With only one exception mentioned below, the organ of Corti explants were dissected from young postnatal rats or wildtype (C57BL6/J) mice at postnatal days 6–12 (P6–12). These are the ages when the OHC electromotility is already easily detectable (Belyantseva et al., 2000). We started this study using both rats and mice, hoping to find evidence for Ca²⁺-dependent regulation of OHC non-linear capacitance (NLC) and continue this study in specific mouse mutants. However, we also wanted to include rat hair cells since they are bigger and better suited for our imaging experiments. Therefore, all experiments with Ca²⁺-induced changes in OHC non-linear capacitance (Figure 5) were performed in mouse OHCs, while the studies of CICR (Figures 1–3) used both mouse and rat OHCs. No differences between the species were found in the latter experiments and the data were combined. After getting no Ca²⁺ effects on NLC, we continued IP₃ uncaging (Figure 4) and stereocilia bundle stiffness (Figure 6) experiments only in rat hair cells at P7–P11 and P4–P7, correspondingly.

The most apical and the most basal (“hook”) regions were cut off before placing the explants in the glass-bottom Petri dishes, where they were held by pre-glued fine glass fibers. The Petri dish was filled with Leibowitz cell culture medium (L-15, Invitrogen) containing the following inorganic salts (in mM): 137.93 NaCl; 5.33 KCl; 1.26 CaCl₂; 0.99 MgCl₂; 1.33 Na₂HPO₄; 0.44 KH₂PO₄; 0.81 MgSO₄. All experiments were performed at room temperature (22–24°C) in the cells located approximately in the middle of the cochlea. All animal procedures were approved by the University of Kentucky Animal Care and Use Committee.

Hair cells were observed using an upright Olympus BX51W1 microscope (Olympus, Center Valley, PA, USA) equipped with a 100× 1.0 N.A. water immersion objective. To access the basolateral plasma membrane of the OHCs and IHCs, a hole in the sensory epithelium was made by suction of a few outermost cells, next to the cell of interest. Then, a new fresh pipette filled with the bath L-15 medium was used to apply positive pressure around the selected cell to prepare a clean surface for patching. Finally, another pipette with an intracellular solution was used to establish whole-cell patch-clamp recordings. For conventional recordings, the pipettes were filled with our standard intracellular solution containing (in mM): CsCl (140), MgCl₂ (2.5), Na₂ATP (2.5), EGTA (1.0), HEPES (5). For experiments with caged Ca²⁺ chelators, the intracellular solution contained (in mM): CsCl (143.2), CaCl₂ (0.7), NP-EGTA (1) and HEPES (10) supplemented with cell-impermeant Ca²⁺ indicator Fluo-4 (0.1) (Invitrogen / Molecular Probes). We found these relative concentrations of NP-EGTA and Ca²⁺ (1 mM vs. 0.7 mM, correspondingly) most effective based on minimal Fluo-4 fluorescence before UV stimulation and a large change in Fluo-4 signal after UV flash. For caged-IP₃ experiments, the intrapipette solution contained (in mM): KCl (12.6), KGlu (131.4), MgCl₂ (2.0), EGTA (0.5), K₂HPO₄ (8.0), KH₂PO₄ (2.0), Mg₂ATP (2.0), Na₄GTP (0.2) plus 100 μM of caged-IP₃ (Enzo Life sciences) and 100 μM of Fluo-4. In all cases, the osmolarity of intrapipette solutions was adjusted with D-glucose to match corresponding values of L-15 (∼325 mOsm) and pH was adjusted to 7.3–7.4. The patch pipettes had a typical resistance of 3–5 MΩ in the bath. Patch-clamp recordings were performed either with the MultiClamp 700B computer-controlled amplifier (Molecular Devices) or with the optical feedback OptoPatch amplifier that has a build-in hardware-based tracking of cell capacitance (Cairn Research, Kent, UK). In both cases, data acquisition was controlled by pClamp 10 software (Molecular Devices). Between the records, HCs were maintained at a holding potential of −60 mV. No correction was applied for liquid-junction potentials or series resistance. During MET recordings, the holding potential was temporarily changed to −80 mV.

Cell-permeable (AM) versions of NP-EGTA and Fluo-4 (both from Invitrogen) were dissolved in DMSO to obtain each stock solution of 4 mM concentration. Immediately before the experiment, the stock solution was diluted to 20 μM in DMEM supplemented with 250 μM of sulfinpyrazone. All specimens were incubated in the prepared solutions for 40 min at 37°C. The specimen was rinsed three times with fresh L-15 and used immediately for the experiment.

