Protective Effects of Molecular Hydrogen on Noise-Induced Auditory and Cognitive Dysfunction

Abstract

Objective:

This study aimed to evaluate the protective efficacy and molecular mechanisms of hydrogen (H₂) against noise-induced damage across the peripheral and central auditory systems.

Materials and Methods:

Mice were exposed to noise with or without H₂ treatment. Auditory function was assessed using the auditory brainstem response (ABR) and distortion product otoacoustic emission (DPOAE). Cochlear morphological changes were examined using a scanning electron microscope. Central effects in the inferior colliculus (IC) were evaluated using hematoxylin and eosin staining, transmission electron microscopy, and proteomic analysis. Behavioral changes were measured by using prepulse inhibition (PPI) and shuttle-box tasks.

Results:

H₂ significantly attenuated noise-induced ABR threshold shifts (P < 0.05) and preserved wave I amplitudes at 80 dB (2.69 ± 0.40 vs. 3.57 ± 0.15 μV, P < 0.05). H₂ attenuated noise-induced reduction of DPOAE (16 kHz: 34.40 ± 2.70 vs. 24.67 ± 4.35, P < 0.05) and mitigated outer hair cell (OHC) loss (survival: 89.92% ± 1.23% vs. 82.10% ± 1.6%, all P < 0.05). In the IC, H₂ reduced neuronal pyknosis by 12.6% (P < 0.05) and preserved the synaptic ultrastructure. Behaviorally, H₂ suppressed the noise-induced elevation of PPI at 75 and 80 dB (75 dB: 40.87% ± 6.32% vs. 58.04% ± 2.47%; 80 dB: 50.01% ± 3.14% vs. 58.04% ± 2.47%, all P < 0.05) and improved performance on the shuttle-box test. Proteomic analysis revealed that H₂ upregulated nicotinamide adenine dinucleotide phosphate hydrogen regeneration-related proteins (e.g., Tigar) and GABAergic synaptic proteins (e.g., Gad1) in the IC.

Conclusion:

H₂ exerts a comprehensive neuroprotective effect against noise-induced damage, spanning from OHCs to central synaptic remodeling. This cognitive protection may be mediated by enhancing the antioxidant capacity and restoring the excitation–inhibition balance in the IC.

KEY MESSAGES

• (1)H₂ inhalation provides comprehensive protection against noise-induced hearing loss by safeguarding both peripheral cochlear hair cells/synapses (wave I) and the structural integrity of the inferior colliculus (IC).
• (2)Beyond direct reactive oxygen species scavenging, H₂ bolsters the endogenous antioxidant system by reinforcing “supply lines” of nicotinamide adenine dinucleotide phosphate hydrogen regeneration and nicotinate/nicotinamide metabolism.
• (3)H₂ bolsters antioxidant defenses to safeguard IC GABAergic neurons, restoring E/I balance and mitigating noise-induced hyperacusis and cognitive decline.

INTRODUCTION

Noise-induced hearing loss (NIHL) is a pervasive global health challenge that affects millions of individuals and accounts for approximately 16% of disabling hearing loss among the global workforce.[1] Beyond elevated auditory thresholds, NIHL is a complex ear-to-brain disorder characterized by peripheral deafferentation and subsequent maladaptive neuroplasticity throughout the central auditory system.[2] These central alterations often manifest as distressing auditory symptoms (such as tinnitus and hyperacusis)[3] and nonauditory sequelae, including emotional instability and cognitive decline,[4,5] which severely compromise quality of life.

As the primary midbrain hub for integrating ascending auditory information and processing complex sound features, the inferior colliculus (IC) is highly susceptible to peripheral deafferentation.[6] Acoustic trauma disrupts the excitation–inhibition (E/I) balance of the IC and triggers “central gain”—a compensatory but maladaptive increase in neuronal excitability following cochlear deafferentation.[2,7] Hyperexcitability in the IC is considered a fundamental neurobiological substrate for hyperacusis, a condition often assessed behaviorally via prepulse inhibition (PPI).[8] Serving as a multimodal integration hub, the IC receives descending projections from the auditory, visual, and motor cortices and the amygdala and projects extensive outputs to the superior colliculus (SC), posterior intralaminar nucleus, and peripeduncular nucleus, which mediate complex behaviors, including auditory fear conditioning and sound-localization learning.[9,10,11,12,13] Consequently, rescuing IC damage following noise trauma is crucial for treating nonauditory sequelae associated with NIHL.

The molecular pathogenesis of noise-induced central damage is multifactorial and involves oxidative stress, neuroinflammation, and synaptic remodeling.[14] Given this complexity, conventional single-target pharmacological interventions often have limited efficacy. There is an urgent need for multitarget therapeutic agents that can simultaneously cross the blood–brain barrier and exert neuroprotective effects across both the peripheral and central compartments. Molecular hydrogen (H₂), a novel medical gas, has gained significant attention due to its ability to rapidly diffuse across biological membranes and the blood–brain barrier.[15] Unlike conventional antioxidants, H₂ selectively scavenges the most cytotoxic reactive oxygen species (ROS) while preserving essential physiological redox signaling.[16] It also exerts potent anti-inflammatory effects by modulating various intracellular signaling pathways.[17] While the protective effects of H₂ on the cochlea have been documented,[18] its potential in mitigating noise-induced central damage—specifically IC neuronal injury, synaptic maladaptation, and associated cognitive-behavioral deficits—remains largely unexplored.

