Developmental Changes in Peripherin-eGFP Expression in Spiral Ganglion Neurons

The two types of spiral ganglion neurons (SGNs), types I and II, innervate inner hair cells and outer hair cells, respectively, within the mammalian cochlea and send another process back to cochlear nuclei in the hindbrain. Studying these two neuronal types has been made easier with the identification of unique molecular markers. One of these markers, peripherin, was shown using antibodies to be present in all SGNs initially but becomes specific to type II SGNs during maturation. We used mice with fluorescently labeled peripherin (Prph-eGFP) to examine peripherin expression in SGNs during development and in aged mice. Using these mice, we confirm the initial expression of Prph-eGFP in both types I and II neurons and eventual restriction to only type II perikarya shortly after birth. However, while Prph-eGFP is uniquely expressed within type II cell bodies by P8, both types I and II peripheral and central processes continue to express Prph-eGFP for some time before becoming downregulated. Only at P30 was there selective type II Prph-eGFP expression in central but not peripheral processes. By 9 months, only the type II cell bodies and more distal central processes retain Prph-eGFP expression. Our results show that Prph-eGFP is a reliable marker for type II SGN cell bodies beyond P8; however, it is not generally a suitable marker for type II processes, except for central processes beyond P30. How the changes in Prph-eGFP expression relate to subsequent protein expression remains to be explored.


INTRODUCTION
The mammalian auditory sensory organ, the cochlea, contains the neurosensory cells specific for the transduction and transmission of sound stimuli to the brain. Bipolar spiral ganglion neurons (SGNs) innervate two different types of mechanosensory hair cells at the periphery and send a process centrally to cochlear nuclei in the hindbrain (De No, 1981). The two types of hair cells, inner hair cells (IHCs) and outer hair cells (OHCs), are innervated by two types of SGNs, myelinated type I and unmyelinated type II, respectively (Brown et al., 1988;Hafidi, 1998;Nayagam et al., 2011). While type I SGNs are the primary auditory neurons for encoding sound stimuli, the function of type II SGNs is to summate and integrate OHC activity to confer cochlear sensitivity and tuning (Guinan and Gifford, 1988). Even though OHCs outnumber IHCs threefold, their innervation by type II SGNs constitutes only about 5-8% of total auditory afferent neurons; the remaining approximately 95% of neurons are type I that innervate IHCs (Spoendlin, 1971). The pattern of innervation delineates the inequality in numbers between the two types of SGNs. A single IHC is innervated by multiple type I fibers, whereas a single type II neuron may innervate 10 or more OHCs (Hafidi, 1998;Simmons and Liberman, 1988;Rubel and Fritzsch, 2002;Weisz et al., 2009;Coate and Kelley, 2013). Centrally, both types I and II SGNs project to three regions of the cochlear nucleus: the anteroventral cochlear nucleus (AVCN), posteroventral cochlear nucleus (PVCN), and the dorsal cochlear nucleus (DCN), maintaining the tonotopic organization reflected in the cochlea (Brown et al., 1988;Nayagam et al., 2011;Fritzsch et al., 2019).
Studying the two different types of SGNs and their projections has been made easier by discovering unique molecular markers (Petitpre et al., 2018;Shrestha et al., 2018;Sun et al., 2018). The prevailing type II neuronal markers are tyrosine hydroxylase (Th), calcitonin-related polypeptide alpha (Cgrpα), and nerve growth factor receptor (Ngfr). However, these markers are not exclusive to type II SGNs, as they have been shown to overlap with type I SGN expression (Vyas et al., 2017(Vyas et al., , 2019Wu et al., 2018). In addition to these markers, an intermediate filament protein, peripherin, is exclusively expressed in mature type II SGNs (Hafidi, 1998;McLenachan et al., 2008;Froud et al., 2015). Peripherin has also been identified in select groups of sensory, motor, and autonomic neurons (Escurat et al., 1990); however, its exact function remains controversial (Lariviere and Julien, 2004). Most of the work on peripherin in type II SGNs has been studied using antibodies. Initially, peripherin protein was shown to be present in all SGNs but becomes restricted to type II SGNs during maturation after birth (Hafidi et al., 1993). Peripherin labeling was detected in both the soma and peripheral and central processes of type II SGNs (Hafidi et al., 1993;Hafidi, 1998;Maison et al., 2016). While antibody work has uncovered some information regarding peripherin protein expression in type II SGNs; not much is known about how peripherin expression changes over time.
To consolidate past research and to expand the range of studies from embryos to adults, we aim to provide a detailed investigation of peripherin expression in type II SGNs during development and aging by demonstrating their peripheral and central projections over time. For this, we used mice with fluorescent labeling of peripherin (Prph-eGFP). Our data show a delayed selective identification of type II neurons, as expected. Furthermore, approximately 2 weeks after birth, while the type II cell bodies retain intense labeling, the peripheral processes begin progressively losing Prph-eGFP expression. In the brain, we show an initial moderate Prph-eGFP expression in the central projection of SGNs that reduces in expression with age.
All animal work was performed as required by the United States Animal Welfare Act and the National Institutes of Health's policy to ensure proper care and use of laboratory animals for research and under established guidelines, supervision, and approved protocols by the Institutional Animal Care and Use Committee (IACUC) of The University of Nevada, Reno, and The University of Iowa.

