Adriamycin nephropathy induces sensorineural hearing loss via blood–labyrinth barrier breakdown in BALB/c mice

Abstract

Sensorineural hearing loss (SNHL) is significantly more prevalent in individuals with chronic kidney disease (CKD) than in the general population. Although a strong independent association has been observed between kidney dysfunction and the risk of hearing loss, the underlying mechanisms linking these conditions remain poorly understood. This study investigated the pathophysiology of hearing impairment using adriamycin nephropathy (AN), a well-established animal model of CKD. AN was induced in male BALB/c mice using 10 or 12 mg/kg of adriamycin (ADR), resulting in severe kidney injury and concurrent hearing loss. ADR-treated mice exhibited significant glomerular injury, podocyte damage, and elevated renal neutrophil gelatinase-associated lipocalin (NGAL), along with increased serum creatinine and blood urea nitrogen levels. Hearing impairment was evident after 4–8 weeks of ADR treatment, as assessed by auditory brainstem response and distortion-product otoacoustic emissions, and was accompanied by cochlear hair cell loss and ribbon synapse degeneration. AN affected cochlear function by altering ion channel expression in the stria vascularis and inducing blood–labyrinth barrier (BLB) hyperpermeability, along with changes in endothelial cells, pericytes, and perivascular-resident macrophage-like melanocytes. AN significantly increased cochlear NGAL and NLRP3 levels at 4 and 8 weeks following ADR administration. NGAL was highly expressed in the tectorial membrane, cochlear neurons, and organ of Corti, while its receptor, 24p3R, was co-localized with NGAL. These findings demonstrated the role of NGAL and BLB disruption in ADR-induced SNHL, providing novel insights into the mechanistic link between CKD and hearing loss.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-32154-z.

Introduction

Hearing loss is defined as an impairment in auditory function, characterized by a sound threshold of ≥ 20 dB in both ears, and represents the third most common cause of disability worldwide¹. Currently, approximately 1.6 billion individuals are affected by hearing loss globally, a figure projected to rise to nearly 2.5 billion by 2050^1–3. This condition is associated with an increased risk of anxiety, depression, dementia, and diminished quality of life, potentially mediated by heightened social isolation and structural changes in the brain related to auditory dysfunction^4,5.

Chronic kidney disease (CKD) and hearing loss share several contributing factors, including the use of ototoxic drugs, hypertension, diabetes, and electrolyte imbalances⁶. At the cellular level, both the kidneys and the cochlea rely on ciliary activity at the apical surface for normal function. Additionally, kidney tubular cells and sensory epithelial cells of the inner ear utilize similar transport mechanisms to regulate luminal fluid composition. Both organs also possess an identical collagen IV basement membrane network, and their stria vascularis exhibit ultrastructural and physiological similarities, which may explain the observed association between hearing impairment and CKD⁷. Although a strong and independent relationship has been established between kidney function and the risk of hearing loss, research exploring the underlying mechanisms remains limited and underfunded.

Analogous to the blood–brain barrier in the brain, the blood–labyrinth barrier (BLB) in the inner ear functions as a selective barrier⁸. The BLB, composed of tightly joined endothelial cells, pericytes, and perivascular macrophage-like melanocytes (PVM/Ms), is essential for maintaining ionic balance and immune protection in the inner ear^8–10. Recent studies have demonstrated that increased BLB permeability facilitates the entry of harmful substances into the inner ear, impairing auditory function^11–15. This permeability, often referred to as a “leaky BLB,” can be induced by various factors, including inflammation, noise exposure, ototoxic drugs, oxidative stress, and ischemic conditions^11–13,16. These disruptions compromise endothelial tight junctions, permitting the infiltration of toxic substances into cochlear tissues^17–19. Such infiltration is known to cause damage to cochlear ribbon synapses, which are crucial for transmitting auditory signals from hair cells to auditory nerve fibers. This process can lead to synaptopathy, hidden hearing loss, and auditory nerve degeneration^20–22. While the impact of specific substances, such as neutrophil gelatinase-associated lipocalin (NGAL), on the blood–brain barrier is well established, the role of NGAL in BLB dysfunction remains inadequately explored and warrants urgent investigation.

The present study aimed to elucidate the intersection between two strongly associated yet understudied conditions, CKD and hearing loss by examining the mechanisms through which adriamycin nephropathy (AN) affects cochlear structure and auditory function. Furthermore, we sought to identify the mechanisms by which nephropathy-induced alterations in the BLB contribute to hearing impairment. We employed an AN model (administering 8 or 12 mg/kg of adriamycin [ADR]) using 6-week-old male Balb/c mice. To the best of our knowledge, ADR ototoxicity has either not been previously detected or has been considered an uncommon phenomenon²³. This allowed us to specifically investigate AN-induced ototoxicity without the confounding effects of direct ADR-induced auditory toxicity. Our findings demonstrated that ADR administration led to kidney disease phenotypes, including glomerulosclerosis, tubulointerstitial fibrosis, podocyte injury, and elevated levels of blood urea nitrogen (BUN), serum creatinine, and renal NGAL. Four to eight weeks after ADR administration, hearing impairment was observed alongside significant BLB disruption and increased levels of cochlear NGAL and its receptor, 24p3r. These results strongly suggested that CKD-associated toxic substances contribute to BLB hyperpermeability, leading to disruptions in BLB components including endothelial cells, pericytes, and PVM/Ms. In turn, this accelerates the infiltration of nephropathy-related toxins into the inner ear, ultimately resulting in auditory dysfunction. To the best of our knowledge, this was the first study to propose an association between NGAL, BLB dysfunction, and hearing impairment in the context of CKD. Our findings suggested that targeting NGAL may serve as a potential strategy for mitigating common morbidities in patients with CKD and hearing loss.

Results

AN induction

AN is induced in animal models, including rats and mice, through a single injection of ADR^24–26, which directly damages the kidneys and leads to chronic proteinuric renal disease, resembling human focal segmental glomerulosclerosis²⁴. To establish an AN/CKD model, 6-week-old male Balb/c mice were administered a single intraperitoneal injection of ADR at doses of 10 or 12 mg/kg, while the control group received 0.09% saline as a vehicle (Fig. 1A). Baseline hearing sensitivity was assessed in 5-week-old male BALB/c mice prior to ADR administration using auditory brainstem response (ABR) testing.

Fig. 1.

