Background
MYO15A, a causative gene of DFNB3 (OMIM 600316) [
1], is a frequently detected deafness gene. Friedman et al., in Bali, first reported that the frequency of autosomal recessive hearing loss caused by the
MYO15A pathogenic variant was about 2% [
2]. Thereafter, the frequency of
MYO15A pathogenic variant was reported in up to 9.9% of deafness cases in Turkey [
3]. In our previous study,
MYO15A pathogenic variant was reported at a frequency of 2.1% in nonsyndromic autosomal recessive deafness, which was the fourth most common deafness gene in Korea, following
SLC26A4, GJB2 and
CDH23 [
4]. Sequentially, several studies have been performed to investigate the effects of
MYO15A pathogenic variants on hearing loss [
5‐
9].
The role of myosin XVA, which is encoded by
MYO15A, includes the graded elongation and maintenance of stereocilia and actin-organization in the inner ear hair cells. These are both essential for normal auditory function. Therefore,
MYO15A pathogenic variants were initially thought to induce congenital severe-to-profound hearing loss [
1,
10‐
15]. However, it was later discovered that the phenotypes of
MYO15A pathogenic variants varied depending on the affected domain. The variation of phenotypes according to the affected domain has been explained by the existence of two isoforms -- a class 1 isoform with the N-terminal domain, which is encoded by exon 2, and a class 2 isoform with no N-terminal domain [
16,
17]. The pathogenic variant in N-terminal domain affects only the class 1 isoform without affecting the class 2 isoform [
18]. The class 2 isoform is present in the human inner ear [
17]. Therefore, pathogenic variants in the N-terminal domain are known to cause minor deficiencies in the inner ear, resulting in amilder auditory phenotype, when compared with other pathogenic variants in
MYO15A [
5‐
7]. Conversely,
MYO15A pathogenic variants that reside in the regions shared by both isoforms are known to cause congenital or prelingual severe-to-profound hearing loss [
7].
Recently, Naz et al. reported that
MYO15A pathogenic variants, which had previously been thought to only cause profound hearing loss, may cause moderate-to-severe hearing loss [
9]. We also found two families with
MYO15A pathogenic variants in the motor and FERM domains, which were expected to cause profound hearing loss. They showed postlingual onset of bilateral symmetrical, partial deafness with significant residual hearing at low frequencies. Based on these results, we suggest that
MYO15A may be a causative gene responsible for the postlingual onset of progressive partial deafness, which in turn requires the expansion of the phenotypic spectrum of
MYO15A pathogenic variants.
Discussion
Although TES of 129 genes in the deafness panel is an efficient and convenient tool for detecting known causative variants, pathogenic variants from a novel deafness gene may not be detected using only TES. To explore and minimize such a possibility, we performed genetic testing via two diagnostic pipelines, TES and WES, in a parallel fashion for the SB246 family. Filtering of candidate variants through these two pipelines indicated the same result, strongly supporting our molecular diagnosis. Moreover, certain genes, such as
OTOF and
STRC, have been reported not to be fully covered by next generation sequencing [
29,
30]. To overcome this, we modified the mapping quality of bioinformatics analysis. Furthermore, we excluded the possibility of structural variations involving
STRC and
CATSPER2, which was reported to be an important molecular etiology of SNHL in Japan [
30]. In the SB224 family, two variants, c.9790C > T (p.Q3264X) and c.10263C > G (p.I3421M), of
MYO15A survived at the final filtering step of TES analyses; c.9790C > T was a nonsense variant causing truncation of the protein, and c.10263C > G has already been reported to be pathogenic. Therefore, it is most likely that hearing loss from SB224–437 was attributed to these two
MYO15A variants. Sequentially, WES was not performed in SB224.
