Altered Outer Hair Cell Mitochondrial and Subsurface Cisternae Connectomics Are Candidate Mechanisms for Hearing Loss in Mice

ABR and DPOAE measurements

ABR experiments were performed as previously published by our group (Reijntjes et al., 2019). All animal experimentation was conducted following SfN’s Policies on the Use of Animals in Neuroscience Research. Equal numbers of males and females were used in the experiments. In cases where odd numbers are reported, females outnumbered male samples. Briefly, mice were anesthetized with tribromoethanol (Avertin; Sigma Millipore), using 0.5 mg/g body weight, injected intraperitoneally. The mice were then placed in a sound-attenuated chamber on a custom holder. Body temperature was maintained at 36.5 ± 0.5°C using a homeothermic blanket control unit (Harvard Apparatus) and rectal probe feedback. Ground and recording electrodes were placed subcutaneously in the scalp, and a calibrated transducer (Intelligent Hearing Systems) placed in the right pinna. At a rate of 20 Hz, with intensity from 0 to 100 dB sound pressure level (SPL; rms for click stimuli) in 10 dB increments, 0.1 ms broadband clicks, and 3 ms pure tone pips at 8, 16, and 32 kHz were presented. The ABR activity was extracted from 128 to 1024 stimuli. The hearing threshold was defined as the minimum sound intensity that elicited a characteristic ABR waveform.

For DPOAE experiments, mice were anesthetized with Avertin as for ABR measurements. After visual inspection to ensure adequate health of external and middle ear was performed, the mice were placed in a sound-attenuated chamber, and a dual acoustic probe/microphone assembly (Etymotic Research) was placed in the ear. Primary tones with an f2/f1 ratio of 1.25 were presented at equal sound pressure levels at 55, 65, and 75 dB SPL over 92 ms, and tones were routed to independent transducers and allowed to mix acoustically in the ear canal. The SPL was sampled and synchronously averaged by a digital signal processor for geometric mean frequencies <20.1 kHz. Above 20.1 kHz, the sampling input was automatically switched to the dynamic signal analyzer for frequency analysis to avoid DSP aliasing artifacts >22.1 kHz. Cubic (2f1–f2) DPOAE levels and corresponding noise floor levels were calculated. DPOAE levels were plotted as a function of the primary tones’ geometric mean frequency. DP thresholds were generated for each stimulus level and were evaluated relative to the noise floor (DPOAE > 5 dB above the noise floor).

Mitochondrial membrane potential

Semi-intact, viable epithelia were dissected from 3, 15, and 24 month mice. Any potential trauma caused by dissection was evaluated using tight-seal current- and voltage-clamp recordings of resting membrane potential, and K⁺ currents as well as nonlinear capacitance recordings. The 24 month cochleae were tested because this age would be the most vulnerable to dissection trauma. Bicarbonate buffered bath solution containing 3 mm glutathione and continuously perfused with 5% CO_2, and 95% O₂ sufficed to maintain stable resting membrane potentials in apical OHCs; 3 month = −52 ± 12 mV (n = 87 OHCs from 35 cochleae), 15 month = −48 ± 14 mV (n = 25 OHCs from 10 cochleae), and 24 month = −49 ± 10 mV (n = 49 OHCs from 14 cochleae). The resting membrane potentials were recorded after 2–3 min after whole-cell configuration. The values were comparable to previous reports from OHCs in the gerbil and guinea pig (He and Dallos, 1999a,b). Only apical OHCs had resting membrane potentials that lasted for ∼20–25 min after whole-cell patches were raptured. OHCs from the apical segments of the cochlea were used for these sets of experiments. Because OHCs at the edge of the severed epithelium may have been damaged, recordings were also only performed from OHCs with intact surrounding supporting cells.

Tetramethylrhodamine, methyl ester (TMRM), was prepared and loaded in isolated organs of Corti from CBA/CaJ mice according to the manufacturer’s protocol (Thermo Fisher Scientific). High-magnification confocal images were collected at 2.5 s time intervals using a SP8 confocal microscope (Leica Microsystems) with a 63× oil immersion lens under the control of LAS X software. The TMRM signal with an absorbance peak at 548 nm and an emission peak at 574 nm was captured with an RFP/TRITC filter. Ptrifluoromethoxycarbonyl cyanide phenyl hydrazone (FCCP) (1 mm) was added 10 min into the imaging as a pharmacologic control because there should be no detectable TMRM signal from mitochondria completely depolarized using FCCP.

