Introduction

Collagen, the most abundant structural protein in mammals, is the major component of the extracellular matrix in various tissues including cartilage, skin, bone, tendon, and ligament (Fukunaga et al. 2003). So far, a total of 28 collagen types have been identified and categorized based on their structure (Ricard-Blum 2011). It is known that collagen proteins contribute to tissue integrity maintenance through interactions with other extracellular matrix proteins (Di Lullo et al. 2002). Many previous studies have reported that collagen deficiency or loss of collagen function through genetic mutation is associated with syndromic diseases such as osteogenesis imperfecta, Alport syndrome, Weissenbacher–Zweymüller syndrome (WZS), and otospondylomegaepiphyseal dysplasia (OSMED) (Barker et al. 1990; Gajko-Galicka 2002; Melkoniemi et al. 2000; Pihlajamaa et al. 1998).

Collagen type XI alpha 2 (COL11A2) also has been reported to cause several syndromic diseases such as WZS and OSMED (Melkoniemi et al. 2000; Pihlajamaa et al. 1998). However, McGuirt et al. discovered COL11A2 mutations that were associated with autosomal dominant non-syndromic hearing loss (ADNSHL) (McGuirt et al. 1999). To date, 6 different pathogenic missense mutations have been reported in patients with autosomal dominant (DFNA13) or recessive (DFNB53) non-syndromic hearing loss (Chakchouk et al. 2015; Chen et al. 2005; Choi et al. 2013; McGuirt et al. 1999). Previous studies have reported that missense mutations in COL11A2 cause non-syndromic disease, whereas nonsense or frameshift mutations lead to severe phenotypic effects, such as OSMED syndrome (Chakchouk et al. 2015; Tokgoz-Yilmaz et al. 2011).

COL11A2, located on the human chromosome 6p21.3, has four transcript isoforms due to alternative splicing, with the longest isoform (isoform 1, NM_080680.2) consisting of 66 exons (http://www.ncbi.nlm.nih.gov/). Because of large exon numbers, previous genetic studies of DFNA13 and DFNB53 have not performed mutation analysis in this gene (Ensink et al. 2001). More recently, however, a causative mutation of COL11A2 was reported for the first time in the Korean ADNSHL population through targeted sequencing (Choi et al. 2013). This suggested the potential that COL11A2 might be a main causative gene of ADNSHL in the Korean population. Nevertheless, there is no published genetic study of COL11A2 in the Korean population, yet. Additional genetic studies targeting COL11A2 are therefore required to investigate the prevalence of ADNSHL associated COL11A2 mutations in the Korean population.

In this study, we performed genetic analysis of COL11A2 in 75 unrelated Korean ADNSHL patients to evaluate its potential as a major causative gene of ADNSHL in the Korean population.

Materials and methods

Subjects

A total of 75 unrelated Korean patients with ADNSHL were recruited from the Department of Otorhinolaryngology-Head and Neck Surgery, Kyungpook National University Hospital, Daegu; Yonsei University Health System Hospital, Seoul; and Soree Ear Clinic, Seoul, Korea. We have used, clinically evaluated and reported on these subjects as part of a previous study (Ryu et al. 2016). An unrelated control group composed of 180 individuals with normal hearing was also evaluated. Written informed consent was obtained from all individuals, and the study was approved by the local ethics committee.

Genetic analysis

We designed 57 pairs of primers covering all exon and exon–intron boundaries using Primer3Plus (http://www.bioinformatics.nl/cgi-bin/primer3plus/primer3plus.cgi) to amplify the coding region of COL11A2. Genomic DNA was extracted from blood using a FlexiGene DNA extraction kit (QIAGEN, Hilden, Germany). Coding regions were amplified by polymerase chain reaction (PCR) using H-Taq DNA polymerase (Solgent, Daejeon, Korea). To confirm the PCR products, gel electrophoresis was carried out using a 1.5 % agarose gel containing ethidium bromide (EtBr). The amplified products were subsequently purified using shrimp alkaline phosphatase (USB, Cleveland, OH, USA) and exonuclease I (USB), and then sequenced using the BigDye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems, Foster City, CA, USA). After ethanol precipitation, PCR products were finally sequenced with a 3130xl Genetic Analyzer (Applied Biosystems, Forster City, CA, USA).

The nucleotide sequences were analyzed using ChromasPro v1.6 (Technelysium, Brisbane, QLD, Australia) and SeqScape software v2.5 (Applied Biosystems). These data were compared with reference sequences (NG_011589.1, NM_080680.2) registered in the NCBI database. The dbSNP (http://www.ncbi.nlm.nih.gov/snp/) and the 1000 genomes (http://www.1000genomes.org/) databases were used as references to investigate the novelty and probable pathogenicity of detected variations. Comparison with conserved protein residues among different species was performed in CLC sequence viewer v6.9 (CLC bio, Aarhus, Denmark). To predict functional pathogenic effects of the variations, 3 types of bioinformatics tools were used: Polyphen-2 (http://genetics.bwh.harvard.edu/pph2/), SIFT and PROVEAN (http://provean.jcvi.org), and Mutation Taster (http://www.mutationtaster.org/).

Results and discussion

COL11A2 mutations have been reported in American, Dutch, Finnish, Northern European, and Korean populations, among others. Currently, 5 missense mutations (p.P779L, p.G808E, p.R1043C, p.G1220D, and G1441E), 3 nonsense mutations (p.R546X, p.R1331X, and G1584X), and 5 frameshift mutations (c.1918-2A>G, c.3032-3033insC, c.3906+5G>A, c.4392+G>A, and c.2775-2801del) have been identified in patients with autosomal dominant diseases (Choi et al. 2013; McGuirt et al. 1999; Melkoniemi et al. 2000; Pihlajamaa et al. 1998; Sirko-Osadsa et al. 1998; van Beelen et al. 2012; Vikkula et al. 1995). To investigate the prevalence of hereditary hearing loss caused by COL11A2 mutations in the Korean population, we performed genetic analysis of COL11A2 in 75 Korean patients with ADNSHL. As a result, we identified 18 variants: 5 non-synonymous variants, 7 synonymous variants, and 6 intronic variants (Table 1).