Adriamycin (ADR) induces sustained renal injury. (A) Schematic of the experimental design. One dose of ADR (10 or 12 mg/kg) was administered i.p. to male Balb/c mice aged 6 weeks. Samples were collected at 4- or 8-weeks after ADR injection and proceeded for biochemical analysis. (B, C) Renal function in ADR-treated mice. Serum creatinine (A) and BUN (B) levels were measured at 4 wk or 8 wk after ADR administration. Data are expressed as mean ± SEM (n = 6). Shapiro–Wilk test for normal distribution (P > 0.05); one-way ANOVA; Tukey’s multiple comparisons test. *P < 0.05; **P < 0.01; ***P < 0.001; ***P < 0.0001. (D) Kaplan–Meier graphs show survival of BALB/c mice treated with a single dose of 0 mg/kg (vehicle-treated), 10 mg/kg, and 12 mg/kg ADR. n = 26, 0 mg/kg; n = 40, 10 mg/kg; n = 20, 12 mg/kg *P < 0.0001.

At the fourth and eighth weeks following ADR administration, kidney function was assessed by measuring serum creatinine and BUN levels, which served as biomarkers of renal injury. ADR-treated mice exhibited significantly elevated serum creatinine (Fig. 1B) and BUN levels (Fig. 1C), confirming kidney dysfunction. A Kaplan–Meier survival curve revealed a time- and dose-dependent response to ADR (Fig. 1D). Mice receiving 12 mg/kg of ADR demonstrated a mortality rate > 50% higher than that of the control group. Therefore, all mice in the 12 mg/kg group were sacrificed at 4 weeks after ADR administration and were excluded in the later time points of this study (indicated by the gray arrow in Fig. 1A). The animals from all other experimental groups were included in the analysis.

Histological analysis using hematoxylin and eosin staining demonstrated significant kidney injury in ADR-treated mice, characterized by glomerular damage, tubular injury, interstitial fibrosis, and brush border detachment (Fig. 2A). Transmission electron microscopy revealed significant mitochondrial damage in the glomeruli of ADR-treated mice (10 or 12 mg/kg), evidenced by rounded mitochondria with disrupted cristae and vacuolization, compared to controls (0 mg/kg) (Fig. 2B). Furthermore, transmission electron microscopy demonstrated extensive foot process effacement and swollen, fused podocyte pedicles in ADR-treated mice (Fig. 2C). These findings indicated that the AN model successfully replicated key features of renal tissue damage and associated pathological changes observed in human kidney disease.

Fig. 2.

ADR induces glomerular injury and podocyte damage. (A) Hematoxylin and eosin staining of kidney sections from vehicle- and ADR-treated mice. At weeks 4 and 8, the glomeruli (*) exhibit segmental sclerosis and vacuolization. The tubulointerstitium at these time points shows evidence of epithelial cell damage (#), brush border detachment (#), swollen epithelial cells (@), and vacuolar degeneration (@). (B) Representative transmission emission tomography (TEM) images of glomeruli reveal significant mitochondrial damage in ADR-treated mice compared to controls. Round mitochondria with disrupted cristae and vacuolization are evident in ADR-affected samples (scale bar = 2 µm). (C) Representative TEM images of kidney tissue from vehicle- and ADR-treated mice. Control glomeruli display well-organized interdigitating foot processes and pedicles in podocytes, whereas ADR-treated mice (12 mg/kg, 4 weeks; 10 mg/kg, 8 weeks) exhibit extensive foot process effacement and swollen, fused pedicles (yellow arrows) (scale bar = 1 µm).

AN impaired hearing function

Linear regression analysis adjusted for age and sex demonstrated that increased serum creatinine levels and elevated blood pressure were both positively associated with higher hearing thresholds, identifying them as risk factors for hearing impairment in CKD²⁷. To investigate the impact of CKD/AN on the inner ear, we examined cochlear structures and hearing function in the AN mouse model.

The inner ear, responsible for hearing and balance, comprises the cochlea and vestibule (Fig. 3A). The cochlea, a spiral-shaped structure, processes sound through its apical and basal turns. Within the cochlea, outer hair cells (OHCs) amplify sound vibrations, while inner hair cells (IHCs) transmit auditory signals to the cochlear nerve via spiral ganglion neurons (Fig. 3A). The stria vascularis, a multilayered epithelial structure located in the cochlear lateral wall, consists of endothelial cells, pericytes, and perivascular macrophages. As part of the BLB, the stria vascularis plays an important role in maintaining the ionic composition of cochlear fluid, ensuring proper hair cell function and auditory signaling (Fig. 3A).

Fig. 3.

Adriamycin-treated mice exhibit impaired hearing function and hair cell loss. (A) Schematic representation of the inner ear and its subregions, including the cochlea, the organ of Corti, and the stria vascularis. (B) ADR-treated mice demonstrated significantly worsened hearing sensitivity compared to vehicle-treated controls. Auditory brainstem response thresholds were measured before and after ADR administration, revealing a substantial decline in hearing function in the ADR group. All graphs represent mean ± SEM (n = 10–12). Two-way ANOVA (a, main effect of hearing, F _{(4, 200)} = 119.0, P < 0.0001; b, main effect of treatment, F _{(3, 200)} = 17.02, P < 0.0001; main effect of interaction, F _{(12, 200)} = 0.9081, P = 0.5398), Tukey’s multiple comparisons test, *P < 0.05; **P < 0.01; ***P < 0.001; ***P < 0.0001. (C) Whole-mount preparations of the auditory epithelium from vehicle- and ADR-treated mice. Hair cells from the apex, middle, and basal turns of the cochlea were visualized using myosin VIIa (red, hair cell marker) and phalloidin (green, F-actin marker) and imaged via fluorescence microscopy. (D, E) Quantification of myosin VIIa-positive outer (D) and inner (E) hair cells. ADR-treated mice exhibited a significant reduction in outer hair cells compared to controls. ADR group displayed a significant loss in OHCs compared to the control group. Data are expressed as mean ± SEM (n = 3). Two-way ANOVA (a, main effect of cochlear turns, P < 0.0001; b, main effect of treatment, P < 0.0001; c, main effect of interaction, P = 0.0055), Tukey’s multiple comparisons test, *P < 0.05; **P < 0.01; ***P < 0.001; ***P < 0.0001. (F) ADR administration exacerbated outer hair cell integrity. Distortion product otoacoustic emissions (DPOAE), an indirect measure of outer hair cell (OHC) function, were recorded from vehicle- and ADR-treated mice. ADR group (10 mg/kg, 8 weeks) showed a significantly decreased DPOAE amplitude. Error bars: standard error of the mean. *P < 0.05; **P < 0.01; ***P < 0.001; ***P < 0.0001. Two-way ANOVA with Tukey’s post hoc test (a, main effect of frequency, P < 0.0001; b, main effect of treatment, P < 0.0001).

