Assessment of auditory and vestibular damage in a mouse model after single and triple blast exposures

Abstract

The use of explosive devices in war and terrorism has increased exposure to concussive blasts among both military personnel and civilians, which can cause permanent hearing and balance deficits that adversely affect survivors’ quality of life. Significant knowledge gaps on the underlying etiology of blast-induced hearing loss and balance disorders remain, especially with regard to the effect of blast exposure on the vestibular system, the impact of multiple blast exposures, and long-term recovery. To address this, we investigated the effects of blast exposure on the inner ear using a mouse model in conjunction with a high-fidelity blast simulator. Anesthetized animals were subjected to single or triple blast exposures, and physiological measurements and tissue were collected over the course of recovery for up to 180 days. Auditory brainstem responses (ABRs) indicated significantly elevated thresholds across multiple frequencies. Limited recovery was observed at low frequencies in single-blasted mice. Distortion product otoacoustic emissions (DPOAEs) were initially absent in all blast-exposed mice, but low-amplitude DPOAEs could be detected at low frequencies in some single-blast mice by 30 days post-blast, and in some triple-blast mice at 180 days post-blast. All blast-exposed mice showed signs of tympanic membrane (TM) rupture immediately following exposure and loss of outer hair cells (OHCs) in the basal cochlear turn. In contrast, the number of inner hair cells (IHCs) and spiral ganglion neurons was unchanged following blast-exposure. A significant reduction in IHC pre-synaptic puncta was observed in the upper turns of blast-exposed cochleae. Finally, we found no significant loss of utricular hair cells or changes in vestibular function as assessed by vestibular evoked potentials. Our results suggest that (1) blast exposure can cause severe, long-term hearing loss which may be partially due to slow TM healing or altered mechanical properties of healed TMs, (2) traumatic levels of sound can still reach the inner ear and cause basal OHC loss despite middle ear dysfunction caused by TM rupture, (3) blast exposure may result in synaptopathy in humans, and (4) balance deficits after blast exposure may be primarily due to traumatic brain injury, rather than damage to the peripheral vestibular system.

Graphical Abstract

1. Introduction

The use of explosive weapons, particularly improvised explosive devices (IEDs), has become a hallmark of recent global military conflicts including the wars in Afghanistan and Iraq. Approximately half of all U.S. casualties in Iraq and Afghanistan were attributed to IEDs (Mann and Fischer, 2019), and explosives caused nearly 80% of injuries sustained in these conflicts (Owens et al., 2008). The high energy content of blast waves can cause lifelong health problems for veterans, including traumatic brain injury (TBI).

Veterans with blast-related TBI consistently report auditory and vestibular deficits (Lew, 2007; Scherer et al., 2007). Additionally, some blast-exposed veterans with clinically normal hearing—as measured by traditional audiograms—nevertheless report difficulty hearing in noisy environments and perform poorly on performance-based auditory processing tests (Saunders et al., 2015). These results suggest the possibility of cochlear synaptopathy, also called hidden hearing loss, which may reduce the listener’s ability to hear signals in noise (Kujawa and Liberman 2009; Kujawa and Liberman 2015), as well as possible damage to the central auditory system (Saunders et al., 2015). These complications are potentially disabling to blast-exposed veterans, hindering societal re-integration and quality of life.

Despite the prevalence of these injuries among soldiers, relatively little is known about the underlying etiology of blast-induced hearing or balance dysfunction. While several groups have used animal models to examine the effect of blast exposure on the inner ear, these studies largely focused on changes to the auditory system alone, and most examined the effect of only a single blast exposure over the course of several weeks. Because military personnel may face multiple blast exposures in the course of training and combat, we sought to characterize the effect of both single and triple blast exposures on the inner ear using a simulator which was specifically engineered to produce highly repeatable blast waves in the range of what an IED produces upon detonation (Varas et al., 2011). Additionally, we examined cochlear and vestibular function and histology over an extended recovery period of up to six months, to investigate the long-term effects of blast exposure on the inner ear.

2. Materials and Methods

2.1. Animals and blast exposures

All animal experiments were conducted under Institutional Animal Care and Use Committee approved protocols at Walter Reed Army Institute of Research (WRAIR), the National Institutes of Health (NIH), and Stanford University School of Medicine, in accordance with the Animal Welfare Act and other federal statutes and regulations relating to animals and experiments involving animals, and adhered to principles stated in the Leader for the Care and Use of Laboratory Animals. All blast exposures were performed at WRAIR in Silver Spring, MD. Adult (8–10 week old) male CBA mice were randomly assigned to sham control, noise control, single blast, or triple blast treatment groups. Animals were anesthetized with 4% isoflurane gas in an induction chamber, then placed inside the blast simulator in a prone position facing the direction of the oncoming blast wave. The blast simulator consisted of a 0.5 ft long compression chamber connected to a 21 ft long expansion section that extends 3.6 ft into the end wave eliminator, which eliminates secondary shock wave. The blast overpressure was generated by rupture of a 0.34 in thick VALMEX® (Mehler Texnologies) membrane between the two sections of the blast simulator. The compression chamber was rapidly pressurized until the membrane ruptured, yielding a shockwave with a peak static pressure of 16 psi with approximately 4 msec positive phase duration and an impulse of ~24 psi*ms. Frequency of the Friedlander-like waveform (positive and negative pressure) is approximately 81 Hz and power spectrum analysis of the wave showed that the impulse is most intense (peak amplitude) at approximately 43Hz (Sajja et al., 2018; Heyburn et al., 2019; Arun et al., 2020; Wang et al., 2020).