To test the hypothesis that H₂ administration represents a multifaceted intervention against noise-induced auditory and cognitive deficits, we systematically evaluated its effects on peripheral and central auditory functions. By integrating multilevel assessments—from auditory and behavioral changes to molecular alterations in the IC—we sought to elucidate the underlying mechanisms. Our findings reveal that H₂ preserves synaptic transmission related to the E/I balance and enhances nicotinamide adenine dinucleotide phosphate hydrogen (NADPH)-related antioxidant metabolism and anti-inflammatory capacity. These insights provide a novel therapeutic perspective for the holistic management of NIHL.

MATERIALS AND METHODS

Animals

All animal procedures strictly followed the National Institutes of Health Guide for the Care and Use of Laboratory Animals,[19] and the study was approved by the Ethics Committee of the Chinese PLA General Hospital (No. S2020-465-01). Eight-week-old male C57BL/6J mice [n = 72; GemPharmatech Co., Nanjing, China; Batch number: SCXK (Beijing) 2023-0008] were used. They were acclimatized to a specific pathogen-free environment maintained at a controlled temperature of 22 ± 2°C and humidity of 55 ± 5% under a standard 12-hour light/dark cycle. Mice were provided ad libitum access to sterile food and water. For the experimental phase, the mice were randomly allocated to three distinct groups: control (Con, n = 20), noise exposure (NE, n = 26), and hydrogen-treated NE (H₂ + NE, n = 26) groups.

The specific allocation and experimental timeline for each group were as follows.

Hearing Assessment

Six mice per group were designated for longitudinal auditory brainstem response (ABR) monitoring (from day 1 to 28 postexposure) and distortion product otoacoustic emission (DPOAE) recording at 28 days postexposure (dpe). Following DPOAE assessment, three mice from this subcohort were randomly selected for scanning electron microscopy (SEM) of the basilar membrane.

Behavioral and Morphological Analysis

Six mice per group were subjected to the shuttle-box test (from day 1 to 7 postexposure). On day 7, three mice were randomly selected for hematoxylin and eosin (H&E) staining after the shuttle-box test, while the remaining three were selected for transmission electron microscopy (TEM). Six mice per group were allocated for PPI testing at 28 dpe.

Molecular Analysis

Two mice per group underwent western blot (WB) analysis on day 7 postexposure. Additionally, six mice from each of the NE and H₂ + NE groups were designated for proteomic analysis at 7 dpe.

Noise Exposure Paradigm

A custom-built acoustic exposure system was used in this study. To model the multifrequency spectral profile of occupational acoustic trauma, mice in the NE and H₂ + NE groups were exposed to field-recorded heavy-duty truck noise at an intensity of 100 dB sound pressure level (SPL) in a sound-attenuating environment for 2 h. This paradigm was selected to induce a robust yet reversible threshold shift, which was consistent with established models of noise-induced hidden hearing loss (HHL).[20] The noise signal was amplified and delivered via loudspeakers, with sound intensity calibrated at the level of the ears of the animals using a 1/4-inch condenser microphone (Brüel & Kjær, Denmark). Control mice were placed in the same chamber for an equivalent duration without NE to prevent the confounding effects of environmental changes, transportation, and restraint on the behavioral outcomes.[21]

H₂ Intervention

To simulate pharmacological intervention against inevitable high-intensity noise, a prophylactic and therapeutic H₂ regimen was implemented using a generator with a flashback arrestor (FC4500C03, CONSIDER Medical Technology, China). Mice in the H₂ + NE group inhaled hydrogen-rich gas (66.7% H₂/33.3% O₂) for 2 h daily, starting 7 days before and continuing for 7 dpe.

To ensure an intrinsically safe and stable gas environment, the procedure was conducted in a grounded, high-performance fume hood (face velocity: 0.6 m/s). The inhalation chamber (15 cm × 15 cm × 20 cm, 4.5 L), constructed from medical-grade anti-static acrylic and free of any internal electrical components, was rigorously grounded to eliminate potential ignition sources and supplied with a continuous flow of the gas mixture at 4.5 L/min (60 air changes per hour), ensuring that the gas mixture rapidly reached a steady state while preventing CO₂ accumulation and maintaining pressure equilibrium. Following an initial 10-min equilibration period, the H₂ concentration was periodically verified at the exhaust port using a thermal conductivity gas detector (XP-3140, New Cosmos Electric Co., Ltd., Osaka, Japan) and consistently confirmed to be a stable concentration of 66.7%. The ambient hydrogen concentration was continuously monitored using a dedicated detector (XP-3110, New Cosmos Electric Co., Ltd., Osaka, Japan) to ensure that it remained well below the lower explosive limit. To control for confinement-induced stress, the NE and Con groups were subjected to sham inhalation (ambient air) in identical chambers for the same duration.

Auditory Brainstem Response Recording

The mice were anesthetized with sodium pentobarbital (Sigma-Aldrich, St. Louis, MO, USA; 60 mg/kg, i.p.) and maintained on heating pads. Subdermal electrodes were placed on the vertex (active), ipsilateral pinna (reference), and tail base (ground). Broadband clicks (0.1 ms, 21.1/s) were delivered using a TDT System III workstation (Tucker-Davis Technologies, Alachua, FL, USA). Click-evoked ABR thresholds were tracked longitudinally at baseline; 0 h; and 3, 7, 14, 21, and 28 dpe (n = 6 mice per group/point). The ABR threshold was set at the lowest intensity (in 5-dB decrements from 90 dB SPL) that elicited a reproducible wave I. To specifically evaluate acute peripheral auditory nerve function, wave I characteristics (peak-to-peak amplitude and latency) were analyzed at a suprathreshold intensity (80 dB SPL) exclusively at the 0 h time point.

