Otoferlin as a multirole Ca2+ signaling protein: from inner ear synapses to cancer pathways

Humans have six members of the ferlin protein family: dysferlin, myoferlin, otoferlin, fer1L4, fer1L5, and fer1L6. These proteins share common features such as multiple Ca2+-binding C2 domains, FerA domains, and membrane anchoring through their single C-terminal transmembrane domain, and are believed to play a key role in calcium-triggered membrane fusion and vesicle trafficking. Otoferlin plays a crucial role in hearing and vestibular function. In this review, we will discuss how we see otoferlin working as a Ca2+-dependent mechanical sensor regulating synaptic vesicle fusion at the hair cell ribbon synapses. Although otoferlin is also present in the central nervous system, particularly in the cortex and amygdala, its role in brain tissues remains unknown. Mutations in the OTOF gene cause one of the most frequent genetic forms of congenital deafness, DFNB9. These mutations produce severe to profound hearing loss due to a defect in synaptic excitatory glutamatergic transmission between the inner hair cells and the nerve fibers of the auditory nerve. Gene therapy protocols that allow normal rescue expression of otoferlin in hair cells have just started and are currently in pre-clinical phase. In parallel, studies have linked ferlins to cancer through their effect on cell signaling and development, allowing tumors to form and cancer cells to adapt to a hostile environment. Modulation by mechanical forces and Ca2+ signaling are key determinants of the metastatic process. Although ferlins importance in cancer has not been extensively studied, data show that otoferlin expression is significantly associated with survival in specific cancer types, including clear cell and papillary cell renal carcinoma, and urothelial bladder cancer. These findings indicate a role for otoferlin in the carcinogenesis of these tumors, which requires further investigation to confirm and understand its exact role, particularly as it varies by tumor site. Targeting this protein may lead to new cancer therapies.

Humans have six members of the ferlin protein family: dysferlin, myoferlin, otoferlin, fer1L4, fer1L5, and fer1L6. These proteins share common features such as multiple Ca 2+ -binding C2 domains, FerA domains, and membrane anchoring through their single C-terminal transmembrane domain, and are believed to play a key role in calcium-triggered membrane fusion and vesicle trafficking. Otoferlin plays a crucial role in hearing and vestibular function. In this review, we will discuss how we see otoferlin working as a Ca 2+ -dependent mechanical sensor regulating synaptic vesicle fusion at the hair cell ribbon synapses. Although otoferlin is also present in the central nervous system, particularly in the cortex and amygdala, its role in brain tissues remains unknown. Mutations in the OTOF gene cause one of the most frequent genetic forms of congenital deafness, DFNB9. These mutations produce severe to profound hearing loss due to a defect in synaptic excitatory glutamatergic transmission between the inner hair cells and the nerve fibers of the auditory nerve. Gene therapy protocols that allow normal rescue expression of otoferlin in hair cells have just started and are currently in pre-clinical phase. In parallel, studies have linked ferlins to cancer through their effect on cell signaling and development, allowing tumors to form and cancer cells to adapt to a hostile environment. Modulation by mechanical forces and Ca 2+ signaling are key determinants of the metastatic process. Although ferlins importance in cancer has not been extensively studied, data show that otoferlin expression is significantly associated with survival in specific cancer types, including clear cell and papillary cell renal carcinoma, and urothelial bladder cancer. These findings indicate a role for otoferlin in the carcinogenesis of these tumors, which requires further investigation to confirm and understand its exact role, particularly as it varies by tumor site. Targeting this protein may lead to new cancer therapies.

Introduction
Interest in otoferlin began with the discovery of mutations in the OTOF gene that were identified as responsible for recessive profound deafness in humans (Yasunaga et al., 1999(Yasunaga et al., , 2000. This deafness is called DFNB9 (for autosomal recessive deafness 9). It represents about 2% of non-syndromic prelingual deafness, i.e., without any other associated health problem, and is the most frequent cause of auditory neuropathy spectrum disorder (ANSD) (Iwasa et al., 2019). To date, about 220 mutations leading to DFNB9 deafness have been identified (Azaiez et al., 2021). Remarkably, some otoferlin mutations, notably in the C2F domain, can lead to exacerbated deafness when body temperature rises during a fever (Varga et al., 2006;Marlin et al., 2010;Matsunaga et al., 2012;Strenzke et al., 2016). This peculiar property underlines the importance of a precise functional 3D architecture of the multi-C2 protein, which is known to interact at the presynaptic active zone of the sensory hair cells with SNARE (soluble N-ethylmaleimide-sensitive-factor attachment protein receptor) proteins such as SNAP25 (Roux et al., 2006;Ramakrishnan et al., 2014;Hams et al., 2017;Calvet et al., 2022) and with Ca v 1.3 Ca 2+ channels (Ramakrishnan et al., 2009;Vincent et al., 2017Vincent et al., , 2018; Table 1 and Figures 1, 2).
DFNB9 is caused by a defect in synaptic neurotransmission from auditory inner hair cells (IHCs) to their contacting spiral ganglion afferent fibers (Roux et al., 2006;Beurg et al., 2010;Michalski et al., 2017). The sound-induced electrical analog signal of the hair cells, also called cochlear microphonic, is incorrectly transduced into nerve impulses at the auditory nerve fibers. We recall that one of the morphological characteristics of the afferent IHC synapses is the presence of a presynaptic ribbon, an electrondense presynaptic structure 300-350 nm wide (Figures 1, 3). Each of the 15-20 presynaptic ribbons per IHC are facing a single postsynaptic auditory nerve structure bearing AMPA (glutamate) ionotropic receptors composed of GluA2, GluA3, and GluA4 subunits (Rutherford et al., 2023). This presynaptic ribbon, composed of RIBEYE proteins arranged in staircase pattern, allows the attachment of numerous synaptic vesicles at the membrane active zone of release. This ribbon structure is essential for spike timing precision of the auditory nerve fibers (Wittig and Parsons, 2008;Becker et al., 2018;Jean et al., 2018).
Otoferlin-induced synaptopathy is due to a defective Ca 2+evoked exocytosis of synaptic vesicles containing glutamate, the main cochlear excitatory neurotransmitter, at the IHC ribbon synapses (Figure 3). Like in most neurosecretory cells, vesicle trafficking and synaptic exocytosis in IHCs are thought to involve interactions of SNARE complex proteins such as SNAP25 (Calvet et al., 2022). Synaptotagmins (Syts), a large family of transmembrane proteins containing tandem Ca 2+ -binding C2domains, confer Ca 2+ sensitivity to SNARE-dependent vesicle fusion in the CNS (central nervous system) (Chapman, 2008). However, the implication of Syts in hair cell synaptic exocytosis has been shown to be limited to immature cochlear developmental stages when hair cells fire action potentials (Beurg et al., 2010). At post-hearing mature stages, when hair cells display gradual depolarization upon sound stimulation, otoferlin becomes the major Ca 2+ sensor triggering synaptic vesicle exocytosis.
DFNB9 is classified as a neuropathy and, more specifically, as an auditory cochlear synaptopathy (Moser and Starr, 2016). In addition to the cochlea, otoferlin expression has also been shown in the brain with unknown function (Schug et al., 2006) and, in the peripheral vestibular organs where hair cell Ca 2+dependent fast synaptic vesicle exocytosis is also impaired (Dulon et al., 2009). In DFNB9 patients, it is important to note that the active electromechanical sound amplification function by the outer hair cells (OHCs) is normal, as indicated by measurements of the distortion products of otoacoustic emissions (DPOAEs). The DFNB9 auditory cochlear synaptopathy is currently overlooked in clinics when the systematic hearing screening performed at birth relies only on DPOAE measurements. Since delays in the management of sensory pathology reduce the quality of language development, it is becoming essential to screen for deafness with automated auditory brainstem recordings (ABRs). For DFBNB9 hearing loss, hearing aids or cochlear implantation are the only effective treatments currently available. Future clinical gene therapies are on the way, as demonstrated by successful rescue viral gene therapies in otoferlin-deficient mouse models (Al-Moyed et al., 2019;Akil, 2020;Rankovic et al., 2020). However, in these studies, for unknown reasons, there was only a partial rescue of ABR wave I amplitude (the neural response associated with the electrical activity of the IHC ribbon synapses), suggesting a nonrestored loss of some ribbon synapses. This incomplete rescuing is likely due to the late developmental stage (postnatal days P1 to P3 in mice) at which AAV treatment was performed in these studies. Indeed, otoferlin seems essential for normal early prenatal development and the maintenance of ribbon synapses (Stalmann et al., 2021).

