Branchio-oto-renal syndrome-1 (BOR1; OMIM#113650), also known as Melnick–Fraser syndrome, is a rare autosomal dominant affected family members disease with an incidence of approximately one in 40,000; it affects 2% of profoundly deaf children (Melnick et al., 1975; Fraser et al., 1980). BOR exhibits a variable spectrum of clinical manifestations that are mainly characterized by the presence of branchial cleft fistulae or cysts, preauricular pits, ear malformations, and hearing loss, along with renal malformations of varying severities (Fraser et al., 1978). In some instances, patients exhibit symptoms similar to those of BOR, with the exception of renal anomalies; they are diagnosed with branchio-oto syndrome-1 (BOS1; OMIM#602588) or branchio-oto syndrome-3 (BOS3; OMIM#608389). The clinical diagnosis of BOR/BOS follows a set of criteria proposed by Chang et al. (2004). Diagnosis of BOR/BOS can be made with at least three major criteria, two major and at least two minor criteria, or one major criterion with at least one first-degree affected family member.

BOR/BOS has marked genetic heterogeneity, and the exact pathogenesis remains unknown in more than half of affected patients. Three genes are currently considered to be associated with this condition: EYA1 (OMIM #601653), SIX1 (OMIM #601205), and SIX5 (OMIM #600963) (Abdelhak et al., 1997; Ruf et al., 2004; Hoskins et al., 2007). Pathogenic variants in the EYA1 gene, the human homolog of the Drosophila “eyes absent” gene, were recognized as a major genetic cause of BOR/BOS; they affect 40% of BOR/BOS patients (Chang et al., 2004). The SIX1 gene, also known as sine oculis homeobox homolog 1, encodes a transcription factor (Six1) that functions as a DNA-binding protein in combination with Eya1, leading to 3.0–4.5% of BOR/BOS cases (Ruf et al., 2004; Wang et al., 2018). The role of the SIX5 gene, the sine oculis homeobox homolog 5, in BOR/BOS is controversial; few variants have been reported (Hoskins et al., 2007; Krug et al., 2011). The EYA1 and SIX1 genes are co-expressed in the developing kidney and ear, beginning during the emergence of basal plates (Xu et al., 1999). Eya1 is epistatic to Six1, and its function is dependent on interaction with Six1. Eya1 does not have direct DNA-binding ability. Instead, it functions as a transcription co-activator and interacts with Six1, thus providing a molecular mechanism for activation of specific target genes that modulate precursor cell proliferation, survival, and differentiation during multiple types of organogenesis (Li et al., 2003; Ruf et al., 2004). In the absence of this interaction, the transcriptional activation of downstream targets required for the development of the branchial, otic, and renal systems is diminished. Functional studies have confirmed that genetically defective EYA1 and SIX1 mice exhibit symptoms similar to BOR/BOS (Xu et al., 1999; Ando et al., 2005).

BOR/BOS is a common form of syndromic hearing loss. Hearing impairment is the most common clinical feature, present in 98.5% of affected individuals; other common features include preauricular pits or tags (83.6%), branchial fistulae or cysts (68.5%), renal anomalies (38.2%), and external auditory canal stenosis (31.5%) (Chang et al., 2004). The forms of hearing loss can be mixed (50%), conductive (30%), or sensorineural (20%), ranging in severity from mild to profound (Abdelhak et al., 1997). Most patients with BOR/BOS present with morphological abnormalities of the middle and inner ear (Chen et al., 1995; Propst et al., 2005), prompting surgeons to explore auditory rehabilitation modalities. Some studies have used middle ear exploratory tympanotomy and ossicular reconstruction to improve mixed and conductive hearing difficulties. However, the postoperative hearing gains have been unsatisfactory (Propst et al., 2005; Song et al., 2013). With the rise of auditory implant technology, cochlear implantation (CI) has provided satisfactory results in some patients with syndromic deafness (e.g., BOR/BOS) (Bajaj et al., 2012).

Here, we screened for EYA1, SIX1, and SIX5 variants in patients with BOR or BOS and then investigated the affected individuals’ clinical manifestations and auditory rehabilitation outcomes. In addition, we analyzed the genotypes and phenotypes of BOR/BOS caused by EYA1 variants through a literature review.