Distortion Product Otoacoustic Emissions Recording

DPOAE was recorded at 28 dpe (Con, NE, and H₂ + NE groups) to assess outer hair cell (OHC) function using the SmartDPOAE system (Intelligent Hearing Systems, Miami, FL, USA). The measurements were performed in a sound-attenuating chamber using a specialized acoustic probe. Two primary tones (f₁ and f₂) were presented with a frequency ratio f₂/f₁ = 1.2 and intensities of L₁ = 65 dB SPL and L₂ = 55 dB SPL. The 2 f₁-f₂ distortion product was recorded over a frequency range of 345 to 22595 Hz. The signal-to-noise ratio (SNR) was calculated by subtracting the noise floor from the DPOAE amplitude. SNR data of the Con, NE, and H₂ + NE groups were analyzed to evaluate the protective effect of hydrogen against truck noise-induced damage.

Scanning Electron Microscopy of the Cochlear Basilar Membrane

The cochlear basilar membranes were harvested to evaluate permanent structural damage using SEM at 28 dpe. The organs of Corti were microdissected after perilymphatic perfusion (2.5% glutaraldehyde) and ethylenediaminetetraacetic acid (EDTA) decalcification. The samples were subjected to graded ethanol dehydration, critical point drying, and gold–palladium sputtering. OHC survival and stereociliary integrity in the basal turns were examined using a Hitachi S-4800 SEM (Hitachi, Tokyo, Japan). This analysis aimed to quantify the long-term structural rescue provided by H₂ intervention.

Histopathological Evaluation of the Inferior Colliculus

The mice were anesthetized and transcardially perfused with 4% paraformaldehyde (Sinopharm Chemical Reagent Co., Ltd., Beijing, China) at 7 days after exposure. The brains were harvested, post-fixed for 24 h, and embedded in paraffin. Coronal sections (3 µm thick) containing the IC were obtained using a microtome, with anatomical localization guided by the Paxinos and Franklin mouse brain atlas. IC sections were stained with H&E following standard protocols. Pathological changes, including nuclear pyknosis and neuronal shrinkage, were evaluated using an Olympus BX53 light microscope (Olympus Corporation, Tokyo, Japan) to determine the neuroprotective effects of H₂ in the central auditory system.

Transmission Electron Microscopy of the Inferior Colliculus

The mice were transcardially perfused with 2.5% glutaraldehyde (Sinopharm Chemical Reagent Co., Ltd., Beijing, China) and 2% paraformaldehyde in 0.1 M PB (pH 7.4, phosphate buffer, Solarbio, Beijing, China) at 7 days after exposure. The IC was micro-dissected into 1 mm³ blocks and post-fixed overnight at 4°C. Following postfixation in 1% osmium tetroxide (Sinopharm Chemical Reagent Co., Ltd., Shanghai, China) (2 h) and embedding in Epon 812 (Sigma-Aldrich, St. Louis, MO, USA), the IC region was localized using toluidine blue-stained semithin sections. Ultrathin sections (80 nm) were double-stained with uranyl acetate and lead citrate (Sigma-Aldrich, St. Louis, MO, USA) and examined via TEM (Hitachi HT7800, Tokyo, Japan). The evaluation focused on neuronal morphology and synaptic ultrastructure.

Prepulse Inhibition Assessment

PPI was evaluated using a startle reflex system (Med Associates Inc., St. Albans, VT, USA) under a constant 65 dB white noise background at 28 dpe. Following a 5-min acclimation period, the session began with Block I, consisting of 10 consecutive startle pulses (105 dB, 20 ms; random interstimulus interval of 10–30 s) to stabilize the initial startle response. Block II comprised 80 trials presented in 10 pseudorandomized sequences, including: (1) pulse-alone trials (105 dB, 20 ms); (2) prepulse-alone trials (70, 75, and 80 dB, 10 ms each); (3) prepulse–pulse pairs (prepulse followed by a 100 ms lead interval before the pulse); and (4) null trials (background noise only). The test chamber was cleaned thoroughly with 75% ethanol to eliminate olfactory cues. The PPI for each prepulse intensity was calculated using the following formula[22]:

PPI (%) = [(Pulse-alone amplitude − Prepulse–Pulse amplitude)/Pulse-alone amplitude] × 100%.

Shuttle-Box Test

To evaluate the associative learning functions, an active avoidance task was performed using a computerized shuttle-box system (Beijing Zhishuduobao, China). This 7-day experiment began 2 h after NE. The apparatus consisted of two interconnected compartments equipped with a grid floor for delivering foot shocks (unconditioned stimulus, US) and a speaker for delivering a 4 kHz, 100 dB pure tone (conditioned stimulus, CS). The experimental paradigm consisted of a 7-day training session, with 50 trials per day for each mouse. Each trial began with a 5-s CS. If the mouse crossed to the opposite compartment during the CS, an “active avoidance” response was recorded and the trial was terminated. If the mouse failed to respond within the CS period, the US was applied for a maximum of 10 s, or until it escaped to the adjacent compartment. The following metrics were analyzed: (1) mean active avoidance rate (AAR), (2) shock duration, and (3) active avoidance latency (AAL).

Proteomic Analysis of the Inferior Colliculus

IC tissues from the Con, NE, and H₂+NE groups were harvested on ice and stored at −80°C at 7 dpe. For each group, three biological replicates were prepared, each consisting of pooled IC tissues from two mice. Total protein was extracted using radioimmunoprecipitation assay buffer containing protease inhibitors and quantified using a bicinchoninic acid assay. Proteins were digested using filter-aided sample preparation protocol. The resulting peptides were analyzed via liquid chromatography–tandem mass spectrometry (LC–MS/MS) using an Orbitrap mass spectrometer (Thermo Fisher Scientific, Waltham, MA, USA). Protein identification and quantification were performed using the Mus musculus database. Differentially expressed proteins (DEPs) were defined by a fold change (FC) of >1.2 or <0.83 with P < 0.05. Functional enrichment was analyzed using Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways.

