Functional hair cell mechanotransducer channels are required for aminoglycoside ototoxicity

Abstract

Aminoglycosides (AG) are commonly prescribed antibiotics with potent bactericidal activities. One main side effect is permanent sensorineural hearing loss, induced by selective inner ear sensory hair cell death. Much work has focused on AG's initiating cell death processes, however, fewer studies exist defining mechanisms of AG uptake by hair cells. The current study investigated two proposed mechanisms of AG transport in mammalian hair cells: mechanotransducer (MET) channels and endocytosis. To study these two mechanisms, rat cochlear explants were cultured as whole organs in gentamicin-containing media. Two-photon imaging of Texas Red conjugated gentamicin (GTTR) uptake into live hair cells was rapid and selective. Hypocalcemia, which increases the open probability of MET channels, increased AG entry into hair cells. Three blockers of MET channels (curare, quinine, and amiloride) significantly reduced GTTR uptake, whereas the endocytosis inhibitor concanavalin A did not. Dynosore quenched the fluorescence of GTTR and could not be tested. Pharmacologic blockade of MET channels with curare or quinine, but not concanavalin A or dynosore, prevented hair cell loss when challenged with gentamicin for up to 96 hours. Taken together, data indicate that the patency of MET channels mediated AG entry into hair cells and its toxicity. Results suggest that limiting permeation of AGs through MET channel or preventing their entry into endolymph are potential therapeutic targets for preventing hair cell death and hearing loss.

Figure 1. Gentamicin causes outer hair cell loss.

A. Low magnification view of an isolated organ of Corti from a postnatal 4-day-old (P4) rat. B. Schematic of two proposed mechanisms of aminoglycoside (red dots) entry into hair cells via mechanotransducer channels (green) on stereociliary bundle or endocytosis (on apical or basolateral membranes). C. P4 cochleae were cultured overnight before treatment with gentamicin (0, 0.1, 0.25, 0.5, and 1.0 mM) for 24 hr, followed by 0-, 24-, or 48-hr recovery periods. Counts of phalloidin- and parvalbumin-labeled hair cells revealed that no significant hair cell loss at the end of a 24-hr treatment. With additional recovery periods, the middle and basal turns were most susceptible to hair cell loss. D. Cochleae were treated with 0.25 mM gentamicin for 1–24 hours followed by a 48-hr, antibiotic-free recovery. After 1-hr treatment, an apical-basal gradient of hair cell loss was observed. Arrow indicates the treatment paradigm chosen for the Panels E–F. Each data point represents an analysis of three to ten experiments. E–F. Representative images of control cochleae and those treated with gentamicin (0.25 mM) for 1 hr, followed by a 48-hr recovery period. Damage to parvalbumin 3-labeled hair cells was noted in the middle and basal turns of the cochlea (notably in OHCs rows 1 and 2). Error bars = S.D. Scale bars = 250 µm in A, 20 µm in C–D.

Figure 2. Gentamicin selectively enters hair cells.

A–B. The dose-response relationship between GTTR and outer hair cell survival was determined. Cochleae were treated with GTTR for 1 hr and allowed to recover for 48 hr in drug free media. Like native (or unconjugated) gentamicin, the basal and middle turns were more susceptible to hair cells loss when treated with GTTR. C–D. GTTR (3 µM) selectively entered inner and outer hair cells and not adjacent prox1-labeled supporting cells (Deiters' and pillar cells). Staining for myosin 7a, a specific marker for outer and inner hair cells, in a P3 mouse cochlea exhibited a similar expression pattern as GTTR. E. NADH signals were detected in both outer and inner hair cells in cochlear organs deemed healthy for analysis for GTTR uptake. F–G. Cochleae cultured in the presence of gentamicin (0.1 mM for 1 hr followed by a 48 hr drug free recovery period) were immunolabeled for gentamicin, which was detected only in myosin 7a-positive hair cells. Error bars = S.D. Scale bars = 20 µm.

Figure 3. Open MET channels are required for GTTR entry into hair cells.

