Hearing loss is the most common sensory deficit in humans, affecting around 1 in 1000 newborns. Its prevalence increases with the age up to 6–8% in the adult population, having a strong impact on the individual’s social isolation [1]. Genetic causes account for 50–60% of newborn hearing loss, 30% of which are nonsyndromic forms of deafness [2]. The genetic etiology of hearing impairment is highly heterogeneous. Up to date, 160 nonsyndromic sensorineural hearing loss (NSSNHL) loci have been mapped, of which 122 genes have been identified [3]. Sixty-seven of these loci and 50 of these genes are associated with autosomal dominant NSSNHL, being, in most cases, rare forms of post-lingual and progressive hereditary hearing impairment.

Autosomal dominant NSSNHL-linked genes encode proteins with a wide variety of functions [4], such as cytoskeleton proteins (ACTG1, DIAPH1, PLS1), adhesion proteins (GJB2, GJB3, GJB6, TJP2), motor proteins (myosins), scaffolding proteins (HOMER2), extracellular matrix proteins (TECTA, COL11A2), proteins involved in ion homeostasis, such as ionic channels (KCNQ4), transcription factors (EYA4, POU4F3), and even a microRNA (MIR96) [5].

The HOMER2 gene maps to chromosome 15q24.3 [6] within the DFNA68 critical interval and consists of nine exons. Two human transcript variants have been described (NM_004839.4 and NM_199330.3) as encoding the short isoform 1 (NP_004830.2, 343aa) and the long isoform 2 (NP_955362.1, 354aa) of HOMER2, respectively. HOMER2 belongs to a protein family encompassing three members: HOMER1 (MIM604798), HOMER2/CUPIDIN (MIM604799), and HOMER3 (MIM604800). Like HOMER2, members 1 and 3 of the family have long and short isoforms due to alternative splicing [7]. HOMER family members are scaffolding proteins that play a key role in Ca²⁺ signaling [8,9,10], mostly at the Post-Synaptic-Densities (PSD) [11], where they interact with G-protein coupled metabotropic glutamate receptors (mGluRs) [12] and regulate excitatory signal transduction and receptor plasticity [7]. In their structure, HOMER proteins present a conserved N-terminal domain known as Enabled/vasodilator-stimulated phosphoprotein (Ena/VASP) homology 1 (EVH1) domain [7,13] that binds to proline-rich sequences (i.e. Pro-Pro-x-x-Phe, Pro-x-x-Phe or Leu-Pro-Ser-Ser-Pro, where x represents any amino acid) and a C-terminal domain that consists of a coiled-coil (CC) structure that includes a CDC42 binding domain (CBD) and two Leucine Zipper (LZA and LZB) motifs [7,14]. This C-terminal fragment mediates self-association with other Homer family members [7] and the interaction with the small GTPase CDC42 [15] through its CBD.

In mice, Homer2 is widespread in the developing and maturing brain [16]. Recently, a study has shown that this gene is also expressed in a wide variety of developing tissues, including tooth, eye, cochlea, salivary glands, olfactory and respiratory mucosae, bone, and taste buds, being highly concentrated at puncta [17]. HOMER2 exhibits overlapping distribution patterns with HOMER1 and HOMER3, although they are distributed at distinct subcellular domains in several cell types [17,18]. Within the inner ear, HOMER2 is expressed in the stria vascularis, the Reissner’s membrane, and the inner and outer hair cells of the organ of Corti, especially in the stereocilia but also in perinuclear puncta and the cytoplasm [17,18]. Mild expression is observed in the vascular endothelium of the cochlea and the spiral ganglion [12]. Mice homozygous for the targeted deletion of Homer2 display early-onset rapidly progressive hearing loss [18].

HOMER2 was firstly associated with hearing loss by Azaiez et al. [18] who identified a heterozygous missense variant (p.Arg196Pro) in a European ascent family. A second mutation in HOMER2 (c.840_840dup; p.Met281Hisfs*9) was identified by Lu et al. [19] by Whole Exome Sequencing (WES) in a Chinese family segregating with hearing loss. Here, we present a third variant in HOMER2 (c.832_836delCCTCA, p.Pro278Alafs*10, NM_199330.3). This variant represents the second truncating mutation that affects the CBD and lead, as in the previous two families, to post-lingual and progressive hearing loss. This mutation is the first one identified in the Spanish population, thus increasing the mutational spectrum of this gene associated with DFNA68, a rare form of autosomal dominant nonsyndromic progressive hearing loss.

Hearing loss caused by HOMER2 mutations is an extremely rare disorder. In this study, we have used OTO-NGS-v2, a custom targeted NGS panel, for the identification of a novel mutation (c.832_836delCCTCA; p.Pro278Alafs*10) in this gene linked to DFNA68 hearing loss. To date, only two different mutations in HOMER2 have been documented to cause hearing loss. The first mutation was a missense substitution (c.587G>C; p.Arg196Pro; NM_199330.3) that affects a highly conserved residue in the coiled-coil (CC) structure, a region that is required for homo/hetero-multimerization to form tetrameric hubs (in which the CC domains align in a parallel fashion) and for interaction through the CBD with Rho family GTPase proteins like CDC42, a GTPase that mediates actin-turnover [28] and is responsible for planar polarity establishment in hair cells [29,30]. Functional assessment of the p.Arg196Pro in zebrafish strongly suggests that this mutation exerts its effect through a dominant-negative mechanism on wild-type protein by either inhibiting multimerization or competing for partner proteins [18]. The second mutation, c.840_840dup, p.Met281Hisfs*9 (NM_199330.3) was identified in a Chinese family with symmetric ADSNHL [19]. This frameshift variant at the CDC42-binding domain leads to the generation of a premature stop codon supposed to produce a truncated HOMER2 protein of 288 amino acids, with an aberrant protein tail of 8 amino acids from position 281 onwards.