Table 1 Genetic variants of COL11A2 in Korean patients with autosomal dominant non-syndromic hearing loss
Full size table

The novel non-synonymous variant p.G829R (c.2485G>C) was found heterozygously in a patient with ADNSHL (Fig. 1a). His average air-conduction level showed moderately severe hearing loss, with a calculated threshold average of 60 and 76.25 dB in the right and left ears, respectively (Fig. 1b). This variant, a nucleotide substitution of guanine with cytosine at nucleotide position 2485 in exon 33, lead to an amino acid change from glycine to arginine at amino acid position 829 in the triple-helical region domain (Fig. 1c). Interestingly, most COL11A2 mutations, including three mutations associated with ADNSHL, were located in the triple-helical region (Fig. 2). This region is mainly composed of the conserved ‘Gly–X–Y’ triple repeat sequences that are a structurally unique feature of the collagen family. Non-glycine positions, X and Y, can be any amino acids, but are often occupied by proline and hydroxyprolin, allowing this region to form a right-handed triple helix consisting of three individual collagen α-chains (Gelse et al. 2003). In this region, the glycine sites are highly conserved. Mutations in glycine sites are considered more likely to cause severe effects than mutations in X or Y positions. For example, in type I collagens (α1 and α2 chains), substitution of a cysteine residue at the X or Y position caused mild phenotypic effects, whereas substitution of glycine with cysteine led to perinatal lethality in a patient with osteogenesis imperfecta. Presumably, decreased structural stability of the triple helix and delayed secretion caused by this substitution results in excessive lysyl hydroxylation and hydroxylysyl glycosylation, leading to collagen degradation (Steinmann et al. 1986). The p.G829R also located in the triple-helical region and the glycine site was highly evolutionary conserved sequence (Fig. 1d). Additionally, it has been predicted to be probably damaging with a score of 1.000 by Polyphen-2, damaging with a score of zero by SIFT & PROVEAN, and disease causing by Mutation Taster. Nevertheless, the p.G829R variant was not co-segregated in this family. This means that the substitution from glycine to arginine at amino acid position 829 may not affect normal collagen type XI functions, despite its location in the triple-helical region and high conservation.

Fig. 1
figure 1

The missense variant, p.G829R (c.2485G>C), was discovered in a Korean family with ADNSHL. a Pedigree of the patient with p.G829R. The arrow indicates the proband of the family, and the filled symbols represent affected individuals. Patients marked with asterisks were analyzed by Sanger sequencing. b The audiograms of the proband (II-3). Red (circle) and black (cross) lines represent the threshold of the right and left ear, respectively. c Chromatograms of nucleotide sequences showing the p.G829R variant in the proband (II-3), his mother (I-2), and a normal control. A black arrow indicates the location of p.G829R. d The evolutionary conservation of amino acids in the vertebrata species. The locations of p.G829R are boxed

Full size image
Fig. 2
figure 2

Schematic representation of COL11A2 protein with signal peptide, followed by laminin G domain, non-helical region, triple-helical region, and C-terminal domain. A total of 5 missense mutations and 3 nonsense mutations were previously reported showing autosomal dominant inheritance

Full size image

Another non-synonymous variant, p.G230W (c.688G>T), was discovered heterozygously in two patients. In this variant, a change from guanine to thymine at nucleotide position 688 in exon 5 lead to substitution of a glycine with a tryptophan at amino acid position 230. Although this variant was already found in Japanese deafness patients (Miyagawa et al. 2013), there was no prediction or investigation about the pathogenicity of this variant. Additionally, this amino acid site was not conserved between other species, and this variant was detected in one of this study’s normal control. This suggests that the p.G230W variant found in Korean and Japanese patients with ADNSHL is likely not a pathogenic mutation. A third non-synonymous variant p.P1422L (c.4265C>T) found heterozygously in one patient was also a single amino acid substitution, from proline to leucine at amino acid position 1422. This non-synonymous variant was previously identified in Korean and Japanese populations (Baek et al. 2012; Miyagawa et al. 2013). However, it was reported that the proline at position 1422 was not conserved in several species and found in individuals with normal hearing (Baek et al. 2012). This suggests that the p.P1422L variant may be a probable polymorphism, and not a pathogenic mutation. The other two non-synonymous variants, p.E276K and p.P894L, were also reported in the 1000 genomes database with high minor allele frequencies (MAFs).

Autosomal dominant hearing loss is a highly heterogeneous genetic disease, with no prominent causative gene of ADNSHL existing in Korean or other East Asian populations. Although COL11A2 was analyzed in more than 100 Korean patients with non-syndromic hearing loss in this study and in previous genetic research by Choi et al., only one missense variant (p.P693L) has been identified as a probable pathogenic mutation (Choi et al. 2013). However, the dbSNP database suggests that the p.P693L variant (rs150877886) is likely to be a benign allele. The MAF of this variant is 0.004, which is higher than the general prevalence of genetic hearing loss (1 in 1000) as a whole, with 9 of 489 individuals possessing this variant heterozygously in the South Asian population (1000 genomes database). In summary, these results indicate that no certain pathogenic mutations of COL11A2 in Korean ADNSHL are known, and COL11A2 may not be a major causative gene of ADNSHL among the Korean population.