High Genetic Heterogeneity in Chinese Patients With Waardenburg Syndrome Revealed by Next-Generation Sequencing

Objective This study aimed to explore the genetic causes of probands who were diagnosed with Waardenburg syndrome (WS) or congenital sensorineural hearing loss. Methods A detailed physical and audiological examinations were carried out to make an accurate diagnosis of 14 patients from seven unrelated families. We performed whole-exome sequencing in probands to detect the potential genetic causes and further validated them by Sanger sequencing in the probands and their family members. Results The genetic causes for all 14 patients with WS or congenital sensorineural hearing loss were identified. A total of seven heterozygous variants including c.1459C > T, c.123del, and c.959-409_1173+3402del of PAX3 gene (NM_181459.4), c.198_262del and c.529_556del of SOX10 gene (NM_006941.4), and c.731G > A and c.970dup of MITF gene (NM_000248.3) were found for the first time. Of these mutations, we had confirmed two (c.1459C > T and c.970dup) are de novo by Sanger sequencing of variants in the probands and their parents. Conclusion We revealed a total of seven novel mutations in PAX3, SOX10, and MITF, which underlie the pathogenesis of WS. The clinical and genetic characterization of these families with WS elucidated high heterogeneity in Chinese patients with WS. This study expands the database of PAX3, SOX10, and MITF mutations and improves our understanding of the causes of WS.


INTRODUCTION
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).
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 Database 1 ), 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 1 www.hgmd.cf.ac.uk/ 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.

Patients
From seven Han Chinese families in the Henan province, 14 patients (Table 1) and nine unaffected family members were recruited for our study and asked to perform audiological and general physical examinations (Figure 1). Furthermore, in family WS04, only WS04-II:1 was recruited because he was adopted and had lost contact with his biological family. Among the seven families, WS01 and WS06 were isolated cases, while the remaining families had multiple affected individuals (Figure 1). Photos and blood were collected after informed consent (Figure 2). This study was conducted according to the Declaration of Helsinki and approved by the institutional review board of the Medical Ethics Committee of The Second Affiliated Hospital of Zhengzhou University (Approval No. 2018008).

Clinical Investigation
All patients (medical history described by parents) received elaborate physical examinations in their hair color and skin pigmentation, joints, skeletomuscular system, digestion, ophthalmology and otology, and intelligence assessment. Patients also underwent audiological examinations, which included auditory steady-state response (ASSR), auditory brainstem response (ABR), and distortion product otoacoustic emission (DPOAE). Additionally, imageological examinations such as computerized tomography (CT) of the temporal bone and magnetic resonance imaging (MRI) were conducted. The characteristics of the patients are summarized in Table 1.

Next-Generation Sequencing-Based Genetic Testing
In this study, WES was applied to identify the potential genetic causes for probands. A standard NGS-based genetic testing, including sample preparation and quantification, library construction, sequencing, and data analyses, was performed as previously described (Pan et al., 2020). Briefly, after library  construction, the resulting libraries were hybridized to the Agilent SureSelect Human All Exon V7. Then, sequencing was carried out on an Illumina HiSeq 4000 sequencer (Illumina Inc., San Diego, CA, United States) to generate paired-end reads of 150 bp. Data analyses were divided into bioinformatics analysis and variant interpretation. Under the framework of bcbionextgen 2 , we used the Burrows-Wheeler Aligner (BWA) (version 0.7.17-r1188) (Li, 2013) to align the sequencing reads to the human reference genome (GRCh37); GATK Haplotype Caller software (version 4.1.2) (McKenna et al., 2010) to 2 https://github.com/bcbio/bcbio-nextgen identify the single nucleotide variants (SNVs) and short indels; DECoN (Fowler et al., 2016) to identify the CNVs; and Vcfanno software (version 0.3.1) (Pedersen et al., 2016) to annotate the VCF files with external database, including Clinvar (Landrum et al., 2018), ExAC (Lek et al., 2016), dbNSFP , 1,000 Genomes (Auton et al., 2015), and gnomAD (Karczewski et al., 2019). The filtered variants were interpreted following the guidelines of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology (ACMG-AMP) (Richards et al., 2015) and the ClinGen hearing loss expert group's recommendation on variant interpretation (Oza et al., 2018). Copy number analysis was performed from NGS data using DECoN with the bam files from the same enrichment panel and sequencing run. Paternity tests were performed on families WS01 and WS06 since the gene tests had shown mutations occurred de novo.

