Waardenburg syndrome (WS) is a congenital developmental disorder, which is mainly characterized by congenital sensorineural hearing loss (SNHL) and abnormal pigmentation of the iris, hair, and skin (manifests as heterochromia iridis and brilliant blue eyes, a white forelock, and premature graying, and hypopigmented skin) (Read and Newton, 1997). WS has an incidence rate of approximately 1/42,000 births and is responsible for 2–5% of cases of total congenital deafness (Read and Newton, 1997; Nayak and Isaacson, 2003). Four different types of Waardenburg syndrome (WS I∼IV) have been described based on genotypic and phenotypic variations (Read and Newton, 1997; Pingault et al., 2010). WS I is distinguished from WS II by the presence of dystopia canthorum, which is lateral displacement of the inner canthus in each eye; WS III (Klein–Waardenburg syndrome) is similar to WS I except with additional upper limb abnormalities; WS IV (Waardenburg-Shah syndrome) is characterized by general WS features as well as Hirschsprung’s disease, a disorder that causes severe blockage of the large intestine. Current research suggests that WS I and WS II are more common than WS III and WS IV (Read and Newton, 1997; Pingault et al., 2010).

Waardenburg syndrome shows a high degree of genetic heterogeneity (Hageman and Delleman, 1977; Read and Newton, 1997; Pingault et al., 2010; Song et al., 2016). Six genes have been linked to this syndrome: paired box 3 (PAX3) (Baldwin et al., 1992; Tassabehji et al., 1992; Hoth et al., 1993), melanocyte inducing transcription factor (MITF) (Tassabehji et al., 1994), SRY-box transcription factor 10 (SOX10) (Pingault et al., 1998; Bondurand et al., 2007), endothelin 3 (EDN3) (Edery et al., 1996), endothelin receptor type B (EDNRB) (Puffenberger et al., 1994), and snail family transcriptional repressor 2 (SNAI2) (Sánchez-Martín et al., 2002). PAX3 is responsible for WS I and WS III (Baldwin et al., 1992; Tassabehji et al., 1992; Hoth et al., 1993). SOX10, MITF, and SNAI2 are associated with WS IV (Tassabehji et al., 1994; Pingault et al., 1998; Sánchez-Martín et al., 2002; Bondurand et al., 2007). SOX10, EDNRB, and EDN3 are found to be involved in WS IV (Puffenberger et al., 1994; Edery et al., 1996; Pingault et al., 1998; Bondurand et al., 2007). Although not currently fully understood, all these genes are involved in a complex network in neural crest cells and other derivatives (Read and Newton, 1997; Bondurand et al., 2000; Pingault et al., 2010). The interaction of these genes during the formation and development of melanocytes could be the pathogenesis of WS and other related diseases (Read and Newton, 1997; Bondurand et al., 2000; Pingault et al., 2010).

Diagnosis of WS can be difficult because all features are not present in every patient (Hageman and Delleman, 1977; Newton, 1990; Tamayo et al., 2008; Pingault et al., 2010; Yang et al., 2013). Even within a single family, patients can display different clinical manifestations due to variations in the expressivity of causative genes (Hageman and Delleman, 1977; Newton, 1990; Tamayo et al., 2008; Pingault et al., 2010; Yang et al., 2013). Therefore, genetic testing is an important method for diagnosing this disease and its subtypes (Hageman and Delleman, 1977; Read and Newton, 1997; Pingault et al., 2010; Tang et al., 2015; Song et al., 2016; Wu et al., 2016; Li et al., 2019). To date, ∼400 mutations including missense/nonsense mutations, frameshift mutations, insertions/deletions, and copy number variants (CNVs) have been identified in genes associated with WS (The Human Gene Mutation Database¹), with most variants in genes PAX3, SOX10, and MITF (Chen et al., 2010; Pingault et al., 2010; Song et al., 2016). Of these variants, ∼100 mutations were identified in Chinese people. Nevertheless, there are still a number of cases unexplained at the molecular level (Pingault et al., 2010; Song et al., 2016). Discovering novel mutations will lead to a better understanding of the genetic causes of WS pathogenesis.