Otoferlin structure and isoforms
Otoferlin is a large 1997 amino acid (aa) protein, that includes a single C-terminal transmembrane domain (TMD) anchoring the protein to the vesicular membrane and six C2 domains (A-F) oriented toward the IHC cytosol (Figures 3, 4). The C2 domains are structures composed of eight antiparallel β-sheets and negatively charged top loops due to the presence of five aspartate residues, constituting the putative binding site for Ca2 + ions (Sutton et al., 1995;Xue et al., 2008). The C2-A domain is the only C2 domain predicted to be unable to bind Ca 2+ due to a shorter top loop, thus lacking aspartates coordinating calcium binding (Helfmann et al., 2011). The main function of C2 domains is to target membrane surfaces following Ca 2+ -binding. C2 domains bind to their target membranes by using a combination of hydrophobic and electrostatic interactions, preferentially with phosphatidylinositol 4,5-bisphosphate (PIP2) or phosphatidylserine (PS) (Corbalan-Garcia and Gómez-Fernández, 2014). Otoferlin also has a FerA domain, a four-helix bundle fold with its own Ca 2+ -dependent phospholipid-binding activity, suggesting that the interaction of this domain with the membrane is enhanced by the presence of Ca 2+ (Harsini et al., 2018).
Otoferlin is produced mainly in the brain and inner ear (cochlea and vestibule), and in smaller amounts in other organs: the heart, liver, pancreas, kidney, and skeletal muscle (Yasunaga et al., 2000). Otoferlin appears as a globular protein when using the AlphaFold protein structure prediction software (Jumper et al., 2021; Figure 4). In humans, a long and short isoform of otoferlin have been identified (Yasunaga et al., 1999(Yasunaga et al., , 2000, with the long one comprising all six C2 domains and a C-terminal trans-membrane domain (TMD) (1997 aa), while the short one expresses only the last three C-terminal C2 domains (C2-DEF) with the TMD (1230 aa). The role of these two isoforms remains unknown. Of note, the short otoferlin isoform is not present in mice, whereas the human long isoform has 98% sequence similarity with murine otoferlin.
The long isoform of otoferlin is encoded by 48 exons, of which exons 6 and 31 can have alternative splicing. Exon 47  Otoferlin expression in auditory inner hair cells. The cochlea (here an open bulla from a guinea pig) is a small spiraled bone structure looking like a snail shell (A) that contains the hearing organ, the organ of Corti. This spiral-shaped sensory epithelium, lying on the basilar membrane between the scala vestibuli and the scala tympani, is composed of two types of hair cells: the outer hair cells (OHCs) which amplify the sound waves within the cochlea, and the inner hair cells (IHCs) which transduce the mechanical sound waves into nerve impulses. IHCs specifically express otoferlin as viewed in green under fluorescence immuno-confocal microscopy of a surface preparation of the organ of Corti (B). Otoferlin is essentially expressed at the basal synaptic pole of the IHCs (at right), where the afferent auditory nerve fibers make synaptic contacts with the presynaptic IHCs ribbons [red, enlarged view in (C)]. An antibody against CtBP2, a constituent of the RIBEYE-B domain, is used to label the synaptic ribbons. Note that CtBP2 is also a nuclear transcription factor, explaining the additional labeling of the cell nuclei. (D) The presynaptic IHC ribbons are closely colocalized with otoferlin, as indicated by the line-scan fluorescence intensity profiles of the ribbon indicated in (C). This is an original figure built from unpublished confocal immunostaining as previously described (Peineau et al., 2021).
also carries an alternative splice site, and upon splicing, exon 48 is expressed and encodes a transmembrane domain called TMD1. If exon 47 is not spliced, it bears a stop codon at its end, so the translation of exon 48 is excluded, and otoferlin has a different TMD domain, TMD2 (Figure 4). In humans and mice, the TMD2 otoferlin isoform (exon 47) is preferentially expressed in the brain, while the TMD1 isoform (exon 48) is specifically found in the cochlea (Yasunaga et al., 2000;Schug et al., 2006). Specific mutations in exon 48 cause hearing loss (Choi et al., 2009), underscoring the importance of TMD. According to Alphafold3D structure predictions, the differences between the two TMD isoforms are partly based on the angulation of the transmembrane domain relative to the rest of the protein ( Figure 4E). These two TMDs only share less than 57% amino acid sequence similarity ( Figure 4D). Notably, the brain TMD2 isoform contains a large peptidic sequence with numerous contiguous Leu residues (Leu-block of more than 9 Leu) in comparison to the cochlear TMD isoform, which contains discontinuous stretches of Leu residues ( Figure 4D). Large hydrophobic Leublocks in TMD peptides confer greater helicity and circumferential hydrophobicity that facilitate biological membrane insertion (Stone et al., 2015). The use of specific TMD should be important for targeting the protein to specific subcellular locations, such as particular glutamatergic excitatory presynaptic active zones, and specifying the interaction with specific organelles and ligands (Sharpe et al., 2010). Mutations in exon 48, such as the in-frame deletion of a conserved isoleucine in the cochlear TMD1 at position 1967 (p.Ile1967del), lead to hearing impairment and underline the importance of this domain in targeting the protein to the endoplasmic reticulum (ER) membrane in hair cells (Vogl et al., 2016). Another isoform, with an alternative splice in exon 31 might explain deafness caused by heat sensitivity, but its presence in the human cochlea remains uncertain (Strenzke et al., 2016). A recent study, by Liu et al. (2023) provides new insights into the role of alternative otoferlin isoforms in auditory function and their modulation by environmental factors such as noise and aging. This study discovered a new short transcript of otoferlin derived from an unannotated exon 6b whose expression is increased in IHCs encoding high frequency sounds and varies under noise and aging conditions.