To assess auditory function, auditory brainstem response thresholds were measured at 4, 8, 16, and 32 kHz, as well as for click sounds, in vehicle- and ADR-treated mice^28–30. Significant increases in auditory brainstem response thresholds were observed at 4, 8, and 16 kHz in ADR-treated mice (Fig. 3B), indicating significant hearing loss associated with AN.

To further investigate AN-induced cochlear damage, cochleae were harvested for whole-mount preparations and stained with auditory epithelial markers (Fig. 3C). Immunofluorescence staining was performed using myosin-VIIa (red) and phalloidin (green) to visualize sensory hair cells and cilia, respectively. Significant OHC loss was observed in ADR-treated mice compared to controls: mice treated with 10 mg/kg of ADR for 4 weeks exhibited OHC loss at the basal turn; those treated with 12 mg/kg of ADR for 4 weeks showed OHC loss at the middle and basal turns; and those treated with 10 mg/kg of ADR for 8 weeks exhibited OHC loss at both the apical and basal turns (Fig. 3D). IHCs remained intact across all groups, except for the 10 mg/kg ADR-treated group at 8 weeks, which exhibited IHC loss at the basal turn (Fig. 3E).

Distortion product otoacoustic emissions (DPOAEs), an indirect measure of OHC function, were recorded from vehicle- and ADR-treated mice^31,32. The ADR group (10 mg/kg, 8 weeks) demonstrated a significant decrease in DPOAE amplitude, confirming OHC dysfunction in AN (Fig. 3F). These results suggested that AN induces OHC damage and impairs hearing function in mice, highlighting a potential link between CKD and auditory dysfunction.

AN damaged cochlear ribbon synapses

The cochlear ribbon synapse is a specialized structure that facilitates the transmission of auditory signals between cochlear hair cells and spiral ganglion neurons. Hair cells release neurotransmitters at synaptic ribbons, which activate the neurons and relay auditory information to the brain via the cochlear nerve (Fig. 4A). To assess cochlear synaptopathy associated with AN, we utilized C-terminal binding protein 2 (CtBP2) as a marker for cochlear ribbon synapses. CtBP2 is a critical component of the synaptic ribbon in IHCs, essential for synaptic integrity and auditory signal transmission. Immunofluorescence staining was performed on cochlear tissues from vehicle- and ADR-treated mice using anti-CtBP2 (for ribbon synapses) and DAPI (for nuclear staining) (Fig. 4B). Quantitative analysis revealed a significant reduction in synaptic ribbons in the 12 mg/kg 4-week and 10 mg/kg 8-week ADR-treated groups compared to vehicle-treated mice (Fig. 4C).

Fig. 4.

ADR mice show cochlear synaptopathy and lower cochlear blood flow. (A) Schematic representing whole mounts of auditory epithelium (black rectangle). (B) Whole mounts of auditory epithelium were double-stained with CtBP2 (red, a marker of synaptic ribbons) and Hoechst (blue, nuclear marker) to visualize synaptic ribbons. Severe synaptic loss was found in the ADR-treated mice compared to the control. (C) A number of pre-synaptic markers (CtBP2, red) per inner hair cell was quantified. ADR induced concentration- and time-dependent decrease in CtBP2 counts. Data are expressed as mean ± SEM (n = 13). One-way ANOVA (a, main effect of treatment, F_{(3, 48)} = 90.28, P < 0.0001); Tukey’s post hoc test, *P < 0.0001. (D) Cochlear blood flow was measured using a 0.1 mm diameter laser Doppler probe placed over the lateral wall of the cochlea of control- and ADR-treated mice. Significantly decreased blood flow was observed in ADR animals. Data are expressed as mean ± SEM (n = 12–15). One-way ANOVA (a, main effect of treatment, F_{(3, 52)} = 4.368, P = 0081); Tukey’s post hoc test, *P < 0.05; **P < 0.01; ***P < 0.001; ***P < 0.0001.

Given the importance of adequate blood supply for the proper function of cochlear hair cells and neurons, we next evaluated cochlear blood flow. A 0.1-mm-diameter laser Doppler probe was positioned over the lateral wall of the cochlea in both vehicle- and ADR-treated mice. ADR-treated animals exhibited significantly reduced cochlear blood flow (Fig. 4D). These results suggested that AN-induced hearing dysfunction in mice is associated with auditory epithelial damage, synaptopathy, and impaired cochlear blood flow.

AN disrupted cochlear BLB and vascular niche

Since the stria vascularis serves as the primary entry point for peripheral substances into the cochlea³³, we next examined the microvasculature of this structure, which plays a critical role in regulating molecular exchange between the bloodstream and the inner ear. Given the structural and functional similarities between the inner ear and the kidney, we first investigated ion channels co-expressed in both tissues, including KCNJ10 (Kir4.1), KCNQ1 (Kv7.1), and ATP1A1 (Na + /K + ATPase).

Following ADR administration for 4 or 8 weeks, cochlear cryosections from vehicle- (0 mg/kg) and ADR-treated groups (10 mg/kg for 4 weeks, 12 mg/kg for 4 weeks, 10 mg/kg for 8 weeks) were double-stained with Hoechst 33342 (nuclear dye) and ion channel markers: KCNJ10 (Fig. 5A), KCNQ1 (Fig. 5B), and ATP1A1 (Fig. 5C). ImageJ analysis revealed a significant reduction in KCNJ10 (Fig. 5A1), KCNQ1 (Fig. 5B1), and ATP1A1 (Fig. 5C1) expression in the stria vascularis following ADR treatment (one-way analysis of variance with Tukey’s post hoc test, vehicle vs. each treatment group). Additionally, mRNA expression levels of these ion channels were significantly reduced in whole cochlea lysates from ADR-treated mice (Fig. 5A2, B2, C2).

Fig. 5.

