Toward the Pathogenicity of the SLC26A4 p.C565Y Variant Using a Genetically Driven Mouse Model

Recessive variants of the SLC26A4 gene are globally a common cause of hearing impairment. In the past, cell lines and transgenic mice were widely used to investigate the pathogenicity associated with SLC26A4 variants. However, discrepancies in pathogenicity between humans and cell lines or transgenic mice were documented for some SLC26A4 variants. For instance, the p.C565Y variant, which was reported to be pathogenic in humans, did not exhibit functional pathogenic consequences in cell lines. To address the pathogenicity of p.C565Y, we used a genotype-based approach in which we generated knock-in mice that were heterozygous (Slc26a4+/C565Y), homozygous (Slc26a4C565Y/C565Y), and compound heterozygous (Slc26a4919-2A>G/C565Y) for this variant. Subsequent phenotypic characterization revealed that mice with these genotypes demonstrated normal auditory and vestibular functions, and normal inner-ear morphology and pendrin expression. These findings indicate that the p.C565Y variant is nonpathogenic for mice, and that a single p.C565Y allele is sufficient to maintain normal inner-ear physiology in mice. Our results highlight the differences in pathogenicity associated with certain SLC26A4 variants between transgenic mice and humans, which should be considered when interpreting the results of animal studies for SLC26A4-related deafness.


Introduction
Recessive variants of the SLC26A4 (Gene ID: 5172) gene are a common global cause of hereditary hearing impairment (HHI) [1]. In certain populations, pathogenic SLC26A4 variants can be identified in approximately 15-20% patients with HHI [2]. SLC26A4 encodes for pendrin, a chloride bicarbonate transporter that is mainly expressed in the thyroid, inner ears, kidneys, lungs, liver, and heart [3,4]. Recessive SLC26A4 variants cause the occurrence of Pendred syndrome (PS; MIM #274600) and nonsyndromic hearing loss, DFNB4 (MIM #600791). DFNB4 is characterized by isolated sensorineural hearing impairment (SNHI), which is usually associated with a common inner-ear malformation known as enlarged vestibular aqueduct (EVA; MIM 603545), whereas patients with PS have goiter in addition to EVA [5]. Clinically, patients with SLC26A4 variants either with the manifestation of DFNB4 or PS usually suffer from progressive or fluctuating SNHI [6].
To date, more than 400 SLC26A4 variants were reported (http://deafnessvariationdatabase. org/ (accessed on 30 June 2020). Corresponding with the extensive variation in the severity of clinical phenotypes, different SLC26A4 variants are associated with different pathogenic consequences in cell-line experiments [7,8]. Some variants, such as p.H723R, p.L236P, and p.T721M, are associated with defective protein expression or trafficking, which leads to the accumulation of pendrin in either the cytoplasm or the perinuclear region. Some variants, such as p.K369Eand p.S166N, are associated with the normal protein expression of pendrin in the cell membrane, but impaired protein function for chloride bicarbonate (Cl -/HCO 3 -) exchange [7,8]. Interestingly, an enigmatic SLC26A4 variant, p.C565Y, was previously reported as pathogenic in humans when present in trans with other SLC26A4 mutations, such as p.Q514R [9,10], p.L236P [11], and p.H723R [12]. However, p.C565Y produced minimal or no functional pathogenic consequences in the HEK 293 and COS-7 cell lines or Xenopus oocytes [8,9]. Therefore, it is crucial to clarify the discrepancy in the pathogenicity between humans and cell lines in order to dissect the pathogenic mechanisms related to SLC26A4 variants.
In addition to cell-line models, transgenic mice were also demonstrated to be a powerful tool for investigating the pathogenic mechanisms of SLC26A4 variants [13][14][15][16][17][18][19]. However, previously established mouse models with different SLC26A4 variants were reported to have diverse auditory phenotypes, ranging from normal to profound hearing loss, none of which perfectly recapitulated the clinical phenotypes in patients. In this study, we generated a knock-in mouse model harboring the p.C565Y variant of Slc26a4, and investigated the associated audiovestibular phenotype and inner-ear pathology. We also generated compound heterozygous mice (Slc26a4 919-2A>G/C565Y ) by intercrossing p.C565Y mice with Slc26a4 919-2A>G/919-2A>G mice. The latter is a transgenic strain previously established in our laboratory with confirmed abolished Slc26a4 function. Heterozygous mice were used to clarify the pathogenicity of p.C565Y in monoallelic, biallelic, or compound heterozygous forms.