Genetic mutations and DFNB9
Approximately 220 mutations causing DFNB9-type deafness have been identified, some of which cause deafness during a febrile episode (Vona et al., 2020;Azaiez et al., 2021). Interestingly, there is no described pathological mutation of the C2A domain ( Figure 5; Stenson et al., 2003). All other areas of the protein can be affected. According to the Human Gene Mutation Database, 41% of the variants are missense, 22% nonsense, 15% deletions, 14% splicing, 6% insertions, and 2% copy number variants. The missense variants alter the folding, stability, or function of the protein. Truncating variants (nonsense, deletion, splicing, insertion, and copy number variation) result in a missing or truncated non-functional protein.
Both types of variants result in severe to profound hearing loss. One of the most common OTOF mutations (p.R1939Q) is located at the junction between the calcium binding C2F domains and the C-terminal TMD (Matsunaga et al., 2012;Kim et al., 2018). As mentioned above, some of the mutations lead to a disorder of the temperature-sensitive auditory neuropathy spectrum. Elevated body temperature causes severe to profound hearing loss, whereas normal body temperature results in normal to moderate hearing. Interestingly, all identified thermosensitive mutations are located between the C2C and C2F domains, underscoring the functional importance of these regions (Zhu et al., 2021).
The mechanisms underlying the temperature sensitive hearing loss remain essentially unknown, and they may vary with their site of location. These mutations probably alter the folding or unfolding properties of the long multi-C2-domain protein, impairing specific Ca 2+ -dependent C2-domain interactions and synaptic vesicle fusion, as temperature rises. Another hypothesis is that there is increased degradation and loss of the protein. This is caused by the heat sensitivity of otoferlin, especially in the presence of variants Otoferlin interaction with Ca V 1.3 Ca 2+ channels. (A) Protein-protein interaction studies have shown otoferlin interaction with the II-III loop of the Ca 2+ channel α subunit (Ramakrishnan et al., 2009(Ramakrishnan et al., , 2014Hams et al., 2017). (B) Hair cells (in particular IHCs) of the auditory organ express several C-terminal splicing isoforms of Ca V 1.3: Ca V 1.3 42L , a long isoform with no or poor current inactivation, Ca V 1.3 43S , and Ca V 1.3 42A short truncated C-terminal isoforms with fast inactivation, presumably carrying about 10% of the total inner hair cell (IHC) calcium current (Vincent et al., 2017). The absence of inactivation of the Ca V 1.3 42L subunit is due to the presence at its extreme C terminus of two α helix proximal (PCRD) and distal C-terminal regulatory domains (DCRD) that interact with each other to form the C-terminal regulatory domain (CTM). This CTM domain interaction slows down the inactivation of Ca v 1.3 channels by competing with the binding of Ca 2+ -CaM on IQ domains, (for review see Striessnig et al., 2014) (figure modified from Vincent et al., 2017). Note that CTM is absent in Ca V 1.3 43S , and Ca V 1.3 42A explaining the fast calcium-dependent inactivation of these channels. (C) As a signature of this interaction between otoferlin and Ca V 1.3 channels, and most likely Ca V 1.3 43S , and Ca V 1.3 42A , IHCs from mice lacking otoferlin display Ca 2+ currents with no fast inactivating component (figure modified from Figure 4B of Tertrais et al., 2019).
such as Ile515Thr (Strenzke et al., 2016). The protein is then degraded more quickly.

Otoferlin tissue expression
The tissue distribution of otoferlin is primarily observed in the brain and inner ear (Schug et al., 2006). In the brain it is mainly found in the amygdala and cerebral cortex (Figure 6). We recall that the two TMD otoferlin isoforms have different distributions: one is almost exclusively found in the cochlea (TMD1, exon 48), and the other is mainly found in the brain (TMD2, exon 47) (Yasunaga et al., 2000). However, this TMD variant distribution in the brain and inner ear, only demonstrated by RT-PCR and Northern blot analysis at the tissue level, needs to be confirmed at the protein and cellular levels.
Otoferlin expression increases in hair cells as they mature, remaining permanently expressed in IHCs and only transiently expressed in immature OHCs (largely decreasing after P4-P6 in mice) (Roux et al., 2006;Schug et al., 2006). This transient expression in OHCs is essential for spontaneous synaptic exocytosis by immature OHCs, a process likely involved in the establishment of frequency tonotopy of the central auditory nuclei during development (Beurg et al., 2008). In IHCs and vestibular cells, otoferlin was localized in synaptic vesicles by immunogold labeling (Roux et al., 2006;Dulon et al., 2009) or in the Golgi apparatus and presynaptic zone by immunohistochemistry (Schug et al., 2006;Heidrych et al., 2008). In HEK cell expression, otoferlin is found at the plasma membrane and colocalizes with some of the trans-Golgi markers (GM130 or TGLON2) (Redpath et al., 2016). In IHCs, otoferlin is distributed throughout the cytoplasm and plasma membrane, except for the apical portion that forms the cuticular plate and tight junctions with neighboring cells. It is also interesting to note that the expression of the protein differs according to its two isoforms, TMD1 and TMD2, when expressed in heterologous cell lines. In HEK293 cell cultures, TMD1 showed a dispersed pattern within the cell, while TMD2 was predominantly found at the plasma membrane (Redpath et al., 2016). Otoferlin is also expressed by type I and type II vestibular utricular hair cells (Dulon et al., 2009). Note that truncation of the TMD of otoferlin alters the development of hair cells and reduces membrane docking (Manchanda et al., 2021). This tail-anchored protein can be inserted Otoferlin as a Ca 2+ -sensor controlling the hair cell synaptic vesicle cycle. (A) The postsynaptic structures, issued from the afferent nerve fibers of the bipolar spiral ganglion neurons (SGNs), are here schematically represented contacting the basal part of an IHC viewed under electron microscopy (EM). Each IHC is contacted by 10-20 nerve fibers, depending on their cochlear position (sound frequency encoding). Presynaptically, facing the afferent nerve boutons, is found an electron-dense 300-350 nm wide structure called ribbon to which several dozens of synaptic vesicles (SV) are attached (inset EM). Viewed under immuno-confocal fluorescence microscopy, each presynaptic ribbon (red) at the basal synaptic pole of the IHC, is attached to an F-actin cage like-structure (purple) and is closely colocalized with a single cluster of Ca V 1.3 Ca 2+ channels (green) (image modified from Vincent et al., 2015). (B) Otoferlin, is a six-C2 domain protein with a single transmembrane domain (TMD) at its C-terminal that allows its anchoring into the phospholipid membrane of synaptic vesicles and active zones of transmitter release. Otoferlin, bearing several C2-Ca 2+ bind domains, is essential for reloading, fusion, and endocytosis of synaptic vesicles at the ribbon synapses (C).