Sham control and noise control animals were anesthetized and handled in the same way as blast-exposed animals. Those in the noise control group were positioned at the same distance from the sound source as blast-exposed animals, but in cages outside the blast simulator, where they were exposed to the sound of the blast but not the shock wave directly. Sham control animals were held in cages in another room entirely. The ABR thresholds of sham control and noise control animals were found to be similar (Suppl. Fig. 1). Therefore, in subsequent analyses, they were combined into a single control group, and are referred to as “control” condition animals for the remainder of this study. Blast-exposed animals received an initial 8 minutes of 4% isoflurane in an anesthesia chamber. They were the transferred into the simulator for blast overpressure exposures, and triple-blast animals were removed and given additional 4% isoflurane for 2 minutes between exposures. Testing for righting reflex time in control and blast-exposed animals indicated that animals remained anesthetized during blast exposures (Table 1). All animals were then maintained for the indicated recovery periods prior to data collection. Whenever possible, physiological data were taken from the same individuals at different time points, to reduce the total number of animals used. The types of data collected and the recovery time points assessed are summarized in Table 2.

2.2. Measures of hearing and balance function, and otoscopy

Auditory brainstem response (ABR) thresholds and the presence or absence of distortion product otoacoustic emissions (DPOAEs) were assessed as described in an earlier study (Morozko et al., 2015). Briefly, for ABR measurements, animals were anesthetized with intraperitoneal injections of ketamine (60 mg/kg) and dexdomitor (0.4 mg/kg), and subdermal electrodes were placed in the scalp between the ears and below the pinna of each ear. Blackman-gated tone-burst stimuli (Blackman gated) of 3 ms with a rise/fall time of 0.75 ms were presented to each ear at 8, 16, 32, and 40 kHz pure tones using hardware (RZ6 Multi I/O processor, MF-1 speakers) and software (BioSigRz, v. 5.1) from Tucker-Davis Technologies (Alachua, FL, USA). The contralateral ear was used as reference. Thresholds were subjectively assessed by visual inspection of waveforms. For data analysis purposes, if ABRs could not be observed at 90 dB SPL, they were coded as 100 dB SPL. For DPOAEs, two tones at f1=65 dB SPL and f2=55 dB SPL were presented to the ear using two MF-1 speakers (as above). The f2 frequency ranged from 4–40 kHz and f2/f1=1.25. The distortion products were recorded using an ER-10B+ microphone (Etymotic, Elk Grove Village, IL). DPOAEs with levels of −5 dB SPL or above were considered present. Some animals were assessed both before blast exposure (baseline), and after set recovery periods ranging from 1 day to 180 days. While most of the physiological data were measured at WRAIR (including all of the long-term data obtained at 90 and 180 days post-exposure), one cohort of mice was evaluated for otoscopy, vestibular evoked potentials, and DPOAE measurements at NIH for time points between 1 and 30 days. For data analysis, the ABR thresholds of the most sensitive ear at baseline or the earliest time point available were used. If data for the chosen ear was missing at a later time point, data from the other ear was substituted. Each animal was represented only once at each time point to avoid pseudoreplication.

Vestibular evoked potentials (VsEPs) were obtained following the protocol described in a previous study (Tona et al., 2019). Briefly, mice were anesthetized in the same manner as described above. Subdermal electrodes were placed in the scalp along the nuchal crest, below the pinna of one ear, and the right hip (ground). Mice were laid supine on a platform and their heads were secured inside a plastic clip coupled to a mechanical shaker (Labworks Inc., Costa Mesa, CA). Shaker movements (1 g/msec rectangular jerk pulses of 2 msec duration, delivered at 17 Hz) were driven by an RMX850 power amplifier (QSC, Costa Mesa, CA). VsEPs were measured from animals at 1 day, 3 days, 14 days, and 1 month after blast treatment. Thresholds were determined by visual inspection of waveforms, and wave P1 peak-to-peak amplitudes and latencies were measured in the BioSigRz software. Otoscopic inspection of tympanic membranes (TMs) was performed using a Sony MCC-500 HD camera attached to a Zeiss Opti Pico S100 surgical microscope prior to tissue collection.

2.3. Immunofluorescence and plastic sectioning

Prior to tissue collection, animals were deeply anesthetized with 4% isoflurane and perfused transcardially with 4% paraformaldehyde or a mixture of 3% glutaraldehyde and 2% paraformaldehyde. Otic capsules were removed from the temporal bones and incubated in 0.25M-0.5M EDTA for 2–7 days to decalcify the bony labyrinth.