Fixation and Tissue Preparation
Mice were anesthetized, culled, and transcardially perfused at various stages (E18.5, P4, P7, P8, P15, P30, P40, and 9 Mo) with 4% paraformaldehyde in phosphate-buffered saline (PBS) (pH 7.6) with 0.3 M sucrose to maintain neuronal structural integrity (Fritzsch, 1979;Cragg, 1980). The head was removed and shipped in 0.4% paraformaldehyde (PFA) with 0.3 M sucrose on ice protected from light. The head was then bisected, and the brain halves and temporal bones were removed. The brain was cryoprotected in 30% sucrose overnight, embedded in Tissue-Tek OCT medium (Sakura Finetek Inc., 4583), and quick frozen in a dry ice ethanol bath. The sample blocks were wrapped in foil and stored briefly at -80 • C until sectioning. To section, the blocks were acclimated to -20 • C, trimmed, and mounted on the specimen holder of a Leica CM1800 cryostat with OCT. The sample was sectioned coronally at a thickness of 50-60 µm, and the sections were collected on Superfrost Plus slides (Thermo Fisher Scientific 12-550-15). The slides were stored briefly at -80 • C until ready to view. To view, the slides were washed in PBS for approximately 4 min to remove the embedding medium and coverslipped using Fluoromount-G with DAPI (Thermo Fisher Scientific 00-4959-52). Care was taken to protect the samples from light at all procedural stages.
The temporal bones were decalcified in 0.25 M ethylenediaminetetraacetic acid (EDTA) solution (RPI E57020) for up to 5 days (>P8) with daily solution changes. Decalcified cochleae were washed in PBS and microdissected, and the tectorial membrane was removed. Cochlear turns were flat mounted in glycerol or Fluoromount-G with 4 ,6-diamidino-2-phenylindole (DAPI) for viewing.

Imaging
Images were acquired using a Leica SP8 scanning laser confocal microscope, analyzed with Leica LAS X software and processed with CorelDRAW graphics suite. Images were taken at 1-6-µm thick optical sections to compile a given stack, in up to four different colors (405, 488, 552, and 638 nm laser lines) using three different magnifications (10× with a 0.6 NA; 20× with a 0.95; 63× with a 1.4 NA).