Western Blot Analysis

Bilateral IC tissues were harvested and pooled from two mice per sample to ensure a sufficient protein concentration at 7 days post-NE. Total protein was extracted, separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis, and transferred onto polyvinylidene fluoride (PVDF) membranes. After blocking with 5% non-fat milk for 1 h at room temperature, membranes were incubated overnight at 4°C with primary antibodies against Aldh1l2 (1:1000, Proteintech, 21391-1-AP, Rosemont, IL, USA), Rbks (1:1000, Proteintech, 31590-1-AP, Rosemont, IL, USA), Tigar (1:1000, SANTA, SC-166290, Dallas, TX, USA), H6pd (1:1000, FineTest, FNab03741, Wuhan, China), ALDOB (1:1000, Abnova, H00000229-D01P, Taiwan, China), and β-actin (1:3000, Servicebio, GB15003, Wuhan, China), followed by horseradish peroxidase-conjugated secondary antibodies. Protein bands were visualized using enhanced chemiluminescence (ECL) chemiluminescence and captured using a ChemiDoc MP Imaging System (Bio-Rad, USA). Band intensities were visually inspected to confirm the expression trends observed in the proteomic analysis.

Statistical Analysis

Data are presented as mean ± standard error of the mean and analyzed using SPSS 27.0 (IBM Corp., Armonk, NY, USA). The normality of the data distribution was assessed by the Shapiro–Wilk test, and the homogeneity of variance was verified using Levene’s test. All statistical hypothesis tests were two-tailed, and a P < 0.05 was considered to indicate statistical significance. Specifically, ABR thresholds, DPOAE, and shuttle-box behavioral data were analyzed by two-way repeated-measures analysis of variance (ANOVA) followed by Tukey’s post-hoc test. Proteomic DEPs were identified using a two-tailed Student’s t-test with Benjamini–Hochberg false discovery rate (FDR) correction. Other multigroup comparisons at a single time point were performed using one-way ANOVA with Tukey’s test.

RESULTS

H₂ Treatment Alleviates Noise-Induced Peripheral Auditory Dysfunction and Cochlear Damage

To evaluate the protective effects of H₂ on hearing function and the frequency-specific integrity of OHCs, ABR thresholds, wave I characteristics, DPOAE, and SEM were analyzed.

The NE group exhibited a significant immediate ABR threshold shift compared with the Con group (53.75 ± 4.04 vs. 12.08 ± 1.14, P < 0.05) [Figure 1a], which persisted until 7 dpe and returned to baseline by 14 dpe. H₂ treatment significantly attenuated this elevation; thresholds in the H₂ + NE group were significantly lower than those in the NE group at 0, 3, and 7 dpe (32.92 ± 4.19 vs. 53.75 ± 4.04; 19.58 ± 3.45 vs. 35.42 ± 4.67; 16.67 ± 2.49 vs. 28.75 ± 3.65, all P < 0.05). Notably, thresholds in the H₂ + NE group returned to near-control levels as early as 3 dpe, indicating that H₂ treatment significantly accelerates hearing recovery following noise-induced trauma. ABR wave I was analyzed to assess cochlear nerve integrity [Figure 1b and c]. The NE group had significantly reduced wave I amplitudes relative to that of the Con group (2.69 ± 0.40 vs. 3.97 ± 0.29, P < 0.05). H₂ treatment significantly attenuated this reduction (3.57 ± 0.15 vs. 2.69 ± 0.40, P < 0.05). Regarding wave I latency, both H₂ + NE and NE groups showed significant prolongation immediately postexposure relative to the Con group (1.83 ± 0.04 vs. 1.67 ± 0.03; 1.85 ± 0.04 vs. 1.67 ± 0.03, all P < 0.05). Although H₂ treatment showed a trend toward shortening this latency, the difference between the H₂ + NE and NE groups did not reach statistical significance (1.83 ± 0.04 vs. 1.85 ± 0.04, P > 0.05).

Figure 1.

Protective effects of H₂ on noise-induced peripheral auditory damage. Note: (a) Time course of auditory brainstem response (ABR) thresholds (0–28 dpe). (b) ABR wave I amplitudes. (c) ABR wave I latencies. (d) Distortion product otoacoustic emission (DPOAE) signal-to-noise ratio (SNR) across frequencies (0.35–22.6 kHz). (e) Representative scanning electron microscopy (SEM) micrographs of the cochlear basal turn (left, scale bar = 10 µm) and quantitative analysis of OHC survival rates (right). Data are presented as mean ± SEM (n = 6 mice/12 ears for a–d; n = 3 ears/9 field of views (FOVs) for Figure 1e). Two-way repeated measures analysis of variance (ANOVA) followed by Tukey’s post-hoc test was used for a and d. One-way ANOVA followed by Tukey’s post-hoc test for b, c, and e. In (a, d): red ^*P < 0.05 (NE vs. Con); blue ^*P < 0.05 (H₂ + NE vs. Con); ^#P < 0.05 (H₂ + NE vs. NE). In (b, c, e): ^*P < 0.05. Con = control, H₂ + NE = H₂ + Noise exposure, NE = noise exposure, ns = not significant, OHC = outer hair cell.