Specimens were observed using an upright Olympus BX51W1 microscope equipped with a 100× 1.0 N.A. water immersion objective and DSU spinning disk confocal unit (Olympus, Center Valley, PA, USA). Before starting the experiment, a bright field image of the cell was examined for the signs of membrane blebbing, swelling of the cell body, and/or excessive movements of intracellular organelles. Only the cells without these signs of deterioration were used. The Fluo-4 fluorescence was excited by a mercury light source illumination passing through a 470/40 nm (center/bandwidth) filter and observed with a 525/50 nm filter. The illumination intensity was always adjusted to minimize photo-bleaching. Images were acquired with the Evolve 512 camera (Photometrics, Tucson, AZ) and MetaMorph software (Molecular Devices). Fast photo release of free Ca²⁺ was achieved by ultraviolet (UV) illumination generated by DPSL355/30 laser (Rapp OptoElectronic, Hamburg, Germany) that was coupled to the microscope through an optical fiber. The 355 nm laser beam was focused to a diameter of ∼5 μm at the specimen. It allowed directing the uncaging UV beam to a particular location within a hair cell that, in most cases, was oriented horizontally and perpendicularly to the beam. The UV light was delivered in pulses of 5–150 ms duration. To account for spatial variations in the total amount of Fluo-4 dye along and within the hair cell, changes in Ca²⁺ signal over time were normalized to the fluorescence before UV illumination (ΔF/F₀) at each pixel. These traces were averaged to obtain the Ca²⁺ signal from an overall cell or a region of interest within a cell. The decay time constants were obtained by fit to a single-exponential curve at the first part of ΔF/F₀ trace after UV flash. To explore the propagation of Ca²⁺ wave along the longitudinal axis of an OHC, two-dimensional plots were constructed to show the changes of a longitudinal profile of ΔF/F₀ signal at different time points.

Perforated-patch recordings were performed in cells loaded with cell-permeable NP-EGTA and Fluo-4. The intrapipette solution additionally contained freshly prepared 120 μM of gramicidin. It took approximately 10–20 min for gramicidin to perforate the membrane. At the end of the perforated-patch experiment, we always broke the patch with negative suction through the pipette and monitored patch clamp parameters and Ca²⁺ fluorescence. Breaking the seal resulted in a decrease of series resistance without significant changes in cell capacitance readings, accompanied by a reduction of Ca²⁺ fluorescence, indicating effective diffusion of Ca²⁺ dye from a cell into the pipette.

Cell capacitance was measured with an Optopatch amplifier in the track-in mode. A 10-mV peak-to-peak sinus wave at 2.5 kHz was superimposed on a normal holding potential of −60 mV, and whole-cell compensation controls for series resistance and cell capacitance were manually adjusted to minimize the sinusoidal component of the whole-cell current. At this point, the built-in lock-in amplifier was turned on, the phase was manually optimized, and the capacitance and resistance dithering circuits were activated to calibrate the system. Rs values typically ranged from 5 to 12 MOhm. Cells displaying higher Rs values were discarded. The track-in feedback circuit was then switched on, and its gain gradually increased to the highest stable value (usually 50). C_m traces were recorded at 1 kHz and filtered online at 20 Hz. Slow ramps of the holding potential from −110 mV to +90 mV were applied to obtain the voltage-dependence of an OHC capacitance. The recorded capacitance was fitted to the first derivative of a two-state Boltzmann function as follows:

C_m = C_v + C_lin, where C_m is the total membrane capacitance, C_v is a voltage-dependent (non-linear) component, and C_lin is a voltage-independent (linear) component.

where Q_max is the maximum nonlinear charge moved, V_pk is a voltage at peak capacitance, V is membrane potential, z is valence, e is electron charge, k is Boltzmann’s constant, and T is absolute temperature.