Quantification
For analysis of the distribution of peripherin-positive neurons within the cochlea, we quantified the number of peripherinpositive neurons in the proximal, middle, and distal regions for the apex, middle, and base. Three 150-µm wide boxes were drawn randomly around the spiral ganglia region, each for the base, middle turn, and apex per animal. The box's length was the distance from the most proximal SGN to the most distal SGN cell body. Each box was divided into equal thirds, separating the proximal, middle, and distal regions, and the number of Prph-eGFP-positive neurons was manually counted within each region. We quantified four P15, four P30, and three 9-monthsold cochleae.
For the analysis of peripherin labeling within the hindbrain, the peripherin signaling intensity level was quantified within the AVCN, PVCN, DCN, and the vestibular nucleus, adjacent to its entry point E18,5, P4, P30, and 9-months-old mice. Three 100 × 100 µm boxes were drawn at random within maximum projection images of coronal sections of the AVCN, PVCN, DCN, vestibular nucleus, and trigeminal nucleus. This size was selected as a compromise between maximizing the area quantified and having the box fit within the different brain regions' boundaries. The trigeminal nucleus was chosen. It had a high level of peripherin expression throughout the entire nucleus and was located within each coronal section containing the AVCN, PVCN, DCN, or vestibular nucleus. In addition, three 100 × 100 µm boxes were drawn randomly in areas of background per image. Images were cropped along the boxes in CorelDraw and exported as individual TIFF files. Individual TIFF files were analyzed using the histogram function in ImageJ software. The resulting analysis calculates the mean fluorescent intensity of the entire TIFF file. The background's mean fluorescent intensity was subtracted from respective images of the AVCN, PVCN, DCN, vestibular nucleus, and trigeminal nucleus. The fluorescent intensity of the trigeminal nucleus (minus background) was set at 100%, and the fluorescent intensities of the AVCN, PVCN, DCN, and vestibular nucleus (minus backgrounds) were calculated as a percent of the portion of the trigeminal nucleus from the image in which they were obtained. The mean intensities were averaged, and the standard error was calculated using Microsoft Excel. Statistical significance was performed with one-way ANOVA with the Tukey honestly significant difference (HSD) post hoc test. The confidence level was set at 95%.

Spiral Ganglion Neurons
All SGN cell bodies are initially positive for Prph-eGFP at E18.5 ( Figure 1A), consistent with previous work using peripherin antibodies (Hafidi et al., 1993;Hafidi, 1998;Nayagam et al., 2011). The level of Prph-eGFP remains relatively high in all SGNs at P4 (Figures 1B,B ), although some SGNs begin to have less expression than others ( Figure 3C). At P8, Prph-eGFP has become restricted to the type II SGNs, with no expression observed in type I cell bodies (Figures 1C-C ). These type II cell bodies continue to express Prph-eGFP through at least 9 months of age ( Figures 1D-F ).
These Prph-eGFP-positive type II cell bodies are not uniformly located throughout the spiral ganglion. Instead, these Prph-eGFPpositive cell bodies appear to be more concentrated along the distal region of the spiral ganglion (Figures 1C-F ), consistent with reports using antibody labeling (Hafidi, 1998;Nayagam et al., 2011;Defourny et al., 2013;Grandi et al., 2020). This asymmetrical distribution was better visualized with NeuN labeling of Prph-eGFP mice to label all SGN nuclei (Figure 2A). To confirm that these Prph-eGFP-positive type II SGNs are indeed asymmetrically located, we quantified the number of Prph-eGFP-positive neurons within 150-µm-wide representative areas of the spiral ganglia, dividing each area into proximal, middle, and distal regions for P15, P30, and 9-month animals ( Figure 2B). This was repeated for each area of the cochlea: base, middle turn, and apex. For the base, middle turn, and apex of all three ages, there were significantly more Prph-eGFPpositive cells in the distal third than in the proximal third regions (ANOVA, p < 0.01, n = 3-4 animals with three measurements per animal) (Figures 2C-C ). In addition, for the base and middle turn, there were significantly more Prph-eGFP-positive cells in the distal third as compared with the middle-third regions for all ages (ANOVA, p < 0.01, n = 3-4 animals with  three measurements per animal) (Figures 2C,C ). However, the asymmetrical distribution of Prph-eGFP-positive cells was less defined at the apex as compared with the near absence of Prph-eGFP-positive cells in the proximal or middle regions at the base or middle turn (Figures 1C-F , 2). While there was a significant difference between distal and proximal regions in all ages, only at P15 and 9 months was there a significant difference between the distal and middle regions in the apex (ANOVA, p < 0.01, n = 3-4 animals with three measurements per animal, Figure 2C ). Interestingly, unlike the asymmetrical distribution of Prph-eGFP-positive cells in the spiral ganglion, there is a wide distribution of Prph-eGFP-positive cells in the vestibular ganglion (Supplementary Figure 1), suggesting that there may be some significance to the distribution.
In summary, these results show a significant asymmetrical distal distribution of mature Prph-eGFP-positive type II SGNs in all cochlea regions, with a more robust distal distribution in the base and middle turn.

