Different rates of endocytic activity and vesicle transport from the apical and synaptic poles of the outer hair cell

Outer hair cells were isolated and handled as previously described [18]. Adult pigmented guinea pigs with a weight of 350–900 g (N = 21) were used for the study. The data derive from 37 cells isolated from 21 cochleae. Animals were anesthetized by intraperitoneal injection of a mixture of 100 mg/kg ketamine and 4 mg/kg xylazine and were killed by cervical dislocation. The dissected temporal bones were placed in an ice-chilled Hanks’ balanced salt solution (HBSS; Biochrom KG, Berlin, Germany), containing: 137 mM NaCl, 5.4 mM KCl, 1.25 mM CaCl₂, 4.2 mM NaHCO₃, 0.81 mM MgSO₄∙7H₂O, 0.44 mM KH₂PO₄, 0.34 mM Na₂HPO₄∙2H₂O, 5 mM glucose, and 5 mM HEPES, with an osmolarity of 310 mOsm/l adjusted with D‑(+)-glucose; the pH was 7.3. HBSS was used as the extracellular fluid throughout the preparation steps and experiments. All chemicals were from MERCK (Darmstadt, Germany), unless otherwise stated. Outer hair cells from the apical third of the cochlea were isolated; their lengths were 74 ± 7 µm.

Single- or double-barrel perfusors (Fig. 1) were used, respectively, to label either both poles or a single pole of the OHC [18]. The tip of the perfusor was positioned near the cell using a micromanipulator (SM5, Luigs and Neumann GmbH, Ratingen, Germany). A peristaltic pump (Ismatec, IDEX Health & Science GmbH, Wertheim, Germany) established laminar flow with a constant flow rate.

Using a single-barrel perfusor enabled homogeneous labeling of the entire PM; the flow rate was 14 µl/min. By contrast, using a double-barrel perfusor enabled labeling of a single pole of the OHC (Fig. 1). The double-barrel perfusor was fabricated from a borosilicate-glass theta capillary with an external diameter of 2 mm (Harvard Apparatus, MA, USA) using a DMZ-Universal Puller (Zeitz Instruments, Augsburg, Germany). To achieve single-pole staining, the perfusor tip was positioned as close as possible to the cell. The coverslip was coated with a cell-and-tissue adhesive (Cell-Tak^TM) to facilitate cell adhesion. The output flow rate was 3 µl/min per barrel. As recently demonstrated [18], the double-barrel perfusion method is an effective tool for exclusively labeling one pole of bipolar cells such as OHCs.

Imaging was performed using a Zeiss LSM 510 confocal laser-scanning system based on a Zeiss Axioskop2 FS mot microscope (Zeiss, Heidelberg, Germany) equipped with a two-photon laser system (Mira^TM 900 Ti:Sapphire Laser pumped by a Verdi V5 Diode-Pumped Laser from Coherent, Santa Clara, USA) and a pinhole diameter of 1 AU. A Zeiss 40 × IR-Achroplan water-immersion objective with NA 0.8 and ZEN2009 software was used.

The fluorescent membrane markers FM1-43 [N-(3-triethylammoniumpropyl)-4-(4-(dibutylamino)styryl)pyridinium dibromide] and FM4-64 [N-(3-triethylammoniumpropyl)-4-(6-(4-(diethylamino)phenyl)hexatrienyl)pyridinium dibromide] were used to visualize endocytosis. FM1-43 is a green-fluorescent, lipophilic styrylpyridinium dye that is commonly used to visualize endocytic and exocytic processes [3], also in cochlear hair cells [15, 16, 18, 23]. Virtually non-fluorescent in aqueous media, the dye rapidly inserts into the PM where it becomes intensely fluorescent; excitation/emission spectral maxima are 473/579 nm, respectively. FM4-64 is a derivative of FM1-43 with similar membrane labeling properties, but exhibits longer wavelength (red) fluorescence making it suitable for double-labeling experiments with lower wavelength dyes [18]; the excitation/emission spectral maxima are 505/725 nm, respectively. Stock solutions of these dyes in a concentration of 10 mM were prepared in DMSO. On the day of the experiment, the dyes were diluted in HBSS to a final concentration of 10 µM.

The fluorescent endoplasmic marker DiOC₆ (3,3′-dihexyloxacarbocyanine iodide) was used to examine possible targets of endocytosed vesicles in the endoplasmic reticulum (ER). DiOC₆ is a cell-permeant, green-fluorescent, lipophilic dye, which at low concentrations is selective for mitochondria of live cells and at high concentrations labels other intracellular structures, such as ER [24, 34]. Excitation/emission spectral maxima are 489/506 nm, respectively. For double-labeling experiments with FM4-64 and DiOC₆, OHCs were first incubated in DiOC₆ for 1 min, the DiOC₆ was washed out with extracellular solution, and then the FM4-64 was applied. This FM dye was used instead of FM1-43 because the fluorescence spectra of FM4-64 and DiOC₆ can be readily separated using bandpass filters. A 1‑mg/ml stock solution of DiOC₆ was prepared in DMSO. Just before the experiments began, it was diluted with HBSS to a final concentration of 0.87 µM [23].