To avoid artifacts of Ca²⁺-dependent changes of overall cell stiffness and associated “slow” motility in OHCs (Dulon et al., 1990; Frolenkov et al., 2003), the effects of intracellular free Ca²⁺ on stereocilia bundle stiffness were studied in cochlear inner hair cells (IHCs), which provided an additional benefit of larger stereocilia that are easier to track with automated image-processing algorithms. All IHCs were patched with our standard intrapipette solution supplemented with cell-impermeable NP-EGTA and Fluo-4. The stereocilia bundles were deflected with a fluid-jet driven by a first-generation analog high-speed pressure clamp (ALA Scientific, Farmingdale, NY, USA). Simultaneously with recordings of the mechano-electrical transduction (MET) currents, the deflections of hair bundles were recorded with an Evolve 512 camera. The digitized video images were analyzed off-line using routines developed with the Image Processing Toolbox of MATLAB (The Mathworks Inc.). The movement quantification was based on the modified algorithm previously developed to quantify OHC electromotility (Frolenkov et al., 1997). Briefly, we calculated a frame-by-frame shift of the intensity profiles along a line perpendicular to the stereocilia rows in a hair bundle. These measurements were repeated over three-to-five locations along the stereocilia row and averaged to obtain the average movement of the stereocilia bundle in this cell. Care was taken to include both central and side stereocilia in the bundle, except the very edges of the bundle. Then, the relationship between bundle deflection (ΔX) and fluid-jet pressure (P) was constructed for each cell before and after Ca²⁺ uncaging. Although the exact stiffness of the hair bundle cannot be determined with fluid-jet stimulation, the inverse of the slope of ΔX(P) relationship is proportional to the stiffness and the changes of this parameter in the same cell could be used as a measure of relative changes in the stereocilia bundle stiffness. The MET currents during these deflections were simultaneously recorded and the current-displacement I(ΔX) curves were constructed. These I(ΔX) curves were fitted to a second-order Boltzmann function to obtain the estimate of the open probability of MET channels at resting bundle position (P_OPEN) as previously described (Peng et al., 2016).

Most experiments required simple assessment of the differences between two groups of the cells or assessment of the changes happening before and after a single treatment (UV-flash). Correspondingly, two-sample or paired-sample t-tests with Welch correction were used. However, statistical significance of the changes of hair bundle stiffness over time and relative to control (Figure 6H) was estimated by two-way ANOVA with post-hoc assessment by t-test. All statistical analysis was performed with OriginPro 2023b (OriginLab Corporation, Northampton, MA, USA).

Next, we systematically investigated the propagation of Ca²⁺ signals along the length of an OHC. Two fundamentally different conditions of Ca²⁺ buffering inside an OHC were studied. First, the cell was loaded with NP-EGTA+Ca²⁺+Fluo-4 via patch pipette simultaneously with whole-cell recordings (Figure 2A). In this case, Ca²⁺ buffering is controlled by intrapipette solution and imaging does not require confocal mode since the recorded cell is the only fluorescent cell in the organ of Corti explant. In the second set of experiments, cochlear explants were loaded with cell-permeable (AM) versions of NP-EGTA and Fluo-4 before imaging (Figure 2B). In the latter case, the actual concentrations of NP-EGTA, Ca²⁺ bound to NP-EGTA, and Ca²⁺-sensitive portion of Fluo-4 inside the cell were all unknown but they are likely to be significantly less compared to the loading through patch pipette. Therefore, Ca²⁺ buffering inside an OHC was closer to the endogenous buffering. However, Ca²⁺ imaging sometimes required confocal mode due to the background fluorescence of other cells within the organ of Corti.

In both patched and intact OHCs, spot UV illumination to the apex of the cell produced an almost instant increase of [Ca²⁺]_i at the illumination site (Figures 2A, B). Then, Ca²⁺ signal propagated along the OHC longitudinal axis as a wave in case of a patched OHC (Figure 2A) and as a fast Ca²⁺ spread in case of an intact OHC (Figure 2B). The amplitude of UV-flash-induced increase in Ca²⁺ signal was significantly larger and the speed of apex-to-base propagation of this signal was considerably faster in intact OHCs compared to patched OHCs, while the difference in time constants of fluorescent signal decay was not statistically significant (Table 2). These results indicate apparently different Ca²⁺ buffering conditions that may affect the amplitude and spatial propagation of Ca²⁺ signals but similar Ca²⁺ extrusion mechanisms that are likely to determine the time constant of [Ca²⁺]_i clearance.

