Apoptosis of type I spiral ganglion neuron cells in Otof- mutant mice

, it remains unclear how the Otof mutation affects spiral ganglions. Thus, we used Otof -mutant mice carrying the Otof tm1a(KOMP)Wtsi allele ( Otof tm1a ) and analyzed spiral ganglion neurons (SGNs) in Otof tm1a/tm1a mice by immunolabeling type I SGNs (SGN-I) and type II SGNs (SGN-II). We also examined apoptotic cells in SGNs. Four-week-old Otof tm1a/tm1a mice had an absent ABR but normal DPOAEs. The number of SGNs was significantly lower in Otof tm1a/tm1a mice on postnatal day 7 (P7), P14, and P28 compared with that of wild-type mice. Moreover, significantly more apoptotic SGNs were observed in Otof tm1a/tm1a mice than in wild-type mice on P7, P14, and P28. SGN-IIs were not significantly reduced in Otof tm1a/tm1a mice on P7, P14, and P28. No apoptotic SGN-IIs were observed under our experimental conditions. In summary, Otof tm1a/tm1a mice showed a reduction in SGNs accompanied by apoptosis of SGN-Is even before the onset of hearing. We speculate that the reduction in SGNs with apoptosis is a secondary defect caused by a lack of otoferlin in IHCs. Appropriate glutamatergic synaptic inputs may be important for the survival of SGNs.


Introduction
Auditory neuropathy spectrum disorder (ANSD) is a hearing dysfunction characterized by an absent auditory brainstem response (ABR) despite preserved distortion product otoacoustic emission (DPOAE) [1][2][3]. OTOF (DFNB9) is the gene for which pathogenic variants are most frequently identified in patients with ANSD [4,5]. Both homozygous and compound heterozygous mutations of OTOF have been reported to cause severe sensorineural hearing loss [6]. Homozygous Otof-mutant mice (Otof -/mice) have a congenital profound hearing impairment with no visible ABR and preserved DPOAEs. Therefore, Otof -/mice are considered model mice for ANSD [7]. Otof -/mice have almost complete abolition of exocytosis in the inner hair cell (IHC) [7], and those at postnatal day 6 (P6) have ribbon synapses with normal structure and number, whereas those at P15 are structurally abnormal and have a reduced number of ribbon synapses, which may be attributed to a secondary defect of synaptic dysfunction [7].
Cochlear implants (CIs) provide sound perception via direct stimulation of the cochlear nerve (i.e., spiral ganglion) without passing through cochlear hair cells [12]. Therefore, the condition of the cochlear nerve greatly affects the effectiveness of CIs [13]. Several previous bedside studies have reported that patients with OTOF variants tend to have good CI outcomes [14][15][16], which are speculated to be because OTOF-related ANSD (DFNB9) is caused by a dysfunction confined to the ribbon synapses of cochlear hair cells [17,18]. Considering recent reports suggesting that earlier CI intervention for DFNB9 is crucial for achieving favorable CI outcomes [14,15,19], determining the impact of the Otof mutation on the spiral ganglion is vital. Spiral ganglion neurons (SGNs) in the mature cochlea are categorized into type I SGNs (SGN-I), which innervate a single IHC, and type II SGNs (SGN-II), which innervate numerous outer hair cells (OHCs). SGNs are composed of 90-95% SGN-Is, and the remaining 5-10% are SGN-IIs [20]. SGN-IIs are thought to activate the enhancement of OHCassociated sound transduction [21,22]. In a previous report, the number of SGNs was reported to be slightly reduced in Otof -/mice compared with Otof +/+ mice at 48 weeks old but not at 8 weeks; however, SGNs were not classified as SGN-Is or SGN-IIs [23]. In the Otof -/mice used in the previous report, deletion of exons 14 and 15 was reported to prevent translation of the Otof messenger RNA (mRNA), because the deletion led to a frameshift and premature stop codon [24].
The aim of the present study was to assess SGNs in Otof-mutant mice by immunolabeling SGN-Is and SGN-IIs. We also examined apoptosis as a possible cause of the SGN reduction in Otof-mutant mice and studied apoptotic cells in SGN-Is and SGN-IIs.

