Hearing loss drives region-specific amyloid accumulation along the auditory pathway in female 5xFAD mice

Abstract

Hearing loss is a major modifiable risk factor for Alzheimer’s disease (AD), but how auditory deprivation engages specific neural circuits, particularly in a sex-dependent manner, to promote amyloid pathology remains unclear. We examined whether sound deprivation alters amyloid-beta (Aβ) deposition, auditory neural activity, and cognition in 5xFAD mice. Notably, sound deprivation using bilateral earplugging from 2 to 4 months selectively increased Aβ plaque burden in the inferior colliculus (IC) of female 5xFAD mice, with no comparable effect in males. This IC-selective pathology was accompanied by increased c-Fos expression in the IC. Microglial density was unchanged, but earplugging increased the fraction of plaque-associated microglia and reduced microglia–plaque distance specifically in the IC, suggesting local amplification of AD-related pathology at the plaque interface. 5xFAD mice without deprivation retained intact novel object recognition (NOR) at 6 months. In contrast, female 5xFAD mice with prior earplugging developed a selective NOR deficit despite no detectable increase in hippocampal plaque burden, whereas males remained unaffected. Together, these findings identify the IC as a female-vulnerable locus where transient sound deprivation drives persistent auditory-circuit hyperactivity, remodels microglia–plaque interactions, and accelerates region-specific amyloid pathology with extra-hippocampal cognitive consequences, highlighting sensory loss as modifiable risk factor in prodromal AD.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1038/s41598-026-64191-7.

Introduction

Sensory impairments associated with aging have been implicated in accelerating neurodegeneration and promoting neuropathological changes, which in turn increase the risk of dementia, including Alzheimer’s disease (AD)¹. Among these, hearing loss is one of the most prevalent and early occurring sensory deficits in older adults and has been consistently linked to an elevated risk of AD². Importantly, hearing loss is the largest modifiable risk factor for dementia, and interventions such as a hearing aid or surgical treatment may help mitigate this risk³. Numerous epidemiological and longitudinal cohort studies have established a strong correlation between age-related hearing loss and cognitive decline, as well as an increased incidence of dementia^4,5. Furthermore, complementary studies using AD mouse models have shown that hearing loss exacerbates cognitive impairments and memory deficits, supporting a causal link between auditory dysfunction and AD progression^6,7. Sustained sensory input is a key driver of brain health: continuous afferent activity maintains synaptic strength, coordinated oscillations, and activity-dependent support that underlie attention and working memory. Indeed, one or more sensory impairments are significantly associated with an increased risk of dementia⁸. In particular, even relatively mild levels of hearing loss—reduced auditory input—can lead to cascading cognitive deficits in perception, comprehension, and memory⁹. However, the mechanistic basis for how reduced auditory input influences neurodegeneration remains unclear.

AD is a progressive neurodegenerative disorder characterized by cognitive decline and neuropathological features, including the accumulation of amyloid-beta (Aβ) plaques, tau tangles, and widespread neuronal loss, which progressively impair cognitive function. The earliest accumulation of Aβ preferentially starts within the highly active Default Mode Network (DMN), which is primarily composed of the posterior cingulate cortex, precuneus, and medial prefrontal cortex¹⁰. The regional vulnerability of this region to Aβ accumulation is largely attributed to its sustained high baseline neuronal activity and metabolic rate, as these DMN regions are highly connected network hubs that show patterns of Aβ accumulation closely aligned with the brain’s functional connectivity map¹¹. However, little is known about whether other brain regions also exhibit similar vulnerability in neural activity-dependent Aβ accumulation in response to changes in external sensory input. This study aims to explore whether alterations in sound-driven neuronal activity can exacerbate Aβ pathology and neurodegeneration in the auditory pathway.

Because sensory input is modifiable, it offers a mechanistic entry point to test reserve and resilience in AD. We therefore use a reversible earplug paradigm to reduce auditory input and determine whether decreased sensory drive is sufficient to exacerbate circuit dysfunction and cognition in an AD-vulnerable brain. Previous hearing loss models rely on irreversible cochlear damage, which is often accompanied by confounding factors such as oxidative stress, inflammation, and synaptopathy within the cochlea^12,13. To isolate the central effects of reduced sensory input, sound deprivation models that preserve cochlear integrity are required. Central auditory dysfunction has been observed in early-stage AD, implicating auditory processing changes as both markers and potential drivers of neurodegeneration¹⁴. In this context, reversible conductive hearing loss models using earplug insertion avoid direct cochlear injury and enable controlled examination of auditory deprivation. This approach allows us to dissect the role of sensory input in modulating AD-related pathology without peripheral inflammation confounds.

Overall, this study investigates the effects of sound deprivation on auditory processing and its contribution to AD progression in 5xFAD mice. By using a non-traumatic, reversible model of auditory deprivation in a genetically optimized strain, we aim to determine how reduced sensory input influences brain pathology and cognition. This work provides new insights into the interaction between auditory experience and neurodegenerative processes, with implications for early interventions in AD.

Results

5xFAD mice show heightened vulnerability to earplug-induced conductive hearing loss

To investigate the impact of auditory sensory deprivation on AD progression, we implemented a conductive hearing loss model in WT and 5xFAD mice by using earplugs at 2 months of age (Fig. 1a). We first validated the effectiveness of earplugging in inducing conductive hearing loss by measuring auditory brainstem response (ABR) thresholds. Before earplug insertion, both WT and 5xFAD mice (at 2-month-old) exhibited similar click-evoked ABR thresholds, indicating normal hearing and no genotype-dependent differences at baseline (female WT vs. 5xFAD; p = 0.9996, male WT vs. 5xFAD; p = 0.8125, Fig. 1b). Earplug insertion significantly elevated ABR thresholds in all groups relative to their pre-earplug values (female WT w/o vs. w/; p = 0.0003, female 5xFAD w/o vs. w/; p < 0.0001, male WT w/o vs. w/; p = 0.0003, male 5xFAD w/o vs. w/; p < 0.0001, Fig. 1b). These results confirm that earplug insertion effectively induced conductive hearing loss in all genotypes and both sexes.