Peripheral Processes to Hair Cells
We next examined the peripheral processes to the hair cells in Prph-eGFP-positive neurons (Figure 3). At E18.5, peripheral processes of both type I and II neurons strongly express Prph-eGFP (Figures 3A-B ). Prph-eGFP-positive processes are observed projecting to OHCs only in the base and middle turn at E18.5 (Figures 3B-B ). Using a tdTomato reporter line driven by Neurod1-cre, we labeled all type I and II SGNs and showed that while there is no Prph-eGFP labeling yet to OHCs in the apex (Figure 3B ), the type II SGNs innervate these OHCs at E18.5 (Supplementary Figure 2), suggesting a delay in Prph-eGFP expression after hair cell innervation. Prph-eGFP-positive processes to IHCs can be seen in each region of the cochlea at this stage (Figures 3B-B ). Four days after birth, the peripheral processes of both type I and II neurons have maintained strong expression of Prph-eGFP (Figures 3C-D ). In addition, at P4, Prph-eGFP-positive processes to OHCs are now observed at the apex in addition to the base and middle turn (Figures 3D-D ). Prph-eGFP-positive processes to IHCs are observed in all areas of the cochlea as well (Figures 3D-D ). We note that at P7, selective antiperipherin-positive peripheral processes can be observed (Supplementary Figure 3), consistent with previous work showing peripherin antibody expression in postnatal stages (Hafidi, 1998;Vyas et al., 2017). At P8, while type I processes to IHCs express a strong level of Prph-eGFP throughout the cochlea, type II processes to basal and middle turn OHCs express less Prph-eGFP than type II processes to apical OHCs (Figures 3E-F ). By P15, the relative level of Prph-eGFP expression in both type I and II peripheral processes at the base and middle turn is less than that in the type II cell bodies, with expression in type II to OHCs being less than in type I to IHCs. Furthermore, expression of Prph-eGFP in type II peripheral processes to apical OHCs has also decreased some, although the expression of Prph-eGFP in type I SGNs to IHCs remains high (Figures 3G-H ). The intensity of expression in type I apical peripheral processes appears similar to that in type II Prph-eGFP-positive cell bodies ( Figure 3G ). By P30, however, the level of Prph-eGFP in type I SGNs to IHCs in the apex has decreased (Figures 3I,J). There is minimal Prph-eGFP expression in peripheral processes at 9 months, and labeled type II processes out to OHCs were not observed (Figures 3K,L). Interestingly, there are highly ramified cells of unknown origin located among the peripheral processes at the base in the 9-months cochlea that express high levels of Prph-eGFP ( Figure 3L). Together, these results suggest that Prph-eGFP expression in the peripheral processes progresses in a basal to apical wave, beginning first with its upregulation at late embryonic stages, followed by its downregulation approximately 2 weeks later, the onset of hearing. After this point, the level of Prph-eGFP in the peripheral processes continues to decrease until it is at a barely detectable level compared with that in the cell bodies by 9 months.