The fluorescent intensities averaged over the whole cell peaked immediately after UV flash and then decayed over time (Figure 2C, right). However, in contrast to patched OHCs, the fluorescent signal in intact OHCs decayed slower in the first ∼3 s after UV flash (Figure 2C, inset). This may be due to the saturation of Fluo-4 indicator by a larger increase of [Ca²⁺]_i. Indeed, when we plot the relationship between ΔF/F₀ values at the time point of 3 s after UV illumination versus maximal ΔF/F₀ values in the same cell, the data from both patched and intact OHCs followed the same trend (Figure 1C). In other words, larger ΔF/F₀ responses had a more prolonged “plateau” in Fluo-4 signal before subsequent exponential decay, which is indicative of Fluo-4 saturation.

Interestingly, the speed of base-to-apex propagation of Ca²⁺ signal (initiated by UV illumination at OHC base) was significantly faster than that of apex-to-base (initiated by UV illumination at the OHC apex) in patched OHCs (Table 2). In contrast, intact unpatched OHCs didn’t exhibit such a big difference in propagation speed between apex-to-base and base-to-apex directions (Table 2). Therefore, slow apex-to-base propagation of Ca²⁺ signal in patched OHCs results from some factors of whole-cell recordings. The most obvious factor is the constant diffusion of “fresh” (unexposed to UV illumination) NP-EGTA from the patch pipette into the cell. Since we always patched the OHCs at the base, this factor produces asymmetric Ca²⁺ buffering, thereby reducing the Ca²⁺ signal coming from the apex to the base but not affecting the propagation of free Ca²⁺ that was uncaged at the base.

Analysis of the changes in the longitudinal intensity profile of ΔF/F₀ Fluo-4 signal after UV illumination may reveal the sites of Ca²⁺-induced Ca²⁺ release along the OHC body. Indeed, if free Ca²⁺ is liberated by UV flash at a particulate spot within a cell, it is expected to diffuse from this uncaging spot along the cell body until fading away by Ca²⁺ extrusion mechanisms or reaching uniform steady-state increase of [Ca²⁺]_i throughout the cell. The peak of the longitudinal intensity profile of ΔF/F₀ signal is expected to decrease throughout time but its location along the length of an OHC should stay at the original source of free Ca²⁺, i.e., the site of Ca²⁺ uncaging. This is exactly what we observed in the first ∼1.5 s after UV illumination to the apex of the patched OHCs (Figure 2D and E, red traces). However, by ∼3 s after UV illumination, the peak of ΔF/F₀ profile suddenly shifted toward the cell nucleus and maintained there with very little signs of dissipation for the remaining ∼5 s of observation (Figure 2D, right and Figure 2E, green traces). This sudden shift of the peak in the intensity profile of Ca²⁺ signal cannot be explained by a simple diffusion. Most likely, free Ca²⁺ diffusing from the apex of the cell reached Ca²⁺-sensitive stores at the base and evoked CICR there. Then, this effect is expected to be even more prominent in the intact OHCs that would not have continuous diffusion of fresh Ca²⁺ chelator into the cell via patch pipette at the cell’s base.

Indeed, while patched OHCs with Ca²⁺ uncaging at their apex (n = 7) exhibited a relatively minor secondary peak at the base of the cell on 2D plots of ΔF/F₀ profile vs. time (Figures 2D, 3A), this peak was significantly larger in intact OHCs (n = 13) in the same experimental conditions (Figure 3B). In a few cells that had a smaller increase of Ca²⁺ signal after uncaging, the ΔF/F₀ signal around the nuclear region became even higher than the average value throughout the cell (data not shown). Thus, our data demonstrates the existence of Ca²⁺-induced Ca²⁺ release stores at the base of the OHCs. However, they cannot yet refute the existence of such stores at the apex of the cell because any additional signal from free Ca²⁺ released from the stores would overlap with uncaged Ca²⁺ when the UV beam is focused directly on the site of CICR. Therefore, we tried to explore a potential CICR at the OHC apex by pointing UV illumination to the base of the cell (Figure 3D, n = 5). We also tried fast confocal imaging (Figure 3C, n = 3), arguing that faster and smaller optical section imaging could help since CICR is expected to be delayed relative to UV flash. In both cases, we did not detect any secondary peaks on 2D plots of ΔF/F₀ profile vs. time that would indicate the presence of functional CICR at the apex of OHCs. However, Ca²⁺ uncaging at the base of intact OHCs caused there a long-lasting and apparently self-sustained increase of [Ca²⁺]_i (Figure 3D). It was even possible to resolve a secondary increase in Ca²⁺ signal after dissipation of uncaged Ca²⁺ (Figure 3D, bottom). We concluded that functional Ca²⁺ stores generating CICR are located at the base but not the apex of OHCs.

