DNAH11 impairs memory via disrupted synaptic plasticity in noise-induced hidden hearing loss mice

Abstract

Although the effects of noise-induced hearing loss (NIHL) on cognitive functions have been widely investigated, the cognitive effects of noise-induced hidden hearing loss (NIHHL), particularly its impact on memory, remain poorly understood. The Dnah11 gene, which encodes a dynein motor protein involved in synaptic development, may play a role in NIHHL-related cognitive impairment. We aimed to investigate whether NIHHL induces memory impairment and explore the role of Dnah11 expression in this process. Behavioral experiments identified the peak of memory impairment at 1 month following noise exposure. To elucidate molecular changes, hippocampal gene expression was analyzed using transcriptomic sequencing, reverse transcription quantitative polymerase chain reaction (RT-qPCR), and immunofluorescence. RNA sequencing revealed significant Dnah11 upregulation, with immunofluorescence confirming DNAH11 overexpression in hyperactivated CaMKIIα-positive excitatory neurons. Stereotaxic injection of recombinant adeno-associated virus (rAAV) vectors to knock down hippocampal Dnah11 expression improved memory performance in NIHHL mice without improving hearing loss. This cognitive improvement was accompanied by partial restoration of synaptic plasticity-related proteins, including SYN and PSD95. These findings indicate that Dnah11 upregulation in hippocampal excitatory neurons contributes to NIHHL-induced memory impairment, and targeting Dnah11 may offer a therapeutic strategy for memory impairment associated with hidden hearing loss.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10565-025-10124-5.

Highlights

• NIHHL induced significant memory impairment in mice at 1 month after noise exposure.

• Overexpression of DNAH11 in hyperactivation of CaMKIIα-positive excitatory neurons of the hippocampus was associated with memory impairment via disrupted synaptic plasticity.

• Dnah11 knockdown in CaMKIIα-positive excitatory neurons rescued memory impairment and restored synaptic plasticity.

Graphical Abstract

Noise-induced hidden hearing loss (NIHHL) can lead to memory impairment in mice through disruption of synaptic plasticity. This effect may be associated with upregulated DNAH11 expression in hyperactivated CaMKIIα-positive excitatory neurons in the hippocampus. Knockdown of the Dnah11 gene in the hippocampus ameliorated memory impairment and restored synaptic plasticity in mice.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10565-025-10124-5.

Introduction

Noise is one of the leading causes of hearing loss, accounting for approximately 16% of all hearing impairment cases in the general population (Coyat et al. 2019; Ha et al. 2021; Sun 2021). Prolonged exposure to moderate-intensity noise can induce auditory deficits characterized by temporary threshold shifts, reduced amplitude, and prolonged latency of auditory brainstem response (ABR) wave I—features collectively referred to as noise-induced hidden hearing loss (NIHHL) (Anna Rita et al. 2019; Eggermont 2017; Guoqiang and Gabriel 2017; Jiang et al. 2024).

While noise-induced hearing loss (NIHL) primarily affects auditory function, it may also disrupt cognitive, emotional, and sleep-awake systems. The auditory system does not operate in isolation but functions within a large-scale neural connectome in which auditory input interacts with sensory, motor, emotional, and cognitive brain regions. Notably, it exhibits strong anatomical and functional connectivity with the hippocampus (Billig et al. 2022; Kari Suzanne and Barbara 2012; Manohar et al. 2020). The hippocampus, a curved structure located deep within the temporal lobe, is densely integrated within neural circuits and serves as a core component of the limbic system. It plays a central role in learning, memory consolidation, emotional regulation, and higher-order cognitive processing in both humans and animals (A et al. 2015; Billig et al. 2022; H 2001; H et al. 1999; Harrison Bush et al. 2015; Howard 2004; Howard 2017; Jue et al. 2002; Manohar et al. 2020; Neil et al. 2002; Wim E and Herbert 2005). The auditory association cortex projects to the hippocampus through both direct and indirect pathways while also receiving indirect feedback from the hippocampus. These reciprocal connections support long-term auditory memory formation and enhance the processing of speech and sound. Numerous studies have demonstrated that hippocampal structure and function are modulated by acoustic input (Billig et al. 2022; Chang et al. 2019; Manohar et al. 2020; Park et al. 2018; Wang et al. 2022), with auditory stimuli inducing dynamic hippocampal plasticity (Angelucci et al. 2007; Goble et al. 2009; Kraus et al. 2010; Liu et al. 2010; Meng et al. 2009). The dentate gyrus (DG)-CA3-CA1 tri-synaptic circuit plays a crucial role in memory encoding. Anterograde tracing studies have revealed that auditory signals from the entorhinal cortex are filtered and processed within this pathway (Kari Suzanne and Barbara 2012; Manohar et al. 2020; Mohedano-Moriano et al. 2007). Auditory stimulation regulates hippocampal information processing, enhances memory function, and mitigates cognitive decline (Pan et al. 2024). This bidirectional interaction between the auditory system and the hippocampus offers a promising therapeutic strategy for auditory-related cognitive disorders.

Previous studies have demonstrated that NIHL impairs learning and memory and may accelerate the progression of Alzheimer’s disease (AD) (Manohar et al. 2020; Pan et al. 2024; Qianru et al. 2024). Notably, each 10 dB increase in hearing loss is associated with a 1.27-fold and 1.20-fold increase in the risk of developing dementia and AD, respectively (Lin et al. 2011; Pan et al. 2024). AD, the most common subtype of senile dementia (Soni et al. 2025), is increasingly affecting younger populations and is clinically characterized by progressive memory decline, cognitive impairment, and behavioral alterations (Ren et al. 2025; Soni et al. 2025; Yuan et al. 2022). Emerging evidence has identified hearing loss as a modifiable risk factor for AD-related dementia (Chang et al. 2019; Gallacher et al. 2012; Griffiths et al. 2020; Kiely et al. 2012; Lin et al. 2011; Lin et al. 2013; Livingston et al. 2017; Manohar et al. 2020; Pan et al. 2024). These studies have elucidated the manifestations and underlying mechanisms of cognitive impairment associated with NIHL caused by high-intensity noise exposure. However, the cognitive effects of NIHHL remain largely underexplored.

Causal relationships between NIHL and cognitive deficits have been demonstrated in murine models, where noise-exposed mice exhibit concurrent auditory and hippocampal dysfunction (Park et al. 2018; So Young et al. 2016). The progression of NIHL is associated with structural and functional remodeling of hippocampal neurons, which correlates with impaired learning and memory performance (Goble et al. 2009; Ha et al. 2021; Liu et al. 2016; Manohar et al. 2020; Park et al. 2018; Shukla et al. 2019). Notably, AD model mice with superimposed hearing loss exhibit accelerated cognitive decline and increased synaptic loss in the hippocampal regions (Chang et al. 2019). Synaptic plasticity—the activity-dependent strengthening of synaptic connections—is a fundamental mechanism underlying learning and memory (Neves et al. 2008). Key proteins involved in synaptic plasticity, including synaptophysin (SYN) and postsynaptic density protein 95 (PSD95), are essential for maintaining synaptic structure and function (Janz et al. 1999; Jiang et al. 2023; Li et al. 2025b; Pan et al. 2024; Wang et al. 2024; Zhang et al. 2024).

Although prior studies have established associations between hearing loss and learning and memory impairment (Goble et al. 2009; Ha et al. 2021; Liu et al. 2016; Manohar et al. 2020; Pan et al. 2024; Park et al. 2018; Shukla et al. 2019), the cognitive consequences of NIHHL and its underlying mechanisms remain poorly understood. We aimed to determine whether NIHHL induces memory impairment and investigate the role of Dnah11 in this process. To address this gap, our study identified elevated DNAH11 expression specifically in hippocampal CaMKIIα-positive excitatory neurons at 1 month following NIHHL induction. Notably, targeted knockdown of DNAH11 in these neurons ameliorated NIHHL-associated memory deficits and partially restored synaptic plasticity-related proteins, including SYN and PSD95, to near-physiological levels. This study highlights the critical involvement of DNAH11 in memory impairment associated with NIHHL, revealing its potential as a viable therapeutic target.

Materials and methods

Animal groups

All procedures were approved by the Institutional Animal Care and Use Committee of the Air Force Medical University, Xi'an, China (Approval No. 20230375). Four-week-old male C57BL/6 J mice were obtained from the University’s Animal Experiment Center and housed under a 12-h light/dark cycle with ad libitum access to food and water. All the mice exhibited normal baseline hearing prior to noise exposure. For the pre-knockdown experiments (Fig. 1A), each time point group included five randomly assigned mice. The control (CON) group received no treatment or noise exposure, while the NIHHL group was exposed to 110 dB sound pressure level (SPL) broadband noise for 2 h to induce NIHHL. For the post-knockdown experiments (Fig. 4A), each group included six randomly assigned mice: CON + shNC (scrambled control virus without noise exposure), CON + shDNAH11 (Dnah11 knockdown virus without noise exposure), NIHHL + shNC (scrambled control virus with NIHHL), and NIHHL + shDNAH11 (Dnah11 knockdown virus with NIHHL).

Fig. 1.