Central Projections of Spiral Ganglion Neurons
Next, we examined the central processes of the cochlear nuclei in Prph-eGFP-positive SGNs. At E18.5, Prph-eGFP-positive processes were observed in each of the three regions of the cochlear nucleus, AVCN, PVCN, and the DCN (Figures 4A-A' ). However, by P4, the Prph-eGFP-positive projections to the DCN were noticeably downregulated compared with E18.5 and compared with the AVCN and PVCN, suggesting that Prph-eGFP expression is downregulated or the Prph-eGFP is no longer being transported along with central processes to the DCN (Figures 4B-B ). Similarly, at P30, while the AVCN and PVCN receive input from processes expressing Prph-eGFP, the DCN does not (Figures 4C-C ). However, by 9 months, central processes to all regions of the cochlear nucleus do not express Prph-eGFP (Figures 4D-D ). In contrast to SGNs, Prph-eGFP-positive vestibular neurons were observed projecting to vestibular nuclei and trigeminal neurons to trigeminal nuclei at all stages (Figure 4). The latter expressed very strong levels of Prph-eGFP throughout development and in the aged, 9months mice. Given the Prph-eGFP labeling in both type I and II peripheral processes beyond when the Prph-eGFP labeling is confined to only type II cell bodies, we cannot be certain of the central processes' origin. However, specific labeled central processes of these labeled type II cells can be observed by P30 and remains at 9 months (Figures 4E ,E ), suggesting that after this time point, the Prph-eGFP-positive central processes are most likely type II fibers. Between P4 and P30, while type II SGNs have a stronger expression of Prph-eGFP, lower expression levels are observed in central processes of type I SGNs (Figures 4E -E ). At E18.5, all central processes are strongly labeled (Figure 4E).
To confirm these observational differences between the different regions of the cochlear nucleus and the vestibular nucleus across development and aging, we determined the mean fluorescent intensity for sample regions within the AVCN, PVCN, DCN, and vestibular nucleus adjacent to its entry point. Since neurons within the trigeminal nucleus expressed high levels of Prph-eGFP continuously, we used it as our reference point. We calculated the relative percent of fluorescent intensity of  afferents innervating the AVCN, PVCN, DCN, and vestibular nucleus relative to the trigeminal nucleus for each time point (Figure 5). In the AVCN, the level of expression of Prph-eGFP remained relatively unchanged in early development, through at least P30, but by 9 months, expression was significantly lower (ANOVA, p < 0.05, n = 3 measurements; Figure 5B). Similarly, the expression of Prph-eGFP in the PVCN at 9 months was significantly lower than the other three earlier time points (ANOVA, p < 0.05, n = 3 measurements; Figure 5B). However, in the DCN, the level of Prph-eGFP significantly dropped at a much earlier time point of P4 (ANOVA, p < 0.05, n = 3 measurements; Figure 5B). While the level of Prph-eGFP in the vestibular nucleus also significantly dropped by P4 (ANOVA, p < 0.05, n = 3 measurements), this decrease was markedly less than in the DCN (Figure 5B). We next wanted to compare differences between the different cochlear nuclei within a given age. At E18.5, there was a slightly significant difference between Prph-eGFP expression between the AVCN and PVCN but no significant difference between either of those with the DCN (ANOVA, p < 0.05, n = 3 measurements; Figure 5B). At P4 and P30, the Prph-eGFP expression levels within the AVCNs and PVCNs were significantly higher than that in the respective DCNs (ANOVA, p < 0.05, n = 3 measurements; Figure 5B), although by 9 months, the level of Prph-eGFP was at an equally low level in all three areas of the CN ( Figure 5B).
Together these results suggest that, like in the peripheral processes to the hair cells, central processes of SGNs to the cochlear nuclei downregulate the expression of Prph-eGFP by 9 months; however, the specific timeline of when this happens for projections to a given region of the cochlear nucleus is variable.

