Inner ear pathologies impair sodium-regulated ion transport in Meniere’s disease

Male mice of the CBA/CaJ strain were purchased from the Jackson Laboratory (Bar Harbor, ME) and were used in this study between 6 and 8 weeks of age. The animals were kept in an in-house animal facility with a uniform diurnal lighting cycle (12 h/12 h) and free access to food and water.

Mice were kept for 7 days on a purified AIN-93 M maintenance diet (TestDiet, St. Louis, MO) with either a standard Na⁺ content (0.14% Na⁺), a low Na⁺ content (0.04% Na⁺), or a high Na⁺ content (4.00% Na⁺), similar to previously established protocols [29]. Animals on standard-Na⁺ and low-Na⁺ diets received tap drinking water, while those on the high-Na⁺ diet had access to 0.9% saline. On the last day of the dietary cycle, all animals were housed individually in metabolic cages. During this 24-h period, fluid intake and urinary output was measured, and 24-h urine samples were collected.

After 7 days on a standard-Na⁺ diet, a low-Na⁺ diet, or a high-Na⁺ diet, mice were killed with an intraperitoneal (i.p.) injection of sodium pentobarbital (100 mg/kg), and cardiac blood samples were collected. All blood samples were collected between 7.30 AM and 8.30 AM, when endogenous plasma levels of ALDO in mice reach their diurnal low [11]. The blood samples were centrifuged, and the plasma was collected in sterile tubes. Urine samples were treated with 1% boric acid. All plasma and urine samples were frozen and kept at − 80 °C. Plasma/urine ALDO levels were determined using an enzyme-linked immunosorbent assay (ELISA; IBL International, Hamburg, Germany) with very low cross-reactivity to corticosterone (> 0.003%, according to the manufacturer).

All animals were killed (sodium pentobarbital (100 mg/kg), i.p.) prior to blood and tissue (temporal bone [TB], kidney) sampling. All tissues that were collected for immunohistochemical analyses were fixed (fixatives are listed in Supplementary Table 1) overnight (ON) at room temperature (RT). The TBs were decalcified in 0.12 M ethylenediaminetetraacetic acid (EDTA) for 1 week at RT. Then, the tissues were dehydrated in an ascending ethanol series, cleared with xylenes (Sigma, St. Louis, MO), and incubated in melted paraffin ON. The blocks were solidified and sliced into 10 μM sections using a rotary microtome (Reichert 2030 Microtome, Bensheim, Germany); the sections were mounted on precoated glass slides (Superfrost™ Plus, Thermo Fisher, Pittsburgh, PA) and stored at RT.

We classified archival TB samples from humans with EH based primarily on the etiology of EH, which can be either secondary or idiopathic. By definition, secondary EH is associated with a history of certain otological or systemic diseases that are believed to cause secondary EH [13], although the detailed pathophysiological connection between the diseases and secondary EH is mostly unknown. In contrast to secondary EH, idiopathic EH, by definition, occurs spontaneously and has unknown etiology [27]. In the present study, patients were classified as having secondary EH when their clinical records mentioned one or more of the diseases that supposedly cause secondary EH (Supplementary Table 2), and patients were classified as having idiopathic EH when their medical history was negative for those diseases. Next, patients with idiopathic EH were classified as having MD when their history of otological symptoms matched the diagnostic criteria for “definite” MD [4]. Patients with secondary EH were classified as having Meniere’s symptom complex when their clinical records fulfilled the diagnostic criteria for “definite” MD [4]. All individuals without a clinical history of MD or Meniere’s symptom complex, respectively, were classified as having had other otological (non-Meniere’s-like) symptoms or no otological symptoms (see also Table 1).

The human TB collection at the Massachusetts Eye and Ear Infirmary contains approximately 2300 histologically processed autopsy specimens as well as the corresponding clinical records. The standardized methods for the histological processing of human TB specimens are described elsewhere [34]. The computer database of clinical records was searched for the key word “endolymphatic hydrops”. A total of 224 specimens were identified. Specimens (114, 50.9%) were excluded from the study for reasons listed in Supplementary Table 3. An investigator blinded to any further histological information classified each case according to the criteria mentioned in the previous paragraph. Control specimens had no history of otological disease (except symmetrical presbyacusis in the corresponding age range), and no obvious histopathological findings in the temporal bone. The severity of EH in each included specimen was rated by an investigator (AHE) according to a previously established four-level (absent, mild, moderate, severe) rating system [46]. The investigator was blind to the ES histopathology and the medical records when assigning the EH ratings. A second investigator (JCA) who was blinded to the severity and group assignment of EH (secondary EH or idiopathic EH), as well as to the medical records, evaluated the integrity of the epithelium in the iES and the eES separately for each case according to a seven-level rating system, which considered the overall epithelial integrity, epithelial cell morphology, and nuclear morphology (Supplementary Table 4).