Behavioral performance of mice at different time points following noise exposure. A Experimental scheme. B Schematic representation of ABR hearing test in mice. C Schematic representation of the MWM. D Latency times of the CON group and the NIHHL group at different time points during the training phase of the MWM. E Average swimming speed of mice in the test phase of the MWM at different time points [1 week (Mann Whitney test, U = 8, p = 0.42), 2 weeks (Unpaired t test, t₈ = 1.39, p = 0.20), 3 weeks (Unpaired t test with Welch's correction, t_4.77 = 1.95, p = 0.11), 1 month (Unpaired t test, t₈ = 0.52, p = 0.62), 3 months (Unpaired t test, t₈ = 1.57, p = 0.16)]. F Latency times of the CON group and the NIHHL group at different time points during the test phase of the MWM [1 week (Mann Whitney test, U = 11, p = 0.84), 2 weeks (Mann Whitney test, U = 7.5, p = 0.33), 3 weeks (Mann Whitney test, U = 7, p = 0.30), 1 month (Unpaired t test, t₈ = 2.55, p < 0.05), 3 months (Unpaired t test, t₈ = 2.91, p < 0.05)]. G The numbers of platform crossings in the test phase of the MWM at different time points [1 week (Unpaired t test, t₈ = 1.02, p = 0.34), 2 weeks (Unpaired t test with Welch's correction, t_4.75 = 1.30, p = 0.25), 3 weeks (Mann–Whitney test, U = 4, p = 0.12), 1 month (Unpaired t-test, t₈ = 2.10, p = 0.07), 3 months (Mann–Whitney test, U = 12, p > 0.99)]. H The route maps of the CON group and the NIHHL group in the MWM during the test phase at 1 month. I Schematic representation of the Y-maze spontaneous alternation test. J Schematic representation of the NOR test. K The total number of entries into the three arms of the Y-maze test at different time points [1 week (Unpaired t-test, t₈ = 1.18, p = 0.27), 2 weeks (Mann–Whitney test, U = 9, p = 0.52), 3 weeks (Unpaired t-test, t₈ = 0.10, p = 0.92), 1 month (Unpaired t-test, t₈ = 1.76, p = 0.12), 3 months (Unpaired t-test, t₈ = 1.75, p = 0.12)]. L The spontaneous alternation rate in the Y-maze test at different time points [1 week (Unpaired t-test, t₈ = 0.61, p = 0.56), 2 weeks (Mann–Whitney test, U = 9, p = 0.55), 3 weeks (Unpaired t-test, t₈ = 1.18, p = 0.27), 1 month (Unpaired t-test, t₈ = 4.02, p < 0.01), 3 months (Mann–Whitney test, U = 0, p < 0.01)]. M The recognition index in the NOR test at different time points [1 week (Mann–Whitney test, U = 11, p = 0.84), 2 weeks (Unpaired t-test, t₈ = 0.49, p = 0.63), 3 weeks (Mann–Whitney test, U = 9, p = 0.55), 1 month (Unpaired t-test, t₈ = 2.48, p < 0.05), 3 months (Unpaired t-test, t₈ = 0.38, p = 0.71)]. Data points represent individual mice, presented as Mean ± SEM. Each group consisted of 5 mice. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. ABR: auditory brainstem response; CON: control; MWM: morris water maze; NE: noise exposure; NIHHL: noise-induced hidden hearing loss; NOR: novel object recognition

Fig. 4.

Behavioral performance and expression levels of DNAH11 protein in mice 1 month after rAAV injection. A Experimental scheme. B Schematic representation of the stereotactic injection into the hippocampal tissues. C The expression of EGFP in the hippocampal tissues of mice injected with rAAV-shNC or rAAV-shDNAH11. Scale bar = 1000 μm. D Latency times of mice in each group during the training phase of the MWM. E Average swimming speed of mice in each group [NIHHL + shDNAH11 group (Unpaired t test, t₁₀ = 0.17, p = 0.87), compared to the NIHHL + shNC group; CON + shDNAH11 group (Unpaired t test, t₁₀ = 0.15, p = 0.88), NIHHL + shNC group (Unpaired t test, t₁₀ = 1.58, p = 0.14), compared to the CON + shNC group] during the test phase of the MWM. F Latency times of mice in each group [NIHHL + shDNAH11 group (Unpaired t test, t₁₀ = 2.28, p < 0.05), compared to the NIHHL + shNC group; CON + shDNAH11 group (Unpaired t test with Welch's correction, t_5.64 = 0.53, p = 0.61), NIHHL + shNC group (Unpaired t test with Welch's correction, t_5.38 = 2.18, p < 0.05), compared to the CON + shNC group] during the test phase of the MWM. G The numbers of platform crossings of mice in each group [NIHHL + shDNAH11 group (Mann–Whitney test, U = 3, p < 0.05), compared to the NIHHL + shNC group; CON + shDNAH11 group (Mann–Whitney test, U = 12, p = 0.42), NIHHL + shNC group (Mann–Whitney test, U = 1.50, p < 0.01), compared to the CON + shNC group] during the test phase of the MWM. H The route maps of mice in each group during the test phase of the MWM. I The total number of entries into the three arms of mice in each group [NIHHL + shDNAH11 group (Unpaired t test, t₁₀ = 1.01, p = 0.33), compared to the NIHHL + shNC group; CON + shDNAH11 group (Unpaired t test, t₁₀ = 1.95, p = 0.08), NIHHL + shNC group (Unpaired t test, t₁₀ = 1.17, p = 0.27), compared to the CON + shNC group] in the Y-maze test. J Spontaneous alternation rate of mice in each group [NIHHL + shDNAH11 group (Unpaired t test, t₁₀ = 2.41, p < 0.05), compared to the NIHHL + shNC group; CON + shDNAH11 group (Unpaired t test, t₁₀ = 1.53, p = 0.16), NIHHL + shNC group (Unpaired t test, t₁₀ = 1.58, p = 0.15), compared to the CON + shNC group] in the Y-maze test. K Recognition index of mice in each group [NIHHL + shDNAH11 group (Unpaired t test, t₁₀ = 3.23, p < 0.01), compared to the NIHHL + shNC group; CON + shDNAH11 group (Unpaired t test, t₁₀ = 0.47, p = 0.65), NIHHL + shNC group (Unpaired t test, t₁₀ = 2.66, p < 0.05), compared to the CON + shNC group] in the NOR test. L The mean fluorescence intensity of DNAH11 in CaMKIIα neurons within the DG [NIHHL + shDNAH11 group (Unpaired t test, t₁₀ = 5.51, p < 0.001), compared to the NIHHL + shNC group; CON + shDNAH11 group (Unpaired t test, t₁₀ = 0.49, p = 0.63), NIHHL + shNC group (Unpaired t test, t₁₀ = 4.22, p < 0.01), compared to the CON + shNC group], CA1 [NIHHL + shDNAH11 group (Unpaired t test, t₁₀ = 7.82, p < 0.0001), compared to the NIHHL + shNC group; CON + shDNAH11 group (Unpaired t test, t₁₀ = 0.20, p = 0.84), NIHHL + shNC group (Unpaired t test, t₁₀ = 6.59, p < 0.0001), compared to the CON + shNC group] and CA3 [NIHHL + shDNAH11 group (Unpaired t test, t₁₀ = 6.34, p < 0.0001), compared to the NIHHL + shNC group; CON + shDNAH11 group (Unpaired t test, t₁₀ = 0.20, p = 0.85), NIHHL + shNC group (Unpaired t test, t₁₀ = 5.42, p < 0.001), compared to the CON + shNC group] regions in the hippocampal tissues of each group. M–O Immunofluorescence images of DNAH11 in CaMKIIα neurons within the DG, CA1 and CA3 regions of the hippocampal tissues. DNAH11 (red) and DAPI (blue) are shown. Scale bar = 100 μm. Data points represent individual mice, presented as Mean ± SEM. Each group contained 6 mice. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. See also Fig. S3A-B. NC: negative control

Noise stimulation

The experimental noise consisted of broadband noise recorded using a helicopter engine. An amplifier (AV 502 BT, BGL, China) delivered the signal to speakers (IT-12, RADIN, China) installed in an acoustic chamber (absorption coefficient > 0.99). During exposure, mice were housed in cages (10.5 cm × 5.5 cm × 5.5 cm) positioned under the speakers. A-weighted SPLs, measured using a sound level meter (HCJYET HT 8352, CHN), varied by less than 3 dB across cages. All animals remained conscious throughout the exposure period (Fig. 1A).

ABR measurement

ABR measurements were performed before noise exposure and again 1 day after exposure. Mice were anesthetized via intraperitoneal injection of 1% pentobarbital sodium (0.3 mL/100 g). Subcutaneous needle electrodes were placed as follows: active electrodes at the cranial vertex, reference electrodes posterior to each pinna, and ground electrodes at the base of the tail. Acoustic transducer tubes were inserted bilaterally into the ear canals to record monaural thresholds (Fig. 1B). Auditory stimuli (clicks and pure tones) were presented starting at 80 dB SPL and decreased in 5 dB steps. Thresholds were defined as the lowest intensity that elicited a reproducible wave III. Wave I amplitude and latency were recorded at 80 dB SPL. Mice with baseline thresholds < 20 dB SPL were included. After behavioral testing, ABR was repeated using the same protocol.

Tissue preparation

Brain tissue

Following all behavioral and ABR assessments, mice were anesthetized via intraperitoneal injection of 1% pentobarbital sodium (0.3 mL/100 g). Once a deep anesthetic state was confirmed, the animals were placed in a supine position, and a V-shaped incision was made below the xiphoid to expose the heart. A perfusion needle was inserted into the left ventricle and connected to an automated pump (flow rate: 250 mL/h), followed by excision of the right atrial appendage. Vascular perfusion was first performed with 20 mL of 0.01 mol/L phosphate-buffered saline (PBS, pH 7.4) to clear the blood, followed by 40 mL of 4% paraformaldehyde (PFA) for tissue fixation. Brains were post-fixed in 4% PFA at 4 °C for 24 h, cryoprotected in 30% sucrose at 4 °C until they sank. Tissue samples were embedded in Optimal Cutting Temperature (OCT) compound (3801480, Leica, Germany) prior to being sectioned.