Fig. 1.

Prolonged sound deprivation with earplugs impairs hearing sensitivity in 5xFAD mice. (a) Schematic diagram illustrating the earplug insertion procedure including insertion and removal, and experimental timeline for ABR, behavioral tests, and brain harvest. (b) Click-evoked ABR thresholds, measured in 2-month-old WT and 5xFAD mice before (without earplug, w/o) and after earplug insertion (with earplug, w/) in female (left, ♀, n = 8 ~ 10/group) and male (right, ♂, n = 7 ~ 9/group). (c) Representative click-evoked ABR traces in 5xFAD in female (left, ♀) and male (right, ♂). Threshold was determined as the lowest level (dB SPL, bold line) eliciting a distinct waveform. (d) Click-evoked ABR thresholds in WT and 5xFAD mice after earplug removal (at 4-month-old) following 2 months of earplug-induced sound deprivation (left; ♀, n = 11 ~ 17/group, right; ♂, n = 8 ~ 10/group). (e) Tone-evoked ABR thresholds across 4, 8, 16, 24, and 32 kHz frequencies (Each frequency comparison, 5x vs. 5x + EP; ^#p < 0.05). All data are presented as mean ± SEM (^*p < 0.05, ^**p < 0.01, ^***p < 0.001), (left; ♀, n = 11 ~ 17/group, right; ♂, n = 8 ~ 10/group).

Next, the effect of chronic conductive hearing loss was evaluated by measuring click-evoked ABR after 2 months of earplugging (Fig. 1c). Interestingly, long-term earplug experience induced changes in hearing sensitivity that differed between females and males. In WT mice, female EP mice showed significantly elevated ABR thresholds compared to untreated controls (Fig. 1d, WT; 30.8 ± 1.83 dB, EP; 47.1 ± 5.92 dB, WT vs. EP; p = 0.0412). In contrast, WT males show no significant change (WT; 31.0 ± 1.63 dB, EP; 33.5 ± 2.12 dB, WT vs. EP; p = 0.9660). However, 5xFAD mice of both sexes displayed robust elevations in ABR thresholds following chronic earplugging (Females: 5x; 29.6 ± 2.38 dB, 5x + EP; 60.4 ± 4.75 dB, 5x vs. 5x + EP; p < 0.0001, Males: 5x; 29.4 ± 3.20 dB, 5x + EP; 58.1 ± 7.79 dB, 5x vs. 5x + EP; p = 0.0002). To further evaluate auditory sensitivity across specific frequency ranges, we performed tone-evoked ABR measurement. In both females and males, 5xFAD + EP mice exhibited more pronounced high-frequency hearing loss compared to their respective controls (5xFAD without earplugs, Females: 5x vs. 5x + EP; p < 0.0001, Males: 5x vs. 5x + EP; p < 0.0001). In WT mice, only females with earplug experience showed elevated threshold (WT vs. EP; p = 0.0066). Taken together, these data indicate that the impact of earplug-induced sound deprivation is largely restricted to females in WT mice, but affects both sexes in 5xFAD mice, demonstrating that 5xFAD mice are particularly susceptible to prolonged conductive hearing loss.

Sound deprivation differentially alters neuronal activity along the auditory pathway in female 5xFAD mice

To investigate how prolonged sound deprivation alters neural activity in the auditory pathway, we quantified c-Fos immunoreactivity as a marker of neuronal activation in key auditory nuclei. Mice were exposed to 2 h of broadband sound prior to tissue collection, and the density of c-Fos–positive cells was calculated as the percentage of cFos+NeuN+ cells among total NeuN+ cells (cFos+NeuN+/NeuN+ × 100) in the medial nucleus of the trapezoid body (MNTB), IC, and auditory cortex (AC) (Fig. 2a). In female 5xFAD + EP mice, sound deprivation resulted in a decrease of c-Fos expression in the MNTB (Fig. 2b, 5xFAD; 67.1 ± 8.08%, 5xFAD + EP; 26.7 ± 4.77%, p = 0.0003), indicating neuronal activities were reduced by sound deprivation. Notably, in the IC, c-Fos+ cells were significantly increased (Fig. 2c and h, 5xFAD; 15.2 ± 2.79%, 5xFAD + EP; 31.7 ± 5.36%, p = 0.0179). These results suggest that reduced MNTB activity may disinhibit and thereby enhance neural activation in the IC, consistent with the predominantly inhibitory MNTB–IC projections. In the AC, c-Fos expression was not significantly altered by earplugging (Fig. 2d and h, 5xFAD; 39.7 ± 7.34%, 5xFAD + EP; 38.6 ± 5.82%, p = 0.7919). In male mice, this region-specific pattern of c-Fos expression was not observed by sound deprivation (Fig. 2e–h). Together, these findings indicate that sound deprivation induces region- and sex-specific alterations in neural activity along the auditory pathway. Notably, these changes are more pronounced in female mice, suggesting heightened susceptibility to auditory functional deficits and greater variability of neural activity in females.

Fig. 2.

Sound deprivation alters neuronal activity in the auditory pathway of female 5xFAD mice. (a) Schematic illustration of coronal brain section for major auditory structure, including medial nucleus trapezoid body (MNTB, ①), inferior colliculus (IC, ②), and auditory cortex (AC, ③). The region highlighted indicates the anatomical locations assessed for c-Fos protein expression. (b–g) Immunostaining of the MNTB (b, e), IC (c, f), and AC (d, g) was performed in 4-month-old 5xFAD mice with or without a 2-month ear-plugging (5x vs. 5x + EP). Images are shown for female (b–d) and male (e–g) mice. Immunolabeling for c-Fos (magenta), NeuN (green), and nuclei (DAPI, blue) in each region from 5xFAD and 5xFAD + EP mice. (h) Graphs show quantification of the number of c-Fos-positive cells normalized to NeuN-positive cells per slice. All data are presented as mean ± SEM, with each dot representing an individual tissue slice (^*p < 0.05, n = 2 ~ 5/group).

