The prevalence and clinical features of MYO7A-related hearing loss including DFNA11, DFNB2 and USH1B

The MYO7A gene is known to be responsible for both syndromic hearing loss (Usher syndrome type1B:USH1B) and non-syndromic hearing loss including autosomal dominant and autosomal recessive inheritance (DFNA11, DFNB2). However, the prevalence and detailed clinical features of MYO7A-associated hearing loss across a large population remain unclear. In this study, we conducted next-generation sequencing analysis for a large cohort of 10,042 Japanese hearing loss patients. As a result, 137 patients were identified with MYO7A-associated hearing loss so that the prevalence among Japanese hearing loss patients was 1.36%. We identified 70 disease-causing candidate variants in this study, with 36 of them being novel variants. All variants identified in autosomal dominant cases were missense or in-frame deletion variants. Among the autosomal recessive cases, all patients had at least one missense variant. On the other hand, in patients with Usher syndrome, almost half of the patients carried biallelic null variants (nonsense, splicing, and frameshift variants). Most of the autosomal dominant cases showed late-onset progressive hearing loss. On the other hand, cases with autosomal recessive inheritance or Usher syndrome showed congenital or early-onset hearing loss. The visual symptoms in the Usher syndrome cases developed between age 5–15, and the condition was diagnosed at about 6–15 years of age.

To date, 882 variants have been reported for MYO7A-associated HL.Most of the reported variants were identified as genetic causes for Usher syndrome, with only 35 variants reported as causative for ADNSHL and only 49 variants reported as causative for ARNSHL 11 .Only a limited number of reports on ADNSHL and ARNSHL patients are available; thus, the detailed clinical features of ADNSHL and ARNSHL patients remain unclear.In addition, most of the previous papers report on only a few cases or only Usher syndrome cases, and the prevalence of ADNSHL, ARNSHL and USH1B in large HL cohorts is also unclear.
In this study, we reported (1) the prevalence of MYO7A-associated HL for each of ADNSHL, ARNSHL, and USH1B in a large Japanese HL cohort, (2) the detailed clinical characteristics of each set of patients including the onset age, severity of HL, progressiveness of HL, and other associated symptoms (tinnitus, vertigo, visual symptoms), and (3) genotype-phenotype correlations for variant type and clinical phenotype.

Identified variants and prevalence of MYO7A-associated hearing loss in a large Japanese hearing loss cohort
As shown in Tables 1 and 2, we identified 70 disease-causing candidate MYO7A variants.Among the 70 variants, 36 were novel variants and 34 were previously reported.Twenty-three of the identified variants were found in ADNSHL patients, with the remaining variants found in ARNSHL or USH1B patients.

