First confirmatory study on PTPRQ as an autosomal dominant non-syndromic hearing loss gene

Hearing loss (HL) is the most common disability of human senses. It is considered to have a monogenic origin in at least half of the individuals developing HL prior to speech acquisition. Similar data are not available for postlingual HL but it is noticeable that familial aggregation of later-onset, progressive HL often follows an autosomal dominant pattern of inheritance. Up to now, 46 genes involved in the development of autosomal dominant non-syndromic HL (ADNSHL) have been identified (https://​hereditaryhearin​gloss.​org; accessed 10/2019). New genes related to ADNSHL are still being discovered and many of the identified variants in known HL genes represent novel changes. This highlights a great genetic heterogeneity of ADNSHL [1, 2].

In 2018, Eisenberger et al. [3] proposed PTPRQ (MIM*603317) as a new candidate gene for ADNSHL. The conclusion was based on the identification of a novel heterozygous nonsense PTPRQ variant p.Trp2294* that cosegregated with progressive HL in a German family. PTPRQ encodes protein tyrosine phosphatase receptor type Q, which catalyzes the dephosphorylation of phosphatidylinositol phosphates and displays a low activity against phosphotyrosine [4, 5]. Expression of PTPRQ in the inner ear is confined to hair bundles in the cochlea and in the vestibule and distributed depending on the hair cell type or its location, primarily in the basal region of hair bundles. In the cochlea, high PTPRQ protein level is detected in its basal turn responding to high frequency sound [6]. The main PTPRQ isoform in the inner ear contains a long extracellular domain, a short hydrophobic transmembrane region and an intracellular domain with a single phosphatase catalytic site [7].

The p.Trp2294* variant is predicted to introduce a premature termination codon (PTC) shortening the PTPRQ protein at the C-terminus by the last six amino acid residues. However, PTPRQ expression at the mRNA level remains unaffected, indicating that the mutant transcript escapes nonsense-mediated decay and supporting the hypothesis that the truncated PTPRQ protein is produced [3]. Performing additional functional studies to confirm the mutant variant mechanism of action was hampered by the large size of the PTPRQ protein and small difference in size between the wild-type and mutant protein. Since the initial report no additional data have been provided to unequivocally confirm PTPRQ as an ADNSHL gene.

Our study provides novel family-based evidence, confirming the pathogenic role of PTPRQ in ADNSHL. Based on the results of clinical exome sequencing, we identified the same PTPRQ variant as Eisenberger et al. in a new five-generation Polish ADNSHL family and found that both families have a common ancestor. Our strong genetic and clinical data constitute an ultimate confirmation for the causative role of PTPRQ in ADNSHL.

Affected family members from the examined Polish family suffered from bilateral, progressive, high-frequency HL with an onset from 4 to 27 years of age. The degree of HL showed a high intrafamilial variability ranging from an almost normal hearing status to severe HL (Table 1). Detailed audiological evaluation revealed severe HL in the proband (IV.2, 32 y/o) and mild to moderate HL in her son (V.1, 12 y/o) at mid and high frequencies that was accompanied by tinnitus. At low frequencies hearing thresholds were within the normal range (Fig. 1a–c). In both individuals, tympanograms were normal and stapedial reflexes were present with ipsi- and contralateral stimulation with the exception of high frequencies where their thresholds were increased or absent. Otoacoustic emissions were bilaterally absent in the proband and recorded in her son up to 1500 Hz in the right and up to 1000 Hz in the left ear. Corresponding with the HL degree, ABRs recordings were lacking waves I and III, while wave V latencies were normal. Combined results of the objective hearing measurements showed bilateral hearing impairment originating from a defective cochlear component of the auditory system. Neurotological examinations showed cVEMP and oVEMP responses and results of bi-thermal caloric irrigation within normal limits. There were no malformations of the auditory system on temporal bone imaging (Additional file 1: Figures S1–S7).