It has been reported that extracellular ATP can generate an increase of [Ca²⁺]_i localized at the apex of the OHC in nominal Ca²⁺-free extracellular medium, apparently by releasing Ca²⁺ from IP₃ gated stores (Mammano et al., 1999). However, IP₃-gated Ca²⁺ release in OHCs has not yet been visualized directly. Likewise, it is not clear whether the putative IP₃-gated stores are functional at the OHC lateral wall and/or the base of the cell. Therefore, we imaged Ca²⁺ responses in the OHCs after IP₃ uncaging. Since IP₃ is available in the cytosol within milliseconds after flash photolysis, we hoped to resolve the location of the IP₃-gated store along the length of an OHC. To separate Ca²⁺ signals in OHCs and neighboring non-sensory Deiters’ cells that are very sensitive to IP₃ (Lagostena and Mammano, 2001), we used 100 μM of cell-impermeable caged IP₃ supplemented with 100 μM of Fluo-4 and delivered them into an OHC through the patch pipette. As expected, IP₃ uncaging at the OHC apex generated a very robust Ca²⁺ response (Figure 4A). However, the increase of [Ca²⁺]_i after IP₃ uncaging was rather slow (Figure 4B), which was very different from Ca²⁺ uncaging in similar experimental conditions (Figure 1F). In average, the amplitude of IP₃-gated increase of ΔF/F₀ signal was 12.9 ± 1.7% and the time constant of this increase was 6.3 ± 1.0 s (n = 8). The uniform increase of [Ca²⁺]_i throughout an OHC was also observed when IP₃ uncaging was performed at the base of the cell (data not shown). The most likely explanation for such a slow rise of [Ca²⁺]_i is that IP₃ quickly diffuses from the site of uncaging throughout the cell and then stimulates slow-activating IP₃ receptors. For comparison, we observed an order of magnitude faster Ca²⁺ responses in the supporting cells that were patched with the same intrapipette solution containing the same concentration of caged IP₃ and stimulated by the same UV flash (Figures 4C, D). It is worth mentioning that we were able to generate Ca²⁺ responses in OHCs only with 100 μM of caged IP₃, a significantly higher concentration than the one that is needed for IP₃-gated Ca²⁺ release in Deiters’ or Hensen’s cells, 8 and 16 μM correspondingly (Lagostena and Mammano, 2001; Lagostena et al., 2001). We concluded that IP₃ receptors in OHCs are less sensitive than in supporting cells, either due to the expression of a different isoform of IP₃R or due to the downregulation of their sensitivity. As to the location of IP₃-gated Ca²⁺ stores along the length of an OHC, our imaging does not reveal any obvious “spots” of Ca²⁺ release, suggesting that they may be distributed throughout the cell. Interestingly, simultaneous patch clamp recordings with K⁺-based intrapipette solution and −60 mV holding potential revealed a small but significant positive shift of the whole-cell current by 6.9 ± 2.8 pA (p < 0.05, paired t-test) after IP₃ uncaging (Figure 4B, bottom), which is consistent with an outward current through Ca²⁺-activated K⁺-channels (Evans, 1996) that are located at the base of the OHCs (Oliver et al., 2000).

Our previous data suggested that OHC motor protein, prestin, could be regulated by Ca²⁺/calmodulin dependent phosphorylation (Frolenkov et al., 2000). Furthermore, subsurface cisternae, a putative intracellular Ca²⁺ store (Schulte, 1993), resides in the proximity of the OHC lateral plasma membrane, where prestin is highly expressed (Belyantseva et al., 2000; He et al., 2010; Holley and Ashmore, 1988; Holley and Ashmore, 1990; Holley et al., 1992). Therefore, a long-standing (but never proved) hypothesis suggests that the release of free Ca²⁺ from subsurface cisternae may regulate the operation of prestin-based OHC plasma membrane motors. To test this hypothesis, we explored the effects of Ca²⁺ uncaging on voltage-dependent (non-linear) capacitance of OHCs, a known “signature” of operation of prestin-based plasma membrane motors in OHCs (Santos-Sacchi, 1991). We measured the changes of OHC non-linear capacitance after establishing whole-cell recordings in the same cell before and 8.5–9.5 s after Ca²⁺ uncaging (Figure 5A). As in the above experiments (see Figure 1), the patch pipette contained Cs⁺-based intrapipette solution, supplemented with Ca²⁺-bound NP-EGTA, and Ca²⁺ indicator Fluo-4. Only the OHCs exhibiting stable recordings of whole-cell current and cell capacitance were stimulated by a UV beam, which was targeted to the upper half of the cell, away from the nucleus region that has a smaller density of plasma membrane motors compared to the later wall region of an OHC (Dallos et al., 1991; Huang and Santos-Sacchi, 1993). In all these cells, UV flash produced a very fast and substantial increase of [Ca²⁺]_i (Figure 5A, insets). Even though our hardware continuously measured OHC capacitance, we didn’t observe any noticeable changes in cell capacitance either immediately or several seconds after Ca²⁺ uncaging (Figure 5A). Statistical analysis revealed only a small rightward shift of C_m(V) curve after Ca²⁺ uncaging (Figure 5B and Table 3). However, a similar rightward shift was observed in control OHCs that were stimulated by UV flash but were not loaded with NP-EGTA+Ca²⁺ and hence were not exposed to the [Ca²⁺]_i increase (Figure 5B, inset and Table 3).