Clinical features of DFNA11, DFNB2 and USH1B patients
The detailed clinical features of DFNA11, DFNB2 and USH1B patients are shown in Tables 3, 4 and 5, respectively.Family segregation analysis results were shown in Supplemental Figure 1, 2 and 3.
For DFNA11, most of the cases in this study showed late-onset progressive HL.About the half of the DFNA11 patients developed or became aware of their HL in their first or second decade; however, about half of the cases experienced HL onset after their second decade (Fig. 1A).Most of the ADNSHL patients showed mild-tomoderate and high-frequency sloping HL.Overlapping audiograms for each age group among the ADNSHL patients showed progressive HL that gradually worsened to flat-type severe-to-profound HL (Fig. 2A).In terms of development, there were no delays observed for the average month at which walking started and at which the neck was supported when sitting (4.11 months and 12.5 months, respectively).
With regard to DFNB2 cases, almost all cases showed congenital or early-onset progressive HL.The onset of HL in DFNB2 cases was in their first decade (Fig. 1B).The severity and audiometric configuration for DFNB2 varied among patients.Overlapping audiograms for each age group among the ADNSHL patients showed progressive HL, and about half of the cases over 30 years of age showed severe-to-profound HL (Fig. 2B).For DFNB2 patients, no developmental delays were observed for the average month at which walking started and at which the neck was supported when sitting (3.62 months and 13.2 months, respectively).Families #12 and #17 showed an AD family history.Patient #12 who carried p.E513Q and p.R1977Q variants, was a 56-year-old female.Her mother is 84 years old and may suffer age-related hearing loss.Unfortunately, we could not obtain detailed audiograms for the mother.However, the p.E513Q variant was identified from four other sporadic cases (family #9, #10, #11 and #13) in combination with several variants.In addition, p.R1977Q was previously reported as a genetic cause of USH1B.Thus, we concluded this case to be DFNB11.Another case was observed in family #17, where a 38-year-old male carried a homozygous p.G1159S variant.From an interview with the proband, his father also had hearing loss, but his father has already died and we could not obtain detailed information.However, a p. G1159S variant was identified from three other autosomal recessive families and three sporadic cases (family #2, #3, #15, #16, # 18 and #19).From this result, we concluded that this patient was also DFNB11.
For USH1B cases, all patients showed congenital or early-onset severe-to-profound HL (Figs. 1C and 2C).All patients aged 10 years or older were diagnosed with RP and/or complained of visual symptoms including night blindness, narrowing of visual field or both.Some patients developed visual symptoms in their first decade.One case was identified with USH1B prior to the onset of visual symptoms based on the results of genetic testing.It is noteworthy that one ARNSHL patient (#19) showed relatively late-onset RP (onset at 22 years old).However, another patient with the identical combination of variants (#20) has not been diagnosed with RP or complained of any visual symptoms at age 38.Thus, the RP observed in case #19 may have been the result of other causes.In terms of development, the average month at which the neck was supported when sitting was normal at 4.85 months, but that of the month at which walking started was significantly delayed at 20.5 months.

Genotype-Phenotype correlations observed in this study
All variants identified from ADNSHL patients were missense or in-frame deletion variants, and no null variants (nonsense, splicing and frameshift variants) were observed (Tables 1 and 3).In ARNSHL or USH1B patients, both missense variants and null variants were identified (Tables 2, 4 and 5).As shown in Table 4, the combination of the identified variants for all ARNSHL patients carried at least one missense variant, and there were no cases with biallelic null variants.On the other hand, almost half of the patients with Usher syndrome carried biallelic null variants (Table 5).
Almost half of the missense variants for ADNSHL and ARNSHL were located in the Myosin head domain, with the few exceptions located in MyTH4 domain, SH3 domain or FERM domain.On the other hand, the majority of missense variants in USH1B families were located in the posterior half region including the MyTH4         www.nature.com/scientificreports/domain, Band4.1 domain and FERM domain.Thus, we estimated that the HL caused by variants located in later half tended to be more severe than that caused by the variants located in the Myosin head domain.

Intervention for HL and outcomes
Most of the ADNSHL or ARNSHL patients with mild-to-severe HL used hearing aids, although some received cochlear implants.On the other hand, almost all Usher syndrome patients showed severe-to-profound HL and most of them received cochlear implants in childhood.The outcome of cochlear implantation (CI) was favorable (Fig. 3), indicating that CI affords a good treatment option for the patients with severe-profound MYO7Aassociated HL in all hereditary forms.