For the NE group, a significant reduction in the SNR was observed at 4004, 7983, 15,979, and 22,595 Hz relative to that of the Con group (10.25 ± 1.61 vs. 20.67 ± 2.02; 23.08 ± 2.92 vs. 33.08 ± 1.65; 24.67 ± 4.35 vs. 39.17 ± 1.77; 8.58 ± 3.71 vs. 22.58 ± 5.67, all P < 0.05), indicating persistent OHC dysfunction at these frequencies despite ABR threshold recovery. For the H₂ + NE group, SNR reduction was localized to 4004, 5640, and 7983 Hz (12.60 ± 1.73 vs. 20.67 ± 2.02; 9.30 ± 3.02 vs. 18.08 ± 2.38; 18.30 ± 3.57 vs. 33.08 ± 1.65, all P < 0.05). Notably, at the high frequency of 15,979 Hz, the SNR of the H₂ + NE group was significantly higher than that of the NE group (34.40 ± 2.70 vs. 24.67 ± 4.35, P < 0.05), suggesting that H₂ treatment specifically attenuated noise-induced damage in this high-frequency region. Furthermore, H₂-treated mice had a narrower range of affected frequencies (three frequencies) than the NE group (four frequencies), demonstrating the protective effect of H₂ against noise-induced subtle OHC impairment [Figure 1d].

Given that the DPOAE results indicated persistent high-frequency impairment, the basal turn of the cochlea was examined using SEM, and the survival rate of the OHCs was determined [Figure 1e]. The Con group displayed highly organized OHCs with intact, V-shaped stereocilia bundles. Sporadic OHC loss was observed for both the NE and H₂ + NE groups (indicated by red asterisks). However, the NE group had more severe damage, a lower OHC survival rate than the Con group (82.10% ± 1.66% vs. 100.00% ± 0.00%, P < 0.05), and pronounced stereocilia lodging (disarray and collapse). In comparison, the H₂-treated mice showed markedly less stereocilia lodging, and the survival rate of OHCs was significantly higher than that of the NE group (89.92% ± 1.23% vs. 82.10% ± 1.66%, P < 0.05). These results suggest that H₂ treatment mitigates permanent noise-induced structural damage to OHCs.

H₂ Mitigates Noise-Induced Histopathological Damage in the Inferior Colliculus

To evaluate the impact of NE and neuroprotective potential of H₂ in the IC, H&E staining, and TEM were performed [Figure 2].

Figure 2.

H₂ mitigates noise-induced neuronal damage and ultrastructural alterations in the inferior colliculus (IC). Note: (a) Representative hematoxylin and eosin (H&E)-stained sections of the IC in the Con, H₂ + NE, and NE groups [scale bar = 200 µm (main images) and 50 µm (magnified). Yellow arrows: pyknotic neurons]. (b) Quantitative analysis of pyknotic neurons in the IC of the control, H₂ + NE, and NE groups. (c) Representative TEM micrographs of IC neurons (scale bar = 2.0 µm; yellow arrows indicate chromatin margination). (d) Representative TEM images of synapses in the IC (scale bar = 500 nm). Red asterisks: presynaptic vesicle accumulation; blue arrowheads: increased electron density in the synaptic cleft). Data are mean ± SEM (n = 3 mice/9 FOVs for Figure 2b), one-way analysis of variance (ANOVA) followed by Tukey’s post-hoc test for 2b, ^*P < 0.05. Con = control, H₂ + NE, H₂ + noise exposure, NE = noise exposure.

Histological analysis of the IC revealed the typical neuronal architecture in the Con group. NE-induced neurotoxic features, including intensely stained, pyknotic nuclei in the H₂ + NE and NE groups [Figure 2a]; however, the number of damaged neurons in the H₂ + NE group was reduced relative to that in the NE group (17.89 ± 0.77 vs. 30.44 ± 1.54, P < 0.05) [Figure 2b].

TEM of the IC revealed pathological alterations in the NE and H₂ + NE groups at 7 dpe, although the Con group maintained normal neuronal and synaptic morphology [Figure 2c and d]. These changes included neuronal chromatin margination and disruption of synaptic double-membrane structures. The NE group demonstrated more severe degenerative features than the H₂ + NE group, which were characterized by excessive presynaptic vesicle accumulation and increased electron density within the synaptic cleft. These findings suggest that H₂ intervention may mitigate noise-induced damage to neurons and synapses in the IC.

H₂ Intervention Attenuates Noise-Induced Compensatory Elevation of Prepulse Inhibition

Previous studies have established that noise-induced peripheral deafferentation can trigger a compensatory increase in central auditory excitability, manifesting behaviorally as an increase in PPI. PPI was assessed to evaluate the neuroprotective effects of H₂ on noise-induced central dysfunction [Figure 3].

Figure 3.

H₂ intervention attenuates noise-induced compensatory elevation of prepulse inhibition (PPI). Note: PPI levels at prepulse intensities of (a) 70 dB, (b) 75 dB, and (c) 80 dB. Data are presented as mean ± SEM (n = 6 mice per group). One-way analysis of variance (ANOVA) followed by Tukey’s post-hoc test was used for analysis. ^* P < 0.05. Con = control, H₂ + NE = H₂ + noise exposure, NE = noise exposure, ns = not significant, PPI = prepulse inhibition.

Our results demonstrated that the NE group had a significant increase in PPI at prepulse intensities of 75 and 80 dB relative to the Con group (58.04% ± 2.47% vs. 39.48% ± 2.07%; 57.98% ± 1.59% vs. 44.55% ± 2.11%, P < 0.05) [Figure 3b and c], indicating a state of central hyperexcitability. For the H₂ + NE group, an upward trend in PPI was observed across 70, 75, and 80 dB intensities relative to the controls [Figures 3a–c). At higher prepulse intensities (75 and 80 dB), the H₂ + NE group showed significantly lower PPI levels than the NE group (40.87% ± 6.32% vs. 58.04% ± 2.47%; 50.01% ± 3.14% vs. 58.04% ± 2.47%, P < 0.05) [Figure 3b and c]. This reduction suggests that H₂ treatment effectively suppresses the maladaptive elevation in PPI. These behavioral findings, which are consistent with the observed ultrastructural preservation in the IC, indicate that hydrogen intervention partially rescues noise-induced central hyperexcitability and exerts a protective effect on the IC.