Otoferlin multirole at the IHC synaptic vesicle cycle
Remarkably, mature inner hair cells (IHCs) lack the Ca 2+sensors synaptotagmin (I and II) and Munc13, which are known to play a crucial role in vesicle exocytosis at conventional central neuronal synapses (Beurg et al., 2010;Reisinger et al., 2011;Vogl et al., 2015). It is worth recalling that mature IHCs, contrary to central neurons, do not fire action potentials but are capable responding to a large dynamic range of sound intensity (>100 dB) by graded depolarization that is encoded into an increasing discharge rate at the postsynaptic nerve fibers. It is now established that otoferlin plays a role at several essential stages of the vesicle cycle in IHCs, including functional docking, priming, fusion, endocytosis, and possibly transport and maturation of vesicles (Figure 3; Michalski et al., 2017;Tertrais et al., 2019). In otoferlin knockout mice, the average tether length between vesicles and the active zone is increased, suggesting a possible defect in vesicle attachment and priming (Vogl et al., 2015). The downstream calcium-triggered exocytosis of vesicles is almost completely blocked in otoferlin knockout mice. Otoferlin is also involved in active zone clearance, including the transport of exocytic material, through interactions with adaptor protein 2 µ, a motor protein, and a GTPase (Jung et al., 2015). Vesicle size may indicate a role for otoferlin in vesicle reformation and maturation, and the functions of otoferlin may be regulated by phosphorylation. Otoferlin structure. (A) C2-domain structures of the two transmembrane domain (TMD) isoforms of otoferlin, TMD1 (cochlea) and TMD2 (brain). (B) AlphaFold prediction of otoferlin 3D-structure as a globular protein with six C2 domains (C2A-F), a Fer domain, and a C-ter TMD (structure built from AlphaFold Protein Structure Database). (C) Brain and inner ear tissues have been reported to express two different specific alternatively spliced forms of the otoferlin TMD, encoded by either exon-47 or exon-48, respectively (Fig adapted from Yasunaga et al., 2000). Exon-47 is skipped in the cochlea, leading to a different carboxy-terminal peptide sequence from that of the brain isoform, which expresses exon-47. In mRNA with exon-47 (neuronal otoferlin), translation terminates in that exon, leaving exon-48 untranslated. (D) Exon-47 of the brain isoform and exon-48 of the cochlear isoform specify the C-terminal TMD 60 amino acids of otoferlin. The amino acid sequences of the two isoforms differ essentially in the predicted transmembrane sequence (underlined). (E) Analysis of protein structure with AlphaFold software online (developed by DeepMind and EMBL-EBI) predicts a different angle of insertion of cochlear TMD1 and neuronal TMD2 otoferlin isoforms within phospholipid membranes. The use of a specific TMD is likely important for targeting the protein to specific subcellular locations, such as particular glutamatergic excitatory presynaptic active zones, and/or for specifying the interaction with ligands. (F) TMD structure analysis and representation using Protter software (Omasits et al., 2014) predicts a 21 aa TMD1 for the cochlear isoform (gi| 154240702| ref| NP_001093865.1| otoferlin [Mus musculus]) and a 23 aa for the brain isoform (gi| 154240679| ref| NP_114081.2| otoferlin [Mus musculus]).

Otoferlin and membrane fusion
Structural analysis and prediction using Protter (Omasits et al., 2014) and Alphafold (Jumper et al., 2021) display otoferlin as a globular protein anchored to the vesicle and/or plasma membrane through its C-terminal domain, while its N-ter Ca 2+ -sensing C2domains are in the cytosol (Figure 4). The structural features of its C2-domains, through their Ca 2+ -binding and phospholipidbinding activities, allow otoferlin to play a role in membrane fusion and trafficking, in endocytic, secretory, and lysosomal pathways. By deleting exons 14 and 15 of the OTOF gene, an OTOF KO mouse model was created (Roux et al., 2006). These otoferlin-deficient mice (Otof −/− ) are profoundly deaf. Patch-clamp recordings for time-resolved change of membrane capacitance in IHCs from Otof −/− mice show a significant impairment of fast and sustained exocytotic activity (Beurg et al., 2010;Michalski et al., 2017). Another mouse model, Pachanga, carrying a missense mutation in the C2-F domain of otoferlin, was obtained by ENU (N-ethyl-N-nitrosourea) mutagenesis (Schwander et al., 2007). Pachanga mice have profound hearing loss, but unlike Otof −/− mice, IHCs still weakly express otoferlin and maintain rapid fusion of the pool of fusogenic vesicles, forming the RRP (for "readily releasable pool of vesicles") (Pangrsic et al., 2010). However, IHC exocytosis cannot be sustained upon repeated stimulations in these Pachanga mutant mice, indicating impaired recruitment of vesicles to IHC fusion sites. These results suggested that  Tissue and brain expression of otoferlin. (A) Tissue specificity of otoferlin RNA expression (data from the Human Protein Atlas). (B) Otoferlin expression (mRNA in situ hybridization) in the mouse brain thalamus and cortex (cortical subplate) from the Allen Brain Atlas portal. This analysis appears to well corroborate previous findings showing otoferlin expression in brain tissues (Schug et al., 2006). otoferlin is also an essential calcium sensor for synaptic vesicle replenishment at IHC active zones. It should be noted that vesicle recruitment is a calcium-dependent process in hair cells (Spassova et al., 2004;Levic et al., 2011). Gradual intracellular Ca 2+ -uncaging experiments in IHCs revealed that otoferlininduced exocytosis has an intrinsically high Ca 2+ sensitivity (affinity) with a mean Kd of 4.0 ± 0.7 µM and responses spanning from 1 to 20 µM Ca 2+ (Vincent et al., 2014). As a comparison, the neuronal Ca 2+ sensors, synaptotagmins, have nearly a one-order of magnitude lower Ca 2+ affinity (higher Kd), a property in good correlation with their specific functions of being activated only by high Ca 2+ concentrations typical for AP-evoked Ca 2+ -elevations near Ca 2+ channels (Bornschein and Schmidt, 2018).
On the other hand, the importance of otoferlin in the maintenance or development of ribbons and the presence of corresponding synapses has been demonstrated in mice (Stalmann et al., 2021). Otoferlin depletion in zebrafish hair cells also results in abnormal synaptic ribbons and altered intracellular calcium levels (Manchanda et al., 2019), suggesting Ca 2+ as an important factor for the maintenance and survival of the ribbons as in aging auditory synapses (Peineau et al., 2021).
Although some of the interacting proteins start to be identified (Table 1), the precise mode of action of otoferlin in membrane fusion is not fully elucidated, particularly in comparison with the neuronal calcium sensor (synaptotagmin1) which regulates exocytosis function as a multimer with oligomerization via a cluster of juxtamembrane linker (Courtney et al., 2021). Note that, in contrast to ferlin proteins, the TMD of synaptotagmins is located at the N-terminal region of the protein for unknown reasons. The homolog of otoferlin, dysferlin, has also been shown to dimerize in living cells, at its transmembrane domain and at its multiple C2 domains (except C2A), as probed by fluorescence resonance energy transfer (FRET) (Xu et al., 2011).
The C2-A domain is the only one of the six otoferlin domains that cannot bind Ca 2+ , which probably explains the absence of pathogenic mutations in this domain (Johnson and Chapman, 2010;Helfmann et al., 2011;Padmanarayana et al., 2014). The idea of a simple repeat between each C2 domain is rejected. An Otof KI mouse model carrying two mutations in C2C (substitution of 2 aspartic acid by alanine, affecting Ca 2+ binding), demonstrated that otoferlin is essential in both rapid fusion and vesicle recruitment (Michalski et al., 2017). Interestingly, the viral expression of truncated forms of otoferlin (C2-EF, C2-DEF, and C2-ACEF) can only partially rescue the fast and transient release component of exocytosis in mouse hair cells lacking otoferlin (Otof −/− ), yet cannot sustain exocytosis after long, repeated stimulations (Tertrais et al., 2019). To note, a C2-EF otoferlin truncated form can also be produced by intracellular calpain-1 digestion in cell lines (Redpath et al., 2014) but it is unknown whether such truncation occurs in auditory IHCs.
Note that otoferlin is also essential for the fast exocytosis of type I vestibular cells. Otof −/− mice show altered vestibular compound action potentials, suggesting impaired vestibular hair cell function (Dulon et al., 2009). In intact utricles ex vivo, otoferlin was found to be critical for a highly sensitive and linear calciumdependent exocytosis, facilitating the linear encoding of lowintensity stimuli at the vestibular hair cell synapse. Surprisingly, mice and humans lacking otoferlin do not have apparent vestibular symptoms. The lack of vestibular phenotype may be explained by some compensatory mechanisms, in particular by stimulus-evoked acidification of the synaptic cleft of vestibular hair cells (protons acting as neurotransmitters) (Highstein et al., 2014) or by the use of another type of neurotransmission via potassium accumulation in the calyx, which depolarizes it (Holt et al., 2007). Although these compensatory modes of transmission are rather slow and nonlinear, they may explain the absence of a vestibular phenotype in OTOF patients and animal models.