Since whole-cell configuration may result in the washout of essential diffusible molecules mediating potential Ca²⁺-dependent regulation of prestin function, we repeated these experiments in OHCs in perforated patch clamp conditions (n = 5). Similar to our previous observations of the larger ΔF/F₀ changes in the intact OHCs (Figure 2C), UV uncaging produced larger Ca²⁺ changes in OHCs in perforated patch conditions (Figure 5C, bottom inset). Yet, we did not observe any statistically significant changes in the non-linear capacitance after Ca²⁺ uncaging, even in perforated patch conditions (Figure 5C). After completing these experiments, we ascertained the perforated patch conditions by breaking the plasma membrane at the tip of the pipette and monitoring the resulting drop in access resistance (Figure 5C, right inset). We concluded that the direct increase of [Ca²⁺]_i does not produce any effects on prestin-based non-linear capacitance, at least in young postnatal OHCs and on a time scale of up to 11 s.

Similar to the putative Ca²⁺ stores in the subsurface cisternae, the function of Ca²⁺ stores at the hair cell apex is also largely hypothetical. However, in contrast to subsurface cisternae, IP₃-dependent Ca²⁺ release at the apex of OHCs has been observed by direct Ca²⁺ imaging in response to extracellular ATP (Mammano et al., 1999). Based on the proximity to the stereocilia bundle, it was hypothesized that the apical Ca²⁺ stores may be involved in the regulation of mechanical properties of the hair bundle (Mammano et al., 1999). This hypothesis is consistent with the old data demonstrating the increase of stereocilia bundle stiffness in chick cochlear hair cells by permeabilizing them with Ca²⁺ ionophores and raising Ca²⁺ concentration in the bath (Pae and Saunders, 1994). Therefore, our next step was to explore whether Ca²⁺ uncaging at the apex of the cell could affect the stiffness of the stereocilia bundle. To avoid confounding effects of the global Ca²⁺-induced softening of cytoskeletal stiffness, which occurs in OHCs (Frolenkov et al., 2003) and is accompanied by an increase of the fluid-jet-induced rocking movement of the cuticular plate (data not shown), we performed these experiments in the inner hair cells (IHCs). IHCs have also the advantage of larger stereocilia that are easier to monitor with automatic tracking algorithms. Using our standard intracellular solution containing NP-EGTA+Ca²⁺+Fluo-4, we established whole-cell patch clamp recording configuration, waited for at least 5 min for the intrapipette solution delivery into the cell, and then recorded MET currents and stereocilia deflections before and after UV illumination to the hair cell (Figure 6A). Although UV flash targeted to a cell through the cuticular plate was less effective in Ca²⁺ uncaging, the train of UV pulses generated a substantial increase in ΔF/F₀ signal (Figure 6B) and a drop in whole-cell current due to activation of Ca²⁺-sensitive ion channels (Figure 6E). Simultaneous video recordings allowed off-line analysis of stereocilia bundle movements (Figure 6C) and estimating the relationship between the amplitude of hair bundle deflection (ΔX) and the fluid-jet pressure (P) in the same cell before and after UV illumination. Data showed that Ca²⁺ uncaging reduces the amplitude of fluid-jet-evoked bundle deflections, which means that the hair bundle becomes stiffer (Figure 6D). Statistical analysis demonstrated that Ca²⁺ uncaging, indeed, decrease the slope of the relationship between bundle deflection (ΔX) and fluid-jet pressure (P), which is proportional to the stereocilia bundle compliance and reciprocal to the bundle stiffness (Figure 6H). Stiffening of the stereocilia bundle was not observed in control experiments that used the same intrapipette solution but without NP-EGTA+Ca²⁺ (Figure 6H).