H₂ Alleviates Noise-Induced Cognitive Impairment in Associative Learning

Given that NIHL and subsequent central nervous system damage are often associated with cognitive decline, a shuttle-box task was employed to evaluate the impact of H₂ intervention on associative learning abilities [Figure 4].

Figure 4.

Effects of hydrogen intervention on associative learning of noise-exposed mice during the shuttle-box task. Note: (a) Active avoidance rate across the 7-day training period. (b) Total shock duration across training period. (c) Active avoidance latency across the training period. Data are presented as mean ± scanning electron microscopy (SEM) (n = 6 mice per group). Two-way repeated measures analysis of variance (ANOVA) followed by Tukey’s post-hoc test was used for analysis. Red ^* P < 0.05 (NE vs. Con); blue ^* P < 0.05 (H₂ + NE vs. Con); ^# P < 0.05 (H₂ + NE vs. NE). Con = control, H₂ + NE, H₂ + noise exposure, NE = noise exposure.

NE resulted in profound impairment of associative learning. During the 1‑ to 7‑day training period, a persistent and significant deficit in the acquisition of avoidance responses was observed in the NE group; the AAR was reduced relative to that of the Con group (1.00 ± 0.68 vs. 26.33 ± 5.04; 3.67 ± 1.50 vs. 47.33 ± 11.12; 9.00 ± 2.37 vs. 70.00 ± 4.84; 11.67 ± 3.28 vs. 72.67 ± 4.67; 26.67 ± 5.00 vs. 75.00 ± 6.92; 26.67 ± 5.00 vs. 75.00 ± 6.92; 33.00 ± 6.23 vs. 89.67 ± 3.40; 39.33 ± 3.17 vs. 94.00 ± 2.63, all P < 0.05; [Figure 4]a). While the H₂ + NE group still had lower AAR levels than the NE group, H₂ intervention significantly attenuated the decrease in AAR on training days 3, 4, and 7 (20.00 ± 2.67 vs. 9.00 ± 2.37; 28.33 ± 4.94 vs. 11.67 ± 3.28; 51.33 ± 2.86 vs. 39.33 ± 3.17, all P < 0.05). The efficiency of escape behavior was similarly compromised for the NE group, as evidenced by a significant prolongation of total shock duration from days 3 to 7 (150.27 ± 22.84 vs. 20.00 ± 3.08; 123.74 ± 15.09 vs. 18.30 ± 3.86; 94.41 ± 7.46 vs. 18.59 ± 9.76; 87.18 ± 9.32 vs. 5.33 ± 2.28; 88.96 ± 6.07 vs. 8.02 ± 3.75, all P < 0.05 vs. Con group; [Figure 4]b). Hydrogen treatment effectively shortened the shock duration on days 3, 6, and 7 compared to the NE group (91.49 ± 6.80 vs. 150.27 ± 22.84; 64.95 ± 8.05 vs. 87.18 ± 9.32; 66.50 ± 2.75 vs. 88.96 ± 6.07, all P < 0.05). Furthermore, the temporal dynamics of associative learning were adversely affected by NE, which significantly increased the AAL on days 1 to 4 and day 7 (4.45 ± 0.12 vs. 3.59 ± 0.24; 4.33 ± 0.37 vs. 3.52 ± 0.07; 4.05 ± 0.12 vs. 3.47 ± 0.07; 4.05 ± 0.12 vs. 3.47 ± 0.07; 3.46 ± 0.12 vs. 2.82 ± 0.21, all P < 0.05 vs. Con group; [Figure 4]c). Notably, hydrogen intervention significantly reduced the AAL on days 3 and 7 relative to that of the NE group (3.67 ± 0.09 vs. 4.05 ± 0.12; 2.95 ± 0.15 vs. 3.46 ± 0.12, all P < 0.05). Taken together, these behavioral results demonstrate that while NE severely compromises associative learning and executive function, H₂ intervention could partially mitigate these cognitive deficits.

Proteomic Profiling and Validation Reveal the Modulatory Effects of H₂ on Redox Homeostasis and Synaptic Transmission in the Inferior Colliculus

We conducted a comparative proteomic analysis of the IC to elucidate the molecular mechanisms underlying the protective effects of H₂ against noise-induced central auditory damage. Using a FC of >1.2 as the threshold, 405 DEPs were identified between the H₂ + NE and NE groups. Among these, 208 were significantly upregulated and 197 were downregulated in the H₂ + NE group [Figure 5a].

Figure 5.

Proteomic analysis of the inferior colliculus (IC) revealing the regulatory effects of H₂ on metabolic and synaptic proteins following noise exposure. Note: (a) Volcano plot illustrating the differentially expressed proteins (DEPs) in the H₂ + NE and the NE groups. (b) Gene Ontology (GO) enrichment analysis of DEPs in the biological process (BP) category. (c) Heatmap showing the expression profiles of DEPs associated with the “nicotinamide adenine dinucleotide phosphate hydrogen (NADPH) regeneration” pathway. (d) Western blot validation of key proteins involved in the “NADPH regeneration” pathway. (e) Kyoto encyclopedia of genes and genomes (KEGG) pathway enrichment analysis of DEPs. (f) Heatmap showing the expression profiles of DEPs associated with the “GABAergic synapse” pathway. (g) Heatmap showing the expression profiles of proteins associated with the “Glutamatergic synapse” pathway. Data are derived from n = 3 biological replicates per group. The red dashed line indicates the significance threshold (P < 0.05). Pathways plotted above this line are considered significantly enriched. Con = control, DEP = differentially expressed proteins, GO = Gene Ontology, H₂ + NE, H₂ + noise exposure, KEGG = Kyoto Encyclopedia of Genes and Genomes, NE = noise exposure.