Cochlear tissue

Following brain extraction, the bulla region was identified, and the cochlea was extracted by gently leveraging curved hemostatic forceps beneath the base of the bulla. The isolated cochleae were immersed in 4% PFA in a glass dish. Under stereomicroscopy, surrounding soft tissue and bone near the oval window were carefully dissected using ophthalmic scissors. Micropipettes were inserted through the oval window and helicotrema to allow gravitational perfusion of the PFA. The cochleae were then post-fixed in 4% PFA at 4 °C for 24 h and decalcified in EDTA solution on a rotator for 72 h.

Immunofluorescence

Hippocampus

Coronal brain sections were cut at 35 μm thickness using a cryostat (Leica CM3050 S, Leica Biosystems, Wetzlar, Germany). Serial coronal sections, starting from bregma and spanning 45 ± 3 layers as defined in the Mouse Brain Atlas (4th ed.), were washed three times in 0.01 mol/L PBS (pH 7.4). Prior to immunostaining, tissue sections were incubated for 1 h in a blocking solution consisting of PBS with 0.3% Triton X-100 and 5% goat serum (AR0009, BOSTER, China). All primary antibodies were diluted in PBS supplemented with 0.01 g/mL bovine serum albumin and 0.3% Triton X-100, and sections were incubated at 4 °C for 72 h. After incubation, sections were washed in PBS and subsequently incubated with a secondary antibody at room temperature for 2 h. Following three additional PBS washes, sections were mounted on slides and coverslipped using a DAPI-containing antifade mounting medium (SL1841, Coolaber, China). Imaging was performed with a confocal microscope (FV 3000, Olympus, Japan) using 20 × and 60 × objectives and excitation wavelengths of 488 nm, 594 nm, and 405 nm. Hippocampal regions (DG, CA1, and CA3) were imaged. For SYN and c-Fos proteins, quantitative analyses were performed to assess the number of puncta and positive cells; for the other proteins, mean fluorescence intensity was used as the primary metric for quantification. The list of antibodies used is provided in Table 1.

Table 1.

Antibodies used in this study

Antibody	Brand and Concentration
Rabbit anti-DNAH11	bs-14360R, Bioss, CHN, 1:100
Mouse anti-CaMKIIα	66843-1-Ig, Proteintech, CHN, 1:200
Mouse anti-cFos	ab11959, Abcam, UK, 1:1000
Rabbit anti-CaMKIIα	13730-1-AP, Proteintech, CHN, 1:200
Mouse anti-GAD1	67648-1-Ig, Proteintech, CHN, 1:200
Rabbit anti-GAD1	41318, Cell Signaling Technology, USA, 1:1000
Rabbit anti-CtBP2	ab128871, Abcam, UK, 1:200
CoraLite488-conjugated Goat Anti-Mouse IgG(H+L)	SA00013-1, Proteintech, CHN, 1:250
CoraLite594-conjugated Goat Anti-Rabbit IgG(H+L)	SA00013-4, Proteintech, CHN, 1:250
Rabbit anti-Synaptophysin	17785-1-AP, Proteintech, CHN, 1:10000
Rabbit anti-PSD95	20665-1-AP, Proteintech, CHN, 1:1000
Mouse anti-GAPDH	MA6007, AntiProtech, CHN, 1:3000
Goat anti-rabbit IgG	EK020, Zhuangzhi Biology, CHN, 1:20000
Goat anti-mouse IgG	EK010, Zhuangzhi Biology, CHN, 1:20000

Cochlear ribbon synapse

Prior to immunofluorescence, the decalcified cochleae were exposed to a 1-h blocking step with 5% goat serum (AR0009, BOSTER, China). The tissues were then probed with a rabbit anti-CtBP2 (ab128871, Abcam, UK) in an overnight incubation at 4 °C. Afterward, tissues were incubated for 2 h at room temperature with the corresponding goat anti-rabbit secondary antibody. The basilar membranes were then mounted on slides using a DAPI-containing antifade medium (SL1841, Coolaber, China). Cochlear ribbon synapses were visualized using a confocal microscope (FV 3000, Olympus, Tokyo, Japan) with a 60 × oil-immersion objective and 594 nm/405 nm excitation. Ribbon synapses were quantified using ImageJ by counting CtBP2-positive puncta per inner hair cell. The antibodies used are listed in Table 1.

Behavioral experiments

Morris water maze (MWM)

The apparatus consisted of a circular plastic pool filled with opaque water maintained at 25 ± 2 °C. The pool was divided into four quadrants: I, II, III, and IV (Fig. 1C). A transparent platform (6 cm in diameter) was submerged 1 cm below the water surface in the designated target quadrant. The experiment included a 4-day training phase followed by a 1-day test phase. During the training phase, mice were allowed 60 s to locate the hidden platform, and the time taken was recorded as the escape latency. Animals that failed to find the platform within 60 s were gently guided to it and allowed to remain for 10 s to facilitate spatial memory consolidation. During the test stage, the platform was removed, and mice were allowed to freely explore the pool for 60 s. The latency to first cross the former platform location and the number of platform crossings were recorded using SMART 3.0 software (Panlab, Harvard, MA, USA).

Y-maze

The Y-maze consisted of three equal-length arms arranged at 120° angles to form a "Y" shape (Fig. 1I). Mice were placed at the central junction and allowed to freely explore the maze for 8 min. Arm entries and their sequences were recorded throughout the session. A spontaneous alternation was defined as consecutive entries into all three different arms. The maximum number of possible alternations was calculated as: total number of entries – 2. The spontaneous alternation rate (%) was determined as: (actual alternations/maximum alternations) × 100%. Data were collected and analyzed using SMART 3.0 software (Panlab, Harvard Apparatus, MA, USA). The maze arms were cleaned with 75% ethanol between trials.

Novel object recognition (NOR)

The apparatus consisted of an opaque cubic chamber measuring 50 × 50 × 50 cm. During the familiarization phase, mice explored two identical objects for 8 min (Fig. 1J). After a 2-h rest, one familiar object was replaced with a novel one, and mice were allowed to explore again for 8 min during the test phase. The time spent exploring each object was recorded, and the recognition index was calculated as follows: recognition index (%) = time spent with a novel object/time spent with novel and familiar objects × 100%. Data were acquired using SMART 3.0 software (Panlab, Harvard, MA, USA). The chamber was cleaned with 75% ethanol between trials.

RNA extraction, library construction, and sequencing

Using the established modeling protocol, four CON and four NIHHL mice were euthanized 1 month post-induction for rapid hippocampal tissue collection. Total RNA was extracted from the hippocampus using the TRIzol reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instructions. RNA quality was assessed using the Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA, USA) and RNase-free agarose gel electrophoresis. Polyadenylated mRNAs were enriched using Oligo(dT) magnetic beads, fragmented, and reverse-transcribed with random primers. Second-strand cDNA synthesis was performed using DNA polymerase I, RNase H, dNTP, and buffer. The resulting cDNA fragments were purified with the QiaQuick polymerase chain reaction (PCR) Purification kit (Qiagen, Venlo, Netherlands), end-repaired, poly(A)-tailed, and ligated to Illumina sequencing adapters. Gene Denovo Biotechnology Co. (Guangzhou, China) performed the final PCR amplification of the size-selected libraries and conducted the sequencing using an Illumina NovaSeq6000 platform. One hippocampal tissue sample from the CON group did not meet the quality control criteria and was excluded.

RNA sequencing data analysis

Differentially expressed genes (DEGs) were identified by comparing RNA expression levels between groups. DEGs were screened based on the thresholds of a p < 0.05 and an absolute fold change ≥ 1.5. To functionally characterize these DEGs, we performed enrichment analyses for Gene Ontology (GO) terms and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways, employing their respective dedicated databases. A false discovery rate (FDR) correction was applied, with FDR ≤ 0.05 considered significant. Bioinformatic analysis was performed using the real-time interactive online platform (http://www.omicsmart.com).

Reverse transcription quantitative polymerase chain reaction (RT-qPCR)

Using the established modeling protocol, six CON and six NIHHL mice were euthanized 1 month post-induction for rapid hippocampal tissue collection. Total RNA was isolated from samples with the TSINGKE TSP413 RNAprep FastPure kit (Beijing Tsingke Biotechnology Co., Ltd.). According to the manufacturers' protocols, this RNA was then reverse-transcribed into cDNA using the MightyScript First Strand cDNA Synthesis Master Mix (Sangon Biotech, Shanghai, China). Quantitative real-time PCR assays were run on a LightCycler 480 system (Roche, Switzerland) with SYBR Green PCR Master Mix (Yeasen, Shanghai, China). Relative gene expression levels were determined via the 2^−ΔΔCt method, normalized to the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) internal reference gene. All reactions were performed in technical triplicates across a minimum of six independent biological replicates. The corresponding primer sequences are provided in Table 2 (synthesized by Tsingke Biotechnology and Sangon Biotech Co.).

Table 2.