This study described nine affected individuals in five Han Chinese families with BOR or BOS. Of the nine patients, all had hearing loss (9/9) and preauricular pits (9/9), five had branchial fistulae or cysts (5/9), and only one had renal abnormalities (1/9), indicating that BOS is the major form. Genetic analysis confirmed five causative heterozygous variants in these unrelated families: c.1425delC (frameshift), c.889C > T (nonsense), c.1050+1G > T (splicing), and c.1140+1G > A (splicing) of EYA1, as well as c.316 > A (missense) of SIX1. To our knowledge, EYA1:c.1425delC and SIX1:c.316 > A have not been previously reported; c.316G > A is also the first reported SIX1 variant in China. Our findings expand the mutational spectrum of BOR/BOS-related genes and demonstrate genetic and phenotypic variability in Chinese BOR/BOS patients.

EYA1 and SIX1 are key genes for mammalian organogenesis; mutations in these genes result in multiorgan malformation that can affect the branchial, ear, and renal systems. In 1997, Abdelhak and others reported novel variants in the EYA1 gene in several families demonstrating typical manifestations of BOR/BOS (Abdelhak et al., 1997). Since then, more than 200 variants have been identified (Figure 3). The EYA1 gene, located at 8q13.3, is a member of the EYA (EYA1–4) family; it contains 18 exons encoding a dual-function transcription factor. Similar to other EYA family members, Eya1 possesses a divergent N-terminal transactivation domain and a C-terminal EYA dephosphorylation region (eyaHR) encoded by 11–18 exons (Abdelhak et al., 1997; Li et al., 2003). Known variants in EYA1 mainly involve exons 6–18, especially exons 11–18, at least for now (Figure 3). These variants mainly cause premature truncation or aberrant splicing of the protein, resulting in the loss of eyaHR function. Screened EYA1 variants in this study were predicted to interfere with this critical function. The protein Six1 encoded by the SIX1 gene, a co-factor of Eya1, contains two evolutionarily conserved domains: a SIX domain that interacts with its co-factors and a homeodomain with DNA-binding ability (Ruf et al., 2004). Of the defects in the SIX1 gene, Six1 functional alterations led to the failed formation of the Six1–Eya1 complex and the DNA–Six1–Eya1 complex (Ruf et al., 2004). Thus far, 14 SIX1 variants are reportedly associated with BOR/BOS, eight of which are in the SIX domain (Figure 3). We also identified a novel missense variant (c.316G > A) in the SIX1 gene; we speculate that this variant interferes with Six1 binding to Eya1, which has been previously reported for variants in the SIX domain (Ruf et al., 2004; Kochhar et al., 2008). Loss of function of either gene in a mouse model led to hearing loss and dysmorphic or missing kidneys, along with developmental abnormalities in other organs (Xu et al., 1999; Li et al., 2003). Our patients harboring EYA1 or SIX1 alterations also had hearing loss, ear deformities, or kidney problems.

BOR/BOS patients demonstrate intrafamilial and interfamilial phenotypic variability in the clinical setting, suggesting a lack of genotype–phenotype correlation. In the present study, proband 5-II-2 harboring c.1425delC exhibited all major symptoms of BOR, while patients 5-II-1 and 5-I-1 only presented with the triad of BOS, demonstrating coexistence of BOR and BOS in this family. The presence of interfamilial phenotypic variability has also been confirmed. A patient from China with the same c.1140+1G > A variant demonstrated no renal involvement; patients from Korea reportedly lack branchial fistulae; and Japanese patients have four main symptoms of BOR (Table 2). Notably, most patients carrying the same or different variants at the same position, such as c.1050+1G > T (Orten et al., 2008), c.1050+1G > C (Henriksen et al., 2004), or c.1050+2T > C (Unzaki et al., 2018), showed renal anomalies; such anomalies were not observed in our patients with the c.1050+1G > T variant. This variability is also present among patients with SIX1 variants. Patients with c.328C > T variant exhibited distinct phenotypes in several unrelated families (Kochhar et al., 2008). In addition, c.316G > A and c.317T > G led to the substitution of amino acids at the same position to methionine and glycine, respectively (Kochhar et al., 2008). Two individuals exhibited similar symptoms of BOS. Specifically, the same or a similar variant can cause variable clinical phenotypes in BOR/BOS, while different variants can cause similar clinical phenotypes.