A specific mechanical tuning
between Ca 2+ , otoferlin, F-actin, and the ribbon could determine the specific firing frequency of each auditory nerve fiber Mammalian auditory inner hair cell (IHC) ribbon synapses have to deal with the great challenge of encoding an extremely wide range of sound intensities, with a dynamic range of more than 100 dB. To perform this challenging task, IHCs have partitioned their synaptic output sensitivity and dynamic range. Indeed, each IHC forms synapses with a pool of 10 to 30 afferent nerve fibers, among which spontaneous activity, acoustic threshold, and dynamic range vary widely. One of the most important questions in auditory neuroscience is to elucidate the functional mechanisms that dictate the synaptic diversity of each type of sensory synapses within a single IHC. Variation in the voltage-gating and spatial  Ricci et al. (2013) and Graydon et al. (2017). The resting size of the hair cells was measured using whole-cell patch clamp recordings of electrical capacitance in absence of stimulation. Data were well fitted with a linear correlation (R 2 = 0.91). Note that hair cells of Otof-KO mice have a decreased number of synaptic ribbons and proportionally smaller cell sizes as compared to WT mice (Vincent et al., 2014(Vincent et al., , 2017Tertrais et al., 2019), suggesting otoferlin as essential for the maintenance of both ribbons and cell sizes, possibly by regulating intracellular hydrostatic pressure and the F-actin membrane network . organization of Ca V 1.3 calcium channels at each presynaptic active zone has been proposed to determine the firing specificity of the auditory nerve fibers (Özçete and Moser, 2021). Other factors, such as a mechanical tuning of each ribbon, could also participate in the firing frequency characteristic of each auditory fiber (Figures 7-9).
Otoferlin has been proposed to control IHC intracellular hydrostatic pressure and in turn exocytosis, presumably via interactions with a synaptic F-actin network  and the Ca 2+ channels Ca V 1.3 at the ribbon (Vincent et al., 2017). Otoferlin is essential for synchronous multivesicular release at IHC active zones; this process underlying fast transient H + -inhibition of Ca V 1.3 Ca 2+ channels .
Remarkably, there is a good linear correlation between the size of the hair cells and their number of synaptic ribbons per cell when comparing the ears of various vertebrates (Figure 7). In Modeling IHC as a hyper-elastic structure whose changes in membrane capacitance upon intracellular Ca 2+ uncaging and otoferlin action behave like the pressure change in an inflating rubber balloon. (A) Typical example of membrane capacitance change upon intracellular Ca 2+ uncaging in an IHC of a WT P20 mouse. (B) Similar experimental conditions in an IHC of a P20 Otof-KO mouse. In (A,B), IHCs were whole-cell voltage-clamped at -70 mV and loaded with the caged-Ca 2+ molecule DM-nitrophen [figures (A,B) are unpublished traces obtained as described in Vincent et al. (2014Vincent et al. ( , 2015 and Tertrais et al. (2019)]. An increase in membrane capacitance reflects the fast fusion of synaptic vesicles at the plasma membrane. Note that the large N-shaped curve of the change in membrane capacitance in WT IHC is absent in IHC Otof-KO. (C) Pressure variation inside an inflating rubber balloon displays a similar N-shaped curve over time, or radius change, with the following relationship: where R is the radius, R0 is the initial radius, K is a constant dependent on material properties, and P is the internal pressure of the balloon (Müller and Struchtrup, 2002;Ajmal et al., 2023). (D) In whole-cell configuration, IHCs were loaded with the fluorescent Ca 2+ indicator molecule OGB-2 and the edge of ruptured plasma membrane inside the patch pipette could be visualized under confocal fluorescence microscopy. As evidence of intracellular pressure changes in IHCs, upon repeated depolarizing pulses, fast intrapipette membrane displacements could be visualized and measured in WT IHCs. These membrane movements are largely reduced in IHCs lacking otoferlin (figure modified from Vincent, 2015). # indicates statistical difference, unpaired t-test with p < 0.01. Overall, these elements suggest otoferlin as a mechanical protein whose interaction with Ca 2+ leads to the development of mechanical forces that change intracellular pressure in IHCs. The changes in intracellular pressure could result from Ca 2+ -induced fast conformational changes of otoferlin and/or the fast, massive, addition of vesicular membrane to the plasma membrane.
IHCs from mice lacking otoferlin, the number of synaptic ribbons is proportionally reduced with the cell size (Tertrais et al., 2019). Also in old P360 C57BL/6J mice, the reduced size of the IHCs is proportionally associated with a lower number of synaptic ribbons (Peineau et al., 2021). This correlation suggests that the number of ribbons may influence membrane exocytosis, and in turn, cell size, possibly by modifying the intracellular hydrostatic pressure.
Numerous studies in various cell type have shown a reciprocal link between cell biomechanics, the submembrane actomyosin network, and exocytosis (Wang and Galli, 2018). Auditory IHCs seem to be able to sense the biomechanical properties of the environment since their exocytotic properties are sensitive to changes in intracellular hydrostatic pressure (Vincent et al., 2014). Membrane addition during hair cell synaptic exocytosis can be visualized by confocalfluorescent microscopy (Hudspeth and Issa, 1996), indicating membrane deformation and a possible local change in hydrostatic pressure.
We propose that otoferlin, by controlling membrane fusion, regulates plasma membrane rigidity and tension of its underlying actomyosin meshwork, whose mechanical properties are likely essential to maintaining active membranous synaptic ribbons. Note that upon UV-flash-photolysis allowing ultrafast intracellular Ca 2+ uncaging from DM-nitrophen, the change in IHC membrane capacitance (produced by massive vesicle fusion to the plasma membrane) behaves like an inflating rubber balloon (Figures 8A-C), suggesting a non-monotonic relation (non-linear stress-strain response) between the elasticity of the IHC plasma membrane and the intracellular pressure. Changes in IHC intracellular pressure can directly be visualized during exocytosis when monitoring membrane displacement during repetitive depolarizing steps (Figure 8D; Vincent, 2015). It is tempting to propose the synaptic ribbons as pulsating microspheres that are activated by Ca 2+otoferlin interactions with the membranous F-actin network; these mechanical interactions being the source generator (hydrostatic force) triggering vesicle membrane fusion. In this model, the mechanical oscillatory frequency of the ribbons would essentially depend on their mass, similarly to a spring-block oscillator, a factor that would then determine the firing frequency of the postsynaptic afferent nerve fibers (Figure 9). In this context, it is interesting to note that amphibian hair cells have been shown to display frequency selectivity in synaptic exocytosis, presumably due to a tonotopic variation of Ca 2+ buffers (parvalbumin 3 and calbindin-28K) that tunes a regime of spontaneously oscillatory vesicle release (Patel et al., 2012).