Paraffin sections were deparaffinized in xylenes and hydrated in a descending ethanol series. Celloidin sections were mounted on microscope slides, and the celloidin was removed according to methods described elsewhere [38]. For heat-induced antigen retrieval (HIAR), sections were immersed in 10 mM sodium citrate (pH 6.0), placed in a pressure cooker, and heated in a microwave oven. The celloidin sections were coverslipped before HIAR to avoid section detachment during the heating phase; a detailed protocol for coverslipping mounted tissue sections for HIAR will be provided in a later publication. All sections were then blocked in 5% normal horse serum (NHS) diluted in phosphate-buffered saline (PBS), and subsequently incubated with primary antibodies that were diluted in 1% NHS/PBS. Primary antibodies were visualized using either chromogenic or fluorogenic detection methods. In the former, sections were incubated with biotinylated secondary antibodies for 1 h, followed by application of an avidin-biotin complex (ABC) reagent (Jackson ImmunoResearch, West Grove, PA) for 1 h; an optional amplification step included incubation with biotinylated tyramine for 10 min, followed by incubation with ABC reagent for 30 min [1]. Then, all sections were incubated in diaminobenzidine/hydrogen peroxide in PBS supplemented with 4% 3,3′-diaminobenzidine (DAB; Sigma) for two to ten minutes. Hematoxylin was used to stain the cell nuclei. The slides were then dehydrated in an ascending ethanol series, cleared with xylenes, and mounted for microscopic analysis (detailed protocols for individual experiments are given in Supplementary Table 1). For immunofluorescent labeling of primary antibodies, the sections were (1) incubated with fluorochrome-conjugated secondary antibodies for 1 h and/or (2) incubated with biotinylated secondary antibodies followed by incubation with ABC reagent for 1 h and then with fluorochrome-conjugated streptavidin for 30 min. The sections were coverslipped in Vectashield mounting medium with DAPI (Vector Laboratories, Burlingame, CA). All primary and secondary antibodies were diluted in 1% NHS/PBS. All incubation steps were performed at RT.

DAB-labeled sections were analyzed using an Olympus BX51 microscope (Olympus, Tokyo, Japan) with an Olympus DP70 digital camera (Olympus). Analysis of fluorescent-labeled sections was performed using a Leica TCS SP5 or a Leica TCS SP8 confocal microscope (Leica, Mannheim, Germany).

Immunolabeled sections of the murine ES were used for counting the numbers of DAB-positive epithelial cells in the eES portion. For each primary antibody, counts were performed on three immunolabeled sections that were derived from different animals, and the mean numbers and standard deviations of labeled cells along the eES were determined. In immunolabeled sections of the human ES, the intensity and area of DAB labeling in the epithelium of the iES and eES portions was compared. Therefore, in each immunolabeled section, three microscopic images were taken in different regions of the iES and three in the eES. The software ImageJ [44] was used to measure the DAB-labeled epithelial area in each image. The same color intensity threshold was used to analyze images that were taken from the same tissue section. Each type of immunolabeling was performed on three nonconsecutive sections from the same specimen in order to determine the mean values and standard deviations of the DAB-stained epithelial area in the iES and eES portions. The ratio of mean DAB labeled epithelium in the eES (A_eES) and the iES (A_iES) is given for each antibody.

Sections were deparaffinized and HIAR was performed using the same protocols that were applied prior to immunohistochemical labeling experiments. Sections were then incubated ON at RT with primary antibodies that were diluted in 1% NHS/PBS. For PLAs, the Duolink Red Fluorescence Kit (Sigma) was used according to the manufacturer’s instructions. Sections were coverslipped with Vectashield mounting medium with DAPI (Vector Laboratories) and analyzed using a Leica TCS SP5 confocal microscope (Leica). Puncta of PLA signal per epithelial cell in the eES portion were counted using the software “BlobFinder” [2]. For each experimental condition (standard Na⁺, low Na⁺, high Na⁺), at least six sections from three different animals were analyzed.

All animal and human procedures were approved by the Institutional Animal Care and Use Committee (IACUC) and the Human Research Protections Program (HRPP) of the Massachusetts Eye and Ear Infirmary, respectively.

