Mechanical Stimulation-Induced Calcium Signaling by Piezo1 Channel Activation in Human Odontoblast Reduces Dentin Mineralization

Odontoblasts play critical roles in dentin formation and sensory transduction following stimuli on the dentin surface. Exogenous stimuli to the dentin surface elicit dentinal sensitivity through the movement of fluids in dentinal tubules, resulting in cellular deformation. Recently, Piezo1 channels have been implicated in mechanosensitive processes, as well as Ca2+ signals in odontoblasts. However, in human odontoblasts, the cellular responses induced by mechanical stimulation, Piezo1 channel expression, and its pharmacological properties remain unclear. In the present study, we examined functional expression of the Piezo1 channel by recording direct mechanical stimulation-induced Ca2+ signaling in dentin matrix protein 1 (DMP-1)-, nestin-, and dentin sialophosphoprotein (DSPP)-immunopositive human odontoblasts. Mechanical stimulation of human odontoblasts transiently increased intracellular free calcium concentration ([Ca2+]i). Application of repeated mechanical stimulation to human odontoblasts resulted in repeated transient [Ca2+]i increases, but did not show any desensitizing effects on [Ca2+]i increases. We also observed a transient [Ca2+]i increase in the neighboring odontoblasts to the stimulated cells during mechanical stimulation, showing a decrease in [Ca2+]i with an increasing distance from the mechanically stimulated cells. Application of Yoda1 transiently increased [Ca2+]i. This increase was inhibited by application of Gd3+ and Dooku1, respectively. Mechanical stimulation-induced [Ca2+]i increase was also inhibited by application of Gd3+ or Dooku1. When Piezo1 channels in human odontoblasts were knocked down by gene silencing with short hairpin RNA (shRNA), mechanical stimulation-induced [Ca2+]i responses were almost completely abolished. Piezo1 channel knockdown attenuated the number of Piezo1-immunopositive cells in the immunofluorescence analysis, while no effects were observed in Piezo2-immunopositive cells. Alizarin red staining distinctly showed that pharmacological activation of Piezo1 channels by Yoda1 significantly suppressed mineralization, and shRNA-mediated knockdown of Piezo1 also significantly enhanced mineralization. These results suggest that mechanical stimulation predominantly activates intracellular Ca2+ signaling via Piezo1 channel opening, rather than Piezo2 channels, and the Ca2+ signal establishes intercellular odontoblast-odontoblast communication. In addition, Piezo1 channel activation participates in the reduction of dentinogenesis. Thus, the intracellular Ca2+ signaling pathway mediated by Piezo1 channels could contribute to cellular function in human odontoblasts in two ways: (1) generating dentinal sensitivity and (2) suppressing physiological/reactional dentinogenesis, following cellular deformation induced by hydrodynamic forces inside dentinal tubules.

