Two novel mutations of PAX3 and SOX10 were characterized as genetic causes of Waardenburg Syndrome.

BACKGROUND: The objective of this study was to investigate the genetic causes of two probands diagnosed as Waardenburg syndrome (WS type I and IV) from two unrelated Chinese families.
METHODS: PAX3 and SOX10 were the main pathogenic genes for WS type I (WS I) and IV (WS IV), respectively; all coding exons of these genes were sequenced on the two probands and their family members. Luciferase reporter assay and co-immunoprecipitation (CO-IP) were conducted to verify potential functional outcomes of the novel mutations.
RESULTS: The first proband is a 9 years old girl diagnosed with WS I. A novel PAX3 heterozygous mutation of c.372-373delGA (p.N125fs) was identified, which results in a frameshift and truncation of PAX3 protein. In family II, a 2 years old girl was diagnosed with WS IV, and Sanger sequencing revealed a de novo SOX10 mutation of c.1114insTGGGGCCCCCACACTACACCGAC (p.Q372fs), a frameshift mutation that extends the amino acid chain of SOX10 protein. Functional studies indicated that the novel mutation of SOX10 had no effects on the interaction of SOX10 and PAX3, but reduced transactivate capacity of melanocyte inducing transcription factor (MITF) promoter. Both PAX3 and SOX10 mutation-induced defects of MITF transcription might contribute to the WS pathogenesis.
CONCLUSION: We revealed a novel mutation in PAX3 and a de novo mutation in SOX10, which might account for the underlying pathogenesis of WS. This study expands the database of both PAX10 and PAX3 mutations and improves our understanding of the causes of WS.

INTRODUCTION

Waardenburg syndrome (WS) is a rare autosomal dominantly inherited disease, which is characterized by hearing loss and pigment alteration in hair, skin, and iris (Read & Newton, 1997). The incidence of WS is estimated 1/40 000; however, it accounts for 2% to 5% of all congenital hearing loss (Dourmishev, Dourmishev, Schwartz, & Janniger, 1999; Song et al., 2016; Zaman, Capper, & Baddoo, 2015), which is the most common clinical feature of WS. WS is divided into four subtypes (WS I‐IV) depending on the presence or absence of additional symptoms such as dystopia canthorum, giant colon, and upper limb abnormalities. WS I (OMIM #193500) and WS II (OMIM #193510) are the most common types, while WS III (OMIM #148820) and WS IV (OMIM #277580) are rare (Read & Newton, 1997). WS is genetically different and six genes have been identified: paired box 3 (PAX3), SRY‐box transcription factor 10 (SOX10), melanocyte inducing transcription factor (MITF), endothelin receptor type B (EDNRB), endothelin 3 (EDN3), and snail family transcriptional repressor 2 (SNAI2) (Otreba, Milinski, Buszman, Wrzesniok, & Beberok, 2013). PAX3 is associated with WS I and WS III. MITF and SNAI2 are found to be involved in the WS II. SOX10, EDNRB, and EDN3 are responsible for WS II and WS IV (Pingault et al., 2010). All these genes play crucial roles in the formation and development of melanocytes.

WS with dystopia canthorum is diagnosed as WS I and PAX3 is the most pathogenic gene. PAX3 encodes the paired box 3 transcription factor and contributes to the development of the central nervous system, skeletal muscle, and melanocytes (Wildhardt et al., 2013). For WS IV, it is characterized by additional feature of megacolon syndrome. In humans, 45%‐55% of WS IV cases are involved in the mutation of SOX10, which works in the development and differentiation of melanocytes (Bondurand et al., 2007). Up to date, dozens of mutations in PAX3 and SOX10 have been identified, including missense mutations, frameshift mutations, and insertions/deletions (Bocangel et al., 2018; Ma et al., 2016). Great efforts are still been made to discover novel mutations in recent studies (Hogan et al., 2019; Ma, Lin, et al., 2019; Ma, Zhang, et al., 2019). Functional analysis also reported that these mutants might induce loss of DNA‐binding ability, failed to transactivate MITF promoter, and inhibited melanin synthesis (Wang et al., 2017; Zhang et al., 2012). Although mutations in PAX3 and SOX10 are the most common genetic cause of WS I and WS IV, respectively, mutational spectrum still needs to be extended to understand the genetic causes of WS pathogenesis.