Single-molecule FISH (smFISH) with RNAscope

Mice (ages 3, 15, 18, and 24 months) were anesthetized with ketamine (100 mg/kg) and xylazine (10 mg/kg) and then transcardially perfused with diethylpyrocarbonate (DEPC)-treated PBS and 4% PFA in 0.1 m phosphate buffer. Cochleae were harvested and kept in a 4% PFA solution (DEPC-treated) overnight on a shaker at 4°C. The cochleae were washed with PBS, and decalcified in 0.35 m EDTA for 5-7 d, depending on age, on a shaker at 4°C and washed with PBS. The cochleae were sequentially dehydrated in 10%, 20%, and 30% sucrose solution at 4°C for 1 h, 2 h, and overnight, respectively. The cochleae were processed as whole-mount preparation or transferred into optimal cutting temperature compound for a minimum of 1 h at 4°C and then snapped frozen, using a dry ice-ethanol mixture. Samples were cryo-sectioned to a thickness of 12 μm, placed onto Superfrost slides, and stored at −80°C until further use.

Probe hybridization closely followed the manufacturer’s instructions (Advanced Cell Diagnostics). Sections were immersed in prechilled 4% PFA for 15 min at 4°C. The sections were dehydrated at room temperature in 50%, 70%, and 100% ethanol for 5 min each and allowed to dry for 1-2 min. Fixation and dehydration were followed by protease digestion, using Protease 4 for 30 min at room temperature. The sections were then incubated at 40°C with the following solutions: (1) target probe in hybridization buffer A for 3 h; (2) preamplifier in hybridization buffer B for 30 min; (3) amplifier in hybridization buffer B at 40°C for 15 min; and (4) label probe in hybridization buffer C for 15 min. After each hybridization step, the sections were washed with wash buffer 3 times at room temperature. For fluorescent detection, the label probe was conjugated to AlexaFluor-488. Probes for Drp1 and Opa1 and a blank negative control were obtained from Advanced Cell Diagnostics. Sequences of the target probes (for the specified genes), preamplifier, amplifier, and label probe are proprietary. Detailed information about the probe sequences can be obtained by signing a nondisclosure agreement provided by the manufacturer.

For subsequent immunofluorescent staining, sections were treated with 10% blocking solution for 10 min at room temperature, incubated with anti-myosin 7A (Novus Biological, 1:300 dilution), overnight at 4°C, washed with PBS 3 times for 5 min each, incubated with the appropriate AlexaFluor secondary antibody (Thermo Fisher Scientific) diluted 1:500 for 2 h at room temperature, and again washed with PBS 3 times for 5 min each. Incubation in DAPI solution for 15 s at room temperature was performed to label cell nuclei. Sections were then mounted in Fluoromount-G and sealed under a coverslip.

Confocal imaging and image analysis

Confocal micrographs were obtained as described previously (Reijntjes et al., 2019). To determine frequency regions, low-magnification micrographs of isolated organs of Corti were obtained using a DM4000B fluorescent microscope (Leica Microsystems). If necessary, the Stitching plugin in ImageJ was used to create a single montage image. Tonotopic maps were then overlaid on the image previously described in the place-frequency map of the mouse cochlea (Muller et al., 2005). High-magnification confocal micrographs were collected using a SP8 confocal microscope (Leica Microsystems) with a 63× oil immersion lens under the control of LAS X software. z stacks of the OHCs from the 8, 16, and 32 kHz regions were collected at a scan speed of 200 Hz and zoom of 1. The step size (optical section thickness) was determined by stepping at half the distance of the theoretical z axis resolution (the Nyquist sampling frequency). Images were acquired in a 1024 × 1024 raster (x = y = 184.52 μm) at subsaturating laser intensities for each channel. Images are presented as z projections through the collected optical stack. All quantitative image analysis was performed on the raw image stacks, without deconvolution, filtering, or γ correction. Individually fluorescently labeled mRNA transcripts appeared as puncta. To quantify the number of mRNA transcripts per OHC, individually fluorescently labeled mRNAs within a given FOV were counted using the spots function in Imaris 6.4 software (Bitplane). mRNA counts were normalized to the number of Myosin 7A-labeled OHCs marked manually in the same FOV.