As schematized in Fig. 1b, the molecular cascade mediating ALDO-responsive control of Na⁺ reabsorption in the kidney’s “aldosterone-sensitive distal nephron” (ASDN) include the mineralocorticoid receptor (MR), which, upon ALDO binding, translocates to the nucleus and upregulates serum/glucocorticoid-regulated kinase 1 (SGK1). SGK1, in turn, increases the expression and activity of the epithelial sodium channel (ENaC), the renal outer medullary potassium channel (ROMK) and the sodium/potassium ATPase (NKA) by inhibiting WNK lysine-deficient protein kinase 4 (WNK4) and E3 ubiquitin ligase NEDD4-2 (NEDD4-2), which, in the absence of ALDO, promote proteolytic degradation of ENaC, ROMK and NKA. The membrane-bound enzyme transmembrane protease serine 3 (TMPRSS3) is an ALDO-independent positive regulator of ENaC. 11β-Hydroxysteroid dehydrogenase isoenzyme 2 (11β-HSD2), the signature enzyme of mineralocorticoid target tissues, prevents overstimulation of MR by constitutively converting the biologically active compound cortisol (F) to the inactive compound cortisone (E).

Immunolocalization of these molecular players in the ALDO cascade was conducted in the endolymphatic duct and the intraosseous ES (iES) and extraosseous ES (eES) of the murine inner ear (Fig. 1a). No immunoreactivity for these proteins was detected in the endolymphatic duct (data not shown). In the iES, only weak and sporadic labeling was observed (Fig. 1c). In the operculum (arrows in Fig. 1c), at the transition from the iES to the eES, the number of labeled epithelial cells abruptly increased, continuing to increase towards the distal eES (Fig. 1d). Quantification of immunolabeled cells confirmed the increased expression of all ALDO-regulated Na⁺ transport proteins along the proximal-to-distal axis of the eES (Fig. 1e). Confocal microscopy analysis demonstrated subcellular localization in the membranous, cytoplasmic, and nuclear domains, consistent with previous reports in other tissues. Figure 1f–f″ shows MR labeling in the nucleus, γENaC labeling in the apical membranes and NKA labeling in the basolateral membranes of the eES, a pattern consistent with transcellular Na⁺ transport across the eES epithelium. The presence of ALDO-regulated Na⁺ transport proteins, in particular ENaC and NKA, in murine and human ES epithelial cells is in agreement with previous reports (reviewed in [25, 35]). Labeling results for other ALDO-regulated proteins in the murine eES and in the murine kidney tissue (positive controls) are shown in Supplementary Figs. 1 and 2, respectively.

Na⁺ transport in ALDO-sensitive epithelia, such as the renal distal nephron, is controlled via MR and its intracellular downstream effectors and inhibitors. Those downstream molecules include SGK1 (effector), as well as NEDD4-2 (inhibitor) and Wnk4 (activator or inhibitor), which regulate the expression, membrane localization, and activity of the ion transport proteins. With elevated plasma Na⁺, a decreased amount of ALDO is released to bind to MR in epithelial cells, and constitutively active NEDD4-2 interacts with ENaC and ROMK channel units, leading to their ubiquitination and degradation from the apical cell membrane. Transepithelial reabsorption of Na⁺ is thereby decreased and excess Na⁺ excreted. In the event of lowered plasma Na⁺, increased MR binding by ALDO leads to activation of SGK1, which directly inhibits NEDD4-2 and thereby prevents degradation of membranous ion transporters. Transepithelial Na⁺ reabsorption from the urine is then increased.

To investigate regulation of Na⁺ transport in the murine eES, we fed mice on diets with different Na⁺ content for 7 days. Measures of metabolic balance and ALDO concentration showed that animals on high Na⁺ had significantly increased fluid intake and urine output and significantly decreased plasma/urine ALDO levels compared to animals on a low-Na⁺ diet; mice on a standard-Na⁺ diet showed intermediate values (Supplementary Fig. 3, A–D). Proximity ligation assays (PLAs) were performed on eES tissue sections from the three experimental groups to quantify protein–protein interactions in the MR downstream cascade. For NEDD4-2 and βENaC, protein–protein interactions (PLA-signal counts) were increased on the high-Na⁺ diet compared to the other two conditions (Fig. 2a), indicating increased NEDD4-2-mediated ubiquitination of βENaC in the eES under high Na⁺ (low plasma ALDO). This change is expected to reduce the membrane abundance of ENaC channels and other ion transport proteins, such as ROMK and NKA. For SGK1 and NEDD4-2, PLA-signal counts were significantly increased under low-Na⁺ conditions (Fig. 2b), indicating increased SGK1-mediated inhibition of NEDD4-2 function. This change is expected to increase the membrane abundance of ion transport proteins due to decreased NEDD4-2-mediated ubiquitination. Negative controls (incubation with NEDD4-2 antibody only) showed PLA-signal counts close to zero under all dietary conditions (Fig. 2c). These results were similar to those reported for the renal distal nephron under different Na⁺ loads [29].