Recently, next-generation sequencing (NGS) has proven to be a potent tool for the identification of pathogenic mutations related to deafness, which can improve the diagnosis of genetic diseases and the detection of mutations in genes associated with different clinical manifestations (Brownstein et al., 2012; Lin et al., 2012; Tang et al., 2012; Li et al., 2019). In this study, WES was used to identify the possible pathogenic mutations of patients with SNHL or WS. A total of seven novel variants in PAX3, SOX10, and MITF were found, and two of them are de novo confirmed by Sanger sequencing of variants in the probands and their parents. Our results show that WS in China has a high degree of genetic heterogeneity and extend the mutational spectrum of WS-related genes.

In this study, we had confirmed seven novel heterozygous variants which are the genetic causes of 14 WS patients from seven unrelated families, including c.1459C > T (nonsense), c.123del (frameshift), and c.959-409_1173+3402del (deletion) of PAX3 (NM_181459.4), c.198_262del (frameshift) and c.529_556del (frameshift) of SOX10 (NM_006941.4), and c.731G > A (missense) and c.970dup (frameshift) of MITF (NM_000248.3) (Table 3). Among 14 patients, seven each were classified as WS I and WS II, respectively, based on their phenotypes and genotypes, showing that WS I and WS II were two major WS subtypes (Read and Newton, 1997; Pingault et al., 2010). While mutations in PAX3 were the major causes for WS I (7/7), SOX10 (4/7), and MITF (3/7) were two major causative genes attributable to WS II. Our findings had extended the mutational spectrum of WS-related genes and revealed high genetic heterogeneity in Chinese WS patients (Yang et al., 2013; Sun et al., 2016; Liu et al., 2020).

To explore the genotype–phenotype correlation, we compared the phenotypes between WS patients with PAX3, SOX10, and MITF mutations (Supplementary Table 2). Several reports had shown that the clinical features of WS II caused by SOX10 and MITF mutations were indistinguishable, except that freckle was frequent in WS II probands with MITF mutation (Chen et al., 2010; Toriello, 2011; Sun et al., 2016). Indeed, in this study, freckle seems to be unique for patients with MITF mutations (1/3) but was absent in those with PAX3 (0/7) or SOX10 (0/4) mutations. Dystopia canthorum is a rebarbative but crucial clinical feature, because of its value in distinguishing WS I and WS II, but it is not completely applicable for Chinese WS I patients (Sun et al., 2016; Morimoto et al., 2018; Suzuki et al., 2018; Minami et al., 2019). Herein, we had an interesting finding that the synophridia, even though a minor symptom, was only present in WS patients (5/7) but absent in WS II patients (0/7). Our results may have shown the clinical differences between WS II patients with SOX10 and MITF mutations, and between WS II and WS I. However, gene test is as necessary as clinical investigation for the accurate diagnosis and subtype confirmation (Hageman and Delleman, 1977; Read and Newton, 1997; Pingault et al., 2010; Song et al., 2016; Li et al., 2019).

Mutations in PAX3, SOX10, and MITF were the most common genetic causes for WS and responsible for almost all Chinese WS patients (Chen et al., 2010; Wu et al., 2016; Liu et al., 2020). Beyond that, to date, several WS cases associated with mutations in ENDRB, EDN3, and SNAI2 had been reported (Sánchez-Martín et al., 2002; Pingault et al., 2010; Xiong et al., 2015; Somashekar et al., 2019), but the situation is a bit different in Chinese. There were two reported cases of WS type I caused by mutations in the EDNRB gene (Cheng et al., 2019; Li et al., 2019), which is different from the cases in other races (WS II or IV) (Pingault et al., 2010; Issa et al., 2017). Of the few reports about WS type II being caused by mutations in the EDN3 gene, there was no Chinese case reported. Only one research group reported SNAI2 mutations caused WS within two unrelated WS II patients (Sánchez-Martín et al., 2002), which had been questioned recently (Song et al., 2016; Mirhadi et al., 2020). To understand the differences of WS among different races, we need further research on the pathogenesis of WS and more accurate diagnostic means.