Discussion
In this study, we showed the prevalence of MYO7A-associated HL in a large HL cohort was 1.36% (137/10,047).The prevalence of MYO7A-assocciated ADNSHL in autosomal dominant or maternal inheritance HL patients was 4.06% (91/2,243), while it was 0.38% (25/6,163) for ARNSHL patients and 0.32% (21/6,163) for MYO7Aassociated Usher syndrome cases among autosomal recessive or sporadic HL patients.This is the first paper reporting the prevalence of all three clinical phenotypes of MYO7A-associated HL identified from a single large cohort.Most previous papers reported only on Usher syndrome cases or a limited number of ADNSHL or ARNSHL patients, so the prevalence as well as the detailed clinical characteristics for ADNSHL and ARNSHL MYO7A-associated HL has been unclear.In previous papers, Sloan-Heggen et al., reported the genetic analysis results for 1119 hearing loss patients (Caucasian, Hispanic, African American, Asian, Middle Eastern, Ashkenazi Jewish and others), with the prevalence of MYO7A-associated HL being 1.79% (20/1119 cases) 12 .Among the 20 cases, one case was DFNA11 (0.7%, 1/141 autosomal dominant HL cases), three cases were DFNB2 (0.36%, 3/830 There was a tendency for onset age to differ depending on variant type.About half of the DFNA11 patients developed or became aware their HL in their first or second decade, whereas HL onset in about half was after twenty years of age.Almost all DFNB2 cases showed congenital or early-onset progressive HL.As for USH1B cases, all patients showed congenital or early-onset severe-to-profound HL. autosomal recessive HL or sporadic cases), and 16 cases were USH1B (1.68%, 14/830 autosomal recessive HL or sporadic cases, one case identified as AD, and one case identified with an unknown family history).Ma et al. 13 , reported the NGS analysis results for 879 Chinese HL patients, with the prevalence of MYO7A-associated HL being 2.39% (21/879 cases).Eleven cases were DFNA11 (6.88%, 11/160 autosomal dominant HL cases) and 10 cases were DFNB11 or USH1B (1.70%, 10/589 autosomal recessive HL or sporadic cases).Abu Rayyan et al. 14 , reported NGS analysis results for 491 Palestinian families with HL, and identified 28 MYO7A-associated HL cases (5.7%, 28/491).All 28 cases in their report were DFNB11 or USH1B.Baux et al. 15 , reported the NGS analysis results for 207 French hearing-impaired patients, and identified 5 cases with MYO7A-associated HL (2.4%, 5/207).The prevalence of MYO7A-associated HL in previous reports varied depending on the sample number and clinical characteristics of the cohort, ranging from 1.79 to 5.7%.The prevalence of MYO7A-associated HL in this study was 1.36%, which was similar to those in previous reports.There are several papers reporting NGS analysis results for hearing loss patients; however, many of them pre-screened GJB2-or SLC26A4-associated HL cases, making it difficult to estimate the true prevalence of MYO7A-associated HL.Our results will shed light on the detailed clinical features of HL, especially for the MYO7A-associated ADNSHL and ARNSHL patients.Interestingly, some of the identified variants from ADNSHL patients in this study were commonly observed even in different families.This result suggested that a common founder mutation or mutational hot spots are associated with this multi-familial identification.The fact that most of the commonly identified variants in this study are only identified from Japanese HL patients or from Japanese and East Asian HL patients supports the notion that these variants were founder mutations.
As for the clinical characteristics of ADNSHL patients, all patients showed delayed onset HL that usually developed after language acquisition (post-lingual onset).In addition, we clearly observed progressive highfrequency HL in a large number of patients.These clinical characteristics were consistent with previous reports 16 , but the patient number in this study is the largest to date and these findings will be useful to our understanding of the clinical characteristics of MYO7A-associated ADNSHL.Most of the DFNA2 variants identified in this study were novel, with few previously reported.In a previous study, Sang et al., reported a DFNA11 family with a c.2003G>A variant 17 .The clinical characteristics for this family were mild-to-severe progressive HL with an onset age of 17-45 years.The clinical characteristics for DFNA2 cases in this previous report were consistent with those in our cases.