Next, we used immunostaining to assess the protein expression levels and subcellular localization of ion transport proteins in the murine eES under different Na⁺ loads. For MR, we detected increased nuclear labeling in the eES in the presence of low Na⁺ and reduced nuclear labeling in the presence of high Na⁺ (Fig. 2d). Since the anti-MR antibody recognizes ALDO-bound protein, label intensity is a semiquantitative measure of binding between MR and ALDO. For αENaC, we saw increased and apically polarized labeling in the eES under low Na⁺ and weaker (almost absent) labeling under high Na⁺ (Fig. 2e). The observed changes were in line with those reported for the kidney [29]. Qualitative immunolabeling patterns for other ALDO-regulated proteins (βENaC, γENaC, ROMK, NKA), as well as corresponding data from murine kidney sections (positive controls) from all three experimental conditions, are shown in Supplementary Fig. 3, E–J.

As in mouse, the human ES is divided into an iES within the vestibular aqueduct and an eES in the posterior cranial fossa (Fig. 3a). The epithelium in the different portions of the human ES exhibits distinct morphologies, i.e., tubular columnar cells in the iES (Fig. 3b), transitional columnar-to-cuboidal cells in the proximal eES (near the operculum; Fig. 3b′), and uniformly cuboidal cells in the distal eES (Fig. 3b″). As in the murine ES, in human, immunoreactivity for ALDO-regulated proteins was detected only in the proximal and distal eES. All proteins exhibited proximal-to-distal gradients within the eES, as demonstrated qualitatively for αENaC (Fig. 3c–c″) and semiquantitatively in Fig. 3d for all ALDO-regulated proteins. Subcellular localization patterns for all ALDO-regulated proteins in the distal eES, as well as immunolabeling results from the human renal tubular epithelium (positive controls), are shown in Supplementary Fig. 4.

From the human pathology collection at the Massachusetts Eye and Ear Infirmary, we identified three groups: 42 specimens from patients with idiopathic EH, 58 with secondary EH, and 20 controls without a history of otological disease. The classification criteria are described in Materials and Methods. In a blinded fashion, epithelial integrity in iES and eES was evaluated in the archival hematoxylin and eosin (H&E)-stained sections on a 7-point scale, from completely intact (+++) to completely absent (− − −). The iES was largely intact (+, ++, +++) in all three groups. Examples for each group and corresponding ratings are shown in Fig. 4a–c. In the eES epithelium, mild to severe epithelial damage (0, −, − −, − − −) was seen in most (72.7%) specimens from patients with idiopathic EH (Fig. 4d). eES pathology mainly comprised shrunken or expelled epithelial cells, pyknotic nuclei, and fibrosis in areas of complete epithelial loss. By contrast, in cases of secondary EH (Fig. 4e) and in control cases (Fig. 4f) the eES epithelium was largely intact (more examples from all three groups in Supplementary Figure 5). In summary, among patients with idiopathic EH, 95.8% (23/24) had a diagnosis of MD, and 54.2% (13/24) exhibited eES degeneration on the affected side(s). Conversely, only 7.7% (3/39) of patients with secondary EH had MD-like symptoms, and only 2.6% (1/39) of these cases showed degeneration in the eES (one patient, both sides affected). None of the controls had a history of otological symptoms or exhibited degeneration of the ES epithelium (see also Table 1).

Degeneration of the eES was also found in a rare adolescent case in the collection. This 13-year-old had symptoms consistent with unilateral definite MD in the early “fluctuating” stage, including hour-long vertigo spells and right-sided fluctuating hearing loss. Consistent with the clinical diagnosis, histopathology of the right inner ear showed severe EH in all endolymphatic compartments. The eES portion exhibited severe degenerative changes (Supplementary Fig. 6). No signs of age-related degenerative change or typical changes associated with late-stage MD were noted in the inner ear.