Odontoblasts play critical roles in dentin formation and sensory transduction following stimuli on the dentin surface. Exogenous stimuli to the dentin surface elicit dentinal sensitivity through the movement of fluids in dentinal tubules, resulting in cellular deformation. Recently, Piezo1 channels have been implicated in mechanosensitive processes, as well as Ca 2+ signals in odontoblasts. However, in human odontoblasts, the cellular responses induced by mechanical stimulation, Piezo1 channel expression, and its pharmacological properties remain unclear. In the present study, we examined functional expression of the Piezo1 channel by recording direct mechanical stimulation-induced Ca 2+ signaling in dentin matrix protein 1 (DMP-1)-, nestin-, and dentin sialophosphoprotein (DSPP)-immunopositive human odontoblasts. Mechanical stimulation of human odontoblasts transiently increased intracellular free calcium concentration ([Ca 2+ ] i ). Application of repeated mechanical stimulation to human odontoblasts resulted in repeated transient [Ca 2+ ] i increases, but did not show any desensitizing effects on [Ca 2+ ] i increases. We also observed a transient [Ca 2+ ] i increase in the neighboring odontoblasts to the stimulated cells during mechanical stimulation, showing a decrease in [Ca 2+ ] i with an increasing distance from the mechanically stimulated cells. Application of Yoda1 transiently increased [Ca 2+ ] i . This increase was inhibited by application of Gd 3+ and Dooku1, respectively. Mechanical stimulation-induced [Ca 2+ ] i increase was also inhibited by application of Gd 3+ or Dooku1. When Piezo1 channels in human odontoblasts were knocked down by gene silencing with short hairpin RNA (shRNA), mechanical stimulationinduced [Ca 2+ ] i responses were almost completely abolished. Piezo1 channel knockdown attenuated the number of Piezo1-immunopositive cells in the immunofluorescence analysis, while no effects were observed in Piezo2-immunopositive cells. Alizarin red staining distinctly showed that pharmacological activation of Piezo1 channels by Yoda1 significantly suppressed mineralization, and shRNA-mediated knockdown of Piezo1 also significantly enhanced mineralization. These results suggest that mechanical stimulation predominantly activates intracellular Ca 2+ signaling via Piezo1 channel opening, rather
Odontoblasts have been reported to express Piezo1 and Piezo2 channels (Khatibi Shahidi et al., 2015;Miyazaki et al., 2019). In addition, Piezo1 channels in rat odontoblasts play an essential role in mechanosensory processes as mechanosensitive ion channels and mediate the intracellular Ca 2+ signaling pathway to release neurotransmitters, establishing neural communication between odontoblasts and the intradental Aδ neurons (Sato et al., 2018). Odontoblastneuron communication generates action potential on the neurons, resulting in the development of dentinal sensitivity (odontoblast hydrodynamic receptor model; Shibukawa et al., 2015;Sato et al., 2018). However, the detailed functional properties of Piezo1 channel in odontoblasts remain to be clarified. To further investigate the functional expression and pharmacological profiles of Piezo channels and their cellular functions, we examined mechanical stimulation-and ligand-evoked intracellular Ca 2+ signaling by Piezo channel activation in human odontoblasts. We also investigated the effects of Piezo1 channel knockdown by gene silencing with short hairpin RNA (shRNA) on the direct mechanical stimulation-induced intracellular free Ca 2+ concentration ([Ca 2+ ] i ) increases. In addition, we examined the role of Piezo1 channels during dentinogenesis using mineralizing assays.

Cell Culture
Human odontoblast cells were obtained from a healthy third molar and immortalized by transfection with the human telomerase transcriptase gene (Kitagawa et al., 2007;Ichikawa et al., 2012;Kimura et al., 2016;Kojima et al., 2017). The resulting cells showed mRNA expression of dentin sialophosphoprotein (DSPP), type 1 collagen, alkaline phosphatase, and bone sialoprotein, and exhibited nodule formation by Alizarin red staining in the mineralizing medium (Kitagawa et al., 2007). Note that these cells were immunopositive for odontoblast marker proteins (Figure 1). The cells were gifted by Dr. Takashi Muramatsu of Tokyo Dental College, Japan. Human odontoblasts were cultured in basal medium [alpha-minimum essential medium containing 10% FBS, 1% penicillin-streptomycin (Life Technologies, Tokyo, Japan), and amphotericin B (Sigma-Aldrich, St. Louis, MO, United States)] at 37°C in a humidified atmosphere of 5% CO 2 for 48 h. The odontoblast suspension was adjusted to a density of 5 × 10 4 cells/ml.

Direct Mechanical Stimulation of Single Human Odontoblast
Direct mechanical stimulation Shibukawa et al., 2015;Nishiyama et al., 2016) was applied using a firepolished glass micropipette with a tip diameter of 2-3 μm, which was filled with standard ECS. Micropipettes were pulled from capillary tubes using a DMZ Universal Puller (Zeitz Instruments, Martinsried, Germany). The micropipette was operated using a micromanipulator (UMp Micromanipulators; Sensapex, Oulu, Finland). The tip was positioned just above the target human odontoblast membrane, and the micropipette was moved vertically downward by 8.0 μm at a velocity of 2.0 μm/s Shibukawa et al., 2015) to generate a focal mechanical stimulation ( Figure 1E). Stimulation was applied for 22 s, after which the pipette was retracted at the same velocity. The stimulations were applied in series no more than three times to avoid unfavorable cell damage.