Congenital or early-onset HL was observed in patients with MYO7A-associated ARNSHL.Most of the USH1B cases showed congenital severe-to-profound HL and first-or second-decade onset RP, which is consistent with previous reports 18 .Unlike DFNA11 variants, many of the DFNB2 and USH1B variants identified in this study were previously reported.The clinical phenotypes for these autosomal recessive cases were dependent on the combination of variants.In previous reports, most DFNB2 cases showed first decade onset severe-to-profound HL 19,20 , and most of USH1B cases showed congenital severe-to-profound HL, with RP diagnosed in the firstto-second decade [21][22][23] .
With regard to genotype/phenotype correlations, all DFNA11 patients carried non-truncating variants (missense or in-frame deletion), and all DFNB2 patients carried at least one non-truncating variants with both the truncating or non-truncating variant in the trans configuration.On the other hand, about half of the USH1B patients carried biallelic truncating variants.These results suggested that the pathogenic mechanism for MYO7Aassociated ADNSHL might be a dominant negative effect, whereas the pathogenic mechanism for MYO7Aassociated ARNSHL and USH1B might be a loss of function.Further, the residual function of each MYO7A variant is thought to be associated with the phenotypic differences in ARNSHL and USH1B.
A similar situation was reported for CDH23-associated HL, which is caused by ARNSHL (DFNB12) or Usher syndrome (USH1D).The DFNB12 phenotype is reported to be associated with biallelic missense mutations, whereas the USH1D phenotype is associated with presumably functional null alleles, including nonsense, splicing, frameshift, or some missense mutations 24 .In addition, it has been reported that patients with compound heterozygous variants of USH1D and DFNB12 show a non-syndromic phenotype 25 .
It is difficult to distinguish between ARNSHL and USH1B, particularly in younger patients, as RP in USH1B patients develops between age 5-15 years and it is difficult to identify RP in younger patients without any visual symptoms.Therefore, it is impossible to strictly distinguish ARNSHL patients from USH1B patients.Thus, some of ARNSHL patients identified in this study may develop visual symptoms later.Of course, follow-up is important for the identification of visual symptoms even in ARNSHL cases.In this study, we collected data for "the month at which the head is supported when sitting" and "the month at which walking started" as indirect evidence with which to elucidate the vestibular dysfunction that characterizes USH1B.Most of the USH1B cases showed delays in balance when sitting and/or walking but there were few cases with such delays among the ADNSHL and ARNSHL patients.Furthermore, the average month at which walking started was delayed in USH1B cases.Loundon et al., reported the average age at which walking starts for USH1B cases was 20 months, suggesting vestibular problems.Our results for the start of walking for USH1B cases showed a significant delayed at 20.5 months, which is consistent with the previous report 26 .
Additionally, we also identified another genotype-phenotype correlation in this study.Almost half of the missense variants in ADNSHL and ARNSHL patients were located in the Myosin head domain, with few located in the MyTH4 domain, SH3 domain or FERM domain.On the other hand, the majority of missense variants in the USH1B families were located in the posterior half region including the MyTH4 domain, Band4.1 domain and FERM domain.Joo et al. 27 , reported that MYO7A-asssociated ADNSHL patients with Myosin head domain variants showed relatively milder HL than did patients with variants in the MyTH4 domain, which appears to support our findings.
In conclusion, next-generation sequencing analysis successfully identified 34 previously reported variants and 36 novel variants in MYO7A-associated HL patients.The estimated prevalence of MYO7A-associated hearing loss in the Japanese hearing loss cohort was 1.36% for all patients, 4.06% for ADNSHL among autosomal dominant or maternal inheritance cases, 0.38% for ARNSHL and 0.32% for USH1B among autosomal recessive or sporadic hearing loss cases.This large cohort study of hearing loss patients provided valuable new insights, particularly with regard to hearing deterioration in MYO7A-associated ADNSHL patients.This information is expected to be useful for the provision of more appropriate intervention for MYO7A-associated HL patients.In addition, understanding the gene involved in hearing loss also opens up possibilities for new future therapies (such as gene therapy).