Interestingly, we did not observe any apparent effects of Ca²⁺ uncaging on the MET responses recorded in the same cell (Figure 6E). In a group of IHCs (n = 9), the average ΔF/F₀ increase was 63 ± 6% (min/max: 34/95%). Yet, in the same cells, we did not detect any statistically significant changes in the open MET channel probability at resting bundle position (P_OPEN) (Figure 6G). Amplitudes of MET currents were lined up along the same MET current-displacement curve, albeit at smaller bundle deflections after Ca²⁺ uncaging (Figure 6I). Thus, we believe that potential Ca²⁺ uncaging inside stereocilia, if present, was too small in our experiments to affect MET responses, while limited amount of liberated free Ca²⁺ from the soma of the cell cannot easily reach the tips of stereocilia. We concluded that a moderate increase of [Ca²⁺]_i within the soma of IHCs can stiffen stereocilia bundles without apparent effects on MET responses.

CICR has been proposed to regulate the OHC response to efferent neurotransmitter acetylcholine. This proposal was based on the experiments demonstrating that the drugs affecting either RyRs or endoplasmic Ca²⁺ ATPases also modulate acetylcholine-induced OHC hyperpolarization and/or outward K⁺ current through Ca²⁺-activated K⁺ channels (Lioudyno et al., 2004). Yet, this proposed CICR has never been visualized in OHC by Ca²⁺ imaging, perhaps due to the fact that it should occur in a narrow gap between synaptic cisternae and the plasma membrane (Fuchs, 2014). In fact, CICR in OHCs has never been demonstrated with direct Ca²⁺ imaging and, therefore, this is the first study to do this. Although our imaging experiments cannot pinpoint the exact site of CICR with nanoscale accuracy, they do show that CICR occurs at the base but not the apex of OHCs. On one hand, it is consistent with the proposed roles of CICR in the modulation of acetylcholine-induced responses in OHCs. On the other hand, it raises the question of why CICR occurs predominantly at the base of the cell and how it is related to the proposed Ca²⁺ stores in the synaptic cisternae. Post-embedding immunogold labeling with antibodies that predominantly recognize ryanodine receptor isoforms 1 (RyR1) and 2 (RyR2) revealed the highest RyR signal in the OHCs in the cytoplasmic region above the nucleus, about 30% smaller labeling at the very base of the cell, and about three-fold smaller labeling at the apex of the cell (Grant et al., 2006). These data are consistent with our imaging experiments (Figures 2, 3). However, they also indicate that the area of functional CICR in OHCs may be larger that is needed only for the modulation of acetylcholine-induced responses. The potential role of CICR in the area above the nucleus remains unknown.

Expression of intracellular Ca-ATPase along the lateral walls of the OHCs (Schulte, 1993; Zine and Schweitzer, 1996) provides a strong argument for the existence of the functional Ca²⁺ stores in the subsurface cisternae. However, Ca²⁺ release from these stores has never been visualized by Ca²⁺ imaging. Here, we demonstrated a slow and uniform Ca²⁺ release throughout the OHC after IP₃ uncaging (Figure 4), suggesting that functional IP₃-gated stores are likely to be distributed along the length of an OHC, including the lateral wall. It is consistent with our previous data showing uniform IP₃R labeling along the lateral wall of the OHCs (Frolenkov et al., 2000). We also did not reveal any signs of CICR in the middle of an OHC, away from the nucleus region. The latter observation suggests that the lateral wall stores may be activated predominantly by IP₃, at least in young postnatal OHCs. That said, elaborated system of the lateral wall subsurface cisternae may not be completely developed in the rodent OHCs at P12, the maximal age of our experiments (Souter et al., 1995). Therefore, functional properties of the lateral wall Ca²⁺ stores may be different in adult OHCs.