In the present study, PAX3 and SOX10 were individually screened in two Chinese probands with WS I and WS IV, as well as their family members. A novel mutation in PAX3 [c.372‐373delGA (p.N125fs)] and a de novo mutation in SOX10 [c.1114insTGGGGCCCCCACACTACACCGAC (p.Q372fs)] were identified by Sanger sequencing. The mutation of N125fs resulted in truncated PAX3 while Q372fs caused structural extension of SOX10. To understand functional consequences of Q372fs‐induced SOX10 extension, transcriptional activity of MITF was detected by dual‐luciferase reporter assay. The effects of mutant SOX10 on the interaction between SOX10 and PAX3 were studied using co‐immunoprecipitation assay.

MATERIALS AND METHODS

Patients

Two Chinese girls with severe‐profound hearing loss and their family members were recruited. Photos and blood were collected before informed consent was signed. Ethics Committees of Beijing Children's Hospital approved our study. After medical history was described by their parents, the two girls got both physical and radiological examination. Auditory evaluations were conducted by play audiometry (PA), auditory steady‐state response (ASSR), auditory brainstem response (ABR), and distortion product otoacoustic emission (DPOAE).

Gene mutation analysis

Blood DNA was extracted from the probands and their family members using Blood DNA Kit (CWBIO, China). Polymerase chain reaction (PCR) was used to amplify fragments covering coding exons of PAX3 (NM_181457.3) and SOX10 (NM_006941.4) with specific primers (Table S1). PCR experiment was done as previously described (Yu et al., 2019). All PCR products were evaluated by 1% agarose gel electrophoresis and sequenced by Sanger sequencing (Applied Biosystems, USA). Gene mutation was identified by Sequencher software.

Plasmids construction

To generate plasmids of pGV230‐PAX3‐HA and pGV362‐SOX10‐Flag, full‐length cDNA of PAX3 and SOX10 were PCR amplified and subcloned into their corresponding vectors. For pGV362‐SOX10‐Q372fs‐Flag construction, frameshift mutation (c.1114ins TGGGGCCCCCACACTACACCGAC) was synthesized and subcloned into the vector. To construct pGV238‐MITF luciferase reporter, MITF promoter region (−1500 bp to +50 bp from transcription start site) was synthesized and subcloned into the luciferase vector. All constructs were verified by direct Sanger sequencing.

Dual‐luciferase reporter assay

Human melanoma A375 cell line was cultured in DMEM with 10% fetal bovine serum. Cells were seeded in 24‐well plates for 24 hr and then transfected by MITF luciferase reporter together with SOX10‐wt‐Flag or SOX10‐Q372fs‐Flag using Lipofectamine 2000 (Life Technologies). Renilla luciferase reporter plasmid was co‐transfected to normalize transfection efficiencies. Luciferase activity was finally measured (CLARIOstar, BMG labtech) using Dual‐Luciferase Reporter Assay system (Promega).

Co‐immunoprecipitation (CO‐IP) and western blotting

SOX10 interacts with PAX3 to regulate MITF expression. To study the interaction between SOX10 and PAX3, Co‐IP was performed. Briefly, PAX3‐HA plasmid was co‐transfected with SOX10‐wt‐Flag or SOX10‐Q372fs‐Flag in A375 cells for 48 hr. Total proteins were extracted and concentration was determined. Equal amounts of proteins were incubated with anti‐HA antibody (Abcam) overnight at 4°C. Commercial protein A agarose beads (Roche, Switzerland) were added and incubated for 4 hr at room temperature. The beads were washed, collected, and resolved repeatedly in chilled GUO HEPES buffer for six times. Finally, the beads were boiled for western blotting detection, according to standard protocol previously described (Yu et al., 2018). Protein bands were obtained by Odyssey CLx imaging system (LI‐COR).