Ferlin family
Otoferlin belongs to the ferlin protein family, which in humans has five members in addition to otoferlin (fer1L2): dysferlin (fer1L1), myoferlin (fer1L3), fer1L4, fer1L5, and fer1L6 Redpath et al., 2016). The first ferlin identified was Fer-1, in Caenorhabditis elegans. Fer-1 is a fertilization factor required for the fusion of specialized vesicles with the plasma membrane during spermatogenesis (Achanzar and Ward, 1997). The main common features of ferlins are: • The presence of multiple C2 domains that share a high sequence similarity of about 100 amino acids, and most often allow binding to Ca 2+ , phosphatidylinositol 4,5-bisphosphate (PIP2) or phosphatidylserine (PS).
• The presence of "FerA" domains also allows binding to phospholipid membrane.
• Their anchoring to the membrane by a single C-terminal domain (tail-anchored proteins).

FIGURE 9
Modeling the IHC synaptic ribbon as a pulsating microsphere at the origin of signal overshoot. (A) Tone-on-tone forward masking: typical recovery timing from cochlear synaptic depression (ABR Wave I amplitude) in a WT mouse when using a forward masking paradigm with increasing time intervals. Note that the overshoot-like responses in the amplitude of the probe-evoked ABR amplitude (also described in gerbils and guinea-pigs, Chatterjee and Smith, 1993;Haddadzade et al., 2021) mimick the overshoot responses described in psychoacoustic studies (Zwicker, 1965;Bacon and Moore, 1987;Bacon, 1990). Although it could reflect in part the compressive input-output characteristic of the basilar membrane, we propose that the damped oscillatory behavior of the Wave I amplitude can be explained by a persistent long-lasting pulsating synaptic activity between hair cells and auditory nerve fibers at the end of the masker stimulation. The red dashed line fits with a single exponential the overall masking recovery of the auditory nerve responses with a time constant of 20 ms, a similar time scale reported for the Ca 2+ -otoferlin-dependent process to refill the hair cell ribbons with synaptic vesicles (Spassova et al., 2004;Levic et al., 2011).
(B) Schematic drawing of an IHC depicting the two main types of postsynaptic afferent fibers regarding their spontaneous activity and threshold response. For clarity only two fibers were depicted. Note that high threshold fibers with low spontaneous activity, contacting large ribbons, are mainly located at the neural (modiolar side) of the IHCs, while low threshold fibers are at the pillar side (Liberman, 1982 ;Liberman et al., 2011). (C) Hypothetical representation of two pulsating ribbons whose mechanical oscillatory frequency will essentially depend on their mass factor as a spring-block oscillator model. The oscillatory mechanical waves likely originate from the interaction of Ca 2+ -otoferlin and the membranous F-actin network (not represented). Small ribbons are predicted to generate a higher oscillating frequency as compared to large ones.
Dysferlin (Fer1L1) is ubiquitously expressed, with high expression in the brain, heart and skeletal muscle (Anderson et al., 1999). Dysferlin is composed of seven C2 domains, the DysF domain and the FerA domain preceding the FerB domain (Figure 10). Dysferlin acts as a primary emergency regulator of membrane repair in a calcium-dependent manner following membrane damage to myofibres (Bansal et al., 2003;Barthélémy et al., 2018). Dysferlin is also involved in Ca 2+ metabolism regulation in injured muscle fibers (Hofhuis et al., 2017). Dysferlin can be cleaved by calpain (1 and 2; calcium-dependent protease) in response to calcium influx following a membrane micro-lesion (Lek et al., 2013). Cleavage releases a "mini-dysferlin-C72" (with C2E, C2F, and TMD). Mini-dysferlin-C72 is selectively recruited and is thought to be a minimal configuration required for physiological membrane repair function (Krahn et al., 2010;Lek et al., 2013;Redpath et al., 2014). Otoferlin and myoferlin can also release a module with the last two C-terminal domains and the TMD after enzymatic cleavage of calpain in vitro (Redpath et al., 2014;Piper et al., 2017). This structure carries a high structural similarity to the synaptotagmin family. Moreover, a phylogenetic study reveals that these two C-terminal C2 domains are the most evolutionarily conserved domains of the ferlin family: there is 90% sequence similarity of C2-EF domains between mammals and mollusks, suggesting a key function .
In the central nervous system, dysferlin has been observed to accumulate in endothelial cells near sclerosis lesions (Hochmeister et al., 2006), as well as within amyloid-beta plaques in people with Alzheimer's disease (Galvin et al., 2006). In addition, in humans, certain genetic variations of the Dysf gene have been associated with an increased risk of developing Alzheimer's disease, and Dysf mRNA expression has been shown to be increased in the brains of people with Alzheimer's disease (Chen et al., 2015). Interestingly, OTOF was identified as a downregulated gene in a humanmouse chimeric model of Alzheimer's disease using genome-wide expression analysis, suggesting an essential function of otoferlin in the central nervous system (Espuny-Camacho et al., 2017).
Myoferlin (Fer1L3) is also ubiquitously expressed in skeletal and cardiac muscle and in the placenta (Davis et al., 2000). Myoferlin is involved in the fusion of myoblasts, in repair and regeneration, and in muscle cell membrane growth (Doherty et al., 2005). Myoferlin may be involved in the maintenance of transverse tubule function (Demonbreun et al., 2014). Between the C2-C and C2D, a domain consisting of two long β-sheets is called the dysferlin (DysF) domain, and a FerA domain, consisting of α-helices, is located before the FerB domain (Figure 10). The exact role of the DysF domain is unknown but must be significant, as mutations in this domain cause myopathies (Sula et al., 2014). Full information about the ferlin family is available in a recent review by Peulen et al. (2019).