A known potential target for free Ca²⁺ released from the lateral wall subsurface cisternae is the cortical cytoskeleton that determines OHC axial stiffness. It is known that the rise of [Ca²⁺]_i causes a decrease of OHC axial stiffness (Frolenkov et al., 2003) likely mediated by Rho GTPases (Kalinec et al., 2000; Zhang et al., 2003) and downstream LIMK/cofilin-dependent actin depolymerization (Matsumoto et al., 2010). These cytoskeletal changes underlie the acetylcholine-induced changes of OHC somatic motility (Dallos et al., 1997) that occur without any changes in the operation of prestin-based plasma membrane motors (Frolenkov et al., 2000; Matsumoto et al., 2010). The experiments of this study show that even a direct increase of [Ca²⁺]_i by Ca²⁺ uncaging is not able to affect prestin motors (Figure 5). Perhaps prestin is regulated through Ca²⁺-independent mechanisms such as phosphorylation via cyclic GMP / protein kinase G (cGMP/PKG) pathway (Deak et al., 2005; Szonyi et al., 1999).

Visualization of ATP-activated Ca²⁺ release from IP₃-dependent stores at the apex of OHCs, right beneath the cuticular plate, led to the hypothesis that these stores may be involved in the regulation of the mechanical properties of stereocilia bundle (Mammano et al., 1999). Unfortunately, we were not able to test this hypothesis directly in OHCs due to the confounding effects of Ca²⁺-dependent softening of the OHC cytoskeleton and the increased rocking motion of the cuticular plate upon fluid-jet stimulation. However, our data show that, at least in IHCs, a rise of [Ca²⁺]_i results in the stiffening of the stereocilia bundle (Figure 6). Although it is yet unclear why the increase of intracellular free Ca²⁺ produces opposite effects on somatic and hair bundle cytoskeletons (softening vs. stiffening, correspondingly), our data are consistent with the previously reported effects of Ca²⁺ on the hair bundle stiffness in chick hair cells permeabilized with ionophore (Pae and Saunders, 1994).

At first sight, it is surprising that Ca²⁺ uncaging did not affect MET responses in our experiments. However, as we already mentioned, the efficiency of UV pulses in liberating free Ca²⁺ was substantially lower when the UV beam was penetrating the cell through hair bundle and cuticular plate, perhaps, due to increased light scattering. Therefore, we do not know whether this experimental arrangement produces any sizable increase of [Ca²⁺]_i within a stereocilium—most of the visible increase in fluorescence came from within a cell (Figure 6B, insets). Second, [Ca²⁺]_i in the tips of stereocilia is expected to be in order of several micromoles (Fettiplace and Nam, 2019). It is very unlikely that our Ca²⁺ uncaging ever reached these levels, bearing in mind that we observed an increase in Fluo-4 fluorescence only within 34–95% range (Figure 6B).

The overall stiffness of a hair bundle is often separated into two components: (i) the stiffness at the stereocilia pivot points; and (ii) the stiffness associated with the tip links and the MET apparatus (Mecca et al., 2022; Tobin et al., 2019). In general, it is at least a very oversimplified assumption due to multiple other links interconnecting stereocilia, especially in OHCs (Goodyear et al., 2005). However, for the purpose of this discussion, we could separate the hair bundle stiffness into “pivot stiffness of stereocilia” and “stereocilia link-associated stiffness.” The fact that Ca²⁺ uncaging changes the hair bundle stiffness in our experiments without apparent effects on MET currents (Figure 6) may indicate that free Ca²⁺ from the soma of the cell cannot reach inside stereocilia and affects only “pivot stiffness.” It could occur through Ca²⁺-dependent modifications of proteins located specifically at the base of stereocilia (Pacentine et al., 2020). This hypothesis would be also consistent with an idea of “self-contained” Ca²⁺ compartment within a stereocilium that is generally isolated from the rest of the cell by specialized Ca²⁺ clearance mechanisms (Dumont et al., 2001; Fettiplace and Nam, 2019; Lumpkin and Hudspeth, 1998). However, we would refrain from making a definitive conclusion. Ca²⁺-induced changes in hair bundle stiffness develop within tens of seconds (Figure 6H). This slow time course allows plenty of time for any possible indirect effects, for example, by immobilizing stereocilia myosin motors and preventing them from entering stereocilia.

Independent of the potential mechanisms, our study, to the best of our knowledge, is the first to explore the changes of stereocilia bundle stiffness in cochlear hair cells in response to physiologically relevant changes of [Ca²⁺]_i inside the cell body. In the future, it would be interesting to explore whether more physiological stimuli, such as low-level extracellular ATP, could produce similar changes in stereocilia bundle stiffness.