The animal experimental protocol was approved by the Ethics Committee of our institute (No. 270301, No. 280301, and No. 290303). This study was performed according to the Guiding Principles for the Care and Use of Animals in the field of physiological sciences, as approved by the Council of the Physiological Society of Japan and the American Physiological Society. All animal experiments followed the guidelines established by the U.S. National Institutes of Health regarding the care and use of animals for experimental procedures, as well as the UK Animals (Scientific Procedures) Act, 1986.

We obtained dental pulp slice preparations from neonatal Wistar rats (6- to 7-days-old) using a previously described method (Shibukawa and Suzuki, 2003; Tsumura et al., 2010, 2012, 2013; Shibukawa et al., 2015; Kimura et al., 2016). Animals were anesthetized using isoflurane (3%) and pentobarbital sodium (25 mg/kg), and the mandible was dissected. The mandible was embedded in an alginate impression material, and 500-μm thick slices were obtained by transversely sectioning tissue through the incisors using a standard vibrating tissue slicer (Dosaka EM, Kyoto, Japan). The mandible was sectioned to the level where the dentin and enamel were directly visible between the bone tissue and the dental pulp. The surrounding impression material, bone tissue, enamel, and dentin were carefully removed from these mandible sections under a stereoscopic microscope, and the residual dental pulp slice was used for further experiments. We selected mandible sections with a thin dentin layer to avoid cellular damage to the odontoblasts, but with enamel and dentin that could clearly be distinguished under a microscope. Pulp slices were treated with a standard Krebs solution containing 0.03% trypsin and 0.17% collagenase at 37°C for 30 min. For patch-clamp recording, enzymatically treated dental pulp slices were plated onto a culture dish, immersed in alpha-minimum essential medium (Thermo Fisher Scientific, Waltham, MA, USA), supplemented with 10% fetal bovine serum and 5% horse serum, and maintained at 37°C in a 5% CO₂ incubator. Primary cultured dental pulp slices were used for patch-clamp recording within 24 h of isolation.

An immortalized cell line derived from dental pulp of the human molar with odontoblastic characteristics (HOB cells; Kitagawa et al., 2007; Ichikawa et al., 2012; Kimura et al., 2016) was cultured in basal media containing alpha-minimum essential medium, 10% fetal bovine serum, and 1% penicillin/streptomycin (Thermo Fisher Scientific).

Whole-cell recordings (Hamill et al., 1981) of odontoblasts in the primary cultured dental pulp slices were conducted through voltage-clamp recording under patch-clamp technique using a conventional method. Cells located in the rim of cultured dental pulp slices positively expressed odontoblast marker proteins of dentin matrix protein-1, dentin sialoprotein, and nestin (Tsumura et al., 2012), indicating that these cells were odontoblasts. Patch pipettes, with a resistance of 3 to 8 MΩ, were made from glass capillary tubes (DMZ-Universal Puller; Zeitz-Instruments, Martinsried, Germany) and filled with an intracellular solution. When the patch pipette was attached onto the plasma membrane, we immediately measured the resistance (seal resistance) between the pipette and membrane. The mean value of initial seal resistance was 2.3 ± 0.7 GΩ (“gigaseal”; N = 51). The membrane resistance of the cells during whole-cell recording was calculated from the current amplitude evoked by a 10 mV depolarizing voltage step from a Vh of –70 mV. The mean value of membrane resistance was 988.1 ± 112.3 MΩ (N = 51). We measured whole-cell currents with an amplifier for patch-clamp recordings (L/M-EPC-7 plus; HEKA Elektronik, Lambrecht, Germany). After digitization of the analog signals at 10 kHz (Digidata 1440A; Molecular Devices, Sunnyvale, CA), current traces were monitored and stored using pCLAMP (Molecular Devices). Data were analyzed with pCLAMP and the technical graphics/analysis program, ORIGIN, on an offline computer (OriginLab Corporation, Northampton, MA, USA). All experiments were performed at 25°C. We calculated the membrane capacitance of odontoblasts using the capacitative transient current induced by depolarizing steps (10 mV) starting from a holding potential (Vh) of 0 mV. Small differences in odontoblast size were accounted for by normalizing the measured capacitance and expressing current amplitudes in terms of current densities (pA/pF).