ADR nephropathy impacts ion channel expression in the inner ear. (A–C) Channels co-expressed in the inner ear and kidney were immunostained. Cochlear stria vascularis was double-stained with ion channel markers (red; K + channels, KCNJ10 (A) and KCNQ1 (B); Na + /K + ATPase (ATP1A1, C)) and Hoechst (blue, nuclear marker). Representative images are shown; quantitative analysis was performed on multiple fields per animal (n = 4 animals, 3 fields per animal), using identical acquisition settings and post-acquisition processing. Mean fluorescence intensity was measured after background subtraction and normalized to the 0 mg/kg control. The scale bar represents 50 μm. (A2–C2) The expression of ion channels significantly decreased in ADR-treated mice compared to control. All graphs represent mean ± SEM. One-way ANOVA with Tukey’s post hoc test; a, main effect of treatment; *P < 0.05. (A3–C3) Gene expression of KCNJ10, KCNQ1, and ATP1A1 measured in the whole cochlea confirmed lower levels of ion channels in ADR animals (10 mg/kg, 8wk) compared to control. Error bars: standard error of the mean. One-way ANOVA with Tukey’s post hoc test, vehicle vs. each treatment group, *P < 0.05; **P < 0.01; ***P < 0.001; ***P < 0.0001.

To further assess whether AN disrupts the endothelial cell barrier and increases vascular permeability in the inner ear, mice were intravenously administered fluorescein isothiocyanate-dextran (MW 75 kDa). Five minutes after injection, the cochlear stria vascularis was dissected and examined via confocal microscopy. ADR-treated groups (10 mg/kg and 12 mg/kg) exhibited a significant dose- and time-dependent increase in fluorescein isothiocyanate-dextran extravasation compared to the vehicle group (Fig. 6A). These findings indicated that ADR significantly increases strial vascular permeability by impairing ion channel expression and disrupting the endothelial cell barrier, potentially facilitating the entry of peripheral toxic substances into the cochlea.

Fig. 6.

ADR nephropathy induces cochlear hyperpermeability and disruption of the blood-labyrinth barrier. (A) ADR caused stria vascularis hyperpermeability. Animals were administered FITC-dextran (MW 75 kDa) at 4- or 8 weeks post-treatment via tail vein. Five minutes later, cochlear stria vascularis was dissected out and observed under confocal microscopy. Optical images showed markedly increased extravasation of FITC-dextran fluorescence post-ADR treatment in comparison to a clear structure of microvessels in control. (B) NG-2 (pericyte marker, green) and Iba-1 (PVM/M marker, red) were observed on the stria vascularis. Vehicle-treated controls exhibited long, densely distributed pericytes and PVM/Ms, whereas ADR-treated cochleae displayed shorter, protruding pericytes and PVM/Ms.

Next, we investigated the impact of AN on vascular niche components, specifically pericytes (NG-2) and PVM/Ms (Iba-1), which are crucial for maintaining BLB integrity. PVM/Ms interact with endothelial cells and pericytes to support vascular stability, while pericytes regulate angiogenesis, blood flow, and tissue fibrosis. Immunofluorescent staining for NG-2 (green) and Iba-1 (red) was performed on stria vascularis tissues harvested 4 or 8 weeks after ADR treatment (Fig. 6B). In vehicle-treated mice (0 mg/kg), NG2- and Iba1-positive cells were evenly distributed along the vasculature, maintaining a well-preserved vascular structure. In contrast, ADR-treated animals exhibited morphological alterations, including a reduction in NG2-positive pericytes and structural changes in Iba1-positive PVM/Ms in a dose- and time-dependent manner.

These findings indicated that AN compromises the structural integrity of the stria vascularis, disrupts the BLB, and damages the vascular niche, specifically by impairing the function and organization of endothelial cells, pericytes, and PVM/Ms.

ADR increased NGAL in both kidney and cochlea

NGAL (also known as lipocalin-2) is widely recognized as a biomarker of renal injury, particularly in acute kidney injury and cisplatin-induced nephrotoxicity, where it plays a critical role in inflammation and stress responses. Given its established function in renal pathology, we hypothesized that NGAL may also contribute to cochlear injury through similar mechanisms.

To test this, NGAL levels were measured in kidney and cochlear tissues from vehicle- and ADR-treated mice. A significant increase in NGAL expression was observed in the kidney following 4 weeks of 12 mg/kg ADR treatment (Fig. 7A, C). Strikingly, NGAL levels in the cochlea mirrored those in the kidney, suggesting its involvement in cochlear injury (Fig. 7B, D).

Fig. 7.

ADR nephropathy increases lipocalin 2 (Lcn2, also known as neutrophil gelatinase–associated lipocalin [NGAL]) both in the kidney and cochlea. (A, B) Western blot of renal and cochlear Lcn2 protein in control and ADR nephropathy. Renal Lcn2 level (A, C) significantly correlates with cochlear Lcn2 protein expression (B, D). (E–G) Cochlear NLRP3 expression was modulated by the progression of ADR-induced nephropathy, whereas IL-1β and TNF-α levels remained unchanged. Data are expressed as mean ± SEM (n = 12–15). One-way ANOVA (a, main effect of treatment, P < 0.05) Tukey’s post hoc test, *P < 0.05; **P < 0.01; ***P < 0.001; ***P < 0.0001. (H, I) The lipocalin receptor, 24p3r, was co-localized with lipocalin protein but not with megalin.

NGAL has been implicated in oxidative stress and cytokine production, with evidence linking it to the activation of the NLRP3 inflammasome, a key regulator of inflammation in response to tissue damage. In line with this, cochlear lysates from ADR-treated mice (10 mg/kg, 8 weeks) exhibited significant upregulation of NLRP3 expression (Fig. 7B, E). However, no significant changes were observed in traditional pro-inflammatory cytokines, including IL-1β and TNF-α, at the tested time points (Fig. 7F, G).

To identify the cochlear subregions expressing NGAL, immunofluorescence staining was performed on vehicle-treated and ADR (12 mg/kg, 4 weeks) animals. NGAL expression was localized to the tectorial membrane, cochlear neurons, and organ of Corti (Fig. 7H). NGAL interacts with two known receptors, lipocalin-2 receptor (24p3R) and megalin (Fig. 7I), prompting further investigation into their presence in the inner ear. Co-immunostaining revealed that NGAL co-localized with 24p3R but not with megalin, marking the first report of 24p3R expression in the cochlea.