When preparing whole-mounts, the cochlea was divided into the apical, middle, and basal turns. Each turn was then trimmed of lateral wall and decalcified bone, and the Reissner’s and tectorial membranes were removed with #55 forceps (Fine Science Tools). The dissected cochlea and utricle were placed in blocking solution containing 0.1% Triton-X, donkey anti-mouse fab fragments at 1:200 (catalog #715-007-003, Jackson Immunoresearch Lab), and normal donkey serum at 1:200 in 1X PBS, and then incubated in different combinations of primary antibodies overnight (rabbit anti-MYO7A catalog #25-6790 from Proteus Biosciences, Inc. at 1:2000, Mouse anti-Pou4f3 catalog #sc-81980 from Santa Cruz Biotechnology, Inc. at 1:200, goat anti-Sox2 catalog #AF2018 from R&D Systems at 1:200, mouse anti-CTBP2 catalog #612044 from BD Transduction Laboratories at 1:500). All immunostaining steps were carried out at 4°C. The following day, samples were washed three times with 0.1% Tween in PBS, and secondary antibodies conjugated to fluorophores (1:1000, Invitrogen Alexa Fluor) were applied for 1–2 hours at room temperature. After three more washes, the samples were mounted in Vectashield (catalog #H-1000-10, Vector Laboratories) or Fluoromount-G (catalog #00-4958-02, Invitrogen) and imaged using a Zeiss 710 confocal microscope. Image analysis was performed using Zen Black software.

For cochleae, the number of IHCs and OHCs were counted in 200 μm stretches in the regions corresponding to the frequency ranges 7–10 kHz, 20–28 kHz, and 48–60 kHz (hereafter referred to as the apex, middle, and basal turns). Areas that were damaged during dissection, and areas with 4 rows of OHCs, were avoided. For synaptic ribbon counts, rectangular regions were drawn around 15 IHC nuclei per turn, and the number of pre-synaptic puncta was quantified by stepping through the Z-stack of a 40x or 63x confocal image. Total puncta per IHC were then averaged within each turn. For the utricle, the number of hair cells in five 50 μm × 50 μm areas were counted and averaged (Figure 7B).

Figure 7. The number of utricular hair cells is unchanged following blast exposure.

a. MYO7A immunostaining of control (left) and triple-blasted (right) utricles at 30 days (top) and 180 days (bottom) post-exposure. The density of hair cells is comparable across samples. While some decrease in striolar hair cells was evident in the triple blast sample at 180 days, a similar decrease was observed in the control. b. Utricular hair cells were counted within five 2500 μm² (50 μm × 50 μm) regions per utricle and averaged for each utricle. Only one utricle per animal was counted. MYO7A, marking hair cell bodies, is stained in magenta. c. Average utricular hair cell counts per 2500 μm² region, plotted by time point. Utricular HC density was not significantly affected by condition or time point. Boxplot whiskers show standard error of the mean. d. Phalloidin staining of stereocilia bundles of a control utricle at the 1-day time point (left)and triple blast utricle at the 180-day time point (right). Bundle density is comparable across samples. The areas boxed in red are shown in higher magnification in the insets.

For plastic sectioning, EDTA-decalcified specimens were serially dehydrated in ethanol, then embedded using the JB-4 Embedding Kit (catalog #0226A-800, Polysciences, Inc.). Sections of 4–8 micron thickness were cut on a rotary microtome (Leica RM2265), stained with thionin, mounted with Permount (catalog #SP15-100, Fisher Chemical), and imaged on a Leica microscope (DM5000 B). Spiral ganglion neurons were counted in only mid-modiolar sections.

2.4. Statistics and data analysis

For each type of data gathered across multiple time points and conditions, outliers (defined using a criterion of three times the interquartile range for each time point and condition) were removed prior to statistical analysis, and each animal was represented in each data set only once to avoid pseudoreplication. For all statistics, a two-way ANOVA with interaction effect was performed using RStudio (version 1.2.5001, RStudio, Inc.), with blast condition and recovery time point as main effects. For ABRs, thresholds were split by test frequency prior to analysis. For cochlear histology, data were split by cochlear turn. In all multiple comparison tests, p-values were adjusted using the false discovery rate (FDR) method. If a main effect was significant at the p ≤ 0.05 level, post-hoc Tukey tests were performed to identify which pairwise contrast(s) contributed to the significant main effect. If the interaction effect between condition and time point was significant at the p ≤ 0.05 level, simple effects tests were performed. This required splitting the data by one main effect, to examine the effect of the other main effect. In this manner, data could be compared across conditions at a single time point, or across time points within a blast condition. P-values returned by RStudio as “<2E-16” (the lowest possible value that R could report with precision) were entered as 2E-16 prior to any necessary multiple corrections adjustments. Any tests with p-values greater than 0.05 were considered non-significant.