Sound deprivation exacerbates Aβ accumulation specifically in the IC of female 5xFAD mice

While previous work has described that elevated neural activity can accelerate Aβ production and AD pathology in vulnerable brain regions^15,16, it is still unclear whether such region-specific neural activity changes in the auditory pathway can affect the development of AD pathology. To address this, we examined the effect of prolonged sound deprivation on Aβ production and accumulation in 5xFAD mice and assessed its association with c-Fos expression changes. We performed immunostaining for Aβ using the 6E10 antibody and Thioflavin S staining for Aβ plaques. The 6E10 antibody recognizes the N-terminal epitope of Aβ (1–16), allowing the detection of both intracellular Aβ accumulation and extracellular plaque¹⁷. We quantified the total 6E10-positive (6E10+) area, as well as the number of 6E10 + deposits, including both 6E10 + cells and plaques, across the central auditory pathway. In parallel, Thioflavin S staining was specifically used to detect and quantify Aβ plaques within these regions. In 4-month-old 5xFAD 3mice without EP (both sexes), 6E10 + cells and plaque accumulation were observed in the AC, IC, but not the MNTB (Fig. 3a-c). Aβ + plaque detection in the auditory system in 5xFAD mice is comparable to previous studies¹⁸.

Fig. 3.

Sound deprivation increases 6E10 + deposits selectively in the IC of female 5xFAD mice. (a–c) Representative immunofluorescence images of 6E10 (magenta), microglia (Iba1, green), and nuclei (DAPI, blue) in the MNTB (a), IC (b, n = 7 ~ 10/group), and AC (c, n = 5 ~ 8/group) of female 5xFAD mice without (5x) and with (5x + EP) auditory deprivation via earplug insertion. Inset, widefield and high-magnification views. The graphs present quantification of 6E10 + area (normalized to total area, %), number of 6E10 + deposits (6E10 + deposit/mm²). (d–e) Thioflavin S staining (green) was used to detect Aβ plaques for IC and AC of female (d, n = 6 ~ 10/group) and male (e, n = 3 ~ 5/group) 5xFAD mice. Insets in all panels show magnified views of representative plaques. The graphs present quantification of Thioflavin S-positive plaque density (plaque number/mm²). All data are presented as mean ± SEM, with each dot representing a biological replicate (^*p < 0.05, ^**p < 0.01, ^***p < 0.001).

Next, we evaluated the effects of prolonged sound deprivation with EP on the Aβ accumulation across the central auditory pathway in female 5xFAD mice, IC exhibited a significant increase in 6E10 + area following sound deprivation in 5xFAD + EP (5x; 1.0 ± 0.16%, 5x + EP; 2.8 ± 0.49%, p = 0.0017), accompanied by a significant increased the number of 6E10 + deposit (5x; 318.1 ± 44.24/mm², 5x + EP; 636.7 ± 46.16/mm², p = 0.0001) (Fig. 3b). While the 6E10 antibody effectively detects both intracellular Aβ and extracellular deposits, it also can recognize APP-derived fragments, limiting its specificity for pathological Aβ plaques¹⁹. In the AC, prolonged sound deprivation does not substantially impact Aβ accumulation (Fig. 3c).

To further validate and assess specifically Aβ plaque formation, we utilized Thioflavin S staining, which selectively labels β-sheet-rich fibrillar Aβ plaques²⁰. Consistent with 6E10 staining results, Thioflavin S stained Aβ plaque number significantly increased in 5xFAD + EP mice (5x; 84.8 ± 17.31/mm², 5x + EP; 186.0 ± 44.45/mm², p = 0.0301, Fig. 3d). In the AC, sound deprivation had no significant effects on Aβ plaque accumulation in the AC of female 5xFAD mice (5x; 59.6 ± 7.93/mm², 5x + EP; 74.4 ± 7.91/mm², p = 0.2345), indicating that minimal impact on Aβ accumulation in the AC. Interestingly, in male 5xFAD mice, Aβ plaque deposition was sparse in both the IC and AC compared with females. Moreover, earplug-induced sound deprivation did not significantly alter Aβ plaque burden in either region, consistent with the absence of sound deprivation-induced changes in c-Fos expression, suggesting a relatively preserved neural activity profile in males (Fig. 3f–g). Together, these results indicate that sound deprivation selectively exacerbates Aβ pathology in the subcortical auditory brainstem, particularly the IC of female 5xFAD mice, highlighting region- and sex-specific vulnerability at an early stage of AD progression.

Sound deprivation selectively increases plaque-associated microglia in the IC

Activated microglia (Iba1 + cells) are known to cluster around Aβ plaques and modulate amyloid pathology²¹. To determine whether sound deprivation alters microglia characteristics, we examined microglial activation and spatial distribution in the auditory pathway. The total number or area of Iba1 + microglial cells did not differ between 5xFAD and 5xFAD + EP mice in both the IC and AC (Fig. 4a). Given that microglia in close proximity to Aβ plaques are referred to as plaque-associated microglia²², we further assessed their spatial proximity to plaques, defined microglia within 10 μm of an Aβ plaque as plaque-associated microglia (Fig. 4b)²³. Sound deprivation significantly reduced the average distance between microglia and Aβ plaques in both the IC (5x; 8.3 ± 0.30 μm, 5x + EP; 6.0 ± 0.28 μm, p < 0.0001) and AC (5x; 19.9 ± 0.80 μm, 5x + EP; 12.8 ± 0.37 μm, p < 0.0001, Fig. 4c), indicating closer microglial apposition to plaques. Notably, the proportion of plaque-associated microglia was increased selectively in the IC (5x; 68.9 ± 6.13%, 5x + EP; 90.2 ± 6.91%, p = 0.0175, Fig. 4d), paralleling the sound deprivation-induced increase in Aβ plaque burden in this region. These findings indicate that sound deprivation specifically enhanced microglia recruitment to plaques and Aβ accumulation in the IC, without causing a generalized increase in microglial density, pointing to region-selective modulation of AD-related pathology.