Measurement of Ca 2+ -Sensitive Dye Fluorescence
Human odontoblasts were incubated for 60 min at 37°C and 5% CO 2 in a standard solution containing 10 μM fura-2acetoxymethyl ester (Dojindo Laboratories, Kumamoto, Japan) and 0.1% (w/v) pluronic acid F-127 (Life Technologies). They were then washed with a fresh standard solution. A dish containing fura-2-loaded human odontoblasts was mounted on the stage of a microscope (IX73; Olympus, Tokyo, Japan), which was installed with an HCImage system (Hamamatsu Photonic, Shizuoka, Japan), an excitation wavelength selector, and an intensified charge-coupled device camera system. Fura-2 fluorescence emission was recorded at 510 nm in response to alternating excitation wavelengths of 340 nm (F340) and 380 nm (F380). The intracellular free Ca 2+ concentration ([Ca 2+ ] i ) was defined as the fluorescence ratio (R F340/F380 ) at two excitation wavelengths and then described as F/F 0 units; the R F340/F380 value (F) was normalized to the resting value (F 0 ). All the experiments were performed at room temperature (28°C). Standard ECS with or without pharmacological Piezo1 channel activator/inhibitors was applied by superfusion using a rapid gravity-fed perfusion system (ValveLink8.2 Controller; AutoMate Scientific, Berkeley, CA, United States; Figure 1E).

Measurement of Intercellular Distance
Human odontoblasts were imaged using an intensified chargecoupled device camera (Hamamatsu Photonic) and microscope (Olympus). The distance from a mechanically stimulated human odontoblast to each neighboring cell was determined from the images obtained (HCImage) by measuring the shortest distance between each pair of cells.

Piezo1 Channel Knockdown by Gene Silencing With Short Hairpin RNA
To generate the odontoblasts transfected by shRNA, including vectors specific for human Piezo1 (shRNA-Piezo1 transfected cells) or an empty vector control (shRNA-Control transfected cells), the Lenti-X™ 293T cell line (Takara Bio Inc., Shiga, Japan) was used to produce lentiviral vector particles. Packaging plasmids pCAG-HIVgp (RIKEN BioResource Research Center, Tsukuba, Japan, RDB04394) and pCMV-VSV-G-RSV-Rev (RIKEN BioResource Research Center, RDB04393) were used for lentiviral vector packaging. Lentiviral vectors specific for human Piezo1 (Sigma-Aldrich) and a TRC2 empty vector control (Sigma-Aldrich) were used in this experiment. Viral supernatants were harvested after 72 h. Human odontoblasts with 60-80% confluency were transfected with lentiviral vector particles including shRNA and then incubated under normal growth conditions for 72 h.

Immunostaining
Human odontoblasts were transferred to eight-well glass chambers (AGC Techno Glass Co., Ltd., Shizuoka, Japan) and maintained under culture conditions as described above. Cells were fixed with 4% paraformaldehyde (FUJIFILM Wako Pure Chemical Co., Osaka, Japan) and washed with 1× PBS (Thermo Fisher Scientific, Tokyo, Japan). After 10 min of incubation with blocking buffer (Nacalai Tesque, Kyoto, Japan) at room temperature, the following primary antibodies were applied for 6 h: mouse monoclonal anti-dentin matrix protein 1 (DMP-1; Immunostaining was observed and analyzed, and photographs were taken (as 8-bit TIFF images) using a fluorescence microscope (BZ-X710; Keyence, Osaka, Japan). For the negative control, the cells were incubated with non-immune antibodies diluted to equivalent concentrations to those of the primary antibodies (data not shown). For shRNA-Piezo1 transfected and shRNA-Control transfected cells, TIFF images obtained with 8-bit of 1,920 × 1,440 pixels were converted to reversed grayscale. Then, the percentage area of the Piezo1-or Piezo2-immunopositive cells was measured using automated thresholds of ImageJ software (NIH, Bethesda, MD, United States).