Ca²⁺ imaging

Mice (ages 3 and 24 months) were deeply anesthetized with ketamine and xylazine mixture and decapitated. The cochleae from the apical aspects of the sensory epithelia were harvested in an oxygenated external solution containing 140 mm NaCl, 2 mm KCl, 10 mm glucose, 10 mm HEPES, 1.3 mm CaCl₂, pH 7.35, osmolarity ∼290 mOsmol. The tissue was incubated in the oxygenated external solution containing the Ca²⁺ dye, Oregon Green 488 BAPTA-1, AM, cell-permeant (OGB) (20 mm) and TMRM according to the manufacturer’s protocol (Thermo Fisher Scientific; Invitrogen) and kept in the dark for 10 min. The OGB concentration and duration of incubation sufficed for dye diffusion into cellular organelles and subsequent esterification for monitoring cytoplasmic Ca²⁺. The OGB dye was chosen because a previous study showed that mitochondrial Ca²⁺concentration is high ∼220 nm (Ivannikov and Macleod, 2013). Therefore, we predicted that a high-affinity Ca²⁺ indicator would be appropriate for our measurements (Paredes et al., 2008). Also, using TMRM as a mitochondrial marker precludes the use of Ca²⁺ indicators, such as Rhod-2/Rhod-2 AM Thus, we selected OGB, which has been used to measure OHC Ca²⁺ previously (Yuan et al., 2010). The use of TMRM to mark mitochondria allowed for comparing the overlap of Ca²⁺ and mitochondrial fluorescent intensities in ROIs in the cytoplasm.

After a thorough washout, the preparation was transferred to a recording chamber under an upright microscope. The experimental chamber was perfused continuously and slowly with an oxygenated external solution at a rate that ensured stable optical recordings for ∼10–15 min. To make a reasonable comparison between the different ages, we focused on apical OHCs. At the beginning of each experiment, the structural preservation of the apical turn was assessed with low magnification and then zoomed in to record from single OHCs. High-magnification confocal images on single OHCs were collected at 2.5 s time intervals using a SP8 confocal microscope (Leica Microsystems) with a 60× water-immersion lens under the control of LAS X software. The TMRM signal with an absorbance peak at 548 nm and an emission peak at 574 nm was captured with an RFP/TRITC filter. For OGB Ca²⁺ imaging, the experiments were performed using an excitation wavelength of 488 nm, and emission was detected at 510 nm. The image processing software allowed for the recording of a selection of an ROI, in this case, the outline of OHCs excluding the hair bundles. The resolution of imaging was insufficient to assess hair bundle Ca²⁺ accurately. Since OGB is not a ratiometric Ca²⁺ indicator, we could not assess the absolute, but rather the relative intracellular Ca²⁺ levels. The preparation was treated with either low or high Ca²⁺ solution in the presence of the Ca²⁺ ionophore A23187 (20 μm). The low Ca²⁺ solution contained 140 mm NaCl, 2 mm KCl, 10 mm glucose, 10 mm HEPES, 1.85 mm CaCl₂, 5.5 mm EGTA, pH 7.3, osmolarity ∼290 mOsmol. The high Ca²⁺ solution contained 140 mm NaCl, 2 mm KCl, 10 mm glucose, 10 mm HEPES, 2.35 mm CaCl₂, 2.25 mm EGTA, pH 7.3, osmolarity ∼290 mOsmol. The Maxcheletor software (https://somapp.ucdmc.ucdavis.edu/pharmacology/bers/maxchelator/CaEGTA-TS.htm) was used to estimate free Ca²⁺ in low and high Ca²⁺ solutions. The Ca²⁺ fluorescence intensity values for mitochondria (using TMRM signal regions) were measured at 15 s time points before and during bath perfusion of external high K⁺ (at t = 10 s) solution, 122 mm NaCl, 20 mm KCl, 10 mm glucose, 10 mm HEPES, 1.3 mm CaCl₂, pH 7.35, osmolarity ∼290 mOsmol and after changing solution to low K⁺ (5 mm KCl) (at t = 75s). After washout with normal external K⁺ solution (5 mm KCl), ∼25% (89 of 359 OHCs) did not recover from the increased fluorescent intensity. OHCs that underwent an irreversible rise in Ca²⁺ were not included in the data analyses. ΔF/F0 = [F(t) − F (0)]/F(0), t is time, F(t) is fluorescence following stimulation (external K⁺) and F(0) is the prestimulus fluorescence computed from the mean of 5–7 images. F(t) and F(0) were corrected for background fluorescence. A 2-D median filter was used to produce 512 pseudo-color indexed images.