Otoferlin-disease management and prospects
Auditory Neuropathy Spectrum Disorder (ANSD) is most often suspected after hearing loss is detected at birth, especially if otoacoustic emissions, temporal bone MRI, and CT scans are normal. This deafness may be detected later if the otoacoustic emissions at birth are falsely reassuring. The diagnosis is confirmed by molecular genetic testing, which shows biallelic pathogenic variants in OTOF. Two phenotypes are possible: OTOF-related ANSD, with severe to profound bilateral hearing loss (>70 dB), and temperature-sensitive ANSD (TS ANSD), with normal to moderate hearing loss (0-69 dB) at normal body temperature and worsening to profound hearing loss during hyperthermia. In these patients, prevention of hyperthermia by early use of antipyretics for infection and avoidance of exposure to high temperatures is important (Azaiez et al., 2021).
As we have seen above, in OTOF-related ANSD, the defects are presynaptic at the level of the hair cells. In these situations, the cochlear nerve is intact and functional. This is a prerequisite for cochlear implantation. It is currently the only option for restoring hearing in patients with severe to profound hearing loss. Early implantation is recommended after the diagnosis of OTOFrelated auditory neuropathy with severe to profound hearing loss. It involves surgery. A foreign body is implanted under the skin, and an electrode array is placed in the cochlea for life (De Seta et al., 2022). It requires the wearing of an external magnetic processor. The risk of infection or failure requiring re-intervention and the limited autonomy of the device are significant limitations of this therapy. We can also point out the potential limitations of speech recognition in noise. In OTOF-related ANSD, audiometric results are known since 2005 (Loundon et al., 2005). They can be considered of high quality since the thresholds for sound perception in silence from 500 to 2000 Hz vary on average from 25 to 45 dB SPL, depending on the series (Loundon et al., 2005;Rouillon et al., 2006;Zhang et al., 2016;Chen et al., 2018;Zheng and Liu, 2020). Regarding speech recognition, which is the main goal of cochlear implantation, discrimination scores are mostly >90%. There is no evidence that cochlear implantation outcomes are correlated with distinct OTOF genotypes (Zheng and Liu, 2020).
Gene therapy for congenital deafness has been expanding in recent years, with the aim of achieving a therapy with better hearing outcomes without the limitations of cochlear implant. A truncated form of the OTOF gene was shown to rescue hearing and balance in zebrafish (Chatterjee et al., 2015). Since then, several studies have been interested in genetic therapy using recombinant dual AAV vectors (adeno-associated viruses) to encode full-length otoferlin (Akil et al., 2019;Al-Moyed et al., 2019;Rankovic et al., 2020). These studies gave very encouraging results by showing significant restoration of hearing in mice. Recently, recombinant dual AAV vectors encoding for the whole human cochlear otoferlin were used to restore hearing in mice to near wild-type levels for 6 months (Tang et al., 2022). This raises the prospect of treatment by gene therapy in humans for DFNB9. The main limitations are the early age of AAV treatment used in most of these studies, which would correspond to an in utero AAV injection in humans, and the weak restoration of the ABR wave-1 amplitude, which could limit speech intelligibility and discrimination, particularly in noise, as the main current defect known for cochlear implants. The weak restoration of wave-1 amplitude during AAV gene therapy is likely due to an uncomplete rescue of the loss of ribbon synapses in the absence of functional otoferlin (Stalmann et al., 2021). Note that a reduction in spiral ganglion neurons with apoptosis is also a secondary defect caused by a lack of otoferlin in IHCs (Tsuzuki et al., 2023). Currently, several biotech companies (SENSORION, NCT05402813 and AKOUOS-NCT05572073) are in the preclinical phase for the development of gene therapy for otoferlin mutations in humans.

Ferlins
As we have seen, ferlins are a group of proteins that play an important role in several membrane-related processes that are crucial for cell survival and signaling. These processes include endocytosis, exocytosis, recycling, and membrane repair (Bansal et al., 2003;Roux et al., 2006;Bernatchez et al., 2009;Redpath et al., 2016). The link between these processes and cancer is through their effect on cell Ca 2+ signaling and development, which allows tumors to form and cancer cells to adapt to a hostile environment. Despite their importance, ferlins have not been extensively studied in the context of cancer. However, recent research has shown that all ferlin genes are modulated in different types of cancer (Peulen et al., 2019). Myoferlin and fer1l4 genes are more commonly upregulated (Barnhouse et al., 2018;You et al., 2020). Myoferlin has been found to be highly expressed in several types of cancer, including kidney, liver, pancreatic, breast, and head and neck squamous cell carcinomas (Wang et al., 2013;Kumar et al., 2016;Blomme et al., 2017;Hermanns et al., 2017;Song et al., 2017).

Myoferlin
The literature on ferlins in cancer only started about 10 years ago, although the proteins have been described for over 20 years. Most of the work has focused on myoferlin, which appears to be the ferlin with the most important role in cancer. The fact that ferlin expression can be either a good prognostic factor (breast cancer) or a poor prognostic factor (kidney, head and neck cancer) is one of the most surprising elements.
In the study of Yadav et al. (2017) which shows that when cells are treated with IL-6, myoferlin dissociates from EHD2 and binds to activated STAT3, a protein involved in cell signaling. The study found that depletion of myoferlin did not affect STAT3 phosphorylation, but blocked its translocation to the nucleus. The study also found that myoferlin knockdown significantly reduced IL-6-mediated tumor cell migration, tumor sphere formation, and the population of cancer stem cells in vitro. In addition, myoferlin knockdown significantly reduced IL-6-mediated tumor growth and metastasis. In 211 patients with head and neck squamous cell carcinoma (HNSCC), Kumar et al. (2016) investigated the association of myoferlin with disease progression and patient outcome. The results showed that nuclear myoferlin expression is associated with poor overall survival and an increased risk of death, as well as tumor recurrence, perineural invasion, extracapsular spread, a higher T stage, and distant metastasis. The study also found a direct association between nuclear myoferlin expression and IL-6 and an inverse association with HPV status. Patients with both nuclear myoferlin expression and high levels of IL-6 and those with HPV-negative/myoferlin-positive tumors had the worst survival. These results suggest that nuclear myoferlin expression independently predicts poor clinical outcomes in these patients.
By modeling the effect of myoferlin on tumor cell invasion through altered regulation of metalloproteinase production, Eisenberg et al. (2011) identified a role for myoferlin in promoting invasive behavior in breast cancer cells. The identification of myoferlin as a key regulator of EGFR (Epidermal Growth Factor Receptor) activity through inhibition of non-clathrin endocytosis in breast cancer cells was important in understanding the molecular mechanisms involved in cancer growth (Turtoi et al., 2013). The production of myoferlin-silenced tumor cells has provided another element that may help explain the role of ferlin: these tumors lacked functional blood vessels, an effect that may be due to a reduction in Survival data linked to otoferlin and oncomodulin expression levels (graphs were built after bioinformatics analysis using OncoLnc, 2023). Year overall survival was better for patients with lower otoferlin gene expression in (A) kidney renal clear cell carcinoma (KIRC) and (B) kidney renal papillary cell carcinoma (KIRP), whereas survival was better for patients with higher otoferlin gene expression in (C) bladder urothelial carcinoma (BLCA). High levels of oncomoduline (OCM), a parvalbumin-family calcium-binding protein highly expressed in hair cells, in KIRP (D) is associated with better 5-year overall survival.
VEGFA exocytosis (Fahmy et al., 2016). This may explain its strong association with cancer prognosis, as an independent prognosis factor in kidney (Song et al., 2017).