RESULTS

Clinical features and evaluation

These two girls were from two unrelated Chinese families. Both of them delayed in speech development and diagnosed as bilateral hearing loss. The first proband was 9 years old and diagnosed with WS I. As shown in Figure 1a, physical examination found iris heterochromia in eyes, a pinch of white hair on forehead, and dystopia canthorum. Symptoms of iris heterochromia in eyes and unilateral severe hearing loss were also found in her mother. For one of her younger brothers, iris heterochromia was in left eye with normal hearing. The other younger brother was normal. Temporal bone CT scan showed that the structure of the inner ear of the girl was normal (Figure 1b). However, she suffered from severe sensorineural hearing loss in both ears (Figure 1c). The second proband was 2 years old diagnosed with WS IV. Both her parents were normal. Physical examination found iris heterochromia in her right eye and her hair was gray (Figure 2a). She also suffered from Hirschsprung disease and got surgery when she was 2 months old. Both CT and MRI scan showed that the structure of the inner ear and cochlear nerve was normal (Figure 2b). Based on audiometry (PA) detection, her both ears got severe hearing loss (Figure 2c). Both probands have received unilateral cochlear implantation (CI) and were in our close follow‐up.

Figure 1

Clinical phenotypes in family diagnosed with WS I. (a) Iris heterochromia in the eyes, dystopia canthorum, and a pinch of white hair on forehead of the proband; Bright blue irides in eyes and unilateral severe hearing loss of the mother; Iris heterochromia in the left eye of one younger brother with normal hearing; Normal younger brother. (b) Temporal bone CT scan showed that the structure of the inner ear was normal. (c) Bilateral play audiometry (PA) detection. The x‐axis indicates frequency in hertz (Hz) and the y‐axis indicates hearing level in decibels (dB nHL)

Figure 2

Clinical features of the proband diagnosed with WS IV. (a) Iris heterochromia in right eye and premature graying of the hair. (b) CT and MRI showed that the structure of the inner ear and cochlear nerve is normal. (c) Bilateral play audiometry (PA) detection

Identification of novel mutations in PAX3 and SOX10

The pedigree chart of family I was described in Figure 3a. As shown in Figure 3b, novel heterozygous mutations in both intron and exon of PAX3 were detected. Based on the fact that genotype was in accordance with clinical appearance, c.372‐373delGA (p.N125fs) in exon 3 might be the pathogenic mutation. This mutation caused a stop codon at position of 143 amino acid and truncated PAX3 (Figure 3c). The protein truncation retains most of the paired domain (PD) but deletes other domains (Figure 3d), which might be crucial for PAX3 function.

Figure 3

Pedigree map and PAX3 (NM_181457.3) mutation detection in family I. (a) Pedigree map. Squares and circles denote males and females, respectively. (b) Mutations in exons and primer‐covered introns of all family members. The “*” indicates homozygous mutation. (c) Protein alignment showed that c.372‐373delGA (p.N125fs) induced a frameshift mutation, caused a stop codon at position of 143 amino acid, resulting in truncation of PAX3. (d) The putative schematic representation of PAX3 protein and N125fs mutation, which lead to a large reduction in the PAX3 protein from 479 amino acids to 142 amino acids

In family II, a fragment of 23bp was inserted in exon 4 of SOX10 at site of 1114bp (Figure 4a,b). It is a de novo mutation because c.1114insTGGGGCCCCCACACTACACCGAC (p.Q372fs) only occurred in the proband but not inherited from her parents. It caused a frameshift mutation of SOX10 from position of 372 to 508 amino acid, which is 41 amino acids longer than wide‐type SOX10 (Figure 4c). Functionally, the structural extension might affect its binding with PAX3. The schematic image of SOX10 protein and its mutant is described in Figure 4d.