Mineralization Assay
Human odontoblasts were grown to full confluency in basal medium, and then transferred to mineralization medium (10 mM β-glycerophosphate and 100 μg/ml ascorbic acid in basal medium) for growth at 37°C in 5% CO 2 . To determine the effects of Piezo1 activity on mineralization, the cells were cultured in mineralization medium without (as control) or with the pharmacological Piezo1 modifier, Yoda1 (2 μM), Dooku1 (10 μM), or Yoda1 (2 μM) and Dooku1 (10 μM) for 28 days. The odontoblasts transfected with shRNA, including vectors specific for human Piezo1, or including an empty vector control (see above), were also grown to full confluency in basal medium, and then transferred to mineralization medium at 37°C in 5% CO 2 for 28 days. During the 28-day culture period, mineralization media with or without pharmacological agents were changed twice a week . To detect calcium deposition, the cells were subjected to Alizarin red staining, and the mineralizing efficiencies were measured using ImageJ software (NIH). Images were obtained with a digital camera (Sony, Tokyo, Japan), converted to 32-bit, and converted into reversed grayscale. Regions of interest (ROIs) were then determined for each whole well to measure the mean luminance intensities of the total pixel numbers (I) of the ROI. Mineralizing efficiencies were normalized and represented as I/I 0 units, and the intensities (I) of Alizarin red staining were normalized to the mean intensity values of areas without cells (I 0 ).

Statistical Analysis
Data are expressed as the mean ± SE or SD of the mean of N observations, where N represents the number of experiments or cells tested. The Kruskal-Wallis test, Friedman test, and Dunn's post-hoc test were used to determine non-parametric statistical significance. Parametric statistical significance was determined using a two-tailed Student's t-test to analyze the percentage areas of Piezo1/2 channel-immunopositive cells in the immunofluorescence analyses. Statistical significance was set at p < 0.05. Statistical analysis was performed using GraphPad Prism 5.0 (GraphPad Software, La Jolla, CA, United States).

Direct Mechanical Stimulation to Human Odontoblasts Transiently Increased [Ca 2+ ] i
We observed distinct DMP-1, nestin, and DSPP immunoreactivity in human odontoblasts (Figures 1A-C). In addition, odontoblasts also expressed Piezo1 immunoreactivity ( Figure 1D). Cells expressing odontoblast marker proteins were loaded with fura-2, and [Ca 2+ ] i was measured. When we applied mechanical stimulations to single human odontoblasts using a glass micropipette moving vertically (8 μm) in a downward direction from a position just above the surface (0 μm; Figure 1E), in the presence of extracellular Ca 2+ , [Ca 2+ ] i was increased to a peak value of 4.20 ± 1.08 F/F 0 units for first stimulation (N = 6; Figure 1F).  Figure 2A) to a peak value of 4.21 ± 1.46 F/F 0 units (N = 38; black circle, Figure 2B). In addition, we observed transient increases in [Ca 2+ ] i in the neighboring human odontoblasts (red lines, Figure 2A) during mechanical stimulation. Note that these cells were not in physical contact with each other. The plots in Figure 2B show F/F 0 values as a function of the distance from the stimulated odontoblast (black circle) to neighboring odontoblasts (red circles). The peak values of [Ca 2+ ] i were 2.52 ± 1.45 (N = 9), 1.59 ± 0.55 (N = 15), 1.34 ± 0.32 (N = 8), 1.07 ± 0.03 (N = 4), 1.11 ± 0.09 (N = 6), 1.06 ± 0.03 (N = 6), and 1.02 ± 0.01 (N = 5) F/F 0 units in human odontoblasts located within distances of 0-10 μm, 11-20 μm, 21-30 μm, 31-40 μm, 41-50 μm, 51-60 μm, and 61-70 μm from mechanically stimulated human odontoblasts, respectively (red circles, Figure 2B). The amplitude of the [Ca 2+ ] i increases in neighboring human odontoblasts was reduced by increasing their distance from mechanically stimulated human odontoblasts. The kinetics of the distance-dependent attenuation of [Ca 2+ ] i were described well by a single exponential function (dotted line, Figure 2B). The spatial constants of the distance-dependent decrease in [Ca 2+ ] i were 6.11 ± 0.76 μm for neighboring odontoblasts. In addition, we observed a delay in the increase in [Ca 2+ ] i in the nearby odontoblasts during the mechanical stimulation of single odontoblasts (time when the mechanical stimulation applied to the stimulated odontoblasts was set to 0 s; vertical dotted line, Figure 2A). The time delay increased with an increase in the distance from the mechanically stimulated odontoblasts ( Figure 2C). The significant time delay in the increase in [Ca 2+ ] i in the neighboring odontoblasts located 51-60 μm and 61-70 μm away from mechanically stimulated odontoblasts was observed compared to that in the cells located at 0-10 μm ( Figure 2C). The time delay was ranged from 0.95 ± 0.47 s at 0-10 μm away to 3.03 ± 1.04 s at 61-70 μm away from mechanically stimulated cells.