Molecular genetic testing showed the presence of a heterozygous NM_001145026.2:c.6881G>A variant in the PTPRQ gene, which fully segregated with ADNSHL in the studied family (Fig. 1a, d, e). The identified variant is located in the last coding exon of PTPRQ and introduces a PTC (NP_001138498.1:p.Trp2294*). The PTPRQ variant has not been reported in population databases but was detected in a single previously published German ADNSHL family [3]. Based on European descent and the close geographic proximity, a common ancestor was assumed. An affected and an unaffected offspring together with their parents were genotyped on a genome-wide single nucleotide polymorphism (SNP) array. The region surrounding the shared PTPRQ variant was analyzed for linkage in each family. The subsequent reconstruction of haplotypes revealed that a SNP haplotype spanning 306 kb (from rs1358476 to rs7978990) is shared by both families (Additional file 2: Figure S8).

In addition to the PTPRQ variant identified in the proband, we have found six other genetic alterations in HL genes that had population allele frequency below 0.001 and were present in protein coding regions. Considering the mode of HL inheritance and literature data, we have selected the TMC1 NM_138691.2:c.1705A>G variant for family segregation study and found that it was present in normal hearing individuals and absent in some affected family members [10]. Its presence did not correspond with a more severe HL phenotype within the family.

Our report is the first confirmatory study on the causative role of PTPRQ in ADNSHL development. Here, we have identified eight previously unreported individuals suffering from progressive, high-frequency ADNSHL caused by the c.6881G>A PTPRQ nonsense variant (p.Trp2294*), which represents the first and so far the only PTPRQ pathogenic variant involved in ADNSHL. Up to now, it was reported exclusively by Eisenberger et al. [3] who, after detailed examination of a German family, proposed PTPRQ as a novel ADNSHL candidate gene (assigned at the DFNA73 locus 12q21.31). Their assumption was based on the results of exome sequencing, supported by linkage analysis, extended sequencing of the alternatively spliced exons and in silico predicted, potential PTPRQ coding regions.

In their paper, Eisenberger et al. have listed variants in genes reportedly associated with hearing impairment and variants located in the mapped candidate region encompassing PTPRQ that were included in the final analysis of potential ADNSHL pathogenic variants in the studied family. We have carefully compared the high throughput sequencing data of the German proband with the sequencing results of our proband and found that the majority of variants were not present in our patient except for variants within SYT1 and TMTC2 genes flanking PTPRQ on chromosome 12. Current data from the gnomAD database showed that both variants have a high population frequency (0.0023 and 0.0097, respectively) and are present in homozygous state in 2 and 43 individuals, respectively. This allowed to unequivocally exclude both genes from further consideration.

To verify whether the detected PTPRQ variant is located within the same inherited chromosomal region or represents an independent event that occurred by chance at the same position in both families, genome-wide linkage analysis was performed. After testing four family members from each family, we found that the PTPRQ c.6881G>A variant is present within a haplotype shared by the affected individuals from both families. It is located almost in the middle of a common chromosomal region containing the distal half of the PTPRQ gene together with the following 3′ region. The data have clearly demonstrated that the c.6881G>A containing PTPRQ chromosomal region did not undergo recombination and HL patients in both families share copies of the same part of the ancestral PTPRQ haplotype.

Here, for the first time we have performed deep phenotyping of the affected individuals and found that the defect caused by the pathogenic PTPRQ p.Trp2294* variant is restricted to the cochlear component of the auditory system. In contrast to individuals with PTPRQ recessive variants who suffer from vestibular dysfunction diagnosed in infancy or early childhood, our patients do not develop a vestibular disorder [11, 12].

The exact function of the PTPRQ protein remains poorly understood. Its large extracellular part forms shaft connectors most probably involved in proper spacing of stereocilia and maintaining their stability. The intracellular fragment containing an active enzymatic site was found to regulate the level of phosphatidylinositol phosphates. It has been also proposed that the PTPRQ C-terminus interacts with a protein complex containing MYO6, RDX, TPRN and CLIC5 [13], encoded by known HL genes. Proper function of the protein complex is required to maintain typical localization of PTPRQ at the base of stereocilia. In MYO6- and CLIC5-deficient mice, PTPRQ cannot be properly compartmentalized and it dissociates to the upper parts of the shafts leading to destabilization of membrane-cytoskeletal attachments, membrane lifting and consequent fusing or loosing of stereocilia [13‐16].

In summary, PTPRQ is better known as a recessive HL gene [11, 12, 17‐21]. The present study strongly reinforces its inclusion to the small set of genes leading to both autosomal recessive and dominant hearing loss. Our data provide the first independent confirmation of PTPRQ causative role in ADNSHL.