DISCUSSION
Our data expand upon previous studies using antibodies (Hafidi et al., 1993;Hafidi, 1998;Nayagam et al., 2011) to describe the gain and loss of Peripherin expression across several time points in development and 9-months-old mice. Using the Prph-eGFP mouse line (McLenachan et al., 2008), we examined expression in SGN cell bodies, their distal processes to hair cells, and the central projections to the cochlear nuclei. Our results here confirm previous antibody labeling, which demonstrated selective peripherin labeling of mature type II SGNs (Hafidi et al., 1993). In our study, Prph-eGFP becomes restricted to type II SGN cell bodies between P4 and P8 (Figure 1). Similarly, peripherin antibody staining also becomes restricted to type II SGN cell bodies after P3 in the rat (Hafidi et al., 1993). As in this previous study with peripherin antibodies, once Prph-eGFP became restricted, specifically type II SGN cell bodies, they remained restricted throughout adulthood (Figure 1), suggesting that Prph-eGFP is a useful marker for mature type II SGN perikarya. The unique localization to the distal region of the spiral ganglia (Figure 2) has been reported (Berglund and Ryugo, 1987;Hafidi, 1998;Nayagam et al., 2011;Maison et al., 2016;Grandi et al., 2020); however, we observed that this primarily distal distribution of type II SGNs was not uniform along the cochlea.
Interestingly, while Prph-eGFP-positive type II neurons are located primarily in the distal third of the spiral ganglion, there was a broader distribution at the apex than the base or middle turn (Figure 2). This broader distribution in the apex was more similar to the overall distribution of Prph-eGFP-positive cells in the vestibular ganglion (Supplementary Figure 1). Like the unmyelinated type II SGNs innervating type II cochlear hair cells, the unmyelinated vestibular afferents also express peripherin (Leonard and Kevetter, 2002). Given that the apex of the cochlea is converted into a vestibular/lagena-like epithelial arrangement with the loss of N-Myc (Kopecky et al., 2011), this might suggest an evolutionary relationship between the distribution of Prph-eGFP-positive cells in the vestibular ganglion and the spiral ganglion at the apex. However, the significance of the distribution of Prph-eGFP-positive perikarya in the two different ganglia is not known. Furthermore, there is heterogeneity in type II SGNs in the specific genes they express depending upon their location in the cochlea (Vyas et al., 2019). Thus, the different genetic makeup of the SGNs located at the base versus the apex may also play a role in their specific distribution, although it remains to be explored.
Remarkably, while Prph-eGFP expression became restricted only to type II SGN cell bodies, the expression of Prph-eGFP in peripheral processes was not (Figure 3). Similarly, peripherin protein expression in type I SGNs peripheral processes to IHCs has also been shown in adult animals (Hafidi, 1998). Furthermore, the expression is not restricted to only type II peripheral processes as it is to the cell body, but the relative level of Prph-eGFP expression in both type I and II SGN peripheral processes changes over time (Figure 3). The expression of Prph-eGFP appears to come on as a wave from base to apex after the peripheral processes have innervated the hair cells (Supplementary Figure 2). Before the onset of hearing, Prph-eGFP expression decreases in the same base to apex progression, first with the type II SGNs followed by type I SGNs. By P30, minimal Prph-eGFP expression in SGN peripheral processes remains (Figure 3). Peripherin protein can be identified in adult rats (Hafidi, 1998); however, the colabeling of peripherin antibodies with other type II SGN markers has shown that some type II SGNs do not express peripherin (Vyas et al., 2017). What remains to be explored is how changes in Prph-eGFP expression in type II peripheral processes correlate with potential protein expression changes, especially within different cochlea regions already known to express different genes (Vyas et al., 2019).
Centrally, at E18.5, all three regions of the cochlear nucleus, AVCN, PVCN, and DCN are innervated by Prph-eGFP-positive type I and II SGNs (Figures 4, 5). This finding is unsurprising given that even before this stage, lipophilic dye labeling has shown that SGNs reach all areas of the cochlear nucleus (Fritzsch et al., 1997;Schmidt and Fritzsch, 2019;Filova et al., 2020). Interestingly, by P4, central processes to the DCN are no longer positive for Prph-eGFP and remain that way through at least 9 months. Whether this has anything to do with the unique function of the DCN from that of the AVCN and PVCN remains to be explored. By 9 months, virtually no Prph-eGFP-positive central projections are detected in any of the cochlear nuclei. However, since at 9 months, only the type II SGNs appear to express Prph-eGFP in their central processes, and given that these neurons only make up about 5% of the total neuronal population (Spoendlin, 1971), the ability to distinguish their projections may be that they are below the level of reliable fluorescence detection in the brain. More likely, since P30 central processes also appear to be specific to type II SGNs and Prph-eGFP expression is detected in the AVCN and PVCN at this stage, the level of Prph-eGFP in the central most aspect of the central processes may decrease from P30 to 9 months as it does in peripheral processes over time.
Collectively, these results show that Prph-eGFP is a reliable marker for type II SGN cell bodies beyond P8; however, it is not completely specific to peripheral or central processes. Still, type II central processes appear to exclusively express Prph-eGFP at P30 and beyond and could serve as a reliable marker in these older animals providing further confirmation of specificity through colabeling with other markers known to be exclusively expressed in mature type II SGNs.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

ETHICS STATEMENT
The animal study was reviewed and approved by Institutional Animal Care and Use Committee (IACUC) of The University of Nevada, Reno, and The University of Iowa.

AUTHOR CONTRIBUTIONS
KE, JK, and BF designed and performed the experiments and wrote the manuscript. JL, IJ, and GP assisted with data collection. EY and BF conceptualized the manuscript. All authors contributed to the article and approved the submitted version.

FUNDING
This work was supported by a grant from the NIH/NIA (R01 AG060504, P01, AG051443, 1R01 DC015135, DC016099, and 1R01 DC015252-01) to EY and BF; by a grant from the Czech Science Foundation (20-06927S), and the Czech Academy of Sciences (RVO 86652036) to GP.