Discussion

This study investigated the intricate relationship between CKD and hearing loss, two conditions that are strongly associated yet remain poorly understood in terms of their underlying mechanisms. Using AN as a representative CKD model, we provided compelling evidence that kidney dysfunction extends beyond renal pathology and significantly compromises cochlear integrity. Our findings demonstrated that AN-induced nephrotoxicity is closely linked to auditory deficits, characterized by hair cell degeneration, synaptopathy, and dysregulated ion channel expression in the stria vascularis, key contributors to sensorineural hearing loss (SNHL). We identified BLB disruption as a critical pathological event, leading to increased cochlear permeability and structural remodeling of endothelial cells, pericytes, and PVM/Ms. These vascular and inflammatory changes coincided with a marked elevation of cochlear NGAL and NLRP3, suggesting an inflammatory response that may exacerbate auditory dysfunction. Importantly, we present the first evidence of NGAL receptor (24p3R) expression in the cochlea, shedding new light on its potential role in CKD-related cochlear injury.

To induce AN, we employed the previously established ADR nephropathy model using BALB/c mice, which mimics human focal segmental glomerulosclerosis³⁴. AN is a murine model of chronic proteinuric renal disease that mimics human focal segmental glomerulosclerosis³⁴. A key consideration in our study was whether hearing impairment resulted from ADR exposure itself or from AN (Fig. 3). The Saito group investigated ADR distribution in tissues at 4 and 24 h after a single intravenous injection (30 mg/kg)³⁵. They observed that the ADR concentration in the inner ear decreased from 2.58 µg/g tissue to 0.99 µg/g tissue within 24 h, suggesting that the clearance of ADR from the cochlea is dependent on BLB integrity and the function of efflux transporters, including mdr1a P-glycoprotein³⁵. In mdr1a-knockout mice, impaired efflux transporter function led to reduced ADR clearance from the cochlea, resulting in greater drug accumulation and prolonged ototoxic exposure. In contrast, wild-type mice with intact mdr1a function exhibited more efficient ADR clearance, reducing cochlear toxicity³⁵. In our study, animals receiving 8 mg/kg ADR showed no signs of renal injury or hearing impairment, indicating that the hearing loss observed following ADR administration is primarily attributable to AN rather than direct ADR ototoxicity.

Previous studies have documented the impact of CKD on auditory and vestibular pathways. Saeed et al. (2018) examined SNHL in CKD patients undergoing hemodialysis in Iraq, reporting that 75.8% of patients developed SNHL after 1 year of follow-up, with hearing impairment predominantly affecting higher frequencies³⁶. Duration of hemodialysis was the only significant independent predictor of SNHL³⁶. Similarly, a study on children with chronic renal failure revealed significant cochlear dysfunction, particularly at low frequencies, as evidenced by reduced transient otoacoustic emissions, with more pronounced deficits in those undergoing hemodialysis³⁷. Three independent studies from South Korea also confirmed an increased prevalence of hearing impairment among individuals with CKD, emphasizing the need for routine auditory screening in CKD patients to facilitate early detection, prompt intervention, and optimized treatment strategies^27,38,39.

The findings of BLB disruption in the stria vascularis warrant consideration as a major mechanism underlying the observed ototoxicity. The pathogenesis of hearing loss can be broadly categorized into two paradigms: direct acoustic or mechanical trauma and systemic disease–mediated insults. While acute acoustic overexposure induces primary, mechanical damage to IHCs and OHCs, our model aligns with the latter. As we previously published in the models of systemic pathologies such as cisplatin-induced ototoxicity or metabolic disorders like diabetes or diet-induced obesity^33,40,41, circulating inflammatory or toxic mediators initially target the vascular supply of the inner ear. This insult compromises the tight junctions of the BLB at the stria vascularis, leading to a failure of the critical protective barrier. This BLB breakdown subsequently allows the secondary influx of toxic substances into the perilymph and endolymph, thereby initiating and accelerating damage to the IHCs, OHCs, and SGNs that ultimately manifests as hearing impairment. Crucially, we also observed marked structural and molecular alterations in the stria vascularis, including the downregulation of key ion channels such as KCNJ10 and KCNQ1. This dysregulation severely impairs K⁺ recycling, leading to a critical drop in the endocochlear potential, which is the essential electrical driving force for hair cell mechanoelectrical transduction^8,42,43. Our histological and functional data, demonstrating early and pronounced structural changes in the stria vascularis concurrent with BLB leakage markers, support the hypothesis that BLB failure is the critical event that mediates the transition from systemic exposure to localized cochlear pathology. Together, the barrier disruption and subsequent ionic imbalance underscore the BLB and strial ion transport machinery as key therapeutic targets for mitigating this class of hearing loss.

In our study, hearing loss emerged as a major outcome of AN, mainly resulting from BLB disruption and vascular injury. A central objective was to elucidate the factors contributing to BLB hyperpermeability and vascular niche damage. Viau et al. demonstrated that NGAL (also known as lipocalin 2 [LCN2]) plays a pivotal role in CKD progression in both mice and humans⁴⁴. Beyond serving as a biomarker, LCN2 actively contributes to renal disease progression, with increased expression correlating with greater renal damage. Notably, Lcn2–/– mice exhibited significantly reduced renal lesion severity. NGAL is also expressed in the brain, where it is produced and secreted by activated microglia and reactive astrocytes, particularly under conditions of neurodegenerative and neuroinflammatory conditions^45–51. NGAL has been implicated in blood–brain barrier dysfunction, neurotoxicity, and neuroinflammation^50,52–55. leading to its identification as a potential therapeutic target for neurodegenerative diseases⁴⁵. Mondal et al. further demonstrated that in a murine model of nonalcoholic steatohepatitis, LCN2 promotes neuroinflammation and blood–brain barrier dysfunction via the liver–brain axis. They showed that elevated Lcn2 levels activate its receptor 24p3R in the brain, triggering oxidative stress through NOX-2 and NLRP3 inflammasome activation, which in turn elevates IL-6 and IL-1β production. This inflammatory cascade reduces Claudin 5 expression, thereby increasing blood–brain barrier permeability and exacerbating neuroinflammation⁵². Moreover, in the brain, NGAL-induced neurotoxicity has been attributed to NF-κB signaling activation by reactive astrocytes. Recent studies have also shown that stress-induced hepatic release of LCN2 contributes to anxiety-like behaviors in mice⁵⁶.