Piezo1 Channel Knockdown by Short Hairpin RNA Reduced the Functional Expression of Piezo1 but Not Piezo2 Channels
We further investigated mechanical stimulation-induced [Ca 2+ ] i increases in human odontoblasts transfected with shRNA with a vector specific for human Piezo1, and an empty vector control.
In the immunofluorescence analysis, we observed immunoreactivity not only of the Piezo1 channels (green, Figures 5A,E), but also of the Piezo2 channels (red, Figures 5B,E) in the cells transfected with an empty vector control. In the cells transfected by shRNA with a vector specific for human Piezo1 (shRNA-Piezo1 transfected cells), we observed a significant

Activation of Piezo1 Channel by the Pharmacological Modifiers Reduced Mineralization
We investigated the effects of Piezo1 channel activity on mineralization induced by human odontoblasts. Alizarin red staining ( Figure 6A) was determined to be indicative of the mineralization levels based on the staining intensity (see Materials and Methods) represented as I/I 0 units; the intensities (I) of the stains were normalized to the mean intensities of adjacent areas without cells (I 0 ; Figure 6B). The application of 2 μM Yoda1, or 2 μM Yoda1 and 10 μM Dooku1, significantly reduced their mineralization levels compared to the level without these pharmacological Piezo1 channel modifiers. In addition, the level in the presence of 2 μM Yoda1 was significantly lower than that in the presence of 10 μM Dooku1.

Piezo1 Channel Knockdown by shRNA Enhanced Mineralization
We also examined the contribution of Piezo1 channel activity to mineralization efficiency. The mineralization level represented by Alizarin red staining (Figure 7A) in the shRNA-Piezo1 transfected cells was significantly higher than that in the shRNA-Control transfected cells (Figure 7B).