Mice
The Otof-mutant mice (Mouse Genome Informatics (MGI):4363757; International Knockout Mouse Consortium (IKMC) Project:28320; Otof tm1a(KOMP)Wtsi mice are referred to here as Otof tm1a ; Fig. 1A) were generated by Knockout Mouse Project (KOMP) (KOMP repository; htt ps://www.komp.org/) in the C57BL/6N genetic background. In Otof tm1a mice, splicing between exon 9 and lacZ trapping element results in the prevention of Otof transcription downstream of the lacZ. Otof tm1a mice were maintained by heterozygous intercrossing or heterozygotes crossed with homozygotes. All animal studies were carried out in accordance with the Guide for the Care and Use of Laboratory Animals (National Institute of Health, USA) and the institutional guidelines approved by the Experimental Animal Care Committee of Keio University School of Medicine (approval number: 08020). Mice of both sexes were used in all animal experiments.

Genotyping
Otof tm1a mice were genotyped by tail polymerase chain reaction (PCR) using a wild-type forward primer (CSD-F), a mutant-specific forward primer (Common-loxp-F), a wild-type reverse primer (CSD-ttR), and a mutant-specific reverse primer (CSD-Otor-SR1; Fig. 1A and B). The wild-type Otof allele was identified as a 520 bp amplicon size, and the mutant Otof tm1a allele was identified as a 391 bp amplicon size.

Quantitative PCR (qPCR)
Disrupted and homogenized tissues of the sensory epithelium and lateral wall (Otof +/+ or Otof tm1a/tm1a mice on P5) were used to quantify the mRNA expression levels of Otof. Total RNA was isolated using the RNeasy Mini kit (Qiagen, Netherlands) according to the manufacturer's protocol. Complementary DNA (cDNA) was synthesized from 1 μg/11 μl of total RNA using the Superscript Ⅳ First-Strand Synthesis System (Thermo Fisher Scientific, USA). The primers used were mouse otoferlin primers designed around the exon 12 at the 3 ′ ends of the mutated region (forward: CCCTGGTGGGTTCCTTCAAA; reverse: ACCA-CAGCGACATCACACTT) and mouse actin primers (forward: GCTCCGGCATGTGCAAAG; reverse: CCATCACACCCTGGTGCCTA). Actin was used as a housekeeping gene for normalization. qPCR was performed using the PowerTrack SYBR Green Master Mix (#A46109 Thermo Fisher Scientific, USA). The reaction mixtures were denatured at 95 • C for 1 min, followed by 40 PCR cycles. Each cycle consisted of two steps: 95 • C for 15 s and 60 • C for 30 s.

Western blotting
Proteins were extracted from the sensory epithelium and lateral wall of either Otof +/+ , Otof +/tm1a , or Otof tm1a/tm1a mice on P5 using radioimmunoprecipitation assay (RIPA) buffer (FUJIFILM Wako Pure Chemical Corporation, Japan). Following sodium dodecyl sulfate-polyacrylamide gel electrophoresis, the proteins were transferred to a polyvinylidene difluoride membrane, followed by blocking with 5% skim milk (FUJIFILM Wako Pure Chemical Corporation, Japan) for 1 h at room temperature (RT). The membrane was incubated with an anti-OTOF antibody (#ABO12302, Abcepta, USA) or an anti-b-actin antibody (#8457, Cell Signaling Technology, USA) overnight at 4 • C. The horseradish peroxidase-conjugated antirabbit immunoglobulin G (IgG) antibody (Abcam, UK) and the ECL prime chemiluminescence system (Cytiva, Japan) were used for detection.

Hearing function
Four-week-old mice (from P28 to P30) were placed under general anesthesia, and their hearing function was measured using the ABR and DPOAE tests. The ABRs were measured using the methods and equipment described previously [25][26][27]. The DPOAEs were measured as described previously [27]. The sound pressure levels of the F1 stimulus (i.e., L1) were decreased in 10 dB steps from 80 to 20 dB sound pressure level (SPL). The sound pressure levels of the F2 stimulus (i.e., L2) were maintained at 10 dB lower than those of L1. The DPOAE levels were estimated using the 2F1-F2 components of distortion. DPOAE thresholds were calculated as L1 producing a 2F1-F2 of 0 dB. If L1 was 80 dB SPL, but 2F1-F2 was<0 dB, the DPOAE threshold was calculated to be 85 dB, whereas if L1 was 20 dB SPL and 2F1-F2 was greater than 0 dB, the DPOAE threshold was calculated to be 15 dB.