3. Results

3.1. Blast exposure causes long-term hearing loss, with limited recovery at low frequencies

Baseline mean ABR thresholds were between 20–35 dB SPL for all frequencies tested. One day after blast exposure, single- and triple-blasted mice exhibited average threshold shifts of ~50–65 dB SPL across all frequencies (Fig. 1a). Most single-blasted mice had no ABR response at the maximum stimulus level of 90 dB SPL at 32 kHz and 40 kHz, but responses were measurable at high levels at 8 kHz and 16 kHz (mean thresholds of ~76 and ~82 dB SPL, respectively). Triple-blasted mice exhibited average thresholds in excess of 90 dB SPL across all frequencies. The effects of condition, time point, and their interaction on ABR thresholds at each frequency were significant (Table 4). When post-hoc tests were performed on the effect of condition within each time point, post-blast ABR thresholds were found to be significantly different from controls at all time points and frequencies. The only exception was the 8 kHz thresholds of single-blast animals, which showed a recovery to control values after 180 days. At early time points (1 day and 7 days post-exposure), thresholds of single- and triple-blasted mice did not significantly differ from each other at any frequency, but significant differences between blast conditions did emerge at or after 30 days post-exposure at 8 kHz, 16 kHz, and 32 kHz. In each case, single-blast animals showed a significant recovery of threshold relative to triple-blast. ABR thresholds at 40 kHz did not differ significantly between single- and triple-blast animals at any time point (Fig. 1a; Table 5).

Figure 1. Blast exposure causes both temporary and long-term threshold shifts.

a. Mean auditory brainstem response (ABR) thresholds at 8 kHz, 16 kHz, 32 kHz, and 40 kHz, for control, single- and triple-blast exposed mice before treatment (baseline), and at 1 day, 7 days, 30 days, 90 days, and 180 days after exposure. Significant threshold shifts were observed at 1 day following either single or triple exposure at all frequencies and time points. Single-blast exposed animals showed a gradual recovery of thresholds over time, although significant differences from control persisted at all frequencies except 8 kHz, even at 180 days of recovery. Threshold shifts in triple-blast exposed mice showed minimal (8 kHz) to no recovery. Error bars are standard error of the mean. b. Percentage of mice with detectable DPOAEs at any frequency at each time point. Sample sizes from baseline to 180 days for control mice: 10, 8, 6, 8, 9, 7; for single blast mice: 14, 15, 14, 14, 7, 6; for triple-blast mice: 8, 12, 6, 5, 5, 4. Symbols within each time point have been laterally spaced for better visualization. *p≤0.05, **p≤0.01, ***p≤0.001.

When post-hoc tests were performed on the effect of time point within a condition, single-blasted animals exhibited significant recovery of thresholds between 1 day and 30 days post-blast at 8 kHz (p=0.02) and between 1 day and later time points at 16 kHz (p=5.94E-03 for 1 vs. 30 days; p=0.03 for 1 vs. 90 days; p=2.4E-03 for 1 vs. 180 days; Figure 1A, Table 6). However, there was no additional significant recovery between 30 days and any later time point. Despite modest reductions over time, the ABR thresholds of single-blast mice at 32 kHz and 40 kHz did not exhibit statistically significant recovery after exposure. Triple-blasted animals showed no significant recovery in threshold at any time point or frequency, except at 8 kHz between 30 days and 1 day post-exposure (p=0.01; Table 6). At 32 kHz and 40 kHz, the thresholds of triple-blasted mice were essentially stable (±5 dB SPL) between 1 day and 180 days after blast. Finally, the only cases of complete recovery, defined as ABR thresholds no longer significantly different from baseline, were observed for single-blast exposure at 8 kHz after 90 and 180 days (Table 6).

3.2. Blast exposure abolishes distortion product otoacoustic emissions (DPOAEs)

To examine outer hair cell function, DPOAEs were recorded from control, single- and triple-blast-exposed mice. DPOAEs were absent at all frequencies in both single- and triple-blasted mice at 1 day and 7 days post blast (Fig. 1b). Small DPOAEs were recorded from some single-blast animals as early as 30 days post-exposure, and more animals showed DPOAE recovery over time. By 180 days post-exposure, low level DPOAEs were detectable in 67% (4 of 6) single blast mice. In contrast, no DPOAEs were detected in triple-blast mice prior to 180 days, and at that time point only 1 of 4 triple-blasted mice had detectable DPOAEs (Fig. 1b).

3.3. Tympanic membranes are damaged by blast exposure

To assess the possible damaging effects of blast exposure on the tympanic membrane (TM), we performed otoscopy on control and experimental mice (Fig. 2). TMs of both single and triple blast-exposed mice showed signs of damage at all time points examined (1 day, 3 days, 14 days, and 30 days post-exposure). TM injury was particularly pronounced one and three days after blast with 100% of blast-exposed ears having perforated TMs at these time points. By one month, TMs appeared to have healed in all blast-exposed ears, although scarring was always present (Fig. 2). TM damage and healing did not differ in any obvious way between single and triple blasted mice.