The fluorescent fluid-phase marker Lucifer yellow (Lucifer yellow carbohydrazide, potassium salt) was used to examine pinocytosis. Lucifer yellow is a well-known dye for studying fluid-phase endocytosis [15]. The dye is hydrophilic, but being anionic cannot permeate the cell membrane by passive diffusion. The potassium-salt of Lucifer yellow has excitation/emission spectral maxima at 428/536 nm, respectively. A 50-mM stock solution of Lucifer yellow was prepared in DMSO. On the day of the experiments, it was diluted with HBSS to a final concentration of 50 µM. This concentration was chosen to be of the same order of magnitude as that of the FM dyes (10 µM).

FM and DiOC₆ were excited with an argon laser with a wavelength of 488 nm and the emitted light was collected with bandpass filters: 505–750 nm for FM1-43, 650–710 nm for FM4-64, and 500–550 nm for DiOC₆. For the double-labeling experiments with FM4-64 and DiOC₆, the emitted light was separated into two channels using a dichroic beam splitter (NFT545). Lucifer yellow was excited by the two-photon laser (840 nm) and the emission recorded below 650 nm using a short-pass filter (KP650).

Except for the Lucifer yellow experiments, fluorescence signals were corrected offline for the average background fluorescence level, which was measured extracellularly just before the start of dye application in a region free of cell remnants.

All dyes were purchased from Thermo Fisher Scientific, Inc. (Waltham, MA, USA).

The detection threshold, used for calculating signal delay, was set to 20% of the saturation intensity from the PM region of interest (ROI) near the cell pole where the dye was applied. Linear regression, with software in Origin (Ver. 7, OriginLab Corporation, Northampton, MA, USA), was used to estimate the time at which the signal reached/crossed the 20% value. Linear regression was also used for calculating the mean speed of vesicle trafficking. Average data are presented as mean and standard deviation. Statistical tests were performed with the Student’s t test and considered significant at p < 0.05. Test results are given as p and t_f values, where f denotes the degrees of freedom.

Endocytic activity was investigated separately at the apical and basal poles of the isolated OHC (Fig. 2). In the first example shown (Fig. 2a–c), the PM marker FM1-43 was applied to the apical pole of the OHC using the double-barrel perfusor (Fig. 2a). Channel A (ChA) contained 10 µM FM1-43 in HBSS while channel B (ChB) contained dye-free HBSS to prevent apical pole labeling. The spatial distribution and the temporal course of the intracellular dye (Fig. 2b) imply that FM-labeled vesicles traffic toward the base of the cell. Fig. 2c shows the time course of the fluorescence signal intensities in the ROIs indicated in Fig. 2a. The dashed line denotes 20% of the saturation intensity from the apically located PM ROI. This 20% value, arbitrarily set as a threshold signal value, was used to calculate the signal delays to the ROIs relative to the PM ROI. Distances of the ROIs from the apical pole are plotted as a function of signal delay in Fig. 2g (triangles) for the average of eight cells. Based on linear regression, the slope yields an average apicobasal speed of 0.11 ± 0.01 µm/s. The distance–axis intercept locates the PM ROI, on average, at 1.52 ± 3.09 µm from the apical pole, which is not significantly different from 0 µm (t₇ = 1.39, p = 0.21).

The same protocol was used to investigate basoapical trafficking, the FM1-43 dye being applied to the basal half of the cell (ChB) and the dye-free HBSS to the apical half (ChA) to prevent access of dye to that region (Fig. 2d–f). Signal delay as a function of distance from the basal pole is shown in Fig. 2g (circles) for five cells. The average apicobasal speed is 0.18 ± 0.01 µm/s. The distance–axis intercept locates the PM ROI, on average, at 6.25 ± 2.73 µm from the basal pole. In other words, the basoapical speed is significantly greater than the apicobasal speed (t₆ = 12.17, p = 9.37.10⁻⁶), indicating more intense trafficking toward the apex of the cell.

For the apical application, the florescent signal in the infracuticular ROI required, on average, 45.1 ± 22.8 s (N = 8) to reach the 20% threshold, whereas for the basal application, on average, only 13.6 ± 6.4 s (N = 5) was required for the infranuclear ROI, relative to the respective PM ROIs at or near the poles. The significantly smaller delay for the infranuclear ROI (t₁₁ = 3.44, p = 0.003) indicates that vesicle formation in the synaptic pole is more intense than that at the apex of the OHC.