Discussion
In this study, we generated a transgenic mouse model harboring the p.C565Y variant of the Slc26a4 gene to examine the pathogenicity of the p.C565Y variant in mice. Phenotypic characterization revealed that both homozygous (i.e., Slc26a4 C565Y/C565Y ) and compound heterozygous (i.e., Slc26a4 919-2A>G/C565Y ) mice exhibited hearing and balance comparable to those of wild-type mice, which was further confirmed by normal inner-ear morphology in histological studies. These findings indicate that the p.C565Y variant of Slc26a4 gene is nonpathogenic for mice.
The SLC26A4 p.C565Y variant was documented as a pathogenic variant in several previous reports, where p.C565Y was exclusively present in trans with another SLC26A4 mutation, including p.Q514R [9,10], p.L236P [11], and p.H723R [12]. With the aid of the Varsome platform [25], which combines a human genetic-frequency database (e.g., gnomAD), a disease variant database, (e.g., Clinvar, Uniprot), and many predictive algorithms to assess variant pathogenicity, p.C565Y was categorized as a likely pathogenic variant in accordance with criteria of the American College of Medical Genetics and Genomics (ACMG) guidelines for PM1, PM2, PM5, PP2, PP3, and PP5 (see definitions, explanations, and classifications in Tables S1-S3, respectively). In addition to being classified as a pathogenic variant by many predictive algorithms (SIFT, DANN, EIGEN, FATHMM-MKL, Mutation-Taster, and others) and databases that include ClinVar, DVD, and Uniprot, p.C565Y also demonstrated low allelic frequency (AF) across populations (Popmax Filtering AF < 0.0001 in GenomAD) and location at a critical domain of pendrin where pathogenic SLC26A4 variants were recurrently detected. Therefore, the pathogenicity of the SLC26A4 p.C565Y variant in humans can be confirmed by several lines of evidence classified in the ACMG guidelines (Supplementary Tables S1-S3), indicating difference between humans and mice.
The mutation landscape of the SLC26A4 gene significantly differs among different ethnic groups [12,[26][27][28]. Previous clinical studies presented inconsistent results in regard to the correlation between SLC26A4 genotypes and phenotypes [29][30][31]. In several studies, the authors did not identify any phenotypic differences among patients with different SLC26A4 variants [32,33]. However, some variants might lead to the progression of hearing loss [34,35], and some SLC26A4 genotypes may be associated with a normal thyroid phenotype and less severe hearing loss [36]. The p.C565Y variant was clinically detected in three patients. All three segregated p.C565Y in the trans configuration with another SLC26A4 variant. One patient harbored the p.Q514R/p.C565Y genotype. This patient was diagnosed as having Pendred syndrome and bilateral EVA, goiter, and abnormal perchlorate discharge value [9,10]. The patient with the p.L236P/p.C565Y genotype also suffered from goiter, abnormal perchlorate discharge value, and early childhood SNHI [11]. The final patient with the p.C565Y/p.H723R genotype was diagnosed as nonsyndromic EVA without thyroid manifestation [12]. Otherwise limited clinical information makes it difficult to delineate a clear phenotype that is specifically associated with the p.C565Y variant.
Consistent with these recent clinical reports, studies based on cell-line experiments demonstrated that different SLC26A4 variants have different pathogenic mechanisms. Some variants, such as p.H723R, p.L236P, and p.T721M, are associated with defective protein expression or trafficking, whereas some other variants, such as p.K369E and p.S166N, are associated with normal protein expression but impaired protein function [7,8]. Notably, even among each subgroup, a gradient of pathogenicity appears to exist. For instance, the trafficking of pendrin in the p.H723R variant could be rescued by salicylate and temperature manipulation, whereas pendrin in the p.L236P variant could not be rescued. The p.C565Y variant seems to be nonpathogenic in cell lines, as pendrin with p.C565Y is expressed correctly in the plasma membranes in both HEK 293 [7,8] and COS-7 cell lines [9], and is also shown to exhibit normal protein function in Xenopus oocytes [9].
Despite the presence of these correlations in the pathogenicity of SLC26A4 variants, cell-line and transgenic mouse experiments still possess some limitations that preclude them from being satisfactory experimental models for SLC26A4-related SNHI. First, some variants that are supported by strong evidence to be clinically pathogenic, such as p.K369E and p.C565Y, did not reflect any abnormality in protein expression or function in cell lines [8,9,37], or audiovestibular phenotypes in transgenic mice [16]. The trafficking or glycosylation process might also significantly differ between mice and humans [38], which could explain the absence of phenotypes in mice with the p.C565Y and p.H723R variants of the Slc26a4 gene. Alternatively, as p.C565Y was documented as pathogenic in humans only when present in trans with other SLC26A4 variants, such as p.Q514R, p.L236P, and p.H723R [9][10][11][12], it cannot be excluded that p.C565Y homozygosity alone is nonpathogenic. An extensive study combining these other variants may be necessary for a comprehensive understanding of the pathogenicity of the p.C565Y variant, considering that human clinical data are limited. Second, the amino acid sequence of pendrin differs across species. For instance, the amino acid sequence of the pendrin C terminus (i.e., amino acids 508-780) shares only 86% identity between mice and humans (https://www.expasy.org/ (accessed on 18 August 2020), which might explain the inconsistent pathogenicity of p.C565Y between mice and humans. To address this, humanized transgenic mice generated with mutant human cDNA sequences could provide a solution, as this approach can better recapitulate clinical phenotypes in humans [19,39]. Third, even for transgenic mice with positive phenotypes, there are still valid discrepancies in the audiovestibular features that are reported between mice and humans. For instance, the congenital profound SNHI reported in Slc26a4 −/− , Slc26a4 loop/loop , hH723R, and Slc26a4 919-2A>G9/919-2A>G mice is too severe when compared to their phenotypes in human counterparts. Although Slc26a4 L236P/L236P and Tg[E]; Tg[R]; Slc26a4 ∆/∆ mice showed some residual hearing, neither could perfectly recapitulate the fluctuating or progressive disease nature in humans. Further studies are required to refine cell-line or transgenic-mouse models to investigate and better understand SLC26A4-related SNHI.
In conclusion, we generated a knock-in mouse model that segregated the deafnessassociated SLC26A4 p.C565Y variant in humans in a genotype-driven approach. Subsequent phenotypic characterization revealed that mice with the Slc26a4 p.C565Y variant exhibit normal audiovestibular phenotypes and inner-ear morphology, indicating that the pathogenicity associated with specific SLC26A4 variants might differ between humans and mice. Therefore, with regard to research on SLC26A4-related SNHI, caution should be taken when translating results of animal studies to humans.