Figure 4

Pedigree map and SOX10 (NM_006941.4) sequence in family II. (a) Pedigree map. Squares and circles denote males and females, respectively. (b) Sequence electropherograms showed that c.1114insTGGGGCCCCCACACTACACCGAC (p.Q372fs) was a de novo mutation, which was not inherited from parents. (c) Q372fs caused a frameshift mutation from position of 372 to 508 amino acid, which is 41 amino acids longer than wide‐type SOX10. (d) The putative schematic representation of SOX10 protein and the extended mutants.

The effect of mutant SOX10 on MITF transcriptional activity

The role of SOX10 in melanocyte is to regulate MITF expression together with PAX3. To identify whether Q372fs affects SOX10 protein function, luciferase assay was performed. As shown in Figure 5a, wild‐type SOX10 induced MITF promoter activity by approximately fourfold than mutant SOX10 in A375 cells. As a result, mutant SOX10 loses its ability to activate MITF promoter.

Figure 5

Functional analysis of SOX10 Q372fs in A375 cells. (a) Transcriptional capacity of wide‐type SOX10 and its mutant detected by luciferase assays. (b) Protein–protein interaction between PAX3 and wide‐type/mutant SOX10, which is performed by Co‐IP assay using anti‐Flag and anti‐HA antibodies. The “*” indicates p < .05

The effect of mutant SOX10 on the interaction between SOX10 and PAX3

We have demonstrated that mutant SOX10 cannot activate MITF expression, but the molecular mechanism is not clear. Since SOX10 interacts with PAX3 to regulate MITF expression (Dai et al., 2019), the interaction changes might help in answering this issue. To investigate whether SOX10 Q372fs affects the interaction between SOX10 and PAX3, Co‐IP assay was performed. As shown in Figure 5b, both wild‐type SOX10 and mutant SOX10 can be co‐immunoprecipitated when co‐expressed with PAX3 in A375 cells. This result suggested that Q372fs‐induced SOX10 protein extension did not affect its interaction with PAX3.

DISCUSSION

Two patients from two unrelated families were diagnosed as WS I and WS IV, which is genetically associated with mutations in PAX3 and SOX10, respectively. To investigate potential pathogenic causes, a genetic and functional analysis was performed. We identified a novel mutation of PAX3 [c.372‐373delGA (p.N125fs)] in proband I, which truncated PAX3 and resulted in protein structure disruption. In proband П, the de novo mutation of SOX10 [c.1114insTGGGGCCCCCACACTACACCGAC (p.Q372fs)] caused structural extension of SOX10. Further in vitro results demonstrated that the mutant SOX10 inhibited transcriptional activity of MITF but not affected the interaction between SOX10 and PAX3.

The first proband in family I was diagnosed as WS I, based on additional symptoms of dystopia canthorum and phenotypes in family members. Genetically, PAX3 is the main causative gene of WS I, and 90% of WS I patients carry PAX3 variants. Recently, many novel PAX3 mutations were reported, such as c.91‐95de1ACTCC, c.808C>G, and c.117C>A (Choi, Choi, & Lee, 2018; Li et al., 2019; Ma, Lin, et al., 2019). To verify the clinical diagnosis in proband I, genetic screening of PAX3 was performed and c.372‐373delGA (p.N125fs) was identified. This frameshift mutation changed PAX3 structure and may result in loss of gene function. Up to March 2019, a total of 164 variants in PAX3 have been reported in the Human Gene Mutation Database (HGMD), among which point mutations and deletions comprised more than 90% (Stenson et al., 2017). As no report of c.372‐373delGA was found in the HGMD, we consider it is a novel mutation. This mutation detected in the proband and his brother came from their mother and was inherited autosomal dominantly.