In this work, we have identified a novel frameshift mutation, c.832_836delCCTCA, p.Pro278Alafs*10 in HOMER2 that represents the second truncating mutation identified in the CDC42-binding domain. The Spanish mutation also creates a premature stop codon that may lead to generate a truncated shorter protein of 286 amino acids with an aberrant tail of 9 amino acids from position 278, thus lacking the canonical C-terminal end. Interestingly, the alignment of both truncated proteins revealed an identical sequence of the last 8 amino acids of the aberrant tails (Figure 2).

Based on the similar effect that the two different frameshift mutations may cause at the protein level, both supposed to truncate the protein at the CDB and resulting in the same aberrant tail, it might be reasonable to suggest that the Chinese and Spanish mutations may share a similar mechanism of pathogenesis. Lu et al. demonstrated that the p.Met281Hisfs*9 mutant protein was less stable than the wild-type protein and it showed an altered subcellular localization in HEK293T and HEI-OC1 cells. Whereas the wild-type proteins were mainly aggregated near the nucleus, the p.Met281Hisfs*9 mutants were more widely distributed throughout the cytoplasm. Furthermore, these authors demonstrated that p.Met281Hisfs*9 showed a decreased ability to oligomerize [19]. It has also been postulated that HOMER2 could play an important role in maintaining stereocilia through its interaction with CDC42 [19], therefore the truncation of HOMER2 in the CBD could eventually prevent its interaction with other HOMER protein family members and with other proteins like CDC42. Indeed, targeted deletion of Cdc42 in murine hair cells causes a progressive hearing loss phenotype that is comparable to the hearing loss phenotype in the Spanish and Chinese families [19]. Another possibility is that the HOMER2 mutation identified in this study may cause nonsense-mediated decay (NMD), a mechanism that affects the processing of the transcripts at different extent depending on the type of mutation and gene involved as previously reported in other pathologies like Neurofibromatosis I [31]. In this regard, and in contrast to mouse mutants homozygous for the targeted deletion of Homer2 that display early-onset rapidly progressive hearing loss, mice heterozygous for the targeted deletion of exon 3 in Homer2 (Homer^−/+) displayed normal hearing levels [18]. It may indicate that a moderate or even low extent NMD might be associated with HOMER2 frameshift mutations thus suggesting that the pathophysiology of the DFNA68 hearing loss in the Spanish and Chinese families would not be mediated by haploinsufficiency, but by a gain-of-function of the Ct aberrant tail or by a dominant-negative mechanism as it has been postulated for the p.Arg196Pro missense mutation [18]. However, more experiments to detect HOMER2 mutant transcripts levels by using a gene-editing tool to mimic actual mutation in cell lines or the generation of knock-in murine models are necessary to fully understand the underlying mechanism of pathogenesis linked to DFNA68 frameshift mutations.

Regarding the clinical phenotype, the hearing loss observed in HOMER2 patients caused by the missense (p.Arg196Pro) and two truncating mutations seems to be quite similar in the three studied families. Affected individuals show progressive hearing loss affecting mainly the high frequencies (downsloping profile) with a typical onset in the first decade of life (7–9 years old), although a more severe phenotype was detected in patients bearing the truncating mutations when age-matched hearing-impaired subjects of the three families were compared. For individuals in a similar age-range, those carrying the missense p.Arg196Pro mutation [18] displayed minor affectation on the entire range of frequencies (i.e., subject IV:2, 8y.o, IV:10, 15y.o, and III:4, 34y.o, European ascent family) than patients carrying the p.Met281Hisfs*9 (subject IV:7, 7y.o; III:10, 34y.o, and III:8, 39y.o in the Chinese family) or the p.Pro278Alafs*10 variant (subject II:2, 8y.o; II:3, 15y.o, and I:2, 39y.o in the Spanish family). The presence of tinnitus or cranial tinnitus, however, has only been displayed by some affected members of the Chinese family and this phenotype was not present in any of the other two families. The identification of other HOMER2 cases is, therefore, necessary to further establish more accurate genotype–phenotype correlations.

In summary, we have identified a third variant in HOMER2; the first one reported in the Spanish population, thus contributing to expanding the mutational spectrum of this gene in other populations. Our study also highlights the importance of using NGS-based diagnostic methods to identify mutations in low-prevalence deafness genes like HOMER2, thus helping to improve our knowledge about the pathophysiology of DFNA68 and to define more accurate genotype–phenotype correlations in this disorder.

The authors sincerely thank the family for their participation in this study.

Conceptualization, M.M. and M.Á.M.-P.; Data curation, M.L., M.V. and I.d.C.; Formal analysis, M.L. and M.M.; Funding acquisition, M.Á.M.-P.; Methodology, M.M.; Supervision, M.Á.M.-P.; Writing—original draft, M.L., M.M. and M.Á.M.-P. All authors have read and agreed to the published version of the manuscript.

This research was supported by grants from the Spanish Institute of Health Carlos III (ISCIII) cofunded with the European Regional Development Fund (ERDF) within the “Plan Estatal de Investigación Científica y Técnica y de Innovación 2017–2020” (PI14-948, PI17-1659 and PI20/0429 to MAMP), the Spanish Center for Biomedical Network Research on Rare Diseases (CIBERER, 06/07/0036 grant, to MAMP) and by the Regional Government of Madrid (CAM, B2017/ BMD3721 grant to MAMP).

The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Institutional Review Board (or Ethics Committee) of Hospital Ramón y Cajal (protocol code 288-17; 29 January 2018).

Informed consent was obtained from all subjects involved in the study.

The data presented in this study are openly available in ClinVar (accession# SCV001499845).

The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.