Fig. 4.

Sound deprivation-enhanced Aβ plaque recruits microglia near plaques. (a) Quantification of Iba1 + area (%), Iba1 + cell number (/mm²) in the inferior colliculus (IC, n = 7/group) and auditory cortex (AC, n = 5/group). (b) Representative image and analysis schematic. Aβ plaques (magenta) and Iba1 + microglia (green) were segmented; the shorter distance was calculated from each microglial centroid to the closest Aβ plaque. “Plaque-associated” microglia were defined as Iba1–Aβ distance < 10 μm. (c–d) Distribution of microglia–Aβ distances (c, violin plots), and percentage of plaque-associated microglia (d). All data are presented as mean ± SEM, and each dot represents an individual biological replicate (^*p < 0.05, ^***p < 0.001).

Anxiety-like behaviors and cognitive performance are unaffected by sound deprivation at an early disease stage

Given that sound deprivation-induced Aβ accumulation appears to be restricted to the auditory brainstem, we next investigated whether this region-specific pathology influences behaviors beyond auditory processing. To address this, we conducted Open field (OF) and Novel object recognition (NOR) tests in WT and 5xFAD following two months of earplug-induced sound deprivation (Fig. 5). In the OF test, average speed and total exploration distance were comparable across all groups, indicating that no detectable changes in locomotor activity at 4-month-old (Fig. 5b). Anxiety-like behaviors, such as freezing and preference for staying in edge areas, were also not observed in 5xFAD and 5xFAD + EP mice (Fig. 5a–d). Thus, prolonged sound deprivation did not alter motor or anxiety-like behaviors in any experimental group, both WT and 5xFAD. Cognitive performance assessed by the NOR test likewise showed no significant differences in discrimination ratios, indicating preserved recognition memory following sound deprivation in either female or male 5xFAD (Fig. 5e, Supplementary Fig.S1). Therefore, despite increased Aβ accumulation in the auditory brainstem of 5xFAD + EP mice, chronic sound deprivation via earplug insertion does not produce detectable deficits in motor, anxiety-like, or cognitive behaviors in WT or 5xFAD mice at this early stage of disease progression.

Fig. 5.

Recognition memory, motor, and anxiety-like behaviors remain unaffected after sound deprivation in 4-month-old mice. (a) Representative heatmaps show the distribution pattern in the OF box of female mice. (b, d) Quantification of motor and anxiety-like behaviors during the OF test (n = 6 ~ 11/group). Summary of the average speed (mm/s), total distance traveled (m), the number of frozen events for 10 min, and total frozen time (sec) (b), and the percentage of time spent in the center and edge areas (d). (c) Schematic illustration of the open field center and edge area parameter. (e) Summary of discrimination ratio in the NOR test, which is the exploration time with the novel object divided by the total interaction time with both objects (n = 7 ~ 11/group). All data is presented as mean ± SEM, and each dot represents an individual biological replicate.

Plaque accumulation persists and progresses in the auditory pathway after hearing restoration in 5xFAD mice

To evaluate the reversibility of auditory deficits and AD-related pathology following sound deprivation, we examined auditory function two months after earplug removal (at 6-month-old, Fig. 6a). At this age, both WT and 5xFAD mice of both sexes exhibited normal hearing without signs of age-related hearing loss, consistent with a previous study¹⁸. Click-evoked ABR thresholds in WT and 5xFAD mice with EP experience fully recovered to control levels, indicating complete restoration of hearing sensitivity (Supplementary Fig. S2a). Similarly, tone-evoked ABR thresholds were restored in both WT + EP and 5xFAD + EP mice, further confirming the recovery of hearing sensitivity following earplug removal. All groups exhibited normal hearing thresholds without signs of age-related hearing loss (Supplementary Fig. S2b).

Fig. 6.

Aβ plaque accumulation persists in the IC after restoration of auditory input in female 5xFAD mice. (a) Illustration of the experimental timeline. Female 5xFAD mice were subjected to bilateral auditory deprivation via earplug insertion from 2 to 4 months of age. Earplugs were removed at 4 months, and assessments including ABRs, behavior test, and brain harvest were performed at 6 months (b, c) In the IC (b) and AC (c), immunofluorescence staining (6E10: magenta; microglia/Iba1: green; DAPI: blue) and Thioflavin-S staining (green) show persistent Aβ pathology at 6 months in earplugged mice (5x + EP) compared to non-deprived controls (5x). Quantification includes 6E10 + area, 6E10 + deposit number (6E10 + deposit/mm²), and Thioflavin S positive plaque density (plaque number/mm²). Dark gray bars represent data from 6-month-old mice; light gray bars with crosshatching indicate 4-month-old data shown in Fig. 2, showing age-dependent progression. Note that each cohort represents independent groups of animals, not longitudinal measurements from the same individuals. All data is presented as mean ± SEM (^*p < 0.05) (n = 3 ~ 9/group).

We next examined whether this recovery of peripheral hearing after earplug removal altered the trajectory of Aβ accumulation in the auditory pathway of female 5xFAD mice at 6 months of age. As expected, Aβ 6E10 staining revealed that total 6E10 + area and the number of 6E10 + deposits exhibited age-dependent increases between 4 and 6 months in the IC and AC for 5xFAD and 5xFAD + EP mice (Fig. 6b and c).