Among the specimens from cases of idiopathic EH, 11 were lacking a discernable eES portion and were, therefore, not included in the analysis in Fig. 4. Closer examination revealed a hypoplastic ES morphology in those 11 specimens. As shown in Fig. 5a and b, a patient with severe left-sided idiopathic EH and a clinical diagnosis of left MD, the left ES terminated prematurely in the operculum, and the entire eES portion was missing (Fig. 5a; arrow marks the distal end of the ES in the operculum). In the unaffected right inner ear, a normal eES portion with intact epithelial cells was noted (Fig. 5b). Evaluation of epithelial integrity showed an abnormally pleomorphic, predominantly squamous epithelium, but no signs of epithelial degeneration in any hypoplastic ES (Fig. 5c, open circles indicate specimens with ES hypoplasia; other data points as shown in Fig. 4d). ES hypoplasia was present as unilateral (e.g., case in Fig. 5a, b, with left ES hypoplasia) or bilateral pathology and was always associated with an ipsilateral MD diagnosis and ipsilateral EH. ES hypoplasia was never seen in specimens from patients with secondary EH or from controls. Thus, in summary, either degeneration of the eES (13/24 cases) or ES hypoplasia (9/24 cases) was found in 95.8% of cases with idiopathic EH (with or without clinical MD) but in virtually no case with secondary EH or no EH (only 1 case with secondary EH showed ES degeneration; see also Table 1).

To investigate expression patterns of ALDO-regulated proteins in ESs affected by degenerative or hypoplastic pathology, we immunostained ES epithelia from selected human cases. A patient with right idiopathic EH and degenerative pathology of the right eES had a clinical history of bilateral sensorineural hearing loss (worse on the right), rare episodes of tinnitus, and intermittent “balance problems”, although the record did not mention MD (Fig. 6a, b′). In the left ear, the eES showed an intact epithelium (Fig. 6a), normal apically polarized immunolabeling for γENaC (and other ALDO-regulated proteins, data not shown), and normal proximal-to-distal immunolabeling gradients (Fig. 6e, box plot labeled a′). By contrast, in the right ear, the eES epithelium showed severe degenerative changes (Fig. 6b); weak, nonpolarized immunolabeling (Fig. 6b′); and loss of the proximal-to-distal immunolabeling gradients (Fig. 6e, box plot labeled B′). Another patient with bilateral idiopathic EH and bilateral ES hypoplasia had a clinical history of bilateral MD (Fig. 6c, d′). In both ears, the most distal ES (in the operculum) showed a pleomorphic, squamous epithelium (Fig. 6c, d); weak, nonpolarized immunolabeling for TMPRSS3 (Fig. 6c′, d′) and ALDO-regulated proteins (data not shown); and loss of the proximal-to-distal immunolabeling gradients (Fig. 6e, box plots labeled c′ and d′). Data for three more cases of idiopathic EH and secondary EH are shown in Supplementary Fig. 7. Immunohistochemical analysis of ALDO-regulated proteins in three controls, four patients with secondary EH, and six patients with idiopathic EH (including Cases #1 and #5; Fig. 6e) showed normal proximal-to-distal immunolabeling gradients (A_eES/A_iES ratios > 1) in all controls and all patients with secondary EH. In contrast, all specimens from patients with idiopathic EH exhibited A_eES/A_iES ratios ≤ 1, indicating loss of the normal proximal-to-distal gradients. Notably, in all unilateral idiopathic EH cases, the unaffected contralateral side (unshaded box-and-whisker plots in Fig. 6e) exhibited an intact eES epithelium, normal proximal-to-distal label gradients and significant differences between the affected and unaffected sides.

We show here that idiopathic EH and associated clinical MD can be subdivided according to ES pathology: degeneration or hypoplasia. Both pathologies were linked with distinct clinical traits, according to our retrospective chart reviews. Degenerative ES pathology was associated with unilateral disease (12 out of 13 cases) significantly (p = 0.0066) more often than bilateral disease, whereas hypoplasia was found in 6 out of 9 bilateral cases (Fig. 7a). EH was significantly (p = 0.021) increased in severity when hypoplastic ES pathology was present (Fig. 7b). Patients with degenerative ES pathology showed significantly (p = 0.038) higher age at onset (58.1 ± 20.5 years) than patients with hypoplastic ES pathology (37.7 ± 19.0 years; Fig. 7c). No significant difference was found in the sex ratio between degenerative (1.6 females/male) and hypoplastic pathology (2 females/male; p = 1.0; Fig. 7f). A positive family history of MD was reported in only two cases of ES hypoplasia; for most cases, this information was not available in the clinical records (Fig. 7e). Among the reported comorbidities, only patients with degenerative ES pathology had a positive history of hematological or cardiovascular disease diagnosed prior to the first onset of MD symptoms (Fig. 7f; detailed diagnoses are listed in Supplementary Table 5); however, the increased age at MD onset in patients with degenerative ES, as well as incomplete clinical records, may confound this analysis.