Immunostaining
The cochleae of P7 mice were dissected and fixed overnight at 4 • C with 4% paraformaldehyde (PFA). P14 and P28 mice were perfused with 4% PFA, and the cochleae were dissected and fixed overnight at 4 • C with 4% PFA. Decalcification with ethylenediaminetetraacetic acid (FUJIFILM Wako Pure Chemical Corporation, Japan) was performed for 3 days on the P7, and P14 cochleae, and for 7 days on the P28 cochleae. Cochleae were stored in 10% sucrose at 4 • C for 1 h, followed by 30% sucrose at 4 • C until embedding. The cochleae were embedded in a Tissue-Tek optimum cutting temperature (O.C.T.) compound (Sakura Finetek, USA) and subsequently frozen for cryosectioning. Frozen blocks were cut into 7 μm sections parallel to the modiolus, and sections every five slices were picked up on slides. For the immunofluorescence study, all slides were incubated for 1 h at RT with 5% donkey serum in phosphate-buffered saline (PBS) overnight at 4 • C with primary antibodies. The sections were rinsed in PBS three times, and incubated in secondary antibodies and Hoechst for 1 h at RT. At each step of immunostaining, 0.1% Triton X-100 was added to the PBS. Images were acquired using a fluorescence microscope (BZ-X710, Keyence, Japan).

Cell counts for SGN quantification
We used ImageJ (NIH) to process and analyze images. Serial three sections on the same slide per one mouse were used for cell counts. The number of SGNs (Hoechst staining coimmunolabeled with beta-III tubulin) and SGN-IIs (Hoechst staining coimmunolabeled with beta-III tubulin and peripherin) in Rosenthal's canal in midbasal cochlear turn was manually counted. The area of Rosenthal's canal in midbasal cochlear turn was also measured. Thereafter, the density of SGNs and SGN-IIs in the Rosenthal's canal was calculated. The number of SGNs and SGN-IIs with immunosignals of cleaved caspase 3, a marker of apoptosis, was also counted manually, and the percentage of apoptotic cells in SGNs and SGN-IIs was calculated. The mean and standard deviation of cell densities and percentage of apoptotic cells were calculated from multiple mice.

Statistical analyses
All statistical analyses were performed using EZR (Saitama Medical The Otof "knockout-first" allele (tm1a) was generated via the insertion of a large DNA cassette composed of 2 flippase recognition target (FRT) flanking an internal ribosome entry site (IRES); beta-galactosidase reporter gene (lacZ) cassette and a human beta-actin promoter (hbactP)driven neo cassette was inserted into the intron 9 of Otof and an additional loxP site downstream of exon 10 and 11. This allele prevents the transcription of Otof into mRNA by splicing the exon 9 into a lacZ cassette. Sizes of exons, introns, and cassettes are not drawn to scale. B) Genotyping primer set and the amplicon size. Wildtype allele: CSD-F/CSD-ttR (520 bp). tm1a allele: Common-loxp-F/CSD-Otor-SR1 (391 bp). C) Quantitative PCR (SYBR® Green) for the RNA from the cochlear sensory epithelium and lateral wall of Otof +/+ and Otof tm1a/tm1a mice. Histograms show the means ± standard deviations of the relative mRNA expression levels of Otof. Actin was used as a housekeeping gene. Otof mRNA expression at the 3' ends of the mutated region was significantly lower in Otof tm1a(KOMP)Wtsi mice than in Otof +/+ mice. D) Western blot analysis with the OTOF antibody revealed that the OTOF protein was not detected in Otof tm1a/tm1a mice. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) Center, Jichi Medical University, Saitama, Japan) [28]. Data are presented as means ± standard deviations. Two-tailed Student's t-tests and Tukey tests were used. A p < 0.05 was considered significant.

Verification of Otof tm1a/tm1a mice by quantitative PCR and western blot analysis
We performed qPCR on the RNA from the cochlear of Otof +/+ mice and Otof tm1a/tm1a mice. After normalizing the cycle threshold (C t ) values of Otof to the C t values of the control actin, Otof mRNA expression in the Otof tm1a/tm1a mice was compared with that in Otof +/+ samples. Transcription of Otof at the 3 ′ ends of the mutated region was found to be ablated in Otof tm1a/tm1a mice (Fig. 1C). In the western blot analysis, the OTOF protein was undetectable in protein samples collected from the cochlear of Otof tm1a/tm1a mice (Fig. 1D).