Primers used in this study

Gene	Primer sequences(5’ to 3’)
Dnah11	Forward:GGTGAAGGTCGGTGTGAACG
	Reverse:CTCGCTCCTGGAAGATGGTG
Gapdh	Forward:CATTCCCTCGCTTCTACTTCATCTC
	Reverse:GTGTTACCTGCTTCGGCTGTG

Stereotactic injection

Six mice per group (CON + shNC, CON + shDNAH11, NIHHL + shNC, and NIHHL + shDNAH11) were stereotaxically injected with viral constructs from BrainVTA (Wuhan, China). The CON + shDNAH11 and NIHHL + shDNAH11 groups received a 1:1 mixture of rAAV-CaMKIIα-CRE-WPRE-hGH polyA and rAAV-CMV-DIO-(EGFP-U6)-shRNA(Dnah11)-WPRE-hGH polyA, respectively. The CON + shNC and NIHHL + shNC groups received a 1:1 mixture of rAAV-CaMKIIα-CRE-WPRE-hGH polyA and rAAV-CMV-DIO-(EGFP-U6)-shRNA(scrambled)-WPRE-hGH polyA. Mice were anesthetized with 1% pentobarbital sodium (0.3 mL/100 g, intraperitoneally), secured in a stereotaxic apparatus, and underwent a scalp incision with soft tissue retraction. A micro-drill was used to expose the skull at the hippocampal injection site (coordinates relative to bregma: posterior − 1.7 mm, lateral ± 1.75 mm, ventral − 2.0 mm). Following the bilateral administration of 200 nL of virus using a 10 μL Hamilton syringe at 50 nL/min, the needle was retained in place for 10 min prior to slow withdrawal to prevent backflow. The scalp was sutured, and the mice were monitored during recovery with appropriate postoperative care. The knockdown of Dnah11 was further validated in S1Tr brain region exhibiting high expression levels in the CON group (Supplementary Fig. 4).

Western blotting

Hippocampal tissues from the CON + shNC, CON + shDNAH11, NIHHL + shNC, and NIHHL + shDNAH11 mice were homogenized. Protein lysates were prepared using radioimmunoprecipitation assay (RIPA) buffer (P0013B, Beyotime, China), supplemented with phenylmethylsulfonyl fluoride (PMSF) (ST506, Beyotime, China). Protein lysates were mixed with 5 × loading buffer (P0015, Beyotime, China), separated by SDS-PAGE, and transferred to polyvinylidene difluoride (PVDF) membranes (IPVH00010, Merck, Germany). Membranes were blocked with NcmBlot Blocking Buffer (P30500, Ncm Biotech, China) for 40 min at room temperature, then incubated overnight at 4 °C with primary antibodies. Following three TBST washes, the membranes were probed with species-appropriate secondary antibodies for one hour at room temperature. Afterwards, a final series of TBST washes was performed. The protein bands were then detected by enhanced chemiluminescence (ECL) substrate (Zhuangzhi Biology, China) and imaged using a FUSION FX SPECTRA imaging system (Vilber, France). For proteins with similar molecular weights, membranes were stripped using an antibody stripping buffer (IC-8013, InCellGene, USA) after signal development, then re-blocked and re-probed. Antibodies used are listed in Table 1.

Experimental design

The experiment comprised three phases: In Phase I, 4-week-old male C57BL/6 J mice (from the Air Force Medical University Animal Center) were randomly assigned to control and experimental groups. Only mice with normal baseline ABR thresholds (< 20 dB SPL) were included. The experimental group was exposed to 110 dB noise for 2 h, followed by ABR testing 1 day post-exposure. The CON group did not undergo any intervention. Behavioral tests and ABR measurements were conducted at 1 week, 2 weeks, 3 weeks, 1 month, and 3 months post-exposure. Cochleae were then harvested for ribbon synapse quantification using immunofluorescence. Based on these results and established criteria, mice were categorized into NIHHL and CON groups. The time point at which the most severe learning and memory impairment was observed was selected for subsequent experiments (Fig. 1A). Phase II: Hippocampal tissues from the CON and NIHHL groups were used for transcriptomic sequencing to identify DEGs. Dnah11 expression was validated using RT-qPCR and immunofluorescence (Fig. 3A). Phase III: rAAV vectors were stereotaxically injected to knock down Dnah11 in targeted groups, followed by behavioral testing, ABR measurements, immunofluorescence verification of knockdown efficiency, and western blot analysis of synaptic plasticity-related proteins (Fig. 4A).

Fig. 3.

Upregulation of DNAH11 expression in the hippocampus 1 month following the occurrence of NIHHL. A Experimental scheme. B Schematic representation of the RNA-seq of hippocampal tissues. C The heatmap of the distribution of up-regulated (red) and down-regulated (blue) genes. D The volcano plot of the position of Dnah11 among the distribution of upregulated (red) and down-regulated (blue) genes. E The Venn diagram of DEGs of GO and KEGG enrichment analyses. F RT-qPCR analysis to test the expression levels of Dnah11 in hippocampus tissues from CON group and NIHHL group (Unpaired t test with Welch's correction, t₅ = 5.22, p < 0.01). G Schematic representation of immunofluorescence analysis in mice brain tissues. H The mean fluorescence intensity of DNAH11 in CaMKIIα neurons within the DG (Unpaired t test, t₁₀ = 4.69, p < 0.001), CA1 (Unpaired t test, t₁₀ = 6.48, p < 0.0001) and CA3 (Unpaired t test, t₁₀ = 12.73, p < 0.0001) regions of the hippocampal tissues. I Immunofluorescence images of DNAH11 in CaMKIIα-positive neurons within the DG, CA1 and CA3 regions of the hippocampal tissues. Dotted line delineates the region of DNAH11 upregulation. DNAH11 (red), CaMKIIα (green) and DAPI (blue) are shown. Scale bar = 100 μm. J Immunofluorescence images showing c-Fos-positive cells within CaMKIIα neurons in the DG, CA1, and CA3 regions of the hippocampus. White arrow indicates the co-localization of CaMKIIα-positive (CaMKIIα⁺) excitatory neurons and c-Fos-positive (c-Fos⁺) cells. CaMKIIα (red), c-Fos (green) and DAPI (blue) are shown. Scale bar = 50 μm. K The number of c-Fos-positive cells within the DG (Unpaired t test, t₁₀ = 3.76, p < 0.01), CA1 (Unpaired t test, t₁₀ = 5.39, p < 0.001) and CA3 (Unpaired t test, t₁₀ = 2.93, p < 0.05) regions of the hippocampal tissues. L The proportion of c-Fos⁺ and CaMKIIα⁺ co-localization neurons in all CaMKIIα.⁺ neurons within the DG (Unpaired t test, t₁₀ = 5.44, p < 0.001), CA1 (Unpaired t test, t₁₀ = 4.52, p < 0.01) and CA3 (Unpaired t test, t₁₀ = 12.76, p < 0.0001) regions of the hippocampal tissues. Data points represent individual mice, presented as Mean ± SEM. Each group consisted of 6 mice. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. See also Fig. S1A-B. CaMKIIα: Calcium/calmodulin-dependent protein kinase II alpha; c-Fos: cellular FBJ osteosarcoma oncogene; DNAH11: Dynein axonemal heavy chain 11

Statistical analysis

Data are presented as Mean ± SEM and were analyzed in GraphPad Prism 9.5 (Insightful Science, CA, USA). Shapiro–Wilk test was employed to assess normality. Normally distributed data were analyzed using two-tailed unpaired or paired Student’s t-tests. Non-normally distributed data were analyzed using the Mann–Whitney U test (unpaired) or the Wilcoxon signed-rank test (paired). For data with unequal variances, appropriate corrections were applied. Statistical significance was defined as p < 0.05.

Results

NIHHL mice exhibited memory impairments

Throughout the entire experimental period of our study, mice in the NIHHL group were uniformly exposed to 110 dB noise for 2 h. To determine whether NIHHL induces learning and memory deficits, we employed classic behavioral experiments to assess cognitive function. In the MWM protocol (Fig. 1C), NIHHL and CON mice were evaluated at multiple post-induction time points (1 week [1W], 2 weeks [2W], 3 weeks [3W], 1 month [1M], and 3 months [3M]) under identical housing conditions. During the 4-day training phase, escape latency decreased similarly across NIHHL and CON groups (Fig. 1D), indicating preserved learning ability and visual function. Swimming speed did not differ significantly between groups at any time point (Fig. 1E), suggesting intact motor performance. During the test phase, NIHHL mice exhibited longer latency to first cross the former platform location than that of the CON group across all time points at 1 and 3 months (Fig. 1F). The number of platform crossings was reduced in NIHHL mice at 1, 2, and 3 weeks, as well as at 1 month, though these differences were not statistically significant. At 3 months, NIHHL mice exhibited a greater increase in platform crossings than that of the CON group, though the difference was not statistically significant (Fig. 1G). Representative swim trajectories at 1 month post-NIHHL compared to the CON group are illustrated (Fig. 1H). These findings indicate impaired spatial reference memory in NIHHL mice, with the most pronounced deficits observed at 1 month.

In the Y-maze test (Fig. 1I), total arm entries did not differ between the NIHHL and CON groups at any time point (Fig. 1K), indicating preserved locomotor and exploratory activity. However, spontaneous alternation rates were significantly lower in NIHHL mice at 1 and 3 months than in the CON group (Fig. 1L). The NOR assay (Fig. 1J) revealed an impaired recognition index in NIHHL mice at 1 month post-induction (Fig. 1M). These findings indicate NIHHL-induced deficits in spatial working memory and short-term recognition memory, with peak impairment at 1 month post-induction. Collectively, these experimental findings demonstrate that NIHHL mice exhibit memory impairment at 1 month post-induction, while their learning capacity remains intact.