Subjects
A total of 10,042 HL patients from 102 institutions across Japan participated in this study.Clinical information and peripheral blood samples were obtained from patients and their relatives.Written informed consent was obtained from all patients (or from their next of kin, caretaker, or legal guardian in the cases of minors or children) and relatives.This study was approved by the Shinshu University Ethical Committee, as well as the Vol:.(1234567890

Next-generation sequencing and bioinformatic analysis
Next-generation DNA sequencing was performed for the 63 target genes reported to cause non-syndromic hearing loss as described in a previous report 28 .In brief, amplicon libraries were prepared using the Ion AmpliSeq Custom Panel, with the Ion AmpliSeq Library Kit 2.0 and the Ion Xpres Barcode Adapter 1-96 Kit (ThermoFisher Scientific) according to the manufacturer's instructions.After amplicon library preparation, emulsion PCR, and next-generation sequencing were performed with an Ion 200 sequencing kit (ThermoFisher Scientific) and Ion PGM sequencer (ThermoFisher Scientific) or an Ion HiQ chef Kit (ThermoFisher Scientific) and Ion Proton sequencer (ThermoFisher Scientific) according to the manufacturer's protocol.The sequence data were mapped against reference human genome sequence (build GRCh37/hg19) with the Torrent Mapping Alignment Program (TMAP).The DNA variants were detected with the Torrent Variant Caller plug-in software (ThermoFisher Scientific).Copy number variation analysis was also performed for all patients by using read depth data according to the copy number variation detection methods described in our previous report 29 ; however, no copy number variations for the MYO7A gene were identified.After variant detection, annotation of identified variants was performed with ANNOVAR software.The missense, nonsense, insertion, deletion, and splicing variants were selected from among the identified variants.For the variants located in the exon-intron border region (including synonymous variants and intronic variants), candidate variants predicted to affect splicing by in silico splicing prediction dbscSNV 30 were also selected.
Variants were further selected as less than 1% of several control population database including the 1000 genome database 31 , the 6500 exome variants 32 , The Genome Aggregation Database 33 , the 1200 Japanese exome data in Human genetic variation database 34 , the 38,000 Japanese genome variation database 35 and the 333 inhouse Japanese normal hearing controls.All filtering procedures were performed using original database software described previously 36 .
The pathogenicity of identified variants was analyzed in accordance with the American College of Medical Genetics (ACMG) standards and guidelines 37 with the ClinGen hearing loss clinical domain working group expert specification 38 .Variants were defined as candidate variants if the following criteria was fulfilled; (1) for the variants previously reported as "pathogenic" or "likely pathogenic" without any contradictory evidence, we applied the same pathogenicity classification, (2) novel variants classified as "pathogenic" or "likely pathogenic" were considered as strong candidates for each case, (3) variants of "uncertain significance" (VUS) identified as only one candidate after the filtering procedure without any candidate variants in the other 62 genes were also included, (4) two variants found in recessive inheritance cases, and (5) there was no contradiction with the family analysis.
We performed Sanger sequencing analysis to validate the identified variants according to the manufacturer's instructions.All PCR and sequencing primers were designed using the web version Primer 3 plus software 39 .

Figure 1 .
Figure 1.Onset age of hearing loss in patients with (A) DFNA11 variants, (B) DFNB2 variants, and (C) USH1B variants.There was a tendency for onset age to differ depending on variant type.About half of the DFNA11 patients developed or became aware their HL in their first or second decade, whereas HL onset in about half was after twenty years of age.Almost all DFNB2 cases showed congenital or early-onset progressive HL.As for USH1B cases, all patients showed congenital or early-onset severe-to-profound HL.

Figure 2 .Figure 3 .
Figure 2.Overlapping audiograms for each age group for each hereditary form.(A) Overlapping audiograms for each age group among DFNA11 patients showed progressive HL that gradually worsened from mild-tomoderate and high-frequency sloping HL to flat-type severe-to-profound HL.For (B) DFNB2 cases, their overlapping audiograms also showed progressive HL and about half of the cases over 30 years of age showed severe-to-profound HL. (C) USH1B patients showed congenital severe-to-profound HL.
. If the patients couldn't have puretone audiometry due to their age (approximately aged four or under) or other reasons, the auditory steady-state response (ASSR), conditioned orientation response audiometry (COR) or play audiometry were performed.The outcome of interventions (HA or CI) was evaluated by PTA.https://doi.org/10.1038/s41598-024-57415-1

Table 5 .
Clinical characteristics of MYO7A-assocciated Usher syndrome patients (USH1B).All variants are indicated on NM_000260.HA, hearing aid; CI, cochlear implant; AD, autosomal dominant; Spo, sporadic; M, male; F, female; N/A, data not available; Rt., right; Lt., left; Bi., bilateral; CP, canal paresis.**This table also includes the patients under 5 years old who carried previously reported biallelic USH1B variants without visual symptoms if they showed delays in sitting and/or walking.
Vol.:(0123456789) Scientific Reports | (2024) 14:8326 | https://doi.org/10.1038/s41598-024-57415-1 respective ethical committees of the other participating institutions, and was conducted in accordance with the Declaration of Helsinki.The estimated inheritance pattern was classified into "autosomal dominant or maternal inheritance" with two or more generations of family members suffering hearing loss, "autosomal recessive" with two or more siblings suffering hearing loss, and "sporadic" in cases without any affected family members.