GO enrichment analysis was performed to further characterize the functional implications of these DEPs. The top 20 significantly enriched terms within the biological process category are shown in [Figure 5]b. Notably, the “NADPH regeneration” pathway was significantly enriched (P = 0.001233 < 0.05). The key proteins involved in this process, including Aldh1l2, Aldob, Tigar, H6pd, and Rbks, were consistently upregulated in the H₂ + NE group [Figure 5c]. To validate these proteomic findings, WB analysis was performed to detect the expression levels of representative proteins involved in the NADPH regeneration pathway (Aldh1l2, Aldob, Tigar, H6pd, and Rbks). Representative blots showed a relative decrease in protein expression levels following acoustic trauma in the NE group. However, H₂ treatment reversed this decline, showing a trend consistent with the protein expression patterns observed in the proteomic data [Figure 5d]. Given the pivotal role of NADPH in maintaining the cellular antioxidant defense system, these findings suggest that H₂ may enhance the metabolic capacity for redox homeostasis in the IC following acoustic trauma.

Furthermore, KEGG pathway analysis identified several signaling cascades modulated by H₂ [Figure 5e]. The analysis revealed the involvement of inflammatory signaling pathways, including the complement and coagulation cascades, efferocytosis, and the transforming growth factor (TGF)-beta signaling pathway, in addition to redox-related metabolic pathways such as nicotinate and nicotinamide metabolism. Of particular interest, the “GABAergic synapse” pathway was significantly enriched and showed a general upregulation in the H₂ + NE group, as illustrated by the expression profiles of associated proteins in the heatmap [Figure 5f]. In contrast, the “glutamatergic synapse” pathway showed a downward trend in the H₂ + NE group, although it did not reach statistical significance [Figure 5g]. This reciprocal modulation of inhibitory and excitatory synaptic proteins suggests that H₂ administration may help restore the E/I balance, which is typically disrupted in the central auditory system after NE.

DISCUSSION

It is increasingly being acknowledged that NIHL extends beyond a simple peripheral cochlear disorder. The pathology escalates into a complex, multisystem process that encompasses both peripheral and central auditory pathways and is frequently associated with cognitive and emotional sequelae.[14]

In the current study, we demonstrate that H₂ inhalation significantly attenuated acute noise-induced temporary threshold shift (TTS). Given that TTS reflects pathological changes spanning the peripheral and central auditory systems,[23] this reduction highlights the comprehensive neuroprotective effect of H₂ throughout the auditory pathway. As wave I amplitude serves as a sensitive marker of IHC-SGN synaptic integrity,[24] its significant preservation in the H₂ + NE group compared to the NE group demonstrates that H₂ inhalation protected the ribbon synapse. Although this likely represents a substantial functional rescue rather than a complete biological restoration, the wave I amplitude of the H₂ + NE group was statistically comparable to that of the control group. This distinction is corroborated by the wave I latency data—a reflection of synaptic transmission efficiency and myelin integrity—[25] showing significant prolongation for the H₂ + NE compared to the Con group. Only a modest shortening trend relative to the NE group was observed. Consistently, structural assessments via DPOAE and SEM at 28 dpe revealed that H₂ treatment significantly attenuated permanent OHC loss, particularly in the high-frequency basal turn. Regarding the decreased high-frequency SNR in the NE group, we acknowledge that early-onset age-related hearing loss may also play a role, as C57BL/6J mice carry the Cdh23 mutation.[26] However, we conclude that the observed deficits are predominantly noise-induced, given that the NE group still demonstrated a significant reduction relative to age-matched controls and the H₂ intervention significantly improved this metric. Furthermore, these results indicate that H₂ exerts a protective effect in the high-frequency region. Additionally, the recovery of ABR click thresholds (which primarily reflect hearing sensitivity in the 2–4 kHz range in rodents)[27] often masks the underlying cochlear damage. Persistent DPOAE SNR impairment at high frequencies suggests OHC dysfunction or loss of the basal turn. Clinically, this correlates with HHL or high-frequency hearing loss, where patients demonstrate normal standard audiograms but struggle with speech recognition in noisy environments.[28] The inherent vulnerability of the basal turn, which stems from its high metabolic rate and mitochondrial activity,[29] renders it particularly susceptible to oxidative stress. We propose that the potent antioxidant capacity of H₂ directly counteracts this vulnerability, thereby preserving high-frequency OHC function. This interpretation is further supported by our proteomic findings (discussed below), which revealed significant upregulation of NADPH regeneration pathways in the IC.

A key finding of this study is that H₂ intervention attenuated the noise-induced elevation of PPI. For the NE group, we observed a paradoxical increase in PPI at 75 and 80 dB, which is a behavioral hallmark of central auditory hyperexcitability or “hyperacusis.”[30] Notably, the direction of PPI change is context-dependent and varies with the underlying pathological mechanism. For disorders such as schizophrenia, reduced PPI reflects a direct deficit in sensorimotor gating function, which is associated with dysfunction in the cortico-striato-pallido-thalamic circuitry.[31] According to the “central gain” theory, a reduction in peripheral neural input (as evidenced by our DPOAE and SEM data) triggers a compensatory increase in the sensitivity of central neurons, particularly in the IC, to maintain a stable neural code.[8,32] However, this compensation is often overshot, leading to pathological hyperacusis. Our histological and TEM data revealed that H₂ significantly reduced neuronal pyknosis and synaptic degradation in the IC. By preserving peripheral input and directly protecting IC neurons, H₂ limits the need for this maladaptive compensatory gain and reduces the noise-induced elevation of PPI.