In the IC, both 6E10 + area and number of 6E10 + deposits exhibit robust age-dependent increase in both 5xFAD and earplug-experienced (5xFAD + EP) mice (Fig. 6b). Aβ + area increased 2-fold from 4 to 6 months in the 5xFAD mice (4mo; 1.0 ± 0.49%, 6mo; 2.0 ± 0.44%, p = 0.0606), and 1.4-fold in the 5xFAD + EP mice (4mo; 2.8 ± 0.49%, 6mo; 3.8 ± 1.05%, p = 0.4517). Similarly, Aβ + deposits increased 1.7-fold in the 5xFAD mice (4mo; 318.1 ± 44.24/mm², 6mo; 525.3 ± 81.53/mm², p = 0.0770), and 1.1-fold in 5xFAD mice (5x + EP, 4mo; 5x + EP; 636.7 ± 46.16/mm², 6mo; 680.7 ± 69.30/mm², p = 0.6262). Notably, while 5xFAD mice showed a trend toward age-dependent increase in both Aβ + area and deposit number, 5xFAD + EP mice exhibited minimal change between 4 and 6 months, suggesting that Aβ burden in the 5xFAD + EP group had already mitigated due to the recovery of hearing. Although the 6E10 + area at 6 months remained 1.9-fold higher in 5xFAD + EP than 5xFAD controls, this difference was smaller than 2.8-fold increase seen in 4 months. The number of 6E10 + deposits also showed a greater relative increase in the EP mice at 4 months (~ 2-fold), which diminished to 1.3-fold at 6 months. These patterns suggest that the impact of prior sound deprivation on 6E10 + burden in the IC persists but becomes less pronounced after hearing recovery. In contrast, the AC displayed a similar age-dependent increase in 6E10 + area and deposit number in both groups, with no significant differences between 5xFAD and 5xFAD + EP mice at either age. Thus, while both regions exhibit age-related Aβ accumulation, the effects of earplug experience are region-specific and largely confined to the IC.

To specifically assess compact plaques, we analyzed Thioflavin S–positive deposits. Plaque numbers in the IC showed a trend toward age-dependent increase in both groups, although this did not reach statistical significance (5x; 4 mo vs. 6 mo, p = 0.1631, 5x + EP; 4 mo vs. 6 mo, p = 0.0870), indicating an ongoing age-dependent accumulation process. 5xFAD + EP mice exhibited a 2.4-fold higher plaque density than controls at 6 months (p = 0.0230). In contrast, AC also showed age-dependent increases in plaque number, but no significant differences between groups (p = 0.4200; Fig. 6c). Overall, these results suggest that while recovery of auditory sensitivity after earplug removal attenuates Aβ production or secretion, plaque accumulation persists and may even be exacerbated, particularly in the IC.

Sex-specific cognitive impairment after sound deprivation in 6-month-old 5xFAD mice occurs independently of hippocampal Aβ pathology

We next examined whether exacerbated Aβ plaque accumulation in the auditory pathway at 6-month-old impacts cognitive behaviors. The OF test revealed no age-related decline in motor behavior in either WT or 5xFAD mice across groups. Although Aβ plaque accumulation was significantly elevated in the auditory pathway in 5xFAD and 5xFAD + EP (at 6-month-old), anxiety-like behaviors were unchanged across groups (Fig. 7a–c). In the NOR task, 5xFAD mice without sound deprivation (EP experience) displayed normal discrimination ratios, indicating preserved recognition memory up to 6 months of age. In contrast, female 5xFAD + EP mice showed a significantly reduced discrimination ratio compared with other experiment groups (Fig. 7d, WT; 0.7 ± 0.02, EP; 0.8 ± 0.04, 5x; 0.8 ± 0.03, 5x + EP; 0.5 ± 0.07, WT vs. 5x + EP; p = 0.0136, EP vs. 5x + EP; p = 0.0025, 5x vs. 5x + EP; p = 0.0028).

Fig. 7.

Persistent AD progression following sound deprivation leads to recognition memory deficits in female 5xFAD mice. (a) Representative heatmaps show the distribution pattern in the open field box of female mice. (b, c) Quantification of motor and anxiety-like behaviors during the OF test (n = 7 ~ 9/group). Summary of the average speed (mm/s), total distance traveled (m), the number of frozen events for 10 min, and total frozen time (sec) (b), and the percentage of time spent in the center and edge areas (c). (d) Summary of discrimination ratio in the NOR test, which is the exploration time with the novel object divided by the total interaction time with both objects (n = 7 ~ 9/group). (e) Representative hippocampal images of Thioflavin S staining (green), a marker of Aβ plaques, in 6-month-old female 5xFAD mice with or without sound deprivation (n = 3 ~ 10/group). Quantification of Thioflavin S-positive plaque density (plaque number/hippocampus). All data are presented as mean ± SEM, and each dot represents an individual biological replicate (^*p < 0.05, ^**p < 0.01).

To determine whether this cognitive impairment was associated with increased hippocampal Aβ pathology, we examined hippocampal Aβ accumulation using Thioflavin S staining. There was no difference in Aβ plaques in the hippocampus between female 5xFAD and 5xFAD + EP mice (Fig. 7e: 5x; 26.9 ± 3.92/hippocampus, 5x + EP; 21.2 ± 7.95/hippocampus, p = 0.3520), indicating that the observed behavioral deficits are unlikely to result from hippocampal Aβ pathology. These findings suggest that in females, Aβ accumulation within the auditory brainstem, most prominently the IC, may contribute to early-stage recognition memory impairment following sound deprivation. By contrast, male 5xFAD + EP mice, which did not show earplug-induced increase in Aβ plaque along the auditory pathway (Supplementary Fig. S3a and b), exhibit no deficits in NOR performance, with discrimination ratio comparable across groups (Supplementary Fig. S4d: WT; 0.7 ± 0.02, EP; 0.7 ± 0.10, 5x; 0.7 ± 0.01, 5x + EP; 0.6 ± 0.05, 5x vs. 5x + EP; p = 0.2434). These findings highlight a female-specific susceptibility to sensory deprivation-induced AD pathology and recognition memory decline.