DISCUSSION
The results showed that human odontoblasts, which showed DMP-1-, nestin-, and DSPP-immunoreactivity, functionally express Piezo1 channels activated by Yoda1 pharmacologically and by direct mechanical stimulation. Both Yoda1 inducedand direct mechanical stimulation-induced-[Ca 2+ ] i increases were sensitive to extracellular Gd 3+ and Dooku1, which are pharmacological non-selective and selective Piezo1 channel inhibitors, respectively. In human odontoblasts, mechanical stimulation-induced-[Ca 2+ ] i increases also mediated intercellular signal networks among odontoblasts, as previously reported in rat studies Shibukawa et al., 2015). These mechanical stimulation-induced [Ca 2+ ] i responses were almost completely abolished in cells when Piezo1 channels were knocked down by shRNA. In addition, Alizarin red staining indicated that pharmacological Piezo1 channel activation by Yoda1 significantly suppressed mineralization, and the shRNA-mediated knockdown of Piezo1 significantly enhanced mineralization.
We previously found that odontoblasts can detect mechanical stimuli as cellular deformation originating from dentinal fluid movement by dentin stimuli and act as sensory receptor cells to generate dentinal sensitivity. The mechanism underlying the generation of dentinal sensitivity has been termed the odontoblast hydrodynamic receptor model (Shibukawa et al., 2015). The stimuli to the dentin surface transformed into dentinal fluid movements and induced cell membrane deformation in odontoblasts. Cell deformation activates mechanosensitive transient receptor potential (TRP) channels, such as TRP vanilloid subfamily member 1 (TRPV1), TRPV2, TRPV4, and TRP ankyrin 1 (TRPA1) channels, increasing [Ca 2+ ] i through Ca 2+ influx Sato et al., 2013Sato et al., , 2018Shibukawa et al., 2015). The [Ca 2+ ] i increase established intercellular odontoblast-odontoblast and odontoblast-neuron signal communication mediated by extracellular signaling molecules, ATP, and glutamate, which are released from mechanically stimulated odontoblasts Shibukawa et al., 2015;Nishiyama et al., 2016). The released ATP (from pannexin-1 channels in odontoblasts triggered by the [Ca 2+ ] i increase) activates ionotropic ATP receptor subtype 3 (P2X 3 receptors) in the intradental Aδ neurons. The activation of P2X 3 receptors induces an action potential in pulpal Aδ neurons, indicative of the generation of dentinal sensitivity (Sato et al., 2018). Additionally, in line with the present results, the application of selective Piezo1 channel inhibitors, GsMTx4, also inhibited synaptic-like responses in TG neurons evoked by the direct mechanical stimulation of odontoblasts. Together with the present and previous results, Piezo1 channels in odontoblasts are involved in detecting cell membrane deformation due to dentinal fluid movements by the stimuli applied to the dentin surface to generate dentinal sensitivity, and in mediating interodontoblast signal networks. Indeed, the inhibition of Piezo1 channels by the systemic administration of GsMTx4 and the somatic elimination of odontoblasts in the dentin-pulp border (using Cre/LoxP-based technology; Zhao et al., 2021) almost completely abolished nociceptive behaviors by cold stimuli to the exposed dentin surface in rats/transgenic mice showing dentinal sensitivity (personal communication from YS).
In line with our results, it has been described that Piezo1 channels are expressed not only in multipotent stem cells from  Continued) human exfoliated deciduous teeth (SHED), but also in odontoblast processes (Miyazaki et al., 2019). In addition to the mechanosensory transduction roles of odontoblasts, Piezo1 channels also mediate hydrostatic pressure-induced differentiation of odontoblasts from SHED (Miyazaki et al., 2019). In the Alizarin red staining of the present study, however, the pharmacological activation of the Piezo1 channel significantly suppressed mineralization efficiency. In addition, the intensity of Alizarin red staining of shRNA-Piezo1 transfected cells was significantly higher than that in shRNA-Control transfected cells. These results indicate that Piezo1 channels in odontoblasts may negatively regulate not only reactionary dentinogenesis induced by cell membrane deformation due to dentinal fluid movements by the stimuli applied to the exposed dentin surface, but also physiological/ developmental dentinogenesis. In contrast, the activation of mechanosensitive-and high-pH sensitive-TRPA1 channels in odontoblasts positively modulates mineralization in physiological, developmental, and pathological conditions [e.g., Ca(OH) 2 or mineral trioxide aggregate (MTA) application on dentin; Kimura et al., 2016]. Although further studies are needed, these convergent results suggest that, among odontoblast-expressing mechanosensitive cation channels, including TRPV1, TRPV2, TRPV4, TRPA1, and Piezo1 channels, their functional roles in sensory transduction and dentin formation are different. (A) Human odontoblasts were cultured for 28 days in mineralization medium with 2 μM Yoda1 (second left), 2 μM Yoda1 and 10 μM Dooku1 (third left), 10 μM Dooku1 (right most), or without these channel modifiers (left most) at pH 7.4, and stained by Alizarin red (red, calcium deposition). (B) Mineralization levels without (control; upper most column) or with 2 μM Yoda1 (second column), 2 μM Yoda1 and 10 μM Dooku1 (third column), and 10 μM Dooku1 (lower column), assessed by Alizarin red staining. The mineralization levels were 4.80 ± 0.13 I/I 0 in the absence of Piezo1 channel modifiers (as controls), while those 1.96 ± 0.04 I/I 0 with 2 μM Yoda1, 2.63 ± 0.12 I/I 0 with 2 μM Yoda1 and 10 μM Dooku1, and 3.11 ± 0.18 I/I 0 with 10 μM Dooku1. Each column denotes the mean ± SE from eight experiments. Statistically significant differences between columns (shown by solid lines) are indicated by asterisks. *p < 0.05.

