely causes ‎disease of retinal dystrophy.‎

Expressions for cdhr1 mRNA in 15 different tissues and 6 different development stages of retina were investigated from mice. The results showed that cdhr1 transcript is very highly expressed in retina, lens, sclera, cornea of eyes and brain; very weakly expressed in skeletal muscle; had no detectable expressions in other nine tissues tested (Figure 5B); and highly expressed in the 6 different development stages/times of retinal tissue (Figure 5C). Very high expression in the retina and ubiquitous expression in different tissues from eye indicated that CDHR1 plays important roles in retina/eye functions.