The deficiency of melanocytes, the neural crest (NC) derivatives, is common to various WS types (Bondurand et al., 2000; Pingault et al., 2010; Song et al., 2016), which is responsible for the phenotypes of pigmentation defects and hearing loss (Steel and Barkway, 1989). PAX3 encodes a DNA-binding transcription factor, consisting of a paired box (PD) encoded by exons 2, 3, and 4, the homeodomain (HD) by exons 5 and 6, C-terminal transcriptional activation domain by exons 7 and 8 (Read and Newton, 1997; Wildhardt et al., 2013). It is indispensable in the development of somites, skeletal muscle, and the neural crest cells (NCC) and their derivatives like melanocytes. It can cooperate with SOX10 to regulate the expression of the MITF promoter (Pingault et al., 2010). PAX3:c.123del and c.959-409_1173+3402del mutations are predicted to activate the nonsense-mediated mRNA decay (NMD) machinery (Khajavi et al., 2006), thereby resulting in haploinsufficiency, which might be the disease-causing mechanism for WS I. PAX3:c.1459C > T mutation is located in the exon 10 and could only influence the isoform PAX3e (Wang, 2006), which most likely pathogenic mechanism is haploinsufficiency (Barber et al., 1999).

SRY-box transcription factor 10 encodes a transcription factor that contains an HMG (high mobility group) DNA binding domain and a C-terminal transactivation domain (Chan et al., 2003). In the early development of NC, SOX10 plays an important role in promoting cell survival and maintaining the multipotency of NC stem cells (Kapur, 1999; Kelsh, 2006; Pingault et al., 2010; Stolt and Wegner, 2010). Besides synergy with PAX3 to regulate the expression of MITF, it also can directly regulate the expression of genes important for melanin synthesis, suggesting the importance for melanocyte differentiation (Bondurand et al., 2000; Lee et al., 2000; Potterf et al., 2000; Verastegui et al., 2000; Jiao et al., 2004; Wegner, 2005; Pingault et al., 2010). It is also crucial for the peripheral nervous system like sensory, sympathetic, and enteric ganglia and along nerves (Bondurand et al., 2000; Pingault et al., 2010). SOX10: c.198_262del and c.529_556del are located in the HMG domain and predicted to activate the NMD machinery, resulting in haploinsufficiency.

Melanocyte inducing transcription factor, a basic helix–loop–helix leucine zipper (bHLHZip) protein, is the key transcription factor of melanocyte development. The bHLHZip structure binds DNA by basic domain, dimerizes through HLH domain, and is stabilized via the Zip domain (Hodgkinson et al., 1993; Steingrímsson et al., 1994). The C-terminal of MITF contributes to defining the target genes by a serine-rich transcriptional activation domain. Mice with MITF mutations show reduced or absent pigmentation, deafness, and small or absent eyes, etc. (Yasumoto et al., 1995; Bertolotto et al., 1998; Steingrímsson et al., 2004; Pingault et al., 2010). MITF:c.970dup is a frameshift mutation and predicted to activate the NMD machinery, leading to haploinsufficiency. MITF:c.731G > A (p.Gly244Glu) is the genetic cause of III:1 and IV:1 of family WS05 and might be responsible for the other seven affected individuals (Figure 1). We had noticed that only one had hearing loss and blue iris while the other eight only had blue iris in this family, although Song et al. (2016) had suggested that nearly 90% of patients with MITF have hearing loss. The Gly244Glu mutation of MITF was found in humans for the first time, while the mouse model with the same mutation (MITF^Mi–b) had been found by Steingrímsson et al. (1996). The phenotype of MITF^Mi–b homozygous animals is mild compared with loss-of-function mi alleles. Gly244 would lie at the very beginning of the second helix, close to the protein-DNA interface. The Gly244Glu alteration is at the junction of the loop and helix 2 of the protein. The MITF^Mi–b protein largely spares dimerization function, while it is defective in its ability to bind DNA. However, the DNA binding function can be partially compensated by a wild-type partner in the dimer, since MITF^Mi–b is capable of forming TFE3 (Transcription factor E3) heterodimeric complexes which had a stronger DNA binding than the MITF^Mi–b homodimers. It may explain why the mutation c.731G > A resulting in a less-severe phenotype in the family WS05.