Cultured HOB cells were grown to full confluency in basal media and then grown in mineralization media, containing 10 mM β-glycerophosphate and 100 μg/mL ascorbic acid (final concentration) in basal media, at 37°C with 5% CO₂. To examine the inhibitory effects of voltage-dependent K⁺ channels on mineralization by odontoblasts, tetraethylammonium chloride (TEA; 2 or 4 mM, N = 6, respectively) was applied to the mineralization medium over a 21 day period. We exchanged the medium once every 3 days. To detect calcium deposits, cells were subjected to alizarin Red and von Kossa staining (Suzuki et al., 2014; Chen et al., 2016; Kimura et al., 2016).

Krebs solution, containing 136 mM NaCl, 5 mM KCl, 2.5 mM CaCl₂, 0.5 mM MgCl₂, 10 mM HEPES, 10 mM glucose, and 12 mM NaHCO₃ (pH 7.4 by Tris) was used as the standard extracellular solution (ECS) and Cl⁻-rich ECS for patch-clamp recording. The Cl⁻-rich intracellular solution (ICS) contained 140 mM KCl, 10 mM NaCl, and 10 mM HEPES (pH 7.2 by Tris). For patch-clamp recording under physiological conditions, we used solutions of Cl⁻-rich ECS and Cl⁻-rich ICS. To record pure K⁺-conductance, we substituted NaCl in the Cl⁻-rich ECS and KCl in the Cl⁻-rich ICS with Na-gluconate and K-gluconate, respectively (gluc-rich ECS/ICS). TEA and 4-aminopyridine (4-AP) were obtained from Wako Pure Chemicals (Osaka, Japan). α-Dendrotoxin (αDTX) was obtained from Alomone Laboratories (Jerusalem, Israel). We prepared stock solutions of these reagents in distilled water. The stock solutions were then diluted with ECS to the appropriate concentration immediately before the experiments. We purchased all other reagents from Sigma Chemical Co. (St. Louis, MO, USA).

We expressed the results as mean ± standard deviation (SD) for an N number of observations. We represented the number of tested cells as N. The Wilcoxon signed-rank test or Steel–Dwass multiple comparisons were used to evaluate non-parametric statistical significance. Values of P < 0.05 were considered significant.

We measured the resting membrane potential (Rm) immediately after establishing the whole-cell recording configuration using conventional method. The averaged Rm value was −56.2 ± 5.3 mV (N = 19) in Cl⁻-rich ECS (with extracellular 5 mM KCl) and Cl⁻-rich ICS. These isolated odontoblasts had a membrane capacitance of 13.1 ± 2.5 pF (N = 19) under physiological conditions.

Voltage steps (400 ms in duration) ranging from −100 to +80 mV in 10 mV increments, from a holding potential (Vh) of −70 mV (upper traces in Figure 1A), elicited time-dependent outward currents in both the physiological Cl⁻-rich ECS/ICS (middle traces in Figure 1A) and gluc-rich ECS/ICS with an extracellular K⁺ concentration ([K⁺]_o) of 5 mM (lower traces in Figure 1A). We obtained current-voltage (I–V) relationships (Figure 1B) by plotting the amplitude of the steady-state component of these currents against the membrane potential in the presence of four different [K⁺]_o ranging from 5 to 100 mM. The outward current activation threshold showed a 10 mV negative shift for the current recorded under gluc-rich conditions, compared to those measured in Cl⁻-rich ECS/ICS (Figure 1B). The I-V relationships for the currents recorded in both the Cl⁻-rich ECS/ICS and gluc-rich ECS/ICS, with various [K⁺]_o (5–100 mM), showed outward rectification (Figure 1B). The amplitudes of outward currents in both the Cl⁻-rich and gluc-rich conditions decreased as [K⁺]_o increased, suggesting the currents were carried by K⁺. Each activation threshold value for the outward currents, in both the physiological and gluc-rich ECS/ICS supplemented with 100 mM [K⁺]_o, also showed a shift toward depolarizing membrane potentials compared to the value recorded with 5 mM [K⁺]_o.