Transmission electron microscopy

The cochlea was isolated, and the perilymph perfused with 1% osmium tetroxide ferricyanide and1% PFA in 0.1 m cacodylate buffer, pH 7.4. They were then immersed in the fixative (30 min, 23°C) and rinsed in the buffer followed by en bloc post-staining with 1% uranyl acetate and Walton’s lead aspartate to enhance contrast. The cochleae were decalcified (120 mm EDTA, 23°C), dehydrated in an ethanol series, and embedded in epoxy resin (Embed 812, EMS). The cochleae were bisected along the modiolus and re-embedded. One-half of the cochlea was sent to National Center for Microscopy and Imaging Research for EM tomography. Mid-modiolar cochlear sections (70 nm) mounted on copper hex grids were obtained from the remaining half cochlea, stained (uranyl acetate-lead citrate), and viewed. Images from the upper basal turn (∼28–32 kHz) were collected and archived using a JEOL 1200 EX transmission electron microscope (JEOL) equipped with an AMT 8-megapixel digital camera and AMT Image Capture Engine V602 software (Advanced Microscopy Techniques). Final figures were assembled using Adobe Photoshop and Illustrator software (Adobe Systems).

Electron microscope tomography

The principles and procedures of EM tomography described previously (Perkins et al., 2015; Sastri et al., 2017) were practiced. Sections from blocks of cochleae upper basal turn sent from tissue processed at Washington University from each age group were cut at a thickness of 400 nm. For each tomographic reconstruction, a double-tilt series of images at 1 degree tilt increment was collected with an FEI titan hi-base electron microscope operated at 300 kV. Images were recorded with a Gatan 4Kx4K CCD camera. The magnification was 11,000×, and the pixel resolution was 0.81 nm. The IMOD software package (https://en.wikipedia.org/wiki/IMOD) was used for alignment, reconstruction, and volume segmentation. R-weighted back projection was used to generate the reconstructions. Samples from 6 animals were used for 3 months and 15 months, and samples from 3 animals were used for 18 months and 24 months. Nine tomographic volumes were generated for 3, 15, and 18 months. Eleven tomographic volumes were generated for 24 months. Each volume had nearly 200 x–y planes and had a size of ∼3.4 µm³. Volume segmentation of the mitochondria, SSC, and their tethers was performed by tracing of membranes in each of the 1.62-nm-thick x-y planes that they appeared to create stacks of contours with the Drawing Tools plug-in in IMOD. Because the mitochondrial membrane profiles did not vary much from plane to plane, the Interpolator plug-in in IMOD was used to accelerate the segmentation of surfaces by automatically generating interpolated contours between user-defined key contours that were manually traced. By using key contours about every 5 planes, accurate membrane segmentation was achieved by generating interpolated contours with straight lines between key contours in each in-between plane. Although the interpolator plug-in was also used to segment the SSC, more manual tracing was required because the SSC meshwork changed more from plane to plane. The mitochondrial-SSC tethers were modeled as rods with a diameter of 10 nm thick, and the crista junctions were modeled as either spheres or tubes with 14 nm diameter. The traced contours were then surface-rendered by turning contours into meshes to generate a 3D model. The mitochondrial reconstructions and surface-rendered volumes were visualized using 3DMOD. Measurements of mitochondrial outer, inner boundary, and cristae membrane surface areas and volumes were made within segmented volumes using IMODinfo. Mitochondrial profile area, volume density (defined as the volume occupied by mitochondria divided by the cytoplasmic volume), number per area, crista density, and crista junction diameter were measured using ImageJ. Crista polarization was determined for each age by using the angle measurement tool in ImageJ with one side along the long axis of a crista and the other side along the SSC chosen so that only acute angles were measured. The measurements were sorted into 10 degree bins from 0 to 90 degrees. Based on visual inspection of many OHC mitochondria, angles >70 degrees were considered as demonstrating cristae alignment facing the SSC (for measurement, see, e.g., Fig. 6A). The degree of polarity was defined as the percent of measures with angles >70 degrees divided by the percent of measurements expected in this angular range if the cristae were randomly oriented. A χ² test was used to test the null hypothesis of random cristae orientation. Movies of the tomographic and surface-rendered volumes were made using the IMOD movie.

Energy calculations

Models of mitochondrial ATP output have lacked structurally realistic input based on 3DEM. The biophysical modeling schema of Song et al. (2013) made steps to remedy this problem by using spatially accurate geometric representations of a crista, inner boundary membrane, and crista junction. From physical principles, this modeling predicted the higher proton motive force on cristae membranes, the effect of crista surface-to-volume ratio on this force, and the effect of crista membrane surface area on the rate of ATP synthesis, all integral to the current paradigm of mitochondrial structure/function relations. We applied this modeling approach to predict ATP production for SSC and PM functions at the lateral wall (graphic associated with Table 1).