Figure 2. Blast exposure causes tympanic membrane (TM) rupture that typically heals by 30 days post-exposure.

In control ears, intact TMs were pearly gray and translucent. A light reflex was visible where the TM reflected the light source (arrowheads, left upper and bottom panels). Light reflexes were not observed in single- or triple-blasted ears 1 day after exposure; instead, these TMs were ruptured, exhibiting torn edges (arrows, middle and right panels of top row). At the 30 day recovery point, light reflexes were present in blast-exposed specimens (arrowheads), but TMs displayed thickening or scarring (asterisks, middle and right panels of bottom row).

3.4. Variable outer hair cell loss following blast exposure

To determine whether blast exposure causes cochlear hair cell loss, the number of hair cells was assessed in different regions of the cochlea at specific recovery time points. Hair cells were labeled with anti-MYO7A and counterstained with phalloidin. Control cochleae retained a normal complement of both inner hair cells (IHCs) and outer hair cells (OHCs) at all time points (Fig. 3a). Similarly, IHCs appeared normal in cochleae from single- and triple-blast mice at all time points and in all cochlear regions. In contrast, OHC loss was evident in the basal and middle turns of triple-blast mice (Figure 3b), which had significantly fewer OHCs than those of control or single-blast mice (Table 7).

Figure 3. Variable OHC loss occurs following blast exposure.

a. Representative confocal images of whole-mount cochlear turns 30 days after blast exposure. MYO7A (green) labels hair cell bodies, and phalloidin (magenta) labels actin in hair and supporting cells. A normal complement of IHCs and OHCs is present the apical and mid cochlear regions for control, single- and triple-blasted animals. In contrast, OHCs are clearly missing in the basal turn of a triple-blasted cochlea. b. Quantification of IHCs (top) and OHCs (bottom) by cochlear turn and condition. Boxplot whiskers show standard error of the mean. Significant decreases in OHCs were observed for both the middle and basal turns following triple blasts. c. Same data as shown in b, plotted to show differences across time points. Both single and triple-blast basal OHC counts were significantly different from those of controls at every time point tested. Significant differences between triple- and single-blast basal OHC counts were present in some, but not all, time points. In addition, at the 7-day time point significant differences were present between triple-blast and single-blast OHCs in the apical turn of the cochlea and between control and triple-blast OHCs in the middle turn. Sample sizes ranged from 3 to 15 ears and averaged 7.5; see Table 3 for specific numbers. *p≤0.05, **p≤0.01, ***p≤0.001.

OHC loss of >50% was observed in triple-blasted mice as early as one day post-exposure, and this loss was maintained at all subsequent time points (Fig. 3b, c). When compared within time points, significant differences in OHC loss were found between basal OHCs of triple-blast and control, and triple-blast and single-blast, mice at every time point (Table 8). Additionally, basal OHCs differed significantly between single-blast and control mice at a subset of time points. At the 7-day time point, middle turn OHCs of triple-blast and control mice differed significantly, as did apical OHCs between triple- and single-blast animals. However, no significant differences in apical or middle turn OHCs were found at any other time points, or among IHCs in any cochlear region (Table 8).

3.5. Blast-exposure does not lead to changes in the number of spiral ganglion neurons, but IHC synaptic ribbons are lost in the upper turns of blast-exposed mice

Although ABR data indicated significant and permanent threshold shifts in blast-exposed animals relative to control animals across all frequencies (Fig. 1a; Table 5), the numbers of IHCs were unchanged across time points in all regions of both single- and triple-blast ears. To determine whether some aspects of the changes in ABR at mid- to low frequencies might be a result of loss of spiral ganglion neurons (SGNs), the number of SGN cell bodies were counted in each cochlear turn of mid-modiolar cross sections of control and triple blast-exposed animals at the 30-day time point (Fig. 4a). Results indicated no significant difference in the total number of SGN cell bodies between control and triple blast-exposed samples (Fig. 4b, Table 12), nor in the density of cell bodies per unit area (Fig. 4c).

Figure 4. The number of spiral ganglion neurons is unchanged 30 days after triple-blast exposure.

a. Representative example of a mid-modiolar section showing spiral ganglion neuron cell bodies (SGNs) within Rosenthal’s canal (RC). The SGNs of the middle cochlear turn are boxed in red and shown in higher magnification in the inset. b. SGNs per section for each cochlear turn. c. SGN density per 10,000 μm² area for each cochlear turn. Boxplot whiskers show standard error of the mean. No statistically significant differences in SGN number of density were found between control and triple-blasted samples.

Next, we assessed whether pre-synaptic components of IHC synapses were affected following blast exposure by labeling pre-synaptic ribbons with anti-CTBP2 antibodies (Fig. 5a). Average numbers of pre-synaptic puncta per hair cell were then determined for cells located in basal, middle and apical turns. Two-way ANOVAs testing the effect of condition, time point, and their interaction effect revealed that the number of pre-synaptic puncta was significantly decreased for IHCs in the middle cochlear region for both single- and triple-blasted mice and in the apical region for triple-blasted mice (Fig. 5b; Table 9).