NIHHL led to long-term loss of ribbon synapses in the mouse cochlea

To investigate NIHHL across different time points, we conducted audiological assessments in mice (Figs. 1B, 2A). Based on our established noise exposure protocol (Jiang et al. 2024), 4-week-old male C57BL/6 J mice with normal baseline hearing (click thresholds < 20 dB SPL) were selected. The experimental group was exposed to 110 dB helicopter noise for 2 h, while the CON group remained unexposed. In earlier studies at 1, 2, and 3 weeks post-exposure, our group validated short-term changes in ABR thresholds, latency of wave I and amplitude of wave I (Jiang et al. 2024). Furthermore, in order to observe the long-term effects of NIHHL and better correlate it with the timeline of memory impairment, we focused here on auditory function and cochlear ribbon synapse morphology at 1 and 3 months post-exposure. ABR measurements performed 1 day post-exposure revealed a significant threshold elevation in exposed mice, with thresholds recovering to near-normal by 1 month (Fig. 2B). In contrast, another experimental cohort exhibited a threshold elevation at 1 day with no recovery at 3 months (Fig. 2E). Additionally, we measured ABR wave I latency and amplitude before and after noise exposure. The experimental group exhibited a prolonged ABR wave I latency 1 day post-exposure, which was normalized by 1 month (Fig. 2C). A separate experimental cohort demonstrated a similar transient latency prolongation, which also recovered by 3 months (Fig. 2F). In contrast, ABR wave I amplitude significantly decreased at 1 day post-noise exposure and failed to return to physiological levels by 1 month (Fig. 2D). Similarly, in the other cohort, amplitude reduction was observed at 1 day and remained suppressed even at 3 months post-exposure (Fig. 2G). These findings indicate that, although ABR thresholds and wave I latency recovered by 1 and 3 months post-noise exposure, the amplitude of ABR wave I remained diminished, suggesting persistent auditory nerve dysfunction and long-term noise-induced hidden hearing loss in mice.

Fig. 2.

Auditory function and cochlear ribbon synapses in mice at different time points following noise exposure. A ABR waveform examples at key time points (NE-Pre, NE-1D, NE-1 M, and NE-3 M). B Hearing thresholds of the NIHHL group at NE-1D (Paired t test, t₄ = 14.23, p < 0.001), NE-1 M (Paired t test, t₄ = 1.63, p = 0.18), compared to the NE-pre. C Latency of wave I of the NIHHL group at NE-1D (Paired t test, t₄ = 7.65, p < 0.01), NE-1 M (Paired t test, t₄ = 0.65, p = 0.55), compared to the NE-pre. D Amplitude of Wave I of the NIHHL group at NE-1D (Paired t test, t₄ = 5.59, p < 0.01), NE-1 M (Paired t test, t₄ = 7.59, p < 0.01), compared to NE-pre. E Hearing thresholds of the NIHHL group at NE-1D (Paired t test, t₄ = 4.76, p < 0.01), NE-3 M (Paired t test, t₄ = 4.09, p < 0.05), compared to the NE-pre. F Latency of wave I of the NIHHL group at NE-1D (Paired t test, t₄ = 4.34, p < 0.05), NE-3 M (Paired t test, t₄ = 0.41, p = 0.71), compared to the NE-pre. G Amplitude of Wave I of the NIHHL group at NE-1D (Paired t test, t₄ = 6.23, p < 0.01), NE-3 M (Paired t test, t₄ = 5.77, p < 0.01), compared to NE-pre. H Immunofluorescence images of ribbon synapse density changes in different turns of basilar membranes of the CON group and NIHHL group at 1 month. Dotted lines outline the IHCs, the yellow arrow indicates the IHCs, and the white arrow indicates the CtBP2-positive ribbon synapses associated with these cells. Scale bar = 10 μm. I Immunofluorescence images of ribbon synapse density changes in different turns of basilar membranes of the CON group and NIHHL group at 3 months. Dotted lines outline the IHCs, the yellow arrow indicates the IHCs, and the white arrow indicates the CtBP2-positive ribbon synapses associated with these cells. Scale bar = 10 μm. J The number of ribbon synapses per IHC at different time points [1 month (Unpaired t test, t₈ = 12.28, p < 0.0001), 3 months (Unpaired t test, t₈ = 4.35, p < 0.05)]. K The number of ribbon synapses per IHC at different locations at 1 month [apical turn (Unpaired t test, t₈ = 2.35, p < 0.05), middle turn (Mann–Whitney test, U = 0, p < 0.05), basal turn (Unpaired t test, t₇ = 15.52, p < 0.0001)]. L The number of ribbon synapses per IHC at different locations at 3 months [apical turn (Unpaired t test, t₄ = 2.08, p = 0.11), middle turn (Unpaired t test, t₄ = 1.81, p = 0.14), basal turn (Unpaired t test, t₄ = 9.38, p < 0.001)]. Data points represent individual mice, presented as Mean ± SEM. Each group consisted of 3–5 mice. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. CtBP2: C-terminal Binding Protein 2; IHC: inner hair cells

Further morphological analysis of cochlear ribbon synapses revealed that noise-exposed mice had significantly fewer ribbon synapses than those of the CON group at 1 month post-exposure (Fig. 2H, J), with persistent reductions at 3 months (Fig. 2I, J). At 1 month, synaptic loss was evident across the apical, middle, and basal turns of the cochlea (Fig. 2K). By 3 months, significant loss remained only in the basal turn, while reductions in the apical and middle turns were no longer statistically significant (Fig. 2L). Collectively, these morphological and audiometric findings confirm sustained NIHHL post-noise exposure, supporting the selection of the 1-month post-exposure time point as the optimal model for subsequent investigations.

DNAH11 expression was upregulated in the hippocampus of NIHHL mice

The hippocampus, a brain region essential for learning and memory, plays a pivotal role in memory formation and regulation. To identify genes associated with NIHHL-induced memory impairment (Fig. 3A), we performed transcriptomic sequencing of hippocampal tissues from NIHHL and CON mice 1 month post-induction (Fig. 3B). Comparative analysis revealed 44 differentially expressed mRNAs—29 upregulated and 15 downregulated—between the two groups (Fig. 3C, Supplementary Table 1). GO enrichment analysis identified three learning- and memory-related genes: Dnah11, Egr2, and Musk (Supplementary Table 2), while KEGG pathway analysis highlighted three neurodegeneration-associated genes: Dnah11, Dnah6, and Dnah10 (Supplementary Table 3). Among these, Dnah11 was ultimately identified as the key candidate gene (Fig. 3E). Its expression profile is presented in the corresponding heatmap and volcano plot (Fig. 3C, D). To validate the transcriptomic findings, RT-qPCR analysis of hippocampal tissues from NIHHL and CON groups confirmed a significant upregulation of Dnah11 expression in NIHHL mice (Fig. 3F). Multiplex immunofluorescence staining of brain sections further revealed greater DNAH11 protein fluorescence intensity in the hippocampal DG, CA1, and CA3 subregions of NIHHL mice than that of the CON group (Fig. 3G-I, Fig. S1A-B). Moreover, the brain regions exhibiting DNAH11 overexpression were predominantly localized to areas enriched in excitatory neurons. To assess CaMKIIα expression levels in these neurons, immunostaining was performed in the hippocampus. Our results revealed that, compared with the CON group, the NIHHL group exhibited upregulated DNAH11 expression, concurrent with a significant increase in the mean fluorescence intensity of CaMKIIα (Fig. 3I, S1A-B). To further investigate neuronal activation, c-Fos co-staining was conducted to evaluate the activation state of CaMKIIα-positive excitatory neurons. The data demonstrated that, relative to the CON group, the number of c-Fos-positive neurons in the hippocampus was markedly increased in the NIHHL group (Fig. 3K, S2A). Furthermore, quantitative analysis of CaMKIIα-positive neurons revealed a significantly higher proportion of c-Fos co-expression within this neuronal subpopulation (Fig. 3J, 3L). Concurrently, we assessed the expression levels of DNAH11 and c-Fos in inhibitory neurons. The results revealed that, compared with the CON group, there were no significant differences in the mean fluorescence intensity of DNAH11 and the number of c-Fos-positive cells in GAD1-positive inhibitory neurons within the hippocampus of the NIHHL group (Fig. S2B-D). Collectively, these results demonstrate upregulation of hippocampal DNAH11 expression at both the transcript and protein levels in NIHHL mice at 1 month post-induction, with preferential localization in hyperactivated CaMKIIα-expressing excitatory neurons.

Knockdown of Dnah11 ameliorated memory impairment in NIHHL mice

To selectively knock down Dnah11 expression in hippocampal CaMKIIα-expressing neurons (Fig. 4A), we performed stereotaxic injection of rAAV vectors (Fig. 4B). Successful viral transduction was confirmed by immunofluorescence staining in hippocampal sections (Fig. 4C, S3A-B). One month following NIHHL induction, behavioral assessments were conducted. During the MWM training phase, all experimental groups (CON + shNC, CON + shDNAH11, NIHHL + shNC, and NIHHL + shDNAH11) exhibited comparable reductions in escape latency (Fig. 4D), indicating preserved intact learning ability and visual function. Swimming velocities did not differ significantly among the groups (Fig. 4E), confirming preserved motor performance. During the probe test, NIHHL + shDNAH11 mice demonstrated a significantly shorter latency to locate the platform than that of the NIHHL + shNC group (Fig. 4F), suggesting improved spatial memory. In contrast, NIHHL + shNC mice exhibited a longer latency than that of the CON + shNC group (Fig. 4F), consistent with observed impairments. CON + shDNAH11 mice exhibited no significant differences in latency compared to the CON + shNC group (Fig. 4F). The number of platform crossings was significantly higher in the NIHHL + shDNAH11 group compared to the NIHHL + shNC group (Fig. 4G), whereas NIHHL + shNC displayed lower crossings than those of the CON + shNC group (Fig. 4G), consistent with previous findings. No significant differences in platform crossings were observed between CON + shDNAH11 and CON + shNC mice (Fig. 4G). Representative swimming trajectories are presented in Fig. 4H.