Discussion

Comparison of various models for hearing loss in the AD research

Despite numerous cohort studies demonstrating the relationship between hearing loss and dementia^3–5,24, the underlying mechanisms by which auditory dysfunction impacts neurodegeneration are poorly understood. Animal models have been instrumental in elucidating the relationship between hearing loss and AD, offering insights into potential mechanisms and therapeutic targets. Age-related hearing loss has been reported in several AD mouse lines^18,25. For example, the APP^NL−G−F mouse model exhibits progressive, high-frequency hearing loss and short-term memory deficits, reinforcing the potential link between auditory dysfunction and cognitive decline²⁶. However, these models have limitations in establishing causal directionality, as it remains unclear whether age-related hearing loss precedes or follows AD pathology. Other in vivo models, such as noise-induced hearing loss, cochlear or tympanic membrane damage^6,27, inevitably cause physical damage to the cochlea, leading to secondary effects such as inflammation and oxidative stress^12,28. For example, noise-induced hearing loss in 3xTg-AD mice has been shown to accelerate hippocampal pathology, including increased tau phosphorylation, increased neuroinflammation, redox imbalance, and synaptic dysfunction in the AC⁶. It has been suggested that noise exposure, acoustic trauma, and aging may induce widespread alterations in brain functions, beyond the sound-evoked neural activities in the auditory system. To specifically investigate the direct effects of auditory sensory deprivation on AD pathology, while avoiding confounding peripheral damage, we have established a sound deprivation model using bilateral earplug insertion for 2 months in young adult mice. Importantly, this earplug-induced sound deprivation did not result in cochlear impairments or neuroinflammatory changes in the auditory pathway, evidenced by the absence of microglial activation, assessed by Iba1 immunostaining. Thus, our approach provides a physiologically relevant model to examine the central consequences of auditory sensory deprivation on AD-related neuropathology.

Region-specific Aβ accumulation and long-lasting effects on AD pathology following auditory deprivation

Our results with c-Fos immunostaining demonstrate that auditory deprivation leads to distinct changes in neural activity within the auditory circuitry of female 5xFAD mice. Specifically, prolonged sound deprivation resulted in a pronounced increase of neuronal activity in the IC (Fig. 2), which was paralleled by a region-specific increase in 6E10 + deposition (Fig. 3). This localized vulnerability of Aβ plaque accumulation is consistent with the findings that hyperactivity promotes Aβ plaque production and aggregation, as previously reported^15,29. Previous studies show that sensory deafferentation (e.g., cochlear ablation) introduces synapse reorganization in the auditory brainstem, leading to an imbalance between excitatory and inhibitory synapses and a shift circuits toward excitation³⁰. Additionally, sound deprivation alters neurotransmitter release and uptake in a region-dependent manner throughout the brain³¹, suggesting that compensatory homeostatic plasticity may underlie these circuit-level changes³². Together, disrupted neural activity and excitatory-inhibitory imbalance may create a permissive environment for Aβ accumulation.

Importantly, we observed that subcortical 6E10+ Aβ/APP-related signal elevations in the IC persisted even after auditory input was restored via earplug removal (Fig. 6; Supplementary Fig. S2). It indicates that prolonged sensory deprivation may initiate neuropathological processes that are not fully reversible. While the reinstatement of auditory stimuli likely reduces ongoing Aβ production by normalizing neural activity levels, it appears insufficient to eliminate or reverse the accumulation of pre-existing plaques. This aligns with previous findings that neural activity influences the formation and aggregation of new Aβ plaques but does not significantly enhance the clearance of existing deposits²⁹. Thus, the persistent Aβ burden observed in the IC of 5xFAD mice following earplug removal suggests that once a critical threshold of Aβ pathology is reached, subsequent normalization of sensory input may be unable to fully reverse the damage. These findings highlight the potential importance of early sensory input and neural activity homeostasis in modulating regional vulnerability to Alzheimer’s pathology and may have implications for therapeutic strategies targeting early circuit-level changes in disease progression^33,34.

Auditory deprivation impairs recognition memory independent of hippocampal Aβ pathology

In addition to region-specific Aβ accumulation, we observed that sound deprivation induced recognition memory deficits in female 5xFAD mice, even though hippocampal Aβ deposition remained unchanged at this age (Fig. 7). This finding is particularly noteworthy, as 5xFAD mice typically exhibit preserved recognition memory performance up to six months of age under normal auditory conditions^35,36. The hippocampus and cortex, while classically implicated in memory consolidation, are not the sole substrates for recognition memory. Auditory input is known to influence hippocampal neurogenesis, synaptic plasticity, and connectivity^37,38. Moreover, the hippocampus itself participates in auditory processing through multisensory integration pathways^39,40. However, in our model, sound deprivation led to Aβ accumulation predominantly in subcortical auditory regions rather than the hippocampus or cortex. Therefore, it is plausible that altered auditory input, particularly subcortical regions such as the IC, may disrupt recognition memory through indirect effects on hippocampal function or associated cognitive networks. This could occur via impaired long-range connectivity or disruption of neuromodulatory systems that support memory processing⁴¹. Additionally, recognition memory measured by NOR task depends not only on the hippocampus but also on a broader network including interconnected regions such as the perirhinal cortex, prefrontal cortex, and thalamic structures^42,43.

Together, these findings raise the possibility that subcortical AD pathology, exacerbated by sound deprivation, may impair recognition memory through either a direct effect on cognitive circuits or indirectly disrupting auditory-hippocampal connectivity. Further studies are needed to determine whether subcortical neurodegeneration directly contributes to memory deficits or whether it indirectly alters hippocampal processing through multisensory circuit disruption.

Sex differences in auditory deprivation-induced AD pathology

Sex differences were strikingly apparent in our model of auditory deprivation-induced AD pathology. Following sound deprivation, female 5xFAD mice exhibited pronounced auditory functional deficits and region-specific increases in Aβ accumulation within subcortical auditory regions, while male counterparts exhibited no significant changes (Fig. 3). This sex-specific outcome is particularly noteworthy considering previous studies suggesting that females are generally less susceptible to peripheral auditory damage, such as age-related hearing loss and noise-induced hearing loss compared to males. Both epidemiological and animal model studies have demonstrated that females typically exhibit a slower progression of age-related hearing loss and less cochlear damage following noise exposure^44–46. For instance, studies using CBA mice have shown that females maintain better auditory thresholds and cochlear integrity with age and after noise trauma, potentially due to the protective effects of estrogen on cochlear structures and auditory pathways^47,48. Similarly, in human cohorts, men consistently show more rapid high-frequency hearing decline than women^49,50.