Figure 5. The number of inner hair cell (IHC) presynaptic puncta is reduced in blast-exposed mice.

a. Labeling of CTBP2 presynaptic puncta in IHC synapses in the basal (top), middle (middle), and apical turns (bottom) of control (left), single- (middle) and triple-blasted ears (right) at the 30 day time point. b. Average number of pre-synaptic puncta from IHCs of control and blast-exposed mice at the indicated time points post-exposure. The number of presynaptic puncta was significantly decreased in the middle region for mice from both blast groups and in the apex for the triple-blast mice. A similar decrease in the average number of puncta per IHC was observed in the basal turn but this result did not reach significance. Boxplot whiskers show standard error of the mean. *p≤0.05, **p≤0.01, ***p≤0.001.

3.6. Blast-exposure does not induce changes in vestibular function or number of utricular hair cells

Blast-treated mice did not exhibit obvious behavioral signs of vestibular dysfunction, such as circling or head-bobbing, at any time after exposure. Consistent with this result, vestibular evoked potentials indicated no statistically significant differences in P1 wave amplitude, latency, or threshold across time points or conditions (Fig. 6; Table 10).

Figure 6. Vestibular function is normal following blast exposure.

Assessment of vestibular evoked potentials (threshold, latency of the P1 wave and P1-N1 amplitude) from control, single- and triple-blasted mice. No significant differences between conditions or recovery time points were identified. Boxplot whiskers show standard error of the mean.

To determine whether blast-exposure leads to vestibular hair cell loss, utricles were dissected, fixed and immunolabeled. The utricles of blast-exposed mice appeared superficially similar to those of control animals (Fig. 7a, b), with no apparent loss of hair cells. This observation was confirmed quantitatively: the average number of hair cells per unit area did not differ significantly among conditions or time points (Fig. 7c, Table 11).

4. Discussion

4.1. Damage to the TM causes acute auditory dysfunction following blast exposure, but TM healing partially restores low-frequency hearing

Consistent with previous studies (Ewert et al., 2012; Cho et al., 2013), we observed significant ABR threshold shifts across all frequencies one day after blast exposure (Fig. 1a). Because TM rupture (Fig. 2) and high thresholds (Fig. 1) were observed in all blast-exposed animals, despite no or minimal loss of cochlear hair cells in the lower turns (Fig. 3), TM damage was likely a major contributor to the initial hearing loss we observed. Several lines of evidence suggest that spontaneous TM healing accounted for most of the recovery in auditory function observed post-blast exposure. Consistent with previous reports of poor high-frequency recovery after large TM perforations in gerbils (Cai et al., 2019; Dong et al., 2019), only at 8 kHz did ABR thresholds of single blast-exposed mice recover to baseline levels (Fig. 1), and all DPOAEs recorded from blast-exposed mice were restricted to frequencies below 16 kHz (data not shown). Additionally, full ABR threshold recovery occurred only after the 30 day time point (Fig. 1a), at which time TM healing appeared mostly complete (Fig. 2). Because DPOAEs require successful transmission of sound across the middle ear space twice, they are more affected by conductive dysfunction than ABRs (Qin et al., 2010). Future work in which DPOAEs are measured after blast exposures should take care to avoid, or account for, TM perforation or scarring, because DPOAEs are not a reliable indicator of OHC function or survival if the middle ear is compromised (Dong et al., 2019). The use of bone-conduction audiometry would provide a way for future blast studies to assess damage to the inner ear independent of middle ear dysfunction.

In a previous study, micro-computed tomography scans performed on mice exposed to 181 kPa blasts revealed no evidence of ossicular dislocation or fluid buildup in the middle ear space, suggesting that TM injury may be the main contributor to middle ear dysfunction after blast exposure. Additionally, three months after exposure, healed TMs were ~3 times thicker, and had ~3 times higher peak density, than control TMs (Cho et al., 2013). This indicates that even after TM perforations have fully closed, TM function may not fully recover after highly traumatic injuries such as blast exposure. Indeed, that we observe permanent ABR threshold shifts, even at frequencies corresponding to areas that exhibited no hair cell loss, suggests that the component of blast-induced hearing loss caused by TM damage can be significant and long-lasting. However, we cannot rule out the possibility that hair cell dysfunction, independent of HC death, may also contribute to the ABR threshold shifts that were observed. Further work is needed to elucidate the impact of thicker TMs on the efficiency of sound transfer at different frequencies. Additionally, replacing damaged or healed TMs with material that recapitulates normal TM properties may allow future studies to evaluate the separate contributions of sensorineural and conductive dysfunction to blast-induced hearing loss.