Generation of Slc26a4 C565Y/C565Y Knock-In Mice
Transgenic mice were generated using the Transgenic Mouse Models Core (TMMC, Taiwan) and the clustered regularly interspaced short palindromic repeat (CRISPR) technologyassociated RNA-guided endonuclease Cas9 to mutate the Slc26a4 gene and further generate the Slc26a4 +/C565Y mouse strain. Specific guide RNAs (sgRNAs) were developed to target exon 15 of the Slc26a4 gene in the C57BL/6 mouse strain. The sgRNA and CRISPR/Cas9 RNA were delivered simultaneously into the zygote of a C57BL/6 mouse to generate the founders. Two male founder mice were obtained, each possessing the p.C565Y (c.1694G>A) variant in the Slc26a4 gene. After germline transmission of the targeted mutation allele, we were able to generate the congenic Slc26a4 +/C565Y mouse line by repeated backcrossing with the C57BL/6 inbred strain for 6-10 generations. Thereafter, the homozygous mice for the Slc26a4p.C565Y variant (Slc26a4 C565Y/C565Y ) were obtained by intercrossing heterozygous mice (Slc26a4 +/C565Y × Slc26a4 +/C565Y ; Figure 1A,B).
Corresponding to the human genotypes, compound heterozygous mice (i.e., Slc26a4 919-2A>G/C565Y ) with both the p.C565Y and c.919-2A>G variants were also generated by intercrossing heterozygous Slc26a4 +/C565Y mice with Slc26a4 919-2A>G/919-2A>G mice [16]. Mice were housed at the National Taiwan University College of Medicine Laboratory Animal Center. The mice were bred and weaned at approximately 3 weeks of age in a vivarium with an alternating 12 h light/dark cycle at controlled temperature (20-26 • C) and humidity (40-70%). Bedding was refreshed once a week. Water and food were always available. All animal experiments complied with animal-welfare guidelines and were approved by the Institutional Animal Care and Use Committee (IACUC) of the National Taiwan University College of Medicine (approval no. 20160337, 1 November 2016). For euthanasia, we intraperitoneally injected barbiturate (100 mg/kg) into each mouse and performed cervical dislocation. All experiments were performed without regard to the sex of mice, and mice were randomly selected for every independent measurement. In total, 75 mice were used in this study, including Slc26a4 +/+ , Slc26a4 C565Y/C565Y , Slc26a4 919-2A>G/C565Y , and Slc26a4 919-2A>G/919-2A>G .