The conservation and pathogenicity prediction study also strengthened our hypothesis that
MYO15A pathogenic variants, c.5504G > A (p.R1835H) and c.10245_10247delCTC (p.S3417del), in the SB246 family and c.9790C > T (p.Q3264X) and c.10263C > G (p.I3421M), in the SB224 family were the causative pathogenic variants of hearing loss. Among the four
MYO15A pathogenic variants discovered here, c.5504G > A and c.9790C > T were novel variants; c.9790C > T was a nonsense pathogenic variant, and c.5504G > A resided in the motor domain. The motor domain is one of the most important domains in myosin XVA. Sequentially, it is reasonable to infer that variants in the motor domain may lead to profound hearing loss. c.10245_10247delCTC and c.10263C > G resided in the FERM domain; they have already been reported to cause congenital profound hearing loss [
31‐
33]. Therefore, four variants detected in this study were expected to cause profound hearing loss; however, they were associated with partial deafness with significant residual or even near-normal hearing at low frequencies.
The establishment of genotype-phenotype correlation is one of the fundament goals of genetics, as it enables personalized and timely management of diseases, leading to great contribution to precision medicine. Personalized and timely auditory rehabilitation is crucial in hearing loss because there is a critical time window for auditory development and degeneration. Therefore, an establishment of genotype-phenotype correlation is a meaningful issue in hearing loss. Sequentially, genotype-phenotype correlations of several genes and their pathogenic variants have been investigated widely. However, genotype-phenotype correlations have generally been based on the observation of phenotypes, rather than prudent analysis of pathogenic mechanism. Previously established genotype-phenotype correlations could be modified. Recently, several genes, including
MYO15A,
CDH23, and
PTPRQ, have been reported to cause different types of hearing loss from those previously postulated in the literature [
9,
31].
MYO15A is one of the genes with a well-documented genotype-phenotype correlation. It has been reported that hearing loss phenotype related to
MYO15A is different according to the affected domain. Specifically, pathogenic variants in the N-terminal domain of
MYO15A have been suggested to be associated with residual hearing [
5‐
8], especially at low frequencies, while pathogenic variants in other domains resulted in congenital severe-to-profound hearing loss [
1,
10‐
15]. With the progress of various auditory rehabilitation technologies, appropriate auditory rehabilitation tailored to the type of hearing loss should be implemented. A moderate hearing loss can be fully rehabilitated with hearing aids; however, cochlear implantation may be necessary in cases of severe hearing loss. Moreover, for subjects with overall profound hearing loss, while retaining significant residual hearing at low frequencies, electroacoustic stimulation (EAS) may be the best option. Therefore, a prediction of phenotypes in accordance with the affected domain in
MYO15A has greatly contributed to personalized and timely auditory rehabilitation. Recently, these previously established genotype-phenotype correlations appear to require an updated modification. Some
MYO15A pathogenic variants affecting the domains other than the N-terminal, have shown to cause moderate-to-severe hearing loss, not profound hearing loss [
9]. We also found two families carrying
MYO15A pathogenic variants, which were expected to cause congenital severe-to-profound hearing loss, but resulted in postlingual onset of progressive partial deafness with residual hearing at low frequencies. Indeed, SB224–437 carrying two
MYO15A mutant alleles in this current study had a shallower angular insertion depth of 393°, by Flex 24 rather than by the longer Flex 28 electrode to minimize the shift in low frequency thresholds from this subject (Fig.
1b).
A milder phenotype may have been influenced by factors like milder pathogenic potential from hypomorphic alleles of MYO15A, genetic modifiers that reduce severity of hearing loss, or environmental factors. A recent progress in the genetic diagnosis technique can also contribute to the expansion of phenotypic spectrum of MYO15 alterations. In the past, genetic hearing loss had been found in consanguineous families using a linkage analysis, especially homozygous mapping. Therefore, it is likely that the severe phenotypes caused by severely pathogenic homozygous pathogenic variants had been preferentially recruited for the genetic study. However, with the development of next generation sequencing, there has been an increase in the frequency of molecular diagnostic testing of small-to-mid sized, non-consanguineous families, leading to the emergence of many compound heterozygotes with varying degrees of pathogenicity. Consequently, it has become possible to discover various phenotypes from the existing genes. Further studies are needed about other potential factors affecting phenotypes, such as genetic modifiers or environmental factors.