Sanger Sequencing
Sanger sequencing was used to confirm the candidate variants detected by NGS and to conduct co-segregation analyses in family members. The specific primers ( Table 2) were designed by NCBI Primer-BLAST and synthesized by Sunya Biotech Co., Ltd. (Zhengzhou, China). Conventional PCR was performed for SNVs and short indels detected in families WS01 to WS06. While long-range PCR (LR-PCR) based on nested-PCR and fragments gel-purified were performed for the CNV of patients WS07-I:1 and WS07-II:1. LR-PCR is a traditional approach to obtain CNV breakpoint junction (Woodward et al., 2005;Zhang et al., 2017), for which several primers were designed from both the proximal and the distal breakpoint regions identified, and used in different combinations until an appropriate size product was generated (Supplementary Figure 1). After PCR amplification, purification, and quality control, Sanger sequencing was run in a SeqStudio Genetic Analyzer (Thermo Scientific, United States) with a mixture of PCR products and BigDyeTM Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems, Foster City, CA, United States). The sequencing results were analyzed by the SnapGene viewer (Figure 3).

Clinical Findings
A total of 14 patients from seven unrelated families were involved in this study. Before genetic testing, patients WS01-II:1 and WS06-II:1 were primarily diagnosed with SNHL, while the other 12 patients were diagnosed with WS. These diagnoses were made by otorhinolaryngologists based on the manifestation of the typical symptom of WS, such as SNHL, abnormal pigmentation, and the presence or absence of dystopia canthorum, musculoskeletal anomalies, and intestinal aganglionosis. After a genetic diagnosis, further examinations were performed on the patients WS01-II:1 and WS06-II:1. We found that the hair color of WS06-II:1 is gray, which was previously ignored, but nothing new with WS01-II:1.
Members in the same family can have different symptoms. WS03-II:1 and WS03-II:2 are identical twins; one has bilateral blue iris and the other has unilateral. According to his adoptive 2 | Primer pairs of the novel mutations of paired box 3 (PAX3), SRY-box transcription factor 10 (SOX10), and melanocyte inducing transcription factor (MITF).

Mutations
Affected family Forward primer sequence (5 -3 ) Reverse primer sequence (  Frontiers in Genetics | www.frontiersin.org father, WS04-II:1 was born with a white forelock that returned to being black (we were unable to collect pictures of the patient with a white forelock). Across four generations, seven members of the family WS05 had unexpectedly brilliant blue eyes without other WS-related symptoms. However, it is impossible to confirm whether they were affected by the same mutation of the proband (WS05-IV:1) since they refused to provide blood samples.

Molecular Etiology
Whole-exome sequencing was performed in probands of these seven families. For each sample, at least 10 Gbp raw data was generated, with more than 82% of bases having a Phred quality score Q ≥ 30 (Q30), 99% of the clean reads can map to the human reference genome (GRCh37/hg19), and the average sequencing depth of target regions was 100×, with 95% of target regions having coverage greater than 20× (Supplementary Table 1).
In this study, the genetic causes for all recruited patients had been confirmed, which contain a total of seven mutations in PAX3, SOX10, and MITF, respectively ( Table 3). To our best knowledge, these mutations, including a nonsense mutation, a missense mutation, a CNV, and four frameshift mutations, have not been reported by previous studies or recorded in any public database. Further analysis by Sanger sequencing of patients and their family members shows that all the variants were present in the affected members and absent in the unaffected ones (Figure 3), and variants of probands WS01-II:1 and WS06-II:1 are de novo. Pathogenicity analysis of variants was according to the standards and guidelines for interpreting genetic variants proposed by the ACMG-AMP. The mutations and pathogenicity analysis were summarized in Table 3.

Genotype-Phenotype Correlation
The phenotypes of WS patients with PAX3 (n = 7), SOX10 (n = 4), and MITF (n = 3) mutations are compared in Supplementary  Table 2. Among the WS patients who participated in this study, all patients with SOX10 mutations have hearing loss, while some patients with PAX3 (3/7) or MITF (2/3) have. Similarly, the symptom of the blue iris could be found in all patients with SOX10 mutations, while it was found in some patients with PAX3 (4/7) or MITF (2/3) variants. Abnormal hair pigmentation is rare in patients with PAX3 (0/7), SOX10 (1/4), and MITF (1/3) mutations. Previous reports suggested that freckles could be observed only in Chinese WS II patients with MITF mutations (Chen et al., 2010;Sun et al., 2016). Indeed, in this study, we found one WS II patient with MITF mutation has freckles. Synophridia is only present in WS I patients with PAX3 mutations.
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, *This mutation was, for the first time, found in humans after being detected and researched by Steingrímsson et al. (1996) in mice, so the criteria "PS3" was given. 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-loophelix 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., 2012Tang et al., , 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.

DATA AVAILABILITY STATEMENT
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/con tent/2019-06/10/content_5398829.htm, http://english.www.gov. cn/policies/latest_releases/2019/06/10/content_28147670894546 2.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.

ETHICS STATEMENT
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.

AUTHOR CONTRIBUTIONS
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.

ACKNOWLEDGMENTS
We sincerely thank all the family members for their participation in this study. We also thank the Supercomputing Center of Zhengzhou University for providing computational and storage resources.