To further elucidate the molecular mechanisms, we performed a proteomic analysis of the IC. Previous studies have established that H₂ exerts significant anti-inflammatory[33,34] and antioxidative stress effects.[35] Recent studies further indicate that H₂ is not merely a direct ROS scavenger; rather, it bolsters the endogenous antioxidant defense system by modulating glutathione (GSH) metabolism and activating the Nrf2/HO-1 signaling axis.[36,37,38] In the present study, inflammatory pathways—including the complement and coagulation cascades, efferocytosis, and the TGF-beta signaling pathway—were similarly enriched in the H₂ + NE group, corroborating the well-established anti-inflammatory properties of H₂. Furthermore, GO and WB analyses revealed that the significant enrichment of the “NADPH regeneration” pathway in the H₂ + NE group (characterized by the upregulation of Aldh1l2, Tigar, etc.) provides a novel metabolic explanation for its neuroprotective effects. NADPH is a crucial cofactor for GSH reductase, which maintains the pool of reduced GSH to combat oxidative stress.[39] Furthermore, the concurrent upregulation of Tigar and Aldh1l2 suggests the coordinated enhancement of NADPH regeneration across subcellular compartments. Tigar-driven cytosolic NADPH supports general antioxidant defense,[40] while Aldh1l2-mediated mitochondrial NADPH specifically combats oxidative phosphorylation-derived ROS.[41] This dual-compartment reinforcement by H₂ treatment provides robust redox homeostasis. Additionally, KEGG analysis identified enrichment of the “nicotinate and nicotinamide metabolism” pathway, which serves as a vital substrate source for NADPH.[42] This suggests that H₂ enhances the antioxidant defense capacity of the IC by reinforcing the “supply lines” of the cellular redox currency system, thereby exerting a protective effect on neurons. The IC is a hub of inhibitory and excitatory integration. NE typically results in the loss of GABAergic inhibition, which is the primary driver of central hyperexcitability.[43] Our KEGG analysis showed that H₂ treatment upregulated the “GABAergic synapse” pathway. We propose that H₂ protects vulnerable GABAergic interneurons from noise-induced damage by enhancing anti-inflammatory properties and NADPH-dependent antioxidant capacity, and this prevents functional decline of the inhibitory system and subsequent compensatory overexcitation of glutamatergic neurons. This restoration of the E/I balance is likely the physiological basis for suppressed PPI elevation.

The IC serves not only as a sensory relay but also plays a vital role in auditory-motor integration and fear-conditioned learning.[9,10,11,12,13] Our shuttle-box experiments demonstrated that NE severely impaired active avoidance learning, whereas H₂ intervention significantly improved the AAR and reduced shock duration. The impairment in the NE group may stem from two factors: (1) distorted auditory perception due to hearing loss and (2) direct neurotoxic damage to the IC-thalamus-amygdala circuitry. By preserving the structural and functional integrity of the IC and preventing the “scrambling” of auditory signals through E/I stabilization, H₂ facilitated more efficient processing of the CS (sound) and its association with the motor response (avoidance), thereby rescuing cognitive-behavioral performance.

Our experiments demonstrate the protective effects of H₂, but the limitations of this study should be acknowledged. The potential confounding effect of 33.3% oxygen in the hydrogen-rich gas mixture cannot be entirely excluded. While this concentration is unlikely to induce oxidative stress, we cannot definitively rule out its contribution. Our observed NADPH enhancement aligns with the established mechanisms of hydrogen, and clinical evidence suggests that ≤36% oxygen is safe.[36,44] Future studies incorporating a 33.3% oxygen control group are warranted to elucidate the specific neuroprotective role of hydrogen.

CONCLUSION

In summary, our study demonstrates that H₂ inhalation exerts multilevel protective effects against NIHL. It not only safeguards the cochlear periphery but also mitigates maladaptive central reorganization in the IC. By enhancing anti-inflammatory and antioxidant defenses, H₂ preserves neuronal function and maintains the E/I balance. This prevents the development of hyperacusis-like symptoms and preserves associative learning abilities. These findings reinforce the potential of H₂ as a systemic, noninvasive therapeutic agent for noise-induced auditory and cognitive deficits.

Availability of data and materials

All relevant data are included within the article. Any additional materials (raw recordings, analysis scripts, or large imaging datasets) are available from the corresponding author upon reasonable request and in accordance with the journal’s data-sharing policy.

Author contributions

Lei Shi and Cong Wei contributed equally to this work. Lei Shi: Conceptualization, Methodology, Investigation, Formal analysis, Writing—original draft. Cong Wei: Visualization, Validation, Supervision. Bo Wang: Formal analysis, Writing—review and editing. Ning Yu: Writing—review and editing, Conceptualization, Funding acquisition. ShiMing Yang: Writing—review and editing, Project administration, Funding acquisition, Conceptualization.

Ethics approval and consent to participate

Approved by the Ethics Committee of the Chinese PLA General Hospital (S2020-465-01).

Conflicts of interest

The authors declare no competing interests.

Funding Statement

This study was supported by the State Key Laboratory of Hearing and Balance Science, the National Key Research and Development Program of China (No. 2020YFC2004002), and the Noise Research Fund (No. BWS14B080 and JDZYY20132).