WS has high genetic heterogeneity and highly variable phenotype expressivity (Hageman and Delleman, 1977; Newton, 1990; Read and Newton, 1997; Tamayo et al., 2008; Pingault et al., 2010; Yang et al., 2013; Song et al., 2016), which makes the diagnosis challenging. NGS of numerous genes is allowed in a single test with lower turnaround time, cost, and higher throughput, which makes it ideal for figuring out the exact genetic mechanism (Brownstein et al., 2012; Tang et al., 2012, 2015). Mutations in PAX3 are responsible for WS I and WS III in most cases; however, using WES, we had detected a heterozygous nonsense mutation of PAX3:c.1459C > T in an SNHL patient (WS01-II:1). To our best knowledge, this is the first mutation found in exon 10 of PAX3 and results in a premature stop codon, which is very close to the normal ending (487/506) (Boudjadi et al., 2018). We suspect that this is why SNHL is the only symptom of WS01-II:1, even though previous studies argued that there was no correlation between genotype and phenotype of WS caused by PAX3 mutations (Tassabehji et al., 1995; Boudjadi et al., 2018). Patient WS06-II:1 was also diagnosed with SNHL initially, before being corrected to WS type II after the genetic testing in which a heterozygous de novo mutation MITF: c.970dup was detected. Besides, the other five families with classic symptoms and clear family histories, these two cases in particular highlight the superiority of NGS in the diagnosis of WS.

In conclusion, the clinical and genetic characteristics of one SNHL patient and six Chinese families of WS had been investigated in this study. Altogether, seven novel pathogenic/likely pathogenic variants in the PAX3, SOX10, and MITF were identified. Our results support that NGS is a useful diagnostic procedure for the diagnosis and subtype differentiation of WS. This report reveals the highly genetic heterogeneity and variable phenotype in Chinese patients with WS and will contribute to a better understanding of the WS by extending the mutational spectrum of WS-related genes.

According to national legislation/guidelines, specifically the Administrative Regulations of the People’s Republic of China on Human Genetic Resources (http://www.gov.cn/zhengce/content/2019-06/10/content_5398829.htm, http://english.www.gov.cn/policies/latest_releases/2019/06/10/content_281476708945462.htm), no additional raw data is available at this time. Data of this project can be accessed after an approval application to the Bio-Med Big Data Center, NODE. Please refer to https://www.biosino.org/node/project/detail/OEP001401 for detailed application guidance. The accession code OEP001401 should be included in the application.

The studies involving human participants were reviewed and approved by the institutional review board of the Medical Ethics Committee of The Second Affiliated Hospital of Zhengzhou University. Written informed consent to participate in this study was provided by the participants’ legal guardian/next of kin.

WT and HX: study design. BC, HL, RL, XL, and BZu: patient phenotypic analysis and genetic counseling. SZ, YT, XH, BZe, HL, and RL: next−generation sequencing and Sanger sequencing. HX, DL, SZ, and YT: data analysis and variant interpretation. SZ, HX, DL, WT, and RT: writing and review of original draft of the manuscript. RT: language editing of original draft of the manuscript. All authors have read and approved the final manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.