Paradoxically, recent findings suggest that despite this relative peripheral resilience, women may be more vulnerable to the central and cognitive consequences of hearing impairment. A large-scale analysis using UK Biobank data revealed that women with hearing impairment experienced greater brain atrophy in auditory and limbic regions and faced a high risk of developing dementia, compared to men with similar hearing thresholds⁵¹. These observations point toward sex-specific neurobiological vulnerabilities. Estrogen and other sex hormones are known to modulate key processes implicated in neurodegeneration, including glial activation, synaptic plasticity, and Aβ metabolism^52–54. In particular, previous studies have demonstrated that estrogen receptor signaling regulates auditory-driven neuronal responses in the central auditory pathway through attenuated GABAergic inhibition⁵⁵. This suggests that estrogen may influence the stability of neuronal activity and increase susceptibility to neurodegenerative disease.

Notably, although both male and female 5xFAD + EP mice also exhibited elevated ABR thresholds compared to 5xFAD following earplug insertion, but c-Fos expression following auditory stimulation diverged markedly between sexes. In female 5xFAD + EP mice, region-specific changes were observed, including decreased c-Fos in the MNTB and increased in the IC, whereas male mice showed no significant differences in c-Fos levels across any auditory brain regions. This suggests that despite peripheral auditory deprivation, male mice may maintain relatively stable central auditory circuit activity, potentially due to sex-hormone-mediated differences in GABAergic inhibitory signals or synaptic plasticity within the central auditory pathway.

Collectively, these findings suggest a dissociation between peripheral auditory resilience and central vulnerability in females. While women may be relatively protected from cochlear damage, they appear to be more sensitive to the downstream neurodegenerative impacts of auditory deprivation. A more comprehensive understanding of the mechanisms underlying this sex-specific susceptibility is essential for the development of personalized preventive and therapeutic strategies targeting AD and related dementias.

In summary, our results demonstrate that auditory deprivation changes in neuronal activity, resulting in persistent, region-specific Aβ accumulation in subcortical brain areas and contribute to cognitive decline, even after hearing restoration. These findings highlight the importance of early intervention to prevent irreversible neuropathological changes.

Materials and methods

Animals

All mouse procedures in this study complied with ethical guidelines and protocols approved by the Institutional Animal Care and Use Committee of the University of Michigan (PRO00011243). The mouse strain used for this research project, B6.Cg-Tg(APPSwFlLon, PSEN1*M146L*L286V)6799Vas/Mmjax, RRID: MMRRC_034848-JAX (5xFAD mice), was obtained from the Mutant Mouse Resource and Research Center (MMRRC) at The Jackson Laboratory, an NIH-funded strain repository, and was donated to the MMRRC by Robert Vassar, Ph.D., Northwestern University¹⁷. To eliminate the effects of age-related hearing loss in the C57BL/6 background of 5xFAD mice, we used F1 hybrid mice generated from a cross between male heterozygous 5xFAD mice and female WT CBA/CaJ (Jax, #000654)^56,57. This hybrid (C57BL/6 × CBA/CaJ, F1) maintains stable auditory thresholds beyond 14 months of age, enabling assessment of auditory deprivation independent of age-related hearing loss⁵⁶. The mice were maintained under standard ventilated conditions under ARRIVE guidelines.

Experimental design

Both male and female WT and 5xFAD mice were randomly assigned to control or sound deprivation groups. All the experiments were performed in a randomized order across groups to eliminate bias. Outcome assessments were conducted in a blinded manner. Earplugs were inserted as described⁵⁸. To bypass confounding age-related hearing loss and intervene before the onset of AD-related pathology, we chose 2-month-old mice with fully developed hearing for the earplug induced sound deprivation model^17,56. For earplug insertion, 2-month-old mice were anesthetized with isoflurane, and small pieces (3 mm diameter × 3 mm length) cut from a human foam earplug were inserted into both external auditory meatus and sealed with silicone. The earplugs were replaced weekly with fresh ones, with at least 1 h of ventilation between changes. Earplugs were maintained for 2 months, and depending on the experimental procedure, mice were harvested either immediately after earplug removal or following a 2-month recovery period. Mice were anesthetized with an overdose of isoflurane, followed by transcardial perfusion with PBS.

In vivo auditory brainstem response (ABR) test

ABR recordings were performed as described⁵⁹. Briefly, mice were anesthetized with 2.5% isoflurane and maintained throughout the recordings. During the procedure, the mice’s body temperature was maintained at 36.5 ~ 37.5 °C using a heating pad. ABR responses were recorded in a soundproof box. The acoustic stimuli were delivered to one ear via an 8 cm tube connected to calibrated speakers. The signals were recorded with subdermal needle electrodes, which were placed at the vertex of the skull (active electrode), the ipsilateral mastoid (reference electrode), and the contralateral mastoid (ground electrode). The acoustic stimuli range from 90 to 10 dB, decrements of 5 dB, 500 stimulus repetitions were recorded using an Auditory Evoked Potentials Workstation (Tucker Davis Technologies, FL). The auditory threshold was defined as the lowest stimulus intensity that evoked a reproducible ABR waveform. The amplitude of each wave was measured by the trough-to-peak amplitudes at 70 dB and 80 dB SPL. Tone burst stimuli were delivered at frequencies of 4, 8, 16, 24, and 32 kHz, ranging from 90 to 10 dB with 10 dB step decrements. The threshold was determined as performed in the click-evoked ABR.