The Y-maze test revealed comparable total arm entries across all groups (CON + shNC, CON + shDNAH11, NIHHL + shNC, and NIHHL + shDNAH11), indicating preserved locomotor and exploratory functions (Fig. 4I). Spontaneous alternation rates were significantly higher in NIHHL + shDNAH11 mice than in NIHHL + shNC mice (Fig. 4J), whereas NIHHL + shNC mice showed lower spontaneous alternation rates than those of the CON + shNC mice (Fig. 4J), consistent with previous results. Spontaneous alternation rates of the CON + shDNAH11 group remained unchanged relative to the CON + shNC group (Fig. 4J). In the NOR test, NIHHL + shDNAH11 mice exhibited a significantly higher recognition index than that of NIHHL + shNC mice (Fig. 4K), whereas NIHHL + shNC mice showed a lower recognition index than that of CON + shNC mice (Fig. 4K), consistent with previous observations. The recognition index of CON + shDNAH11 mice was comparable to that of CON + shNC mice (Fig. 4K).

Immunofluorescence of brain sections revealed significantly lower DNAH11 protein expression in the hippocampal DG, CA1, and CA3 subregions of NIHHL + shDNAH11 mice than that of the NIHHL + shNC group (Fig. 4L-O). However, DNAH11 levels remained higher in NIHHL + shDNAH11 mice than in CON + shNC mice (Fig. 4L-O), consistent with previous observations. DNAH11 expression in CON + shDNAH11 mice was comparable to that in CON + shNC mice across all hippocampal subregions (Fig. 4L-O), demonstrating that Dnah11 knockdown ameliorated NIHHL-associated memory impairment without adverse effects in healthy controls.

Knockdown of Dnah11 did not improve hearing loss in NIHHL mice

To investigate auditory and morphological changes following Dnah11 knockdown, we performed ABR measurements and cochlear basilar membrane immunofluorescence. The results showed that NIHHL + shDNAH11 mice exhibited higher click-ABR thresholds 1 day post-noise exposure than pre-exposure levels, with thresholds progressively recovering to physiological levels by 1 month (Fig. 5A). Similarly, NIHHL + shNC mice demonstrated an acute threshold elevation at 1 day post-exposure, followed by progressive recovery by 1 month (Fig. 5A). NIHHL + shDNAH11 mice showed longer ABR wave I latency at 1 day post-exposure than pre-exposure levels, which decreased by 1 month (Fig. 5B). A similar transient latency prolongation was observed in NIHHL + shNC mice, with normalization at 1 month (Fig. 5B). In contrast, NIHHL + shDNAH11 mice exhibited a reduced ABR wave I amplitude at 1 day post-exposure without recovery to pre-exposure levels at 1 month (Fig. 5C). NIHHL + shNC mice showed a comparable amplitude reduction at 1 day post-exposure, with no subsequent recovery at 1 month (Fig. 5C).

Fig. 5.

Auditory function and cochlear ribbon synapse in mice from each group 1 month after rAAV injection. A Hearing thresholds of the NIHHL + shNC group [NE-1D (Wilcoxon matched-pairs signed rank test, W = 21, p < 0.05), NE-1 M (Wilcoxon matched-pairs signed rank test, W = 10, p = 0.13), compared to NE-pre] and NIHHL + shDNAH11 group [NE-1D (Wilcoxon matched-pairs signed rank test, W = 21, p < 0.05), NE-1 M (Wilcoxon matched-pairs signed rank test, W = 10, p = 0.13), compared to NE-pre]. B Latency of wave I of the NIHHL + shNC group [NE-1D (Paired t test, t₅ = 2.68, p < 0.05), NE-1 M (Paired t test, t₅ = 4.17, p < 0.01), compared to NE-pre] and NIHHL + shDNAH11 group [NE-1D (Paired t test, t₅ = 3.22, p < 0.05), NE-1 M (Paired t test, t₅ = 3.81, p < 0.05), compared to NE-pre]. C Amplitude of Wave I of the NIHHL + shNC group [NE-1D (Paired t test, t₅ = 7.41, p < 0.001), NE-1 M (Paired t test, t₅ = 4.33, p < 0.01), compared to NE-pre] and NIHHL + shDNAH11 group [NE-1D (Paired t test, t₅ = 15.68, p < 0.0001), NE-1 M (Wilcoxon matched-pairs signed rank test, W = 21, p < 0.05), compared to NE-pre]. D Immunofluorescence images of ribbon synapse density changes in different turns of basilar membranes of mice in each group. Dotted lines outline the IHCs, the yellow arrow indicates the IHCs, and the white arrow indicates the CtBP2-positive ribbon synapses associated with these cells. Scale bar = 10 μm. E The number of ribbon synapses per IHC of mice in each group [NIHHL + shDNAH11 group (Unpaired t test, t₁₀ = 1.26, p = 0.24), compared to the NIHHL + shNC group; CON + shDNAH11 group (Unpaired t test, t₁₀ = 0.29, p = 0.78), NIHHL + shNC group (Unpaired t test, t₁₀ = 7.81, p < 0.0001), compared to the CON + shNC group]. F The number of ribbon synapses per IHC of mice in each group at apical [NIHHL + shDNAH11 group (Unpaired t test, t₁₀ = 1.26, p = 0.24), compared to the NIHHL + shNC group; CON + shDNAH11 group (Unpaired t test, t₁₀ = 1.52, p = 0.16), NIHHL + shNC group (Unpaired t test, t₁₀ = 2.69, p < 0.05), compared to the CON + shNC group], middle [NIHHL + shDNAH11 group (Unpaired t test, t₁₀ = 0.60, p = 0.56), compared to the NIHHL + shNC group; CON + shDNAH11 group (Unpaired t test, t₁₀ = 0.93, p = 0.38), NIHHL + shNC group (Unpaired t test, t₁₀ = 4.08, p < 0.01), compared to the CON + shNC group], and basal [NIHHL + shDNAH11 group (Unpaired t test with Welch's correction, t_4.45 = 1.24, p = 0.28), compared to the NIHHL + shNC group; CON + shDNAH11 group (Unpaired t test, t₉ = 1.82, p = 0.10), NIHHL + shNC group (Unpaired t test, t₉ = 5.65, p < 0.001), compared to the CON + shNC group] turns. Data points represent individual mice, presented as Mean ± SEM. Each group contained 5–6 mice. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

Cochlear ribbon synapse counts were not significantly different between the NIHHL + shDNAH11 and NIHHL + shNC groups at 1 month (Fig. 5D-E). NIHHL + shNC mice had fewer synapses than those of CON + shNC mice (Fig. 5D-E), consistent with NIHHL pathology. CON + shDNAH11 mice showed synapse counts comparable to those of CON + shNC mice (Fig. 5D-E). Further regional analysis revealed no significant differences in ribbon synapse counts between the NIHHL + shDNAH11 and NIHHL + shNC groups across the apical, middle, or basal cochlear turns (Fig. 5D, F). NIHHL + shNC mice exhibited lower synapse counts than those of CON + shNC mice in all cochlear regions (Fig. 5D, F), consistent with NIHHL pathology. Synapse counts in the CON + shDNAH11 group did not differ from those in the CON + shNC group at any cochlear turn (Fig. 5D, F). These results demonstrated that Dnah11 knockdown failed to rescue hearing loss, as cochlear ribbon synapse numbers remained unrecovered, consistent with the auditory outcomes observed at 1 month post-NIHHL induction.

Knockdown of Dnah11 ameliorated SYN and PSD95 expression in NIHHL mice

To investigate changes in synaptic plasticity-related proteins in the hippocampus following Dnah11 knockdown, we performed Western blotting and immunofluorescence to quantify SYN and PSD95 expression. Western blot analysis revealed significantly lower SYN expression in NIHHL + shNC mice than in CON + shNC mice (Fig. 6A-B), while NIHHL + shDNAH11 mice exhibited greater SYN expression than that of NIHHL + shNC mice (Fig. 6A-B). SYN expression levels in the CON + shDNAH11 mice remained comparable to those in the CON + shNC group (Fig. 6A-B). Similarly, NIHHL + shNC mice exhibited significantly lower PSD95 expression than that of the CON + shNC mice (Fig. 6A-B). NIHHL + shDNAH11 mice showed higher PSD95 expression than that of the NIHHL + shNC mice (Fig. 6A-B). PSD95 levels in the CON + shDNAH11 group remained comparable to those in the CON + shNC group (Fig. 6A-B).

Fig. 6.