Our data on the time course of hearing recovery suggests that multiple blasts may cause greater TM damage and/or slower healing than single blasts. We began to detect DPOAEs in the ears of some single-blast animals after 30 days, while triple-blast mice did not have detectable DPOAEs until 180 days (Fig. 1b). Furthermore, relative to one day post-exposure, the ABR thresholds of single-blast mice were significantly reduced at both 8 kHz and 16 kHz by the 30-day time point, while those of triple-blast mice were significantly reduced only at 8 kHz over the same period (Table 6). While TMs appeared largely intact by 30 days after exposure in both single and triple blast mice (Fig. 2), our otoscopic photos were not of sufficiently high resolution to perform detailed quantitative analysis of TM healing. A recent study found that perforations of 50% or larger by area resulted in no detectable DPOAEs, and that perforations larger than 25% disproportionately affected high frequencies (Dong et al., 2019). Additionally, Cho et al. (2013) performed detailed analyses of TM perforation size and healing in mice after single blast exposures, and found that fewer than 50% of TMs were completely healed after three months of recovery time. Together with these studies, our data on hearing recovery suggest that single-blast mice sustained TM perforations that healed more quickly than those of triple-blast mice. It is also possible that the TMs of mice exposed to triple blasts were thicker and denser after healing than those of single-blast mice, but further studies employing micro-CT scans, as in Cho et al. (2013), are needed to evaluate differences in TM healing between animals exposed to single and multiple blast exposures.

4.2. Loss of OHCs contributes to long-term hearing loss at high frequencies, and may continue due to secondary effects long after blast exposure

Consistent with other studies, we found that OHCs were more susceptible to blast-induced cell death than IHCs (Ewert et al., 2012; Cho et al., 2013; Kim et al., 2018). This was not surprising, as OHCs are intrinsically more sensitive to damage, particularly those in the basal turn (Sha et al., 2001). However, the specific contribution of this extensive loss of basal OHCs to higher ABR thresholds and reduced DPOAEs could not be assessed in this study due to the aforementioned TM rupture and subsequent likely scarring and thickening, which would affect both measures of hearing function.

Previous studies which employed multiple blasts at pressures low enough to leave TMs intact have reported that OHC loss occurred across the cochlea—not just at the basal end—and that increasing numbers of blast exposures led to greater OHC loss (Ewert et al., 2012; Hickman et al., 2018). In the current study, and a previous one (Cho et al., 2013), all blast-exposed mice had ruptured TMs and sustained OHC loss primarily in the basal turn. This suggests that blast-induced TM rupture reduces the transfer of blast pressure to the inner ear, such that most damage to OHCs is confined to the basal turn immediately adjacent to the oval and round windows. Nevertheless, we did observe greater loss of OHCs in triple-blast mice than in single-blast mice, suggesting significant energy transfer to the cochlea even after TM rupture. Because damaging levels of sound were conducted to the inner ear presumably by other routes, such as bone conduction, our data suggest that in-ear hearing protection may not be sufficient to protect workers or soldiers exposed to high amplitude impulse sounds.

Finally, while significant HC loss was not observed in the mid- or apical- turns, we cannot rule out the possibility that sublethal injury to HCs may also have contributed to the loss of hearing function. Unfortunately, methods for assessing defects in HC function are limited and were not performed in this study.

4.3. Peripheral neuropathy is not a likely major contributor to hearing loss in our blast model

To determine whether a neural mechanism might also contribute to blast-induced hearing loss, we examined the number of SGNs and IHC pre-synaptic structures in control and blast-treated animals. We did not find changes in the number or density of SGN cell bodies at 30 days post-exposure (Fig. 4b), consistent with reports of SGN death occurring over an extended period following injury to IHCs or cochlear nerve fibers after noise exposure (Spoendlin, 1975; Liberman and Kiang, 1978; Kujawa and Liberman, 2006; Fernandez et al., 2015). It is possible that some SGN loss may occur at later time points, but the results clearly indicate that the hearing loss present in both single and triple-blasted mice at 30 days of recovery is not a result of decreased SGNs. A previous study did report significant loss of SGNs just 7 days after blast exposure (Cho et al., 2013), possibly due to stronger blast exposures (181 pKa/26.25 psi, compared to the 110 kPa/16 psi blasts used in this study).

A significant loss of pre-synaptic CTBP2 puncta was observed in the upper turns of single- and triple blast-exposed mice (Fig. 5b), a result that is consistent with previous studies (Cho et al., 2013; Kim et al., 2018). Given that OHC loss was limited to the basal turn, it is surprising that IHC synaptic ribbon loss was variable in this region. It is possible that loss of OHCs and, presumably, IHC stimulation, is actually protective for IHC synaptic ribbons. Notably, pure-tone audiometry is largely insensitive to neural damage. Normal thresholds have been measured in cats and humans even after removal of significant portions of the auditory nerve (Dandy, 1934; Neff, 1947; Schuknecht and Woellner, 1955; Mikaelian and Warfield, 1970), and only small changes in audiometry were observed in chinchillas that had lost ~80% of their IHCs due to carboplatin treatment (Lobarinas et al., 2013). Therefore, blast-induced synaptopathy is unlikely to be a significant factor in the elevated ABR thresholds reported here. However, the results also suggest that synaptopathy could be a long-term concern for some individuals exposed to blasts.

