EYA4 is located at 6q23.2, contains 21 exons, encoded from exon 2, a total of 640 amino acids, and contains two structural functional domains: The two domains, eya variable region (eyaVR) and eya homologous region (eyaHR), each having different functions. EyaVR is located at the N-terminus and functions as a transcriptional activator and a threonine phosphatase. EyaHR is located at the C-terminus, on the one hand, eyaHR is involved in the formation of co-transcription factors into the nucleus, on the other hand, eyaHR also has tyrosine phosphatase activity[3].
As a transcriptional coactivator, EYA4 plays a key regulatory role in the development of multiple organs, including the eye, muscle, ear, heart, lung, endocrine glands, placode, pharyngeal pouch, craniofacial skeleton, and parathyroid gland [4]. It has two clinical phenotypes: (1) autosomal dominant deafness type 10 (DFNA10), a non-syndromic sensorineural hearing loss. (2) Dilated cardiomyopathy type 1 J (OMIM: 605362): AD: Dilated cardiomyopathy (DCM) is a disease characterized by ventricular enlargement and cardiac systolic dysfunction, which leads to difficult cardiac congestion and irregular heartbeat, and patients are at risk of sudden death. The clinical phenotype in this study was DFNA10, and none of the families had any heart-related diseases.
In this study, whole-genome sequencing (WGS) and Sanger sequencing were used to comprehensively analyze the genetic characteristics of a Chinese family with autosomal dominant NSHL. One heterozygous mutation, c.1745_1748del, was found in EYA4, resulting in a frameshift mutation (p.Glu582ValfsTer6). According to ACMG guidelines, the variant was initially determined to be a likely pathogenic PVS1 + PM2_Supporting: PVS1. This variant is a null variant (frameshift mutation) that may lead to loss of gene function. PM2_Supporting: The frequency in the normal population database was. This mutation has not been previously reported or found in the ClinVar database. Co-segregation of the gene variant and the phenotype of the family were confirmed by pedigree verification analysis.
EYA4 is not a common pathogenic gene in ADNSHL, and most mutations in EYA4 have been identified in Asian populations. To date, 15 mutations have been identified in Japan, 10 in China (including the mutation reported in this study), and four in Korea. In Europe, most mutations have been reported in Spain, where eight types have been identified (Fig. 3, Table 1). No hotspot mutations in EYA4 were found, and most mutations did not recur. Figure 4 and Table 1 summarize the mutation profiles of EYA4 reported to date. Among the 52 pathogenic mutations, there were 17 missense mutations, 12 frameshift mutations, nine nonsense mutations, six splicing mutations, and eight copy number variations (CNV) (Fig. 4). The novel heterozygous variant c.1745_1748del, which is a 4-bp deletion that causes a frameshift mutation, was found in exon 19 of the EYA4 gene. This variant is a null variant that may lead to loss of gene function.
In a large-scale genetic screening for deafness in Japan, the prevalence of EYA4-related hearing loss in ADNSHL was 0.90% (121334cases) in the Japanese population[5]. By screening 531 Spanish familial cases for mutations, Morin[6] identified eight variants of the EYA4 gene and concluded that the global prevalence of EYA4 in the DFNA10 hearing-impaired population in Spain was approximately 1.5% (8 of 531). Kim et al. [7]identified four pathogenic EYA4 mutations in 87 patients with ADNSHL in South Korea. In contrast, Choi et al.[8] reported a case of DFNA10 deafness caused by an EYA4 mutation in 14 Korean families with ADNSHL and moderate hearing loss. They therefore estimated a prevalence of 4.6–7.14% in Korea. In a genetic screening of 116 cases of sensorineural hearing loss in the Chinese Han population, only one EYA4 mutation was found, with a prevalence of 0.86%[9].
The audiological features of EYA4 mutations are highly heterogeneous, with identical or similar mutations causing different clinical phenotypes and variants causing similar clinical phenotypes[10]. The main forms of hearing loss are late-onset, post-language, progressive, and bilateral symmetrical sensorineural deafness[5]. The most commonly reported hearing loss is mild to moderate mid-frequency hearing loss at onset, which is called "Cookie-bite" audiogram. Subsequent hearing loss also affects high frequencies, resulting in moderate to severely flat or downward-sloping audiograms[11]. The proband in this study showed mild-to-moderate hearing loss with medium- and high-frequency descending hearing loss. His acoustic immittance was of the bilateral As type, and DPOAE was not detected bilaterally, indicating that EYA4 gene variation affected the eustachian tube function and cochlear function of the patient. At the same time, the decline in speech recognition rate indicated that the patient's auditory function was affected.
It was initially proposed that truncation of the C-terminal end of the EYA4 domain causes non-syndromic ADSNHL, whereas upstream truncation of the N-terminal variable region causes dilated cardiomyopathy-type hearing loss[12]. However, this was subsequently challenged, as several families have been reported to have truncating mutations in the variable region, without a DCM phenotype (Fig. 4, Table 1). Many studies have shown that large deletions in EYA4 have no significant genotype-phenotype correlation with dilated cardiomyopathy[13]. Shinagaw[5] compared hearing thresholds in patients with truncating variants to those in patients with non-truncating variants (missense mutations). Patients with truncating variants exhibited flat hearing loss, whereas those with non-truncating variants exhibited high-frequency hearing loss. Zhang[14] summarized the current type of EYA4 mutation and degree of hearing loss in 2020, suggesting that the severity of hearing loss is not related to the type or location of the variant. Kim et al. [7] suggested that there was no genotype-phenotype correlation in EYA4-related hearing loss. Based on the genotype-phenotype analysis of EYA4 in this case and in previous reports, we believe that the genotype-phenotype correlation of EYA4 is unclear.
The precise molecular pathological mechanism by which EYA4 mutations cause hearing impairment has not been defined but may be related to reduced gene dosage or reduced protein activity[15]. It has been shown that EYA4 gene-deficient mice develop hereditary otitis media, suggesting that EYA4 regulation is essential for the development and function of the middle ear cavity and Eustachian tube[16]. EYA4 may be involved in the development and maintenance of hair cells in the inner ear through regulation of Na+ /K+ ATPase[17]. Similarly, EYA4 mutations impair the maintenance of ciliated cells that line the inner wall of the eustachian tube, resulting in eustachian tube defects and conductive hearing loss[18]. The novel EYA4 exon 19 mutation we report here, c.1745_1748del, similar to previously studied EYA4 mutations, is expected to affect the eyaHR domain, which is essential for the continued function of the mature organ of Corti after development.
In conclusion, a novel EYA4 mutation, c.1745_1748del (p.Glu582ValfsTer6), was identified in a Chinese family with non-syndromic ADNSHL using whole genome sequencing (WGS) and Sanger sequencing. This expands the spectrum of variants in the EYA4 gene and autosomal dominant non-syndromic hearing loss in the Chinese population and emphasizes the importance of genetic testing in patients with familial progressive sensorineural hearing loss for earlier intervention. This study provides a reference for clinical genetic counseling and prenatal diagnosis and lays a foundation for gene therapy research.