Proteins associated with synaptic plasticity in mice from each group 1 month after rAAV injection. A Representative immunoblot images of SYN and PSD95 in the hippocampal lysates from mice in each group. B Quantitative analysis of SYN [NIHHL + shDNAH11 group (Unpaired t test, t₁₀ = 3.34, p < 0.01), compared to the NIHHL + shNC group; CON + shDNAH11 group (Unpaired t test, t₁₀ = 1.12, p = 0.29), NIHHL + shNC group (Unpaired t test, t₁₀ = 6.47, p < 0.0001), compared to the CON + shNC group] and PSD95 [NIHHL + shDNAH11 group (Unpaired t test, t₁₀ = 2.35, p < 0.05), compared to the NIHHL + shNC group; CON + shDNAH11 group (Unpaired t test, t₁₀ = 1.34, p = 0.21), NIHHL + shNC group (Unpaired t test, t₁₀ = 7.55, p < 0.0001), compared to the CON + shNC group] in hippocampal lysates from mice in each group. C Immunofluorescence images of SYN in DG, CA1, and CA3 regions of the hippocampus of mice in each group. The dotted lines delineate the region containing CaMKIIα-positive excitatory neurons, and the white arrow indicates SYN puncta on CaMKIIα neurons. CaMKIIα (green), SYN (red) and DAPI (blue) are shown. Scale bar = 20 μm. D Immunofluorescence images of PSD95 in DG, CA1, and CA3 regions of the hippocampus of mice in each group. The dotted lines delineate the region containing CaMKIIα-positive excitatory neurons. CaMKIIα (green), PSD95 (red) and DAPI (blue) are shown. Scale bar = 20 μm. E The number of SYN puncta in CaMKIIα neurons within the DG [NIHHL + shDNAH11 group (Unpaired t test, t₁₀ = 6.56, p < 0.0001), compared to the NIHHL + shNC group; CON + shDNAH11 group (Unpaired t test, t₁₀ = 1.61, p = 0.14), NIHHL + shNC group (Unpaired t test, t₁₀ = 6.45, p < 0.0001), compared to the CON + shNC group], CA1 [NIHHL + shDNAH11 group (Unpaired t test, t₁₀ = 4.98, p < 0.001), compared to the NIHHL + shNC group; CON + shDNAH11 group (Unpaired t test, t₁₀ = 1.02, p = 0.33), NIHHL + shNC group (Unpaired t test, t₁₀ = 7.70, p < 0.0001), compared to the CON + shNC group] and CA3 [NIHHL + shDNAH11 group (Unpaired t test, t₁₀ = 8.21, p < 0.0001), compared to the NIHHL + shNC group; CON + shDNAH11 group (Unpaired t test with Welch's correction, t_6.08 = 0.26, p = 0.80), NIHHL + shNC group (Unpaired t test with Welch's correction, t_6.06 = 5.16, p < 0.01), compared to the CON + shNC group] regions in the hippocampal tissues of mice in each group. F The relative fluorescence intensity of PSD95 in CaMKIIα neurons within the DG [NIHHL + shDNAH11 group (Unpaired t test with Welch's correction, t_6.34 = 4.17, p < 0.01), compared to the NIHHL + shNC group; CON + shDNAH11 group (Unpaired t test, t₁₀ = 0.53, p = 0.61), NIHHL + shNC group (Unpaired t test, t₁₀ = 3.64, p < 0.01), compared to the CON + shNC group], CA1 [NIHHL + shDNAH11 group (Unpaired t test with Welch's correction, t_5.38 = 5.68, p < 0.01), compared to the NIHHL + shNC group; CON + shDNAH11 group (Unpaired t test, t₁₀ = 0.64, p = 0.54), NIHHL + shNC group (Unpaired t test, t₁₀ = 5.84, p < 0.001), compared to the CON + shNC group] and CA3 [NIHHL + shDNAH11 group (Unpaired t test, t₁₀ = 2.73, p < 0.05), compared to the NIHHL + shNC group; CON + shDNAH11 group (Unpaired t test, t₁₀ = 0.44, p = 0.67), NIHHL + shNC group (Unpaired t test, t₁₀ = 5.08, p < 0.001), compared to the CON + shNC group] regions in the hippocampal tissues of mice in each group. Data points represent individual mice, presented as Mean ± SEM. Each group contained 6 mice. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. See also Fig. S5A. GAPDH: glyceraldehyde-3-phosphate dehydrogenase; PSD95: postsynaptic density protein 95; SYN: synaptophysin

Immunofluorescence analysis of brain sections revealed that the number of SYN puncta in the hippocampal DG, CA1, and CA3 regions was more increased in the NIHHL + shDNAH11 group than in the NIHHL + shNC group (Fig. 6C, E). No significant differences in the number of SYN puncta were observed between the CON + shDNAH11 and CON + shNC groups across these hippocampal subregions (Fig. 6C, E). In contrast, the NIHHL + shNC group exhibited lower number of SYN puncta in the DG, CA1, and CA3 regions than that of the CON + shNC group (Fig. 6C, E). Similarly, PSD95 protein expression was higher in the DG, CA1, and CA3 regions of NIHHL + shDNAH11 mice than in those of the NIHHL + shNC group (Fig. 6D, F). No significant changes in PSD95 expression were observed between the CON + shDNAH11 and CON + shNC groups across these regions (Fig. 6D, F). The NIHHL + shNC group displayed downregulation of PSD95 proteins in the DG, CA1, and CA3 regions relative to the CON + shNC group (Fig. 6D, F). Collectively, these findings indicate that Dnah11 knockdown markedly ameliorated synaptic plasticity deficits in the hippocampus of NIHHL mice without eliciting detectable adverse effects in the CON group.

Discussion

Rapid societal development and modernization have increased occupational noise exposure globally, with widespread upgrades to industrial equipment. As more regions undergo industrialization, the harmful effects of noise exposure have emerged as a global public health concern (Qianru et al. 2024). Although hearing protection devices are commonly used, workers often remain chronically exposed to moderate-to-low-intensity noise (Tepe et al. 2017; Zheng et al. 2025). This type of exposure typically does not raise hearing thresholds but can result in speech-in-noise difficulties, tinnitus, and auditory hypersensitivity—collectively referred to as NIHHL (Anna Rita et al. 2019; Eggermont 2017; Guoqiang and Gabriel 2017; Jiang et al. 2024). Hearing plays a critical role in perceiving environmental sounds and adapting to everyday life. Chronic auditory deprivation reduces sensory stimulation and linguistic input, thereby increasing the risk of cognitive impairment (Qianru et al. 2024). While NIHL has been associated with cognitive deficits (Griffiths et al. 2020; Livingston et al. 2017; Qianru et al. 2024), the phenotypic and molecular mechanisms underlying NIHHL remain largely unclear. In this study, NIHHL mice exhibited memory impairment at 1 month post-exposure, accompanied by upregulated DNAH11 expression in hippocampal CaMKIIα-hyperactivated excitatory neurons. Targeted knockdown of Dnah11 in these neurons alleviated memory deficits and restored synaptic plasticity protein expression to physiological levels.

Learning and memory are fundamental cognitive capacities in mammals. Animal models of NIHL have demonstrated significant and persistent impairments in long-term memory. Epidemiological studies have shown that hearing loss (HL) not only increases the risk of dementia but also accelerates AD progression. Various factors—such as aging, noise exposure, and ototoxic medications—can lead to HL (Gallacher et al. 2012; Kiely et al. 2012; Lin et al. 2011; Lin et al. 2013; Nieman and Oh 2020; Sun 2021). The hippocampus, a key brain region involved in learning and memory (Ha et al. 2021; Harrison Bush et al. 2015), is frequently studied in the context of cognitive decline. HL can be anatomically classified as conductive, sensorineural, or mixed (Wu et al. 2001). Notably, ototoxic drug-induced sensorineural HL has been shown to induce hippocampal degeneration, impair spatial memory, and promote tau protein hyperphosphorylation (Shen et al. 2021). Recent studies have further demonstrated that cochlear ablation or ototoxic drug-induced HL exacerbates cognitive decline in AD mouse models (Pan et al. 2024). A major challenge in animal memory research lies in selecting reliable and sensitive assessment tools. The classic MWM evaluates spatial reference memory by testing the recall of the hidden platform location. The Y-maze spontaneous alternation test measures spatial working memory, while the NOR test assesses short-term recognition and non-spatial memory (Hodges 1996; Lanke et al. 1993; Li et al. 2025a). In our previous work, NIHHL was validated only at 1 day, 1 week, and 2 weeks after exposure (Jiang et al. 2024). In this study, we extended the observation period to include 3 weeks, 1 month, and 3 months to identify the optimal time points for assessing memory function. Our findings revealed that memory impairment progressively worsens over time following NIHHL. While spatial memory impairment revealed a progressive trend at 1, 2, and 3 weeks post-exposure, statistical significance was observed only at 1 month and persisted through 3 months. Notably, learning capacity remained preserved throughout the experimental period. However, platform crossings did not differ significantly between groups—potentially due to the limited platform dimensions—consistent with reports suggesting that NIHL and its associated cognitive decline may develop gradually over extended periods (Cui et al. 2012; Ha et al. 2021). Y-maze results further demonstrated reduced spontaneous alternation rates 1 month post-NIHHL, which persisted for 3 months, indicating a sustained impairment in spatial working memory. Similarly, NOR tests revealed significant deficits at 1 month, reflecting compromised short-term recognition and non-spatial memory. These findings align with previous studies reporting that NIHL induces hippocampus-dependent memory impairment (Kraus et al. 2010; Liu et al. 2016; Liu et al. 2018; Manohar et al. 2020; Park et al. 2018; Shukla et al. 2019). Critically, 1 month represents a pivotal time point at which deficits emerge across spatial reference memory, working memory, and recognition/non-spatial domains. These findings confirm that both NIHL and NIHHL can induce memory impairment, consistent with recent studies on NIHHL (Jagersma et al. 2024, 2025). Based on our established NIHHL model, we propose that NIHHL-induced memory impairment provides novel insights into the diagnosis and treatment of AD-related dementia. ABR measurements and ribbon synapse quantification confirmed that this noise exposure paradigm induces persistent NIHHL, with neither auditory thresholds nor synaptic counts recovering by 3 months, aligning with previous findings that NIHHL represents an irreversible auditory pathology (Budak et al. 2021; Jiang et al. 2024; Liberman and Kujawa 2017; Liu et al. 2019; Seo et al. 2022).

Previous studies have revealed that acute noise exposure alters hippocampal gene expression (Lee et al. 2020), prompting further investigation into the molecular mechanisms underlying NIHHL-induced cognitive dysfunction. To investigate gene expression changes related to memory impairment in NIHHL, we conducted transcriptomic sequencing of hippocampal tissues from the CON and NIHHL groups at 1 month post-induction. Furthermore, we screened for DEGs and performed GO functional enrichment analysis, focusing on biological processes related to learning and memory. Given that most human neurodegenerative disorders are characterized by cognitive impairment, KEGG pathway analysis was used to identify genes overlapping with neurodegeneration-related pathways. This integrative analysis ultimately identified Dnah11 as a key candidate gene.