NGAL has been reported to facilitate intracellular iron delivery, leading to iron overload, oxidative stress, cellular degeneration, and increased levels and binding of advanced glycation end-product⁵⁷. Additionally, NGAL has been shown to activate metalloproteinase-9 by forming a stable complex with it⁵⁸, a process that may contribute to increased vascular permeability through the proteolytic degradation of tight junction proteins⁵⁹. In animal models of metabolic inflammation, type 2 diabetes mellitus, and nonalcoholic steatohepatitis, elevated LCN2 expression promotes inflammation by recruiting inflammatory cells, such as neutrophils, and inducing proinflammatory cytokines⁶⁰. During lipopolysaccharide-induced inflammation, neutrophils infiltrate the spiral ligament but not the stria vascularis in the cochlea⁶¹. Our findings demonstrated that LCN2 is expressed in type 4 and 5 fibrocytes and Vas spirale, suggesting a potential role in cochlear pathology.

A key observation in our study was the strong correlation between renal injury and cochlear pathology, as evidenced by the similar expression patterns of NGAL in both the kidney and cochlea at corresponding time points and ADR concentrations (Fig. 7H). This finding raised the critical question of whether NGAL is synthesized locally within the cochlea or predominantly originates from the kidney and subsequently infiltrates the inner ear. Given that AN was shown to compromise BLB integrity, leading to increased capillary permeability, it is highly plausible that elevated NGAL secreted from the damaged kidney traverses the BLB and accumulates in the cochlea. This pathological influx of NGAL may intensify cochlear inflammation, oxidative stress, and cellular damage, ultimately contributing to hearing impairment. However, our current data could not definitively establish the primary origin of NGAL in the cochlea, highlighting the need for further investigation. Future studies employing advanced tracer techniques and targeted gene expression analyses will be crucial to determining the source of NGAL and elucidating its precise mechanistic role in CKD-associated hearing loss.

Despite histological evidence of outer hair cell (OHC) loss, particularly in the basal and apical turns, discrepancies were observed between structural and functional outcomes. While ABR thresholds exhibited significant elevations at 8–16 kHz in the 12 mg/kg (4-week) and 10 mg/kg (8-week) groups (Fig. 3B), DPOAE amplitudes showed limited reduction within this range, with a notable decline only at 8.844 kHz in the 10 mg/kg (8-week) group (Fig. 3F). This divergence can be explained by the distinct physiological bases of these two auditory assessments: ABR reflects the integrated function of inner hair cells (IHCs), spiral ganglion neurons (SGNs), and central auditory pathways, whereas DPOAE specifically measures OHC motility and cochlear amplification^62–65. Consequently, neural or synaptic dysfunction may elevate ABR thresholds without proportionate changes in DPOAE amplitude. Moreover, DPOAE responses are known to exhibit a substantial functional reserve, requiring up to 40–50% OHC loss before a measurable amplitude reduction occurs^66–69. Thus, the remaining functional OHCs in our study likely preserved DPOAE amplitudes despite partial OHC loss observed histologically. Tonotopic mapping of the mouse cochlea³⁰ indicates that the apical turn (0–40%) corresponds to low-to-mid frequencies (4–16 kHz), while the basal turn represents frequencies above 30–35 kHz—beyond the measurable range of our ABR and DPOAE systems (< 32 kHz). Accordingly, the OHC loss identified in the basal region could not be fully reflected in our functional recordings. Therefore, the apparent discrepancy between histological and functional findings likely arises from the combined influence of frequency mapping constraints, measurement sensitivity, and the intrinsic functional redundancy of OHCs.

Over the decades, the association between CKD and hearing impairment has been well-documented; however, the underlying mechanisms connecting these two conditions remain insufficiently explored. In this study, we successfully established a robust animal model and used a suite of validated assays to systematically investigate hearing loss in the context of CKD. Our findings not only confirmed the susceptibility of cochlear function to CKD-related pathology, but also provided critical insights into the vascular changes that may contribute to auditory dysfunction. By identifying key structural and functional alterations in the cochlea, our study highlighted the urgent need for routine auditory monitoring in CKD patients. These findings lay the foundation for future research aimed at elucidating therapeutic targets and developing clinical strategies to mitigate hearing loss in individuals with CKD, ultimately improving their overall quality of life.

Translational statement

This study investigated the often-overlooked complication of sensorineural hearing loss (SNHL) in chronic kidney disease (CKD). Although CKD-related hearing impairment is well-documented, its underlying mechanisms remain unclear, and auditory dysfunction is rarely considered in CKD management. Our findings emphasized the need to recognize hearing loss as a systemic consequence of CKD, shedding light on how kidney dysfunction affects cochlear health. Using the adriamycin nephropathy model, we demonstrated that CKD induces significant cochlear pathology, including blood–labyrinth barrier disruption and hair cell loss. Additionally, we identified neutrophil gelatinase-associated lipocalin (NGAL) as a potential mediator of these effects, with increased expression observed in both the kidney and cochlea. These findings highlighted the systemic nature of CKD and suggested NGAL as a promising biomarker or therapeutic target for CKD-associated SNHL.

Methods

ARRIVE

This study was conducted in accordance with the ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines to ensure transparent and comprehensive reporting of research involving laboratory animals.

Accordance

We confirmed that all experiments in this study were performed in accordance with the relevant guidelines and regulations.

Ethical statement

All procedures involving animals were approved by the Institutional Animal Care and Use Committee (IACUC) of Chungnam National University Hospital (CNUH) under reference numbers CNUH-2019012ACNU-169.

Animals and drug administration

Male BALB/c mice (20 mice per group, total of 80 mice) were acquired from Jackson Laboratory (Bar Harbor, ME, USA). All animals were housed under controlled conditions (temperature 22 °C, humidity 45–55%) with a 12-h light/dark cycle (0700 to 1900 h) and provided with ad libitum access to pelleted food (Envigo, #2018C Teklad Global 18% Protein Rodent Diet) and water. Adriamycin (Doxorubicin hydrochloride, Sigma #D1515) was administered via intraperitoneal injection at doses of 10 mg/kg and 12 mg/kg to six-week-old male BALB/c mice, with the control group receiving 0.09% saline. Mice within each group were randomly assigned to experimental conditions, with each animal considered an independent experimental unit, except for molecular assays, where a single cochlea was considered a unit. To minimize potential confounders, treatment order, measurement timing, and animal/cage locations were randomized and counterbalanced across groups.