Measurements of the crista shape factor, defined as the crista membrane surface area divided by the crista volume, were made from 3DEM mitochondrial volumes using ImageJ tools. The rate of ATP production found in column 2 of Table 1 was derived from Song et al. (2013, their Fig. 5B), which plots the rate of ATP production as a function of the crista surface area for differing values of the crista shape factor. The Gibbs free energy of ATP hydrolysis varies from −28 to −34 kJ/mol, depending on the concentration of the cation Mg²⁺. We took the middle value of −31 kJ/mol for our power calculation in column 3 of Table 1. The estimation of the number of mitochondria in the OHC volume of interest in column 4 of Table 1, indicated by the black bars in Figure 3A was made by first modeling that portion of the volume as a cylinder, which is also how it is described in the literature. Only EM images were chosen that appeared to be longitudinal (see Fig. 3A) or cross-sectional (see Fig. 3E) in the volume of interest. The mitochondria near the SSC were counted in these images, which accounts for >90% of the mitochondria in the volume (somewhat less in aged OHCs as described in Fig. 8). Thus, the model is a cylinder decorated on the inside with mitochondria; that is, the few mitochondria separated from the SSC were not used because we desired to model the rate of ATP production available for use by the SSC and lateral cell wall components. This constraint allows for an estimation of the number of mitochondria decorating the cylinder by multiplying the number of mitochondria in a cross-sectional EM cut by the number of mitochondria in a longitudinal EM cut. But, because the cross-sectional and longitudinal EM images were from different cells, the multiplication can be made using one cross-sectional measurement with “n” longitudinal measurements, thus producing multiple combinations. Longitudinal EM images are as follows: n = 10, 3, 10, and 4 for 3, 15, 18, and 24 months, respectively. Cross-sectional EM images are as follows: n = 4, 9, 3, and 6 for 3, 15, 18, and 24 months, respectively. Mitochondrial number measurements are as follows: n = 40, 27, 30, and 24 for 3, 15, 18, and 24 months, respectively.

There are two important caveats with the calculations presented in Table 1 and the Results. First, this analysis does not take into account the ATP produced by glycolysis. However, a quantitative analysis of the relative contributions of glycolysis and oxidative phosphorylation in the OHC determined that its metabolism is primarily aerobic (Puschner and Schacht, 1997). Thus, ATP production comes mostly from mitochondria; hence, OHC fate is subject to mitochondrial dysfunction as much as other cell types. Second, mitochondrial respiration can be governed by various states, the most common being State 3 (active) and State 4 (resting), and State 3 respiration can have an ATP production rate up to 4 times that of State 4 (Perkins and Ellisman, 2011). It was not clear from the work of Puschner and Schacht (1997) which state (or combination of states) their OHC mitochondria were in. Taking into consideration the estimates and uncertainties, it is important to understand that the calculations of ATP production rate and power (energy per second) availability may have “order-of-magnitude” accuracy.

Experimental design and statistical analyses

The number, ages, and sex of the mice used in this study, the number of OHCs, tomographic volumes, and mitochondria are reported in Materials and Methods, Results, or figure legends, as appropriate. The statistics package of Microsoft Excel was used for statistical comparisons. The mean, SD, SEM, median, and first and third quartiles were calculated for all reported measurements and shown in box-and-whisker plots or other plots, as appropriate. Individual data points were shown, often in “bee-swarm” configuration. The whiskers (vertical lines) extend from the ends of the box to the minimum value (bottom) or maximum value (top). A data point was considered an outlier if it exceeded 1.5 times the interquartile range below the first quartile or above the third quartile. ANOVA was used with Bonferroni’s correction post hoc using the two-tailed unpaired Student’s t test between groups. Pearson’s χ² test was used to determine whether there was a significant difference between the null hypothesis of the random orientation of cristae to the SSC and a polarized cristae orientation. The significance from combined statistical tests, as used in Table 1, columns 4, 5, and 6, followed the procedure given by https://pdfs.semanticscholar.org/e232/ec9efda181afaeb3e99f4c9993d5c6ea79f2.pdf.

The reporting “p value” used is R₂ = p₁ × p₂ (1 + log (1/ (p₁ × p₂)), where p₁ is the p value from the statistical test reported for the first column (e.g., column 4), and p₂ is the p value from the statistical analysis reported for the second column (e.g., column 5). p or R < 0.05 was considered to be statistically significant, and the values of all significant p or R values were reported in the figure captions or Table 1, as appropriate.