The mechanism by which mutations produce phenotypic variability is undefined in BOR/BOS. Known EYA1 variants mainly comprise loss-of-function mutations; such variants typically imply haploinsufficiency through reduced gene dosage and expression (Figure 3) (Zhang et al., 2004). These results suggest that the inconsistency of phenotypes is partially influenced by the dosage effect. The Eya1 protein activates target genes controlling the development of the branchial arch, ear, or kidney only when a specific threshold is exceeded (Wang et al., 2018). Otherwise, environmental factors and genetic modifiers might also modify the phenotypes of BOR/BOS. Notably, we found that reported SIX1 variants mainly comprise missense (12/15). In the future, further exploration of the role of the dominant-negative effect (Shah et al., 2020) is also warranted.

BOS appears to be the main manifestation in China. In 1995, a study indicated that the estimated prevalence of renal abnormality was 67% in 21 patients with BOR in the United States; renal agenesis occurred most often, followed by hypoplasia, renal dysplasia, ureteral-pelvic junction obstruction, calyceal cyst/diverticulum, caliectasis, and hydronephrosis (Chen et al., 1995). In 2004, Chang et al. reported the renal abnormality frequency of 38% based on analyses of 40 families, but no detailed renal phenotypes were described (Chang et al., 2004). In France, Kurg et al. also identified 53% of the prevalence of renal anomalies in BOR patients harboring EYA1 variants; there is a wide range of abnormalities, namely, renal hypoplasia, multicystic kidney dysplasia, agenesis, abnormal pyelo-ureteral junction, and kidney malrotation, to name the most common (Krug et al., 2011). In our study, only one patient presented with left-sided renal hypoplasia, hydronephrosis, and pyelo-ureteral separation (1/9). We then calculated the frequency of renal malformations in East Asian populations through a literature review (Figure 4B). The frequencies were 14% in China, 39% in Japan, and 41% in Korea. Comparatively, the morbidity of renal anomalies appears to be lower and statistically different, suggesting that patients may primarily present with BOS in China. The elucidation of population differences in renal anomalies requires further exploration.

The clinical management of BOR/BOS requires individual management and multidisciplinary collaboration because of the multi-systemic symptoms. Deafness, the most prevalent manifestation of BOR/BOS, requires close attention and early treatment. Patients with BOR/BOS exhibit multiple structural malformations of the middle ear, especially ossicular malformations. Surgeons perform ossicular reconstruction to repair mixed or conductive hearing loss (Cremers et al., 1981), but the results are not always satisfactory. Our therapeutic findings for probands 1-II-3 and 5-I-1 were consistent with previous results. Thus, we presume that middle ear surgery is not optimal for patients with BOR/BOS. Hearing aids are beneficial to most patients, but this treatment should be individually tailored. Proband 1-II-3 received soft-banded bone-conduction hearing aids, while proband 2-II-1 wore air-conduction hearing aids after fitting. Their hearing abilities were substantially improved, which facilitated the fulfillment of learning and communication needs. Miyagawa et al. performed bone-anchored hearing aid implantation on a patient with BOS, which led to an improved hearing threshold (Miyagawa et al., 2015). Therefore, bone-conduction hearing device implantation may be considered when cranial bone thickness is at least 4 mm. CI may be a good option when hearing aids do not substantially improve hearing. Kameswaran et al. performed CI on a BOR patient with multiple inner ear malformations (Kameswaran et al., 2007); Bajaj et al. analyzed the surgical effect of CI in syndromic children, including children with BOR/BOS (Bajaj et al., 2012); their results indicated that CI was feasible in patients with BOR/BOS. In the present study, proband 5-II-3 underwent CI in the right ear after a comprehensive evaluation of the preoperative findings. CI can be successful through careful intraoperative operation, and the postoperative hearing ability is satisfactory with rehabilitation.

In conclusion, the symptoms of BOR or BOS in this study were attributed to EYA1 or SIX1 alterations. BOR/BOS exhibited some genetic and phenotypic variability. The outcomes of auditory rehabilitation reiterated that middle ear surgeries are generally unsatisfactory in patients with BOR/BOS; CI may be a feasible option when patients cannot benefit from hearing aids. Genetic testing contributes to the diagnosis and future genetic consultation; it has vital roles in therapy and intervention.