Transmission electron microscopy (TEM)

Animals were deeply anesthetized with ketamine/xylazine and then euthanized for subsequent experimental procedures. Kidney samples from mice were immediately prefixed in a solution of 2.5% glutaraldehyde and 2% paraformaldehyde in 0.15 M sodium cacodylate buffer (pH 7.4) at 4 °C. After washing with sodium cacodylate buffer, the tissue samples were postfixed in 2% osmium tetroxide–1.5% ferrocyanide in 0.15 M sodium cacodylate buffer (pH 7.4) for 1 h. The samples were then incubated with 1% TCH for 30 min and treated with 2% OsO4 for an additional 30 min. Following this, the samples were en bloc stained overnight at 4 °C with 1% uranyl acetate and post-stained with lead citrate for 30 min at 60 °C. After dehydration, the tissues were embedded in an Epon 812 resin mixture through an ethanol and propylene oxide series, followed by polymerization at 70 °C for 24 h. Sections of 200 nm were cut using an ultramicrotome (Ultra Cut-UCT, Leica, Vienna, Austria) and collected on 100-mesh copper grids. The sections were then analyzed using conventional Transmission Electron Microscopy (TEM, JEM-1400Plus) at 120 kV and Bio-HVEM (JEM-1000BEF, JEOL, Tokyo, Japan) at 1000 kV.

Histology

Kidney and cochlear samples from mice were harvested and fixed in 4% paraformaldehyde in phosphate-buffered saline (PBS) at room temperature. The samples were then isolated and processed for routine histology (H&E staining) and biochemical analysis.

Auditory brainstem response (ABR) and distortion product otoacoustic emissions (DPOAE)

All auditory assessments were performed in a sound-attenuated booth (Sontek, South Korea) on mice anesthetized via intramuscular injection of zolazepam HCl (40 mg/kg, Zoletil, Virbac Animal Health, Carros, France) and xylazine (10 mg/kg, Rompun, Bayer Animal Health, Monheim, Germany). Auditory Brainstem Response (ABR) and Distortion Product Otoacoustic Emissions (DPOAE) were recorded using the TDT System-3 (Tucker Davis Technologies, Gainesville, FL, USA) with the RZ6 Processor and BioSig RP software (ver. 4.4.1; Tucker Davis Technologies, https://www.tdt.com/product/biosigrz-abr-dpoae-software/). ABR thresholds were measured across frequencies from 4 to 32 kHz and click sounds, using computer-generated tone pips and subcutaneous needle electrodes placed at the skull vertex and infraauricular regions. Tone bursts of 4 ms duration, with a 1 ms rise–fall time, were presented at frequencies of 4, 8, 16, and 32 kHz, along with clicks. Stimulus intensity was reduced in 5-dB decrements, with the contralateral ear unmasked due to stimulus delivery through a sealed earphone. ABR waveforms were analyzed with the researcher blinded to the treatment groups, and the threshold was defined as the lowest intensity eliciting a wave III response > 0.2 μV.

For DPOAE testing, a single acoustic assembly (ER-10B + (Etymotic, Elk Grove Village, IL, USA)) connected to TDT MF-1 transducers was inserted into the external ear canal to ensure an unobstructed path to the tympanic membrane. Two primary tones (L1 = 65 dB SPL and L2 = 55 dB SPL) were presented at f2 frequencies ranging from 0.545–35.344 kHz (f2/f1 = 1.2). The amplitude of the DPOAE at 2f1 − f2 was recorded along with background noise from 6 surrounding spectra. Following DPOAE recordings, a closed-field TDT MF-1 speaker was placed in the left ear of each mouse, and subdermal needle electrodes (Rhythmlink, Columbia, SC, USA) were positioned at the vertex (noninverting), under the test ear (inverting), and at the tail base (ground). Tone-burst stimuli were presented at a rate of 29.9/s at 4, 8, 16, and 32 kHz starting at 90 dB SPL, with the stimulus level gradually decreasing in 20-dB steps and 5-dB steps near threshold. Waveforms from 1024 presentations were averaged, amplified (20 ×), filtered (0.3–3 kHz), and digitized (25 kHz). Auditory threshold was determined as the lowest stimulus level that generated a reproducible waveform with identifiable peaks.

Cochlear dissection and immunostaining

Cochlear tissues were collected and fixed in 4% paraformaldehyde in phosphate-buffered saline (PBS) for 30 min at room temperature, as previously described⁴⁰. The individual cochlear turns were isolated and prepared for whole-mount immunostaining.

For whole-mount immunostaining, the dissected cochleae were blocked in 0.3% Triton X-100 (Sigma-Aldrich, St. Louis, MO, USA) and 5% normal goat serum (Vector Laboratories, Burlingame, CA, USA) for 1 h, then incubated with primary antibodies (listed in Table 1) in the blocking solution overnight at 4 °C, along with Hoechst 33,342 for 1 min. After rinsing six times in PBS for 10 min, the tissues were incubated with the respective secondary antibodies in PBS for 2 h (antibody list, Table 1). The specimens were mounted on glass slides using CrystalMount (Biomeda, Foster City, CA, USA) and observed under an epifluorescence microscope (Zeiss Axio Scope A1, Zeiss, Oberkochen, Germany) with a digital camera.

Table 1.

Antibody list.