A limitation of the current work is that we did not evaluate the contribution of blast-induced TBI or neuropathy in the auditory CNS to the hearing loss we observed. Previous work in rats has shown that blast exposure damages the central auditory system (Mao et al., 2012; Sosa et al., 2013; Kallakuri et al., 2018), reduces the sensitivity and spontaneous discharge rates of vestibular afferents (Yu et al., 2020), and can even alter the tonotopic organization of the auditory cortex (Masri et al., 2018). Additional studies are needed to evaluate the relative contributions of peripheral and central damage to auditory and vestibular deficits caused by blast exposure.

4.4. Vestibular function was not affected by blast exposures

We observed no overt behavioral changes such as spinning, head bobbing, or hyperactivity, that would indicate vestibular dysfunction following blast exposure, consistent with other studies. Additionally, the VsEP measurements of blast-exposed mice did not significantly differ from those of control mice. We did observe that some blast-exposed animals had higher VsEP thresholds and lower P1 amplitudes than controls (Fig. 6). Similar VsEP results have reported in mouse mutants with otoconial abnormalities, even in heterozygous animals (Jones et al., 2004). This raises the interesting possibility, which has not previously been examined to our knowledge, that blast exposure might displace the otoconia of the vestibular organs. Otoconial displacement frequently occurs after head injuries in humans, resulting in benign paroxysmal positional vertigo, and has been documented in multiple studies of blast-exposed veterans (Fausti et al., 2009; Akin et al., 2017). Mouse mutants with otoconial abnormalities perform poorly on VOR and behavioral vestibular tests (Harrod and Baker, 2003; Jones et al., 2004; Dror et al., 2020). Future blast studies would benefit from examining otoconial displacement to account for its possible role in measures of vestibular function.

Several groups have reported vestibulomotor deficits in blast-exposed rodents using rotarod or righting reflex tests. These tests require behavioral responses that involve other sensory systems, like vision and proprioception, and depend on motor function. Performance deficits in these tasks therefore cannot be ascribed specifically to vestibular dysfunction, and measurements made after the acute-recovery period may be confounded by habituation or central compensation (Lien and Dickman, 2018; Studlack et al., 2018; Arun et al., 2020). Strikingly, Jones and colleagues reported that some mutant mice with no detectable VsEPs nevertheless performed adequately on a swimming test, and suggested that behavioral tests alone may not reliably indicate vestibular deficits (Jones et al., 2004). We employed VsEPs to avoid these complications, as they directly measure evoked responses following stimulation of utricular HCs.

Consistent with our functional results, no significant difference in utricular HC density was observed between blast-exposed and control animals at any time point (Fig. 7c). Other studies have reported loss of utricular stereocilia bundles following blast exposure (Lien and Dickman, 2018; Yu et al., 2020), and while we did not specifically examine damage to stereocilia bundles, we routinely labeled specimens with phalloidin and saw no obvious stereocilia loss in the utricles of blast-exposed animals (Fig. 7d). Blast-induced otoconial displacement, as mentioned above, could provide a mechanistic explanation for vestibular bundle damage.

One final consideration is that, unlike the cochlea, the mature utricle possesses a limited capacity to regenerate HCs by transdifferentiation of supporting cells (SCs), although these HCs are morphologically and physiologically immature (Lin et al., 2011; Bucks et al., 2017; Sayyid et al., 2019). Therefore, it is possible that blast-exposure could have induced a mild loss of hair cells that were subsequently replaced through regeneration. To address this hypothesis, future studies examining the effect of blast exposure on the vestibular periphery could consider using a transgenic mouse line in which SCs are labeled prior to exposure, so that transdifferentiated SCs could be distinguished from endogenous HCs.

5. Conclusion

Overall, our results demonstrate that blast exposure causes long-term hearing loss with both sensorineural and conductive components. Unfortunately, because TM rupture is inevitable at pressures that recapitulate blast injury to the inner ear, and leads to loss of middle ear function that affects the hearing assessments we employed, the relative contributions of OHC loss and TM rupture could not be separately evaluated in detail. However, our data show that damaging levels of sound are still transferred to the inner ear after TM rupture, as animals subjected to triple blasts lost >50% of their basal OHCs, while single-blast animals retained most of their OHCs. Moderate levels of pre-synaptic ribbon loss was observed in the apical and middle turns of blast-exposed mice, but we did not find evidence of SGN loss 30 days after blast exposure. Our analysis did not reveal damage to the vestibular periphery, nor did we find significant differences in VsEP characteristics between blast-exposed and control animals. Because balance is a multi-modal sense that is subject to central compensation, future research on long-term effects of blast exposure should employ direct assessments of vestibular function after blast exposure, such as VsEPs.

Highlights.

• Blast exposure causes hearing loss with sensorineural and conductive components

• Multiple blast exposures caused greater OHC loss and slower hearing recovery

• Cochlear outer hair cells sustain damage even after tympanic membrane rupture

• Blast-exposed mice sustained moderate loss of pre-synaptic ribbons

• Vestibular function and histology was not affected at the blast levels employed