In the human genome, the Dnah11 gene encodes a component of the ciliary outer dynein arm and belongs to the dynein heavy chain family. It is primarily known for its role in regulating respiratory ciliary motility, with mutations linked to primary ciliary dyskinesia (PCD), congenital heart defects, heterotaxy syndrome, and asthenozoospermia (Dougherty et al. 2016; Kurokawa et al. 2022; Kurokawa et al. 2021). Beyond its canonical role, Dnah11, as a member of the dynein heavy chain family, is also involved in neuronal migration, axonal sprouting, and neurite outgrowth through cytoskeletal remodeling (Tang et al. 2019). Recent studies have reported upregulated Dnah11 expression in the hippocampus of AD model mice (Cai et al. 2023), suggesting a potential—though as yet unvalidated—mechanistic role in memory impairment. Concurrently, p50 dynamitin overexpression disrupts dynactin function in mice, providing evidence that impaired axonal transport contributes to neurodegeneration (Fujiwara et al. 2012). Additional studies have demonstrated that the hippocampus is functionally subdivided into dorsal and ventral regions, with the dorsal hippocampus predominantly mediating learning and memory processes (El Mahmoudi et al. 2023; Faust et al. 2021; Michael S and Hong-Wei 2010). Immunostaining has revealed that DNAH11 protein distribution aligns with granule cell localization, and hippocampal granule cells are primarily excitatory neurons (Zhang et al. 2022). In rats, granule cells in the dorsal hippocampus are critically involved in learning and memory (J et al. 1998), while the CA3 and CA1 regions are largely composed of pyramidal neurons—findings consistent with our observations (Zhang et al. 2022). Furthermore, studies have reported that HL induces hyperactivation of the central auditory pathways (Jafari et al. 2020; Park et al. 2020), suggesting a link between auditory dysfunction and hippocampal excitability. Our previous study demonstrated overactivation of CaMKIIα-positive excitatory neurons of the auditory cortex at 2 weeks post-NIHHL. To investigate whether this phenomenon also occurs in the hippocampus and characterize CaMKIIα and c-Fos expression patterns across neuronal subtypes, we performed immunofluorescence assays. These results confirmed that these neurons in the NIHHL group not only displayed elevated CaMKIIα expression but also exhibited marked hyperactivation. Hyperactivation of CaMKIIα-positive excitatory neurons in the NIHHL hippocampus exhibited concomitant DNAH11 upregulation, while non-CaMKIIα-positive excitatory neurons did not show increased DNAH11 expression. Notably, in GAD1-positive inhibitory neurons, no upregulation of DNAH11 expression and hyperactivation of inhibitory neurons was observed. While DNAH11 was ubiquitously expressed across all neurons, its expression was significantly increased only in hyperactivation of CaMKIIα-positive excitatory neurons, suggesting a potential functional correlation between CaMKIIα signaling and DNAH11 expression. Previous studies have also reported elevated hippocampal glutamate levels after chronic noise exposure (Bo et al. 2009), potentially associated with CaMKIIα-mediated neuronal hyperactivation. Increased glutamate may induce excitotoxicity, thereby contributing to cognitive impairment. It is worth noting that our study is the first to report the subcellular localization of DNAH11 in the central nervous system and its pronounced upregulation specifically in hyperactivation of CaMKIIα-positive excitatory neurons. However, the precise molecular mechanisms linking CaMKIIα signaling to DNAH11 function remain unclear.

The CA3 and CA1 hippocampal subregions play critical roles in memory encoding and retrieval (Mostafa et al. 2025; Paulson et al. 2025). To determine whether NIHHL-induced cognitive and auditory deficits could be reversed, we selectively knocked down Dnah11 in CaMKIIα-positive excitatory neurons located near these subregions. While memory function improved following knockdown, auditory function in NIHHL mice remained unaltered, despite a modest shortening of ABR wave I latency—a change that may be attributable to surgical variation rather than a true therapeutic effect. Notably, Dnah11 knockdown significantly ameliorated NIHHL-induced memory impairment, an unexpected but important finding. Control mice exhibited no cognitive deficits following knockdown, indicating that DNAH11 plays a pivotal mechanistic role in NIHHL-related memory impairment. This represents protein-level validation of the proposed link between DNAH11 and memory impairment. Previous studies have identified HL as a major risk factor for hippocampal synaptic degeneration and cognitive impairment (Chang et al. 2019; Mun et al. 2021; Shukla et al. 2019). Additionally, aberrant hippocampal activation may suppress protein synthesis and disrupt memory reconsolidation (Ressler et al. 2021). Degeneration of the central auditory pathways following HL may further sensitize hippocampal synapses to damage (Chang et al. 2019; Gröschel et al. 2018; Gröschel et al. 2010). Based on these findings, we observed significantly reduced expression of the synaptic plasticity proteins SYN and PSD95 in the hippocampus of NIHHL mice. These changes are consistent with NIHL models demonstrating cognitive impairment and synaptic dysfunction (Kurioka et al. 2021; Pan et al. 2024; Qianru et al. 2024). Notably, Dnah11 knockdown restored SYN and PSD95 levels to near-physiological levels, although the precise molecular mechanisms linking Dnah11 to synaptic plasticity and memory impairment require further investigation. Fortunately, transcriptomic profiling suggests that Dnah11 may participate in retrograde vesicular transport pathways, which have been implicated in human neurodegenerative diseases (Lee et al. 2024).

Concurrent studies have indicated BDNF-mediated synaptic development during retrograde transport of endosomes in dendrites is dynein-dependent (Moya-Alvarado et al. 2022). Given that DNAH11 is a component of the dynein motor complex, we hypothesize that it may serve as a regulatory component in this process. Our studies primarily focus on the impact of excessive upregulation of the DNAH11 protein on retrograde vesicular transport, which may disrupt the normal trafficking of BDNF. Given that BDNF is a critical neurotrophic factor involved in regulating synaptic plasticity (Moya-Alvarado et al. 2022), its aberrant expression in neurons may be a key molecular mechanism underlying memory impairment in NIHHL. We speculate that its overexpression may disrupt the precise trafficking of BDNF in hippocampal neurons, thereby impairing synaptic plasticity and leading to memory impairment. This provides a plausible mechanistic explanation for the observed reduction in synaptic plasticity proteins in the hippocampus of NIHHL mice. Collectively, these findings support the use of a well-characterized NIHHL model as a sensorineural HL paradigm to investigate the molecular pathways through which HL may promote AD-related dementia.

This study has some limitations. First, although this study demonstrated that Dnah11 mediates memory impairment in NIHHL, however, the relationship between DNAH11 upregulation in the hippocampus and hyperactivation of CaMKIIα-positive excitatory neurons remains poorly understood. Second, while Dnah11 knockdown improved cognitive outcomes, we did not assess the effects of Dnah11 overexpression, which may serve as a complementary model to mimic noise-induced molecular changes. This approach aligns with previous studies in which overexpression of target genes successfully recapitulated experimental phenotypes (Pan et al. 2024). These aspects will be addressed in future work.

Conclusions

In this study, we demonstrated that NIHHL induced significant memory impairment in mice at 1 month after exposure. Overexpression of DNAH11 in CaMKIIα-positive excitatory neurons of the hippocampus was associated with memory deficits and reduced expression of synaptic plasticity-related proteins. Crucially, Dnah11 knockdown in these neurons rescued memory dysfunction and restored SYN and PSD95 expression, although auditory function remained unimproved. These findings establish a foundation for further investigation into the Dnah11-mediated molecular mechanisms underlying NIHHL-associated cognitive impairment.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

AD

Alzheimer’s disease

ABR

Auditory brainstem response

BDNF

Brain-derived neurotrophic factor

CaMKIIα

Calcium/calmodulin-dependent protein kinase II alpha

c-Fos

Cellular FBJ osteosarcoma oncogene

CON

Control

CtBP2

C-terminal Binding Protein 2

DNAH11

Dynein axonemal heavy chain 11

DEG

Differentially expressed gene

EGFP

Enhanced green fluorescent protein

GAD1

Glutamate decarboxylase 1

GAPDH

Glyceraldehyde-3-phosphate dehydrogenase

GO

Gene ontology

HL

Hearing loss

IHC

Inner hair cells

KEGG

Kyoto Encyclopedia of Genes and Genomes

MWM

Morris water maze

NE

Noise exposure

NIHHL

Noise-induced hidden hearing loss

NIHL

Noise-induced hearing loss

NOR

Novel object recognition

PCD

Primary ciliary dyskinesia

PSD95

Postsynaptic density protein 95

rAAV

Recombinant adeno-associated virus

RT-qPCR

Reverse transcription quantitative polymerase chain reaction

SYN

Synaptophysin

Author contributions

XCW, TC, and XRW conceived and designed the experiments. YF, YHJ, and BW performed the majority of the experiments, including data collection and analysis, and revised the final version of the manuscript. MZ, JZ, BR, XTZ, and KW contributed to partial experimental work and provided valuable suggestions. JYZ, TC, ZYZ, and JQ evaluated the model, conducted hippocampal tissue sampling, and performed RT-qPCR experiments. All authors are responsible for the integrity of the data and the accuracy of the data analysis.

Funding

This study was supported by the National Natural Science Foundation of China (No. 82373610), the Key Research and Development Projects of Shaanxi Province (2024SF-YBXM-348), Joint Founding Project of Innovation Research Institute, Xijing Hospital (LHJJ24KQ04), and the Natural Science Basic Research Program of Shaanxi Province (2025JC-QYCX-069).

Data Availability

Raw RNA-seq data of all samples were uploaded to the Gene Expression Omnibus (GEO) of the National Center for Biotechnology Information (GSE300555), and RNA-seq data will be made publicly available on January 1, 2028.

Declarations

Ethics approval and consent to participate

All animal experiments conducted in this study were approved by the Institutional Animal Care and Use Committee of Xi'an Air Force Medical University, China (No. 20230375).

Consent for publication

Not applicable.

Clinical trial number

Not applicable.

Competing interests

The authors declare no competing interests.