Antibody	Company	Catalog #	Dilution
β-actin	Cell signaling	4967 s	1:5000
ATP1a1	Thermofisher	MA3-928	1:200
CtBP2	BD Bioscience	Cat#612044	1:200
Hoechst 33342	Invitrogen	H3570	1:1000
Iba1	Abcam	ab234437	1:200
KCNJ10	Thermofisher	PA5-103644	1:200
KCNJ10, potassium channel	Alomone labs	APC-035	1:200
KCNQ1	Thermofisher	PA5-37284	1:200
MyoVIIa	Proteus	25-6790	1:200
NG2	Santa Cruz	sc-33666	1:200
NLRP3/NALP3	Novus	NBP2-12446	1:500
NGAL/Lipocalin-2	R&D Systems	AF1857	1:1000
PECAM	EMP Millipore	MAB1398Z	1:200
Alexa Fluor 594 Phallodin (F-actin probe)	Invitrogen	A12381	1:500
Alexa Fluor 594, Goat anti-Rabbit IgG (H + L) Cross-Adsorbed ReadyProbes™ Secondary Antibody	Invitrogen	R37117	1:500
Alexa Fluor 594, Goat anti-Rabbit IgG (H + L) Highly Cross-Adsorbed Secondary Antibody	Invitrogen	A11037	1:500
Alexa Fluor 488, Goat anti-Rabbit IgG (H + L) Highly Cross-Adsorbed Secondary Antibody	Invitrogen	A11034	1:500
Alexa Fluor 488, Goat anti-Mouse IgG (H + L) Cross-Adsorbed Secondary Antibody	Invitrogen	A11001	1:500
Alexa Fluor 488, Goat anti-Chicken IgY (H + L) Secondary Antibody	Invitrogen	A11039	1:500

For section immunostaining, 4% PFA-fixed, and 120 mM EDTA-decalcified, paraffin-embedded tissues were stored at − 20 °C prior to sectioning at 4 µm thickness. Blocking and antibody hybridization steps were performed in PBS containing 1.5% normal chicken serum and 0.3% Triton X-100. Section immunostaining was visualized using epifluorescence microscopy (BX53F2, Olympus, Tokyo, Japan).

Measurement of cochlear blood flow

Under anesthesia, the left tympanic bulla of each mouse was carefully exposed and opened as previously described⁴¹. At 14 weeks of age, cochlear blood flow was measured while the mouse was positioned on a stereotaxic apparatus, using a 0.1-mm diameter laser Doppler probe connected to a Laser Doppler Flowmeter (Transonic Systems, Ithaca, NY, USA). Blood flow was quantified by analyzing the intensity oscillations, which were derived from the frequency of the oscillations.

Protein extraction and quantification

Kidney and cochlear tissue samples were collected and homogenized following previously described protocols^33,41. Protein concentrations were quantified using the BCA protein assay kit (Thermo Scientific, #23225) according to the manufacturer’s instructions, with absorbance measured at 562 nm using a microplate reader (Tecan US Inc., Durham, NC, USA).

Western blotting

Protein samples prepared in sample buffer were separated by electrophoresis on 4–15% precast Tris–HCl SDS–polyacrylamide gels (BioRad, Hercules, CA, USA), and then transferred to PVDF membranes (Millipore, Burlington, MA, USA). The membranes were incubated with primary and secondary antibodies (Table 1) in TBS containing 3% protease-free BSA (Sigma, St. Louis, MO, USA), and protein bands were detected using Immobilon Western Chemiluminescent HRP Substrate (Millipore, Burlington, MA, USA). The resulting chemiluminescent signals were captured and quantified using the Azure 300 Chemiluminescent Western Blot Imaging System (Azure Biosystems, Dublin, CA, USA).

Quantitative real-time polymerase chain reaction (qRT-PCR)

Quantitative RT-PCR was conducted as previously described^70,71. Briefly, tissues were collected, immediately frozen in liquid nitrogen, and homogenized. Total RNA was extracted using TRIzol reagent (Thermo Fisher Scientific, Waltham, MA, USA) following the manufacturer’s protocol, and RNA concentration was determined using a NanoDrop (NanoDrop Technologies, Wilmington, DE). cDNA was synthesized with the cDNA synthesis kit (Roche, Branchburg, NJ, USA). Real-time PCR was performed on a CFX Connect Real-Time PCR Detection System (BioRad, Des Plaines, IL, USA), using a reaction mixture containing SYBR Green as the fluorescent dye (Applied Biosystems, Waltham, MA, USA), a 1/10 volume of cDNA as template, and 250 nM of each primer (primer list in Table 2). The fold change in the target gene expression relative to the endogenous control gene (glyceraldehyde 3-phosphate dehydrogenase, GAPDH) was calculated using the method described previously^72,73.

Table 2.

Primer list.

Symbol	Forward	Reverse
Gapdh	CTGGAGAAACCTGCCAAGTA	TGTTGCTGTAGCCGTATTCA
18S	ATTCCGATAACGAACGAGACT	AGCTTATGACCCGCACTTACT
KCNJ10	GAAGCCTTGCCTTATGATCC	TTTGAAGCAGTTTGCCTGTC
KCNQ1	TCATTGACCTCATCGTGGTT	ATAGCTGATGTGGCGAACAC
Atp1a2	GACAGATTGGCATGATCCAG	GAAACCCATTCTCTGCCAGT

Assessment of capillary permeability

Animals were intravenously injected with FITC-dextran (75 kDa, 40 mg/ml, Sigma, USA) via the tail vein. After 5 min, cochlear stria vascularis was swiftly collected and fixed in 4% paraformaldehyde for 30 min at room temperature. The specimens were mounted and examined using a confocal microscope (ZEISS, LSM 900, Germany). Image contrast, image superimposition, and colorization of monochrome fluorescence images were performed using Adobe Photoshop CS6.

Image processing and statistical analysis

All experimental procedures were conducted in a double-blinded manner, ensuring that both the personnel administering the treatments and those performing outcome assessments were unaware of group allocations to minimize bias. The normality of the data was assessed using the D’Agostino–Pearson omnibus normality test when the sample size was adequate. All measurements were obtained from independent samples, and sample sizes were determined without prior assumptions regarding effect size. A one-way ANOVA was conducted, followed by planned comparisons. Statistical analyses were performed using GraphPad Prism 10.6.1 (GraphPad Software, San Diego, CA, USA, https://www.graphpad.com/features).

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

AG, SJ, and AL designed and conducted experiments, performed data analysis, and helped write the manuscript. AL, SJK, THK, and SS performed experiments. YHH and ARJ performed and generated transmission electron microscope images. FS interpreted data and edited the manuscript. YP and MJP designed the studies, interpreted the data, wrote the manuscript, and approved the final version of the manuscript for publication.

Funding

This research was supported by National Research Foundation (NRF) of Korea Grants to A-R. L. (2022R1C1C2007705); Y.-H.P. (NRF-RS-2024-00334688, NRF-RS-2024-00406568), and Chungnam National University Hospital Research Fund 2023-CF-028.); and M.J.P. (NRF-2020R1I1A3052557).

Data availability

The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Competing interests

The authors declare no competing interests.