Otof tm1a/tm1a mice had profound hearing loss but a normal DPOAE response
The average ABR hearing thresholds of Otof tm1a/tm1a mice were at ceiling levels at all frequencies and significantly higher than those of the Otof +/+ and Otof +/tm1a mice ( Fig. 2A).

Otof tm1a/tm1a mice had a reduction in SGNs and an increase in apoptotic cells in the spiral ganglion
We calculated the density of SGNs in the Rosenthal's canal on P7, P14, and P28 (Fig. 3A). The number of SGNs (beta-III tubulin-positive cells) included SGN-Is (beta-III tubulin-positive and peripherinnegative cells) and SGN-IIs (cells positive for both beta-III tubulin and peripherin). The number of SGNs was significantly lower in Otof tm1a/tm1a mice than in wild-type mice at each stage (Fig. 3B). The density of SGNs in Otof tm1a/tm1a mice was not significantly different between P7, P14, and P28 (p = 0.274, Tukey's test). Additionally, significantly more SGNs with immunosignals of cleaved caspase 3 were observed in Otof tm1a/tm1a mice than in wild-type mice at all stages (Fig. 3C).

The number of SGN-IIs between Otof +/+ and Otof tm1a/tm1a mice were not significantly different
We calculated the density of SGN-IIs in the Rosenthal's canal on P7, P14, and P28 ( Fig. 4A and B). The density of SGN-IIs was not significantly different between Otof tm1a/tm1a and wild-type mice at any stage (Fig. 4B). No SGN-IIs with immunosignals of cleaved caspase 3 were observed under our experimental conditions.