Otoferlin
Currently, there is limited information available on otoferlin. Involvement of IL-6 in inner ear damage during noise trauma or cisplatin treatment has been reported (So et al., 2007;Wakabayashi et al., 2010), but a possible link with otoferlin function during disease progression has not yet been explored. Using OncoLnc, a tool for exploring correlations between the expression of mRNAs, mRNAs, and IncRNAs, we analyzed 5-year overall survival in 21 cancer types (Anaya, 2016). In three cancers, we found a significant impact of otoferlin expression on survival: in renal clear cell, papillary cell carcinoma, and in bladder urothelial carcinoma (Figure 11). Interestingly, otoferlin expression was either a protective or a risk factor depending on the histology of the cancer. Given the differences in 5-year survival, it seems appropriate to consider using otoferlin expression as a prognostic biomarker to guide therapeutic decisions.
In clear cell renal cell carcinoma, Cox et al. (2021) showed in a study of 79 patients that otoferlin was an indicator of tumor staging and a prognostic biomarker for cancer-specific survival. In 25 patients with oral squamous cell carcinoma, Kraus showed complete suppression of otoferlin expression in malignant tissue (Kraus et al., 2022). Taken together, these results suggest a role for otoferlin in the carcinogenesis of these tumors, which deserves investigation to confirm and understand its exact role, especially as it appears to vary according to tumor site. Studies are needed to ensure that otoferlin is an independent prognostic factor for overall survival. This could lead to new cancer therapies targeting this protein, which are already underway for myoferlin (Zhang et al., 2018;Li et al., 2019;He et al., 2021).

Oncomodulin
Hair cells, notably OHCs, highly express another protein, which has also been used as a prognostic marker in cancer: oncomodulin (OCM). OCM is a small EF-hand Ca 2+ -binding protein (CaBP) of approximately 12 kDa. It belongs to the parvalbumin family. It is the β-isoform of parvalbumin. It shares at least 53% sequence identity with alpha-parvalbumin (PVALB) (Climer et al., 2019). OCM has an unusually restrictive post-embryonic expression pattern in mammals, mostly restricted to subsets of sensory hair cells in the inner ear and more recently found in certain subtypes of immune cells. Oncomodulin is essential for normal hearing function since its deletion in mice leads to progressive deafness due to alterations in the mechanical amplification function of OHCs (Tong et al., 2016). We recall that OHC lateral mechanical stiffness and motility have long been recognized as regulated by calcium acting on the actin-based cortical cytoskeleton (Dulon et al., 1990). Oncomodulin, acting as a CaBP, may be essential in sculpting calcium signals important for F-actin remodeling and membrane mechanics in OHCs, a role somewhat similar to that of otoferlin in IHCs, as proposed in our review.
The initial discovery of OCM as an oncoprotein in cancer tissue and its similarity to calmodulin as a CaBP led to the term "oncomodulin" (Durkin et al., 1983). OCM was considered oncogenic because no expression was detected in normal postembryonic tissue. However, after its initial discovery, OCM was identified as a major protein in sensory cells of the cochlea. Studies have suggested that OCM may play a role in cell proliferation, and recently, Yin et al. (2006) identified oncomodulin as a potent macrophage-derived growth factor for retinal ganglion cells (RGCs) and other neurons. Oncomodulin also stimulates the outgrowth of peripheral sensory neurons.

Discussion-conclusion
Otoferlin, a protein mainly produced in the brain and inner ear, plays a key role in hearing. More than 220 mutations causing DFNB9-type deafness have been identified so far. Many of these mutations alter the folding, stability, or function of the protein, resulting in severe to profound hearing loss associated with a defect in synaptic transmission between the IHC and the nerve fibers of the auditory nerve.
An intriguing question is: why have auditory and vestibular hair cells selected otoferlin as a key Ca 2+ -sensor to control exocytosis of synaptic vesicles instead of the classical synaptotagmins found in central synapses? The need to encode tiny graded microphonic potentials spanning over a large range of amplitude (several dozens of dB) by graded multivesicular release at each ribbon synapses and the requirement to sustain an extremely high rate of vesicle release by an efficient vesicular replenishment are probably the two main factors explaining the use of a large multi-C2 Ca 2+ sensing protein such as otoferlin. Another important property of otoferlin is its capability to function as a high affinity largerange Ca 2+ sensor as compared to synaptotagmins. In our review, we propose otoferlin as a Ca 2+ -dependent mechanical interactor between the membranous F-actin network and the synaptic vesicles at the hair cell ribbon. This protein would confer Ca 2+ -dependent oscillatory movement of the ribbon, whose resonant frequency would depend on their size (mass), tuning the firing frequency of the postsynaptic fibers.
In this context, it is important to note that otoferlin belongs to the ferlin protein family in humans: dysferlin, myoferlin, fer1L4, fer1L5, and fer1L6, all of which being expressed in mechanically active or contracting cells. These proteins share common features such as the presence of multiple C2 domains, FerA domains, and membrane anchoring through their C-terminal domain, and are thought to play a role in calcium-triggered membrane fusion and trafficking. Although otoferlin is expressed in the central nervous system, in particular in the cortex and amygdala, its role in brain tissues remains unknown. Interestingly, a genomewide expression analysis of a human-mouse chimeric model of Alzheimer's disease identified OTOF as a downregulated gene, suggesting a critical function of otoferlin in the central nervous system. Dysferlin has been observed to accumulate in endothelial cells adjacent to sclerotic lesions and within amyloid-beta plaques in individuals with Alzheimer's disease. These findings highlight the need for further research into the role of ferlin proteins in the central nervous system, particularly their potential involvement in neurodegenerative diseases such as Alzheimer's disease.
Also, otoferlin, like many of the other ferlins, plays a crucial role in membrane-related processes that are important for cell survival and signaling. Studies have suggested a link between ferlins and cancer through their effect on cell signaling and development, allowing tumors to form and cancer cells to adapt to a hostile environment. Despite their importance, ferlins have not been extensively studied in the context of cancer. Data show that otoferlin expression is significantly associated with survival in specific cancer types, including clear cell and papillary cell renal carcinoma, and urothelial bladder cancer. These results suggest a role for otoferlin in the carcinogenesis of these tumors, which needs further investigation to confirm and understand its exact role, especially as it appears to vary according to tumor site. This could lead to new cancer therapies targeting this protein.

Author contributions
J-CL and DD wrote the manuscript. Both authors contributed to the article and approved the submitted version.