A. Following incubation in control media for 24 hr, P3 rat cochleae treated with GTTR alone (3 µM for 1 hr at 37°C) showed robust uptake in inner (asterisk) and outer hair cells (OHC) in the middle and basal cochlear turns. B. Co-treatment with curare (1 mM) effectively prevented GTTR uptake. C–D. Quantification of GTTR fluorescent intensity demonstrated that uptake is significantly reduced by blocking MET channels with curare in both the inner and outer hair cells throughout the cochlear turns. Error bars = S.D. Scale bars = 100 µm.

Figure 4. Two-photon imaging for GTTR uptake in live cochlear hair cells.

P3 cochleae were cultured for 24 hr before being treated with GTTR or FM1-43 for an hour at room temperature. A–H. GTTR (3 µM) rapidly entered hair cells, whereas Texas Red alone (3 µM) did not. Both high dose GTTR (30 µM) and hypocalcemic conditions promoted GTTR entry into hair cells. In these 2 conditions GTTR was rarely observed outside hair cells in a delayed (>10 min) fashion (arrowheads). E–F. The amphiphatic dye FM1-43 has been shown to be a permeant blocker of the MET channel. The entry of FM1-43 (0.75 µM) (E–F in pink background) into hair cells was more rapid (note different time scale) than GTTR (A–D in yellow background) and was also promoted by increasing its concentration (7.5 µM). Error bars = S.D. Scale bar = 20 µm.

Figure 5. Kinetics of GTTR uptake with blockers of MET channel or endocytosis.

A–G. Both curare (1 mM) and amiloride (0.3 mM) reduced GTTR uptake into outer hair cells, but an initial rise in GTTR uptake, like that seen when cochleae were treated with GTTR alone, was still observed despite amiloride co-treatment. Quinine (1 mM) was the most effective in limiting GTTR uptake among all the drugs tested. No significant change in GTTR uptake was seen with concanavalin A (2.6 µM) treatment. Error bars = S.D. Scale bar = 20 µm.

Figure 6. Blockers of MET channels, but not endocytosis, protect hair cells from gentamicin toxicity.

A–G. Cochleae were harvested from P4 rats and cultured in antibiotic-free media for 24 hr before treated with gentamicin (0.1 mM) for 1 hr, followed by a 48-hr antibiotic-free recovery period. Tissues were fixed and labeled with parvalbumin 3 antibody. Representative confocal images from each region of the cochlea are shown. When applied concurrently with gentamicin, both curare (1 mM) and quinine (0.5 mM), improved hair cell survival. No significant protection was observed with a different MET channel blocker, amiloride (300 µM), or when blockers of endocytosis (concanavalin A 2.6 µm or dynosore 80 µM) were co-administered. Scale bar = 20 µm.

Figure 7. Quantitative analysis of outer hair cell survival.

A. Curare, quinine, amiloride, concanavalin A, and dynosore were administered individually to rat cochlear cultures and found not to affect outer hair cell survival (determined by parvalbumin 3 labeling). B. Both curare and quinine provided significant protection to outer hair cells from gentamicin damage in both the basal and middle turns of the cochlea. No significant protection from gentamicin toxicity was observed when cochleae were co-treated with amiloride, concanavalin A or dynosore. C. Curare significantly improved outer hair cell survival in the basal turn of cochleae exposed to various aminoglycosides, including streptomycin and dihydrostreptomycin, but not neomycin. Error bars = S.D. * = p<0.001, ** = p<0.01.

Figure 8. Curare protected hair cells in long-term cultures.

A–C. Cochleae treated with gentamicin (0.1 mM) with or without curare (1.0 mM) remained in gentamicin-free media for four additional days. Survival of outer hair cells (determined by parvalbumin 3 labeling) in the curare-treated cochleae was comparable to that in untreated, cultured cochleae. D. Quantitative analysis of the gentamicin-treated cochleae demonstrated a slight decline in outer hair cell survival 96 hr versus 48 hr after treatment. Cochleae co-treated with curare and gentamicin have comparable numbers of outer hair cells at both time points. Each data point represents an analysis of three to twenty-two experiments. E. Schematic of possible entry pathways for aminoglycoside first into the scala media from the stria vasculars or Reissner's membrane, then into sensory hair cells in the inner ear. Error bars = S.D. Scale bar = 50 µm.