A B
FIGURE 7 | Piezo1 channel knockdown by shRNA enhances the mineralization levels. (A) Human odontoblasts transfected by shRNA including a vector specific for human Piezo1 (right; shRNA-Piezo1 transfected) or including an empty vector control (left; shRNA-Control transfected) were cultured for 28 days in mineralization medium, and subjected to the Alizarin red staining (red indicative of calcium deposition). (B) The mineralization levels in the cells of shRNA-Control transfected (upper column) and shRNA-Piezo1 transfected (lower column), assessed by Alizarin red staining. The mineralization levels were 5.95 ± 0.36 I/I 0 in the shRNA-Control transfected cells, and 8.30 ± 0.59 I/I 0 in the shRNA-Piezo1 transfected cells. Each column indicates the mean ± SE from three experiments. Statistically significant differences between columns (solid line) are indicated by asterisks. *p < 0.05.
Recent evidence have indicated that mouse odontoblasts also express Piezo2 channels (Khatibi Shahidi et al., 2015;Krivanek et al., 2020). In line with these results, we successfully observed Piezo2-immunoreactivity in both shRNA-Piezo1 transfected and shRNA-Control transfected cells. Thus, we could not exclude the possibility that Piezo2 channels are also involved in mechanosensory transduction in odontoblasts. However, further study will be needed to examine how Piezo2 channels contribute to the sensory transduction sequences, and what differences exist between the roles of Piezo1 and Piezo2 channels in cellular function. In addition, in endothelial cells, it has been demonstrated that Piezo1 channels regulate TRPV4 channel openings. In the present study, the pharmacological Piezo1 inhibitors Gd 3+ (1 μM) and Dooku1 (10 μM) almost completely suppressed direct mechanical stimulation-induced [Ca 2+ ] i increases in single human odontoblasts. We also only observed small residual responses in mechanical stimulation-induced [Ca 2+ ] i increases in shRNA-Piezo1 transfected cells compared to those in the shRNA-Control transfected cells. Therefore, together with the present results, Piezo1 channels may act upstream of mechanosensitive cation channels, including TRPV1, TRPV2, TRPV4, and TRPA1 channels, as well as Piezo2 channels, in odontoblasts.
In conclusion, we demonstrated that human odontoblasts functionally express Piezo1 channels, and mechanical stimulationinduced-[Ca 2+ ] i increases mediate the intercellular signaling networks among them. Piezo1 channels predominantly contribute to the detection of cellular deformation, such as mechanical stimulation, induced by dentinal fluid movements in dentinal tubules, during various stimuli applied to the exposed human dentin surface, such as enamel lesions including dental caries and/or acid erosion. However, Piezo1 channel activity negatively regulates not only reactionary dentinogenesis, but also physiological/developmental dentinogenesis. This suggests that Piezo1 channel activity may physiologically prevent the lifelong loss of volume in the pulp chamber by excessive dentin formation induced by mechanical stimuli applied on the teeth, such as mastication.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/supplementary material; further inquiries can be directed to the corresponding author.