Sound stimulation

Mice were exposed to acoustic stimulation consisting of 80 dB broadband sound (8–24 kHz, centered at 16 kHz) for 2 h before brain harvest. All groups removed earplugs before sound stimulation. Each mouse was individually placed in a partitioned mesh compartment located 10 cm from the speaker within a sound-attenuating chamber (Med Associates, VT). Stimulation was delivered by an Auditory Evoked Potentials Workstation (Tucker Davis Technologies, FL). Sound signals consisted of a series of amplitude-modulated square waves (duration, 0.1 ms; repetition rate, 16/sec) generated and presented through multifield magnetic speakers (Tucker Davis Technologies, FL).

Immunofluorescence staining

The Aβ formation and plaques were assessed using anti- Aβ (6E10) immunofluorescence or Thioflavin S staining. Mouse brains were collected following cardiac perfusion with PBS and 4% paraformaldehyde. The tissues were cryoprotected with sucrose and embedded in OCT (Fisher Scientific, #23730571) and sectioned at a thickness of 30 μm using cryostat (Leica, CM3050). For staining, free-floating sections were blocked with 10% goat serum (Sigma, G9023) in 0.1% PBS-T solution, then incubated with primary antibodies against anti- Aβ (6E10) (Biolegend, #803001), c-Fos (Proteintech, #66590-1-Ig), Iba1 (Wako, #019–19741), and NeuN (Millipore, MAB377) antibodies. After rinsing with 0.1% PBS-T, the sections were incubated with goat anti-mouse IgG1 conjugated to Alexa Fluor 647 (Fisher Scientific, #A21240) or anti-rabbit IgG conjugated to Alexa Fluor 488 (Fisher Scientific, #A11008). Following immunostaining, sections were mounted on slides and coverslipped with a mounting medium containing DAPI (Fisher Scientific, #50–112-8966). For Thioflavin S staining, sections were mounted on coated slides and dried thoroughly. The tissues were gradually rehydrated through a graded ethanol series, each for 3 min. Subsequently, sections were incubated in a 1% aqueous Thioflavin S solution (Fisher Scientific, #ICN15071105) for 10 min at room temperature. After rinsing with distilled water, the tissues were dehydrated through a graded ethanol series and coverslipped. Fluorescence images were acquired using a confocal microscope (Zeiss, #LSM980) or, for the large-field images, a super-resolution microscope (Nikon, #N-SIM S). Quantification of Aβ (6E10), Iba1, and Thioflavin S was performed using FIJI/ImageJ (ver.1.54q) and Python and normalized to tissue area and expressed as plaque density (numbers/mm²).

Open field (OF) test

The OF test was performed to assess locomotor activity and anxiety-like behaviors. Prior to the test, mice were acclimated to the testing room for at least 1 h. Then, each mouse was placed in the center of a square testing box (40 cm × 40 cm × 30 cm) and allowed to explore freely for 10 min. The behavior was recorded using an overhead camera. The average speed, total distance traveled, the number of freezing episodes, and the percentage of time spent in the center and edge areas were quantified. The testing box was thoroughly cleaned with 70% Ethanol between trials. The recordings were analyzed using ToxTrac software⁶⁰.

Novel object recognition (NOR) test

The NOR test was performed in the same arena as the OF test on the following day, to evaluate recognition memory. The test was conducted in the same testing box over two consecutive days and consisted of two phases: training and testing. During the training phase, two identical objects were placed in the box, and each mouse was allowed to explore for 10 min. On the following day, one object was replaced with a novel object that differed in shape and color. Mice were again allowed to explore for 10 min. The discrimination ratio was calculated as the time spent exploring the novel object divided by the total time spent exploring both objects.

Statistical analysis

All the graphs were generated and statistically analyzed using GraphPad Prism software. Outliers were screened using the ROUT method (Q = 1%) during the data analysis phase. No data points that were identified as statistical outliers, and thus no data was excluded from the final analysis. Graphs display individual data points (biological replicates) as dots, with bars or lines representing the mean ± standard error of the mean (SEM). The normality of data distribution was tested using the Shapiro-Wilk test. For data sets with two groups, comparisons were performed using either parametric (unpaired Student’s t-test) or non-parametric (Mann-Whitney U test) tests, as appropriate, based on data normality. For ABR threshold analyses, differences between pre- and post-earplug conditions (Fig. 1b), were assessed using a paired t-test, where each animal served as its own control. For comparisons among all groups data for ABR and behavioral test (OF, NOR), one-way ANOVA followed by Tukey’s post hoc test was performed. For tone frequency-dependent ABR analyses, a two-way repeated measures ANOVA followed by Tukey’s post hoc test was performed. For behavioral analyses including ABR and NOR test, a post hoc power analysis was conducted using G*Power (version 3.1.9.7) to confirm the adequacy of the sample sizes used. For paired t-test analyses, effect size was calculated as Cohen’s d, derived by dividing the mean of differences by the standard deviation of differences obtained from GraphPad Prism. For one-way or two-way ANOVA analyses, effect sizes were calculated from eta squared (η²) obtained from GraphPad Prism. Power analysis was performed for analyses that yielded statistically significant results. Across all significant analyses, the power analysis confirmed that all groups achieved a statistical power exceeding 0.8 at α = 0.05, supporting the statistical reliability of the sample sizes used in this study.

For immunofluorescence data sets with more than two groups that met the assumption of normality, parametric tests (one-way or two-way ANOVA) were used, followed by Tukey’s post hoc test for multiple comparisons. For non-normally distributed data, non-parametric tests were performed, and post hoc comparisons were corrected for multiple testing using the Benjamini, Krieger and Yekutieli method to control the false discovery rate (FDR). For all analyses, p-values < 0.05 were considered significant.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

J.H.K. and J.P. contributed to the conception and design of the study. J.P. performed all experiments, analyzed data, and prepared the manuscript. S.Y.J. performed immunohistochemistry. J.H.K. supervised the experiment, prepared the manuscript, and approved the submission.

Funding

This work was supported by the National Institute of Health R01 DC018797; and Alzheimer’s Association 23AARG-NTF-1026470.

Data availability

All data generated or analyzed during the study are included in this published article.

Declarations

Competing interests

The authors declare no competing interests.