To investigate ion selectivity of the outward currents, we analyzed tail currents and measured their reversal potentials at four different [K⁺]_o: 5, 10, 50, and 100 mM. In each experiment, tail current families were elicited by voltage steps (upper traces) from Vh of −70 mV to +80 mV before hyperpolarizing in 10 mV increments from −110 to +40 mV under both Cl⁻-rich (middle traces) and gluc-rich conditions (lower traces) (Figure 2A). We measured tail current amplitudes 50 ms after the start of the hyperpolarizing pulse (dashed arrows in Figure 2A). We plotted the amplitudes as a function of the applied membrane potential and subsequently obtained the reversal potentials. The mean reversal potential values under physiological conditions were −44.5 ± 10.2 mV in 5 mM [K⁺]_o, −38.8 ± 6.6 mV in 10 mM [K⁺]_o, −25.3 ± 2.1 mV in 50 mM [K⁺]_o, and −13.8 ± 2.4 mV in 100 mM [K⁺]_o (filled circles in Figure 2B; N = 6). The mean reversal values in the gluc-ECS/ICS were −68.3 ± 7.0 mV in 5 mM [K⁺]_o, −53.3 ± 5.6 mV in 10 mM [K⁺]_o, −30.0 ± 3.2 mV in 50 mM [K⁺]_o, and −12.5 ± 2.5 mV in 100 mM [K⁺]_o (open circles in Figure 2B; N = 6).

Semilogarithmic plots of the reversal potentials against [K⁺]_o (5–100 mM) demonstrated that the mean reversal potential values in the gluc-rich ECS/ICS were similar to those expected for a current showing K⁺-selectivity, as estimated by the Nernst equation at 25°C (Figure 2B, solid line) by assuming an intracellular K⁺ concentration of 140 mM (see Materials and Methods). We also observed deviations of the reversal potentials recorded under physiological conditions from those estimated by the Nernst equation for pure K⁺ conductance. The Cl⁻ permeability was estimated according to the Goldman-Hodgkin-Katz equation from recorded reversal potentials under physiological conditions, where the extracellular Cl⁻ concentration ([Cl⁻]_o) was 147 mM, the intracellular Cl⁻ concentration was 150 mM, and the [K⁺]_i was 140 mM with various [K⁺]_o concentrations. Under these conditions, we estimated a Cl⁻ permeability of 0.15 in 5 mM [K⁺]_o, 0.17 in 10 mM [K⁺]_o, 0.03 in 50 mM [K⁺]_o, and 0.42 in 100 mM [K⁺]_o. The overall mean Cl⁻ permeability value was 0.19 ± 0.16 (K⁺ permeability was set to 1.0), with the estimated reversal potential values being −41.4 mV ([K]_o = 5 mM), −37.8 mV ([K]_o = 10 mM), −19.5 mV ([K]_o = 50 mM), and −6.9 mV ([K]_o = 100 mM) (dashed line in Figure 2B).