Auditory Evaluations
For audiological evaluations, 28 day postnatal mice (P28) were intraperitoneally anesthetized with pentobarbital (35 mg/kg; Sigma-Aldrich, St. Louis, MO, USA) and maintained in an acoustically and electrically insulated, and grounded test room [16]. The thresholds of the auditory brainstem response (ABR) in mice were measured with an evoked potential detection system (Smart EP 3.90; Intelligent Hearing Systems, Miami, FL, USA). Click sounds and 8, 16, and 32 kHz tone bursts at varying intensity were given to evoke the ABRs of subject mice. To determine ABR thresholds, sound-pressure levels (SPL) between 15 and 70 dB SPL were used. Response signals were detected using subcutaneous needle electrodes.

Vestibular Evaluations
For vestibular evaluations, the mice were subjected to a series of tests (all performed at 8 weeks of age), including a swimming test and the rotarod test. For the swimming test, the swimming performance of mice was scored from 0 to 3, with 0 being normal swimming and 3 being underwater tumbling [40]. For the rotarod test, mice were placed on a rotating rod for a maximum of 240 s. The speed of the rod's rotation was accelerated from 5 rpm to the maximal speed of 20 rpm for 1 min. The duration for which each mouse was able to remain on the rotating rod was recorded [41].

Inner-Ear Histology Studies
To perform light-microscopy analysis, tissue samples were subjected to hematoxylin and eosin (H&E) staining, and the morphology of each sample was determined using a Leica optical microscope (Leica). First, inner-ear tissue from adult mice (3 months) was harvested and fixed using perilymphatic perfusion with 4% paraformaldehyde (Bio Basic Inc., Markham, ON, Canada), prepared in phosphate-buffered saline (PBS). Specimens were decalcified for 1 week. Subsequently, samples were dehydrated and embedded in paraffin. Serial sections of 7 µm thickness were obtained and stained with H&E.
For the whole-mount studies of mouse inner ears, specimens were prepared as previously described [16]. Tissue samples were stained with rabbit anti-Myosin-VIIA primary antibody (1:200; Proteus Bioscience Inc., Ramona, CA, USA). Later, tissue sections were incubated with DAPI (1:5000; Thermo Fisher, Waltham, MA, USA) and Alexa Fluor 568-conjugated goat antirabbit IgG (H + L) secondary antibodies (1:200; Thermo Fisher, Waltham, MA, USA) at 4 • C overnight. After washing with PBS to remove any residual antibodies, tissue samples were mounted using the ProLong Antifade Kit (Molecular Probes, Eugene, OR, USA) for 20 min at room temperature. Images of mice inner-ear tissue samples were captured using a laser-scanning confocal microscope (LSM880, Zeiss, Oberkochen, Germany).

Statistical Analyses
Data are presented as mean ± SD. Statistical analyses were conducted using unpaired Student's t-test with Bonferroni correction for continuous variables. A p value < 0.05 indicates significance. All analyses were performed using SPSS/Windows software 15.0 (SPSS Inc., Chicago, IL, USA).