Discussion
In the present study, we investigated SGN counts and apoptotic cells in SGNs using Otof tm1a/tm1a mice on P7, P14, and P28 (i.e., before and after hearing onset and adulthood). The Otof tm1a/tm1a mice had a lower number of SGNs than the wild-type mice. SGNs are composed of SGN-I and SGN-II [20], and no apoptotic SGN-IIs were observed under our experimental conditions. Therefore, the reduction of SGNs in Otof tm1a/ tm1a mice was accompanied by apoptosis of SGN-Is (beta-III tubulinpositive and peripherin-negative cells).
Although the mechanisms of SGN degeneration are not yet fully understood, primary and secondary factors have been reported. In temporary noise-induced hearing loss, primary degeneration of SGNs may be preceded by the loss of synapses, despite the presence of intact hair cells [29,30]. The absence of normal hair cells and supporting cells, which results in a reduction in neurotrophin, leads to secondary degeneration of SGNs [29,31].
The Otof immune signal of the spiral ganglion and OHC in mice is transiently and faintly labeled on embryonic day 19.5 (E19.5), P0, and P2 but not on P12 or P20 [5,7]. Therefore, there is a possibility that apoptosis in SGNs occurs as a primary defect due to the Otof mutation. However, it is reasonable to speculate that apoptosis in SGNs is a secondary defect due to a lack of otoferlin in IHCs because Otof signals are primarily expressed in mouse IHCs from E16 to adulthood [7]. Similar to IHCs, ribbon synapses are formed in the retina by retinal photoreceptors and bipolar cells [32]. It has been reported that zebrafish larvae, in which Ribeye a (a protein specific to synaptic ribbons) is depleted, have many apoptotic bipolar cells, and it has been speculated that this apoptosis may be a primary defect due to the inhibition of developing bipolar cells or a secondary defect due to the lack of transmitter release [32]. In this study, the reduction in SGNs was observed in P7 mice, which are prehearing mice, because hearing in mice emerges around P12 [10]. Synaptic exocytosis triggered by spontaneous Ca 2+ spiking activity is observed in IHCs before hearing onset in mice, and Ca 2+ -triggered exocytosis in IHCs shifts from an otoferlin-independent to an otoferlindependent mechanism on P4 [33]. In the present study, the density of SGNs and the percentage of SGNs with immunosignals of cleaved caspase 3 on P4 were not significantly different between wild-type and Otof tm1a/tm1a mice (data not shown). In other words, the reduction in SGNs was observed after the transition to otoferlin-dependent exocytosis. Therefore, apoptosis in SGN-Is may be a secondary defect due to the lack of otoferlin-dependent neurotransmitters, regardless of sound stimulus transmission.
In a previous report, Otof -/mice exhibited a slightly lower number of SGNs than Otof +/+ mice at age 48 weeks [23]. In addition, the reduction in the number of SGNs was observed at an earlier stage in Otof tm1a/tm1a mice than in Otof -/mice. The reason is not clear, but it is possible that different mutations in the Otof gene lead to changes in the timing of decreases in SGN number. In the Otof tm1a/tm1a mice used in the present study, insertion of a large DNA cassette into intron 9 prevented transcription of Otof into full length mRNA due to splicing of exon 9 to a lacZ cassette (Fig. 1A). In fact, qPCR in the present study showed that shorter mRNA at the 5 ′ ends of the mutated region was produced in Otof tm1a/tm1a mice (data not shown). The shorter mRNA may be translated to produce truncated protein, which may have resulted in the earlier reduction in the number of SGNs in Otof tm1a/tm1a mice than in Otof -/mice, although the presence of truncated proteins in Otof tm1a/tm1a mice was not confirmed in the present study. Further investigation will be required to elucidate the mechanism responsible for the reduction in the number of SGNs in Otof-mutant mice.
SGN degeneration in Otof tm1a/tm1a mice in this study is similar to the SGN degeneration reported in vesicular glutamate transporter 3 (VGLUT3) knockout (KO) mice. Although the detailed mechanism of SGN degeneration in VGLUT3 KO mice remains elusive, VGLUT3 KO mice have profound hearing loss due to impaired glutamate release from hair cells at the synapse and exhibit a reduction in SGNs that becomes evident by P10 [34]. Moreover, abnormal ribbons are observed in the IHCs of both otoferlin and VGLUT3 KO mice [34], and otoferlin KO and VGLUT3 KO mice similarly lack normal glutamate release at IHC afferent synapses. Therefore, SGN degeneration in Otof tm1a/tm1a mice and VGLUT3 KO mice may be induced by similar mechanisms.
Previous studies have reported the mechanisms by which synaptic degeneration and neural cell death occur. Mattson et al. reported that various apoptotic signals act at synaptic terminals, including glutamate, reduced trophic support, oxidative stress, amyloid, and ischemia [35,36]. Verhage et al. reported that the neurons of mice lacking Munc18-1, an essential gene for neurotransmitter release, initially differentiate normally but subsequently undergo massive apoptosis due to the lack of synaptic transmission [37]. However, it remains unknown whether a lack or decrease in glutamatergic synaptic activity is a neuronal apoptotic signal.
The differences between wild-type and Otof tm1a/tm1a mice were examined using two-tailed Student's t-tests. Error bars represent standard deviations. *p < 0.05 and **p < 0.005. n = number of sections. SGN, spiral ganglion neuron. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) for neuronal survival [40], and sufficient glutamatergic synaptic activity mediated by AMPA receptors may also be required for SGN survival. We observed that the number of SGN-IIs in Otof tm1a/tm1a mice was not significantly different from those in wild-type mice. The fate decisions of SGN-Is and SGN-IIs are determined before birth [41]; moreover, neurotransmission at the OHC-SGN-II afferent synapse is glutamatergic [42]. Further investigations are required to determine the association between maintenance of SGN-IIs, preserved DPOAE, and glutamatergic OHC-SGN-II afferent synapses in Otof-mutant mice.
Previous studies have reported that patients with OTOF variants have excellent CI performances, because the SGNs in patients with OTOF variants are presumed to be preserved [16]. Although various human OTOF phenotypes have been reported [4], it is not yet clear whether distinct OTOF genotypes correlate directly with CI outcomes. Otof tm1a mice cannot resemble all of the various human OTOF phenotypes. Additionally, it has been also reported that there is no significant correlation between the number of SGNs and speech perception in CI users, and a small number of SGNs may be sufficient for good CI outcomes [43]. To our best knowledge, there are no reports of temporal bone histopathology in patients with OTOF variants. Elucidation of SGN condition in patients with OTOF variants may be important for improving CI outcomes.
In conclusion, we speculate that the reduction in SGN number with apoptosis in Otof tm1a/tm1a mice is likely to be a secondary defect caused by a lack of otoferlin in IHCs, because otoferlin is primarily expressed in IHCs but not in SGNs. Appropriate glutamatergic synaptic inputs may be required for SGN survival. Further investigation is required to elucidate the mechanism underlying the reduction in the number of SGNs in Otofmutant mice. This may enable improved CI outcomes and new drug treatments for patients with OTOF variants in the future.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability
Data will be made available on request.