We examined the voltage dependence of the steady-state inactivation kinetics of outward K⁺ currents in gluc-rich ECS/ICS. We obtained a family of currents (middle traces in Figure 2C) by a voltage step to a membrane potential of +80 mV from various starting Vh values, which varied from −90 to +20 mV (upper traces in Figure 2C). In each recording, the Vh was maintained for 30 s before applying the depolarizing step. As expected, the outward current amplitude decreased as the holding potential became more positive. Current amplitudes for a given holding potential (I) were normalized to the amplitude at the holding potential of −90 mV (I_max). The normalized values (I/I_max) were subsequently plotted against the selected Vh values (Figure 2C). The voltage dependence of the steady-state inactivation of outward K⁺ currents was obtained by fitting the data using a Boltzmann function: (1)I/Imax=1/{1+exp[(Vh−V0.5)/k]} where V_h is the holding potential, V_0.5 is the membrane potential at which the channels are inactivated by 50%, and k is the slope factor. The best fits of V_0.5 were −38.0 ± 1.0 mV in 5 mM [K⁺]_o (N = 6).

To analyze activation kinetics, we measured the time to peak current amplitudes at membrane potentials from 0 to +80 mV in gluc-rich ECS/ICS with the addition of a range of [K⁺]_o (5–100 mM). The time to peak current amplitude values significantly decreased as the membrane potential became more positive. Each time to peak current amplitude for activation with different [K⁺]_o vs. membrane potentials showed strong dependence on the membrane potential (Figure 2D). There were no significant differences in the values using the different [K⁺]_o (5–100 mM).

We further examined the effect of pharmacological K⁺ channel blockers to identify the α-subunit of the time- and voltage-dependent K⁺ channels (Kv). Outward K⁺ currents (Figure 3) in the gluc-rich ECS/ICS supplemented with 5 mM [K⁺]_o were activated by depolarizing voltage steps (400 ms in duration) ranging from −100 to +80 mV, in 10 mV increments, from a Vh of −70 mV (most upper traces) before, during, and after application of Kv inhibitors. TEA and 4-AP, which are both non-selective Kv blockers, attenuated outward current amplitudes at a membrane potential of +80 mV in a dose dependent manner. Peak current amplitudes at +80 mV were significantly inhibited by 10 mM TEA (75.0 ± 1.8%) and 100 μM 4-AP (69.4 ± 6.1%) (Figure 3B; N = 6). An inhibitor of Kv1.1, Kv1.2, and Kv1.6, αDTX (100 nM), significantly suppressed the peak current amplitudes at a membrane potential at +80 mV to 51.0 ± 2.4% of the peak current amplitude (Figure 3B; N = 6). The pattern of these results suggested that odontoblasts express K⁺ currents via Kv1.1, Kv1.2, and/or Kv1.6. The effects of antagonists were evaluated by fitting the data according to the function: (2)I/Icont=(I/Icont−I/Icont min)/[(1+([antagonist]o/IC50))                +I/Icont min where I/I_cont was the normalized value (of current amplitudes with various concentrations of extracellular antagonist ([antagonist]_o) normalized to those without antagonist); I/I_{cont min} was minimal value of I/I_cont; and IC₅₀ was the half maximal (50%) inhibitory concentration of the antagonists. The IC₅₀ was determined to be 0.49 ± 0.39 mM for TEA (N = 6; I/I_{cont min} was 0.24), 6.15 ± 1.65 μM for 4-AP (N = 6; I/I_{cont min} was 0.30), and 4.77 ± 0.91 nM for αDTX (N = 6; I/I_{cont min} was 0.45). In addition, reversal potentials recorded by 400 ms depolarizing voltage steps in the gluc-rich ECS/ICS with 5 mM [K⁺]_o showed a shift toward depolarizing membrane potentials to −21.8 ± 8.1 mV by 10 mM TEA, −28.2 ± 1.6 mV by 100 μM 4-AP, and – 34 ± 3.6 mV by 100 nM αDTX (N = 6 each).

Using alizarin red (each right image, Figure 4) and von Kossa (each left image, Figure 4) staining, we further examined the effects of Kv activity on mineralization. The continuous application of TEA (2 mM or 4 mM) to mineralization medium containing human odontoblasts over a 21 day period showed a dose-dependent decrease in the mineralization efficiency compared to the cells without TEA (as control).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.