Until now, targets of vesicles formed at the synaptic pole of OHCs had not been demonstrated. Therefore, here, the endoplasmic reticulum of OHCs was labeled with DiOC₆ (Fig. 3a, green) and compared with vesicle targets of the fluorescence PM marker FM4-64 applied to the basal pole (Fig. 3b, red). The yellow signal in Fig. 3c indicates colocalization of DiOC₆ and FM4-64. Similar observations were made in a total of 14 cells. These observations imply that basally endocytosed vesicles are transcytosed to the endoplasmic reticulum.

To investigate the maximum possible molecular weight (MW) for a substance to be internalized by fluid-phase endocytosis (pinocytosis), Lucifer yellow (MW = 522 Da) was applied to the OHC and the development of intracellular fluorescence recorded (Fig. 4). In these experiments (N = 10), a single-barrel perfusor was used to apply the stain homogeneously around the entire cell. To avoid recording of scattered emitted light from extracellular dye, fluorescence was induced and recorded with two-photon confocal microscopy. Despite the presence of a constant and high-intensity extracellular signal and an extensive application time of the dye (>30 min), an intracellular fluorescence signal was not detected above the background noise level (Fig. 4). Importantly, the displayed intracellular signals represent noise derived from the extracellular signal because the intracellular signal dropped to zero when the Lucifer yellow application was terminated and extracellular solution applied (not illustrated). In the absence of a detectability problem, this result implies that anionic molecules with an MW greater than 500 Da cannot be internalized by pinocytosis.

It has been shown that FM1-43 can penetrate mechanoelectrical transducer (MET) channels in hair cells of immature cochlea cultures [12], hair cells of embryonic mice [13], developing hair and sensory cells [26], cultured chick auditory papilla hair cells [32], cultured zebrafish larvae lateral line organs [31], and Xenopus larvae lateral line organs [28]. In addition, Crumling et al. [5] used hair cells of chicks 5–10 days post-hatch and showed PPADS- and suramin-dependent AM1-43 entry, implying that FM dyes can also penetrate P2X receptors. To avoid this type of nonspecific labeling, recently two techniques were developed: (1) photo-oxidation electron microscopy [22], whereby the FM dye taken up by the organelles is converted to a dark precipitate visible electron microscopically, and (2) instead of using an FM dye, a novel endocytotic probe impermeable to MET-channels, called membrane-binding fluorophore-cysteine-lysine-palmitoyl group, was used [30]. For both techniques, experiments with IHCs from mice at postnatal day 14–18 yield vastly different patterns of endocytotic signals compared with those found for when uptake is also through the MET channels. Therefore, there is still ongoing controversy about whether FM1-43 does indeed reliably report endocytosis in (P14–18) mice.

However, for the mature cochlea of the guinea pig, there is no evidence of FM-1-43 uptake through the MET channels or P2X receptors. For the inner hair cell (IHC), it has been demonstrated, using an in-situ cochlea preparation from young adult guinea pigs, that neither the MET-channel blockers dihydrostreptomycin or D‑tubocurarine nor the lesioning of tip links with BAPTA treatment influences FM1-43 labeling [15]. Similar results were reported for OHCs in situ using the same animal and preparation as in the IHC study; namely, that FM1-43 labeling was not influenced by these MET-channel blockers or BAPTA [16]. By applying FM1-43 to isolated OHCs from the mature, pigmented guinea pig cochlea, we also demonstrated that block of MET channels by dihydrostreptomycin or of P2X receptors by PPADS had no influence on the FM1-43 labeling [23]. Therefore, for the experimental model on which the present results are based, there is no evidence of FM1-43 entering through MET channels or P2X receptors of hair cells of the functionally mature cochlea. In other words, the available evidence convincingly suggests that in the present experiments FM1-43 uptake was via endocytic activity.

Plasma-membrane Ca²⁺ ATPase (PMCA) activity is crucial for removing Ca²⁺ from the cytosol of the stereocilia and, therefore, for protecting hair cells from Ca²⁺ overload [4]. It has been proposed that apical endocytic activity of hair cells is involved in recycling of PMCA molecules of the hair bundle [14].

Receptors and ion channels of the efferents and afferents located at the synaptic pole of the OHC may require continuous replacement. However, the colocalization data with DiOC₆ and FM4-64 show that vesicles formed at the synaptic pole of OHCs traffic basoapically, targeting the endoplasmic reticulum. Therefore, in addition to PM recycling, proteins such as acetylcholine receptors might also be involved.

It has been proposed that aminoglycoside uptake might be regulated by endocytosis [7]. Although it has been suggested that aminoglycosides enter hair cells through transducer channels in cultured organs from newborn rats [1], the significance of endocytosis-dependent aminoglycoside uptake at the synaptic pole cannot be excluded because aminoglycosides such as gentamicin have a relatively low MW (450–478 Da), which is below the MW limit suggested in this study for pinocytic internalization. With the aid of the double-barrel perfusor, experiments with fluorescent-labeled gentamicin could address the possibility of its endocytic internalization.