PAX3 locates on chromosome 2q35, encoding a 479 amino acids protein, which is a transcriptional factor from the paired box (PAX) family (Boudjadi, Chatterjee, Sun, Vemu, & Barr, 2018). PAX3 contributes to the migration and differentiation of melanocytes, which originate from the embryonic neural crest. In melanoblast, PAX3 is associated with the expression of markers for melanocyte development, including MITF (Boudjadi et al., 2018; Dye, Medic, Ziman, & Coombe, 2013). Due to the fact that color change is a typical characteristic for WS, PAX3 function in pigmentation defects of the hair, skin, and eye might account for the pathogenesis of WS I (Bocangel et al., 2018). Previous functional study has demonstrated that PAX3 R270G mutation failed to activate MITF promoter but retained abilities of nuclear distribution and DNA‐binding (Niu et al., 2018). However, another study reported that PAX3 H80D can retain partial activity (Zhang et al., 2012). According to the guidelines from ACMG (American College of Medical Genetics and Genomics), one missense variant is known to be pathogenic in most cases. However, frameshift mutation was a null variant, which may disrupt gene function (Richards et al., 2015). Therefore, the new reported frameshift mutation of c.372‐373delGA in PAX3 might be the major molecular pathogenesis of WS I in this family.

With regard to the second proband in family II, a de novo mutation of c.1114insTGGGGCCCCCACACTACACCGAC (p.Q372fs) was found, which induced structural extension of SOX10. In humans, half of WS IV cases are associated with SOX10 mutation and more than 160 variants have been reported in HGMD till March 2019 (Stenson et al., 2017). SOX10 encodes a transcription factor and acts as a transcriptional activator to regulate MITF expression by forming a protein complex with PAX3 (Kamachi, Cheah, & Kondoh, 1999; Seberg, Van Otterloo, & Cornell, 2017). Consistent with previous studies (Dai et al., 2019; Wang et al., 2017), our results demonstrated that Q372fs SOX10 lost the ability to activate MITF expression. Generally, SOX10 is localized in the nucleus and promotes target DNA transcription (Seberg et al., 2017). Functional study reported that SOX10 mutation of p.L141P affected DNA or protein‐binding capacity, and inhibited MITF expression by inducing aberrant cytoplasmic and nuclear localization (Wang et al., 2017). However, mutant SOX10 was also reported to reduce MITF transcription but not affect nuclear localization and DNA‐binding capacity (Dai et al., 2019). Therefore, SOX10 function is not localization‐dependent and Q372fs SOX10‐induced suppression of MITF might be the genetic cause of WS IV.

Although we have demonstrated that mutant SOX10 affected MITF expression, the molecular mechanism is not clear. Functionally, SOX10 was reported to interact with PAX3 to regulate MITF expression (Dai et al., 2019). To determine whether the reduced MITF transcription was caused by defects in the interaction between PAX3 and SOX10, Co‐IP was performed but the results showed that Q372fs SOX10 did not affect the ability to interact with PAX3. Although the structure of SOX10 is not fully characterized (Pingault et al., 1998), it structurally contains a DNA‐binding HMG (high mobility group) domain, a dimerization region right upstream to the HMG, a conserved domain in the center, and a transactivation (TA) domain at the extreme C‐terminus (Pingault et al., 2010). Herein, the frameshift mutation of Q372fs made SOX10 protein extended in structure, destroying the TA domain but not affecting other domains. This might be the reason why Q372fs SOX10 failed to transactivate MITF, but still can interact with PAX3.

In summary, we identified a novel mutation in PAX3 and a de novo mutation in SOX10 in two unrelated probands diagnosed with WS I and WS IV, respectively. In family I, the mutation of N125fs in PAX3 leads to a frameshift mutation and truncates PAX3 protein. In family П, the de novo mutation of Q372fs caused structural extension of SOX10. Further in vitro results demonstrated that the Q372fs SOX10 inhibited transcriptional activity of MITF, but not affected the interaction between SOX10 and PAX3. Our finding is expected to expand the mutation spectra of PAX3 and SOX10, which might be the genetic causes of WS pathogenesis.

CONFLICT OF INTEREST

The authors declare they have no actual or potential competing financial interests.

Table S1

Click here for additional data file.