Cell Deformation at the Air-Liquid Interface Evokes Intracellular Ca2+ Increase and ATP Release in Cultured Rat Urothelial Cells

Urothelial cells have been implicated in bladder mechanosensory transduction, and thus, initiation of the micturition reflex. Cell deformation caused by tension forces at an air-liquid interface (ALI) can induce an increase in intracellular Ca2+ concentration ([Ca2+]i) and ATP release in some epithelial cells. In this study, we aimed to examine the cellular mechanisms underlying ALI-induced [Ca2+]i increase in cultured urothelial cells. The ALI was created by stopping the influx of the perfusion but maintaining efflux. The [Ca2+]i increase was measured using the Ca2+ imaging method. The ALI evoked a reversible [Ca2+]i increase and ATP release in urothelial cells, which was almost abolished by GdCl3. The specific antagonist of the transient receptor potential vanilloid (TRPV4) channel (HC0674) and the antagonist of the pannexin 1 channel (10panx) both diminished the [Ca2+]i increase. The blocker of Ca2+-ATPase pumps on the endoplasmic reticulum (thapsigargin), the IP3 receptor antagonist (Xest-C), and the ryanodine receptor antagonist (ryanodine) all attenuated the [Ca2+]i increase. Degrading extracellular ATP with apyrase or blocking ATP receptors (P2X or P2Y) with pyridoxalphosphate-6-azophenyl-2',4'-disulfonic acid (PPADS) significantly attenuated the [Ca2+]i increase. Our results suggest that both Ca2+ influx via TRPV4 or pannexin 1 and Ca2+ release from intracellular Ca2+ stores via IP3 or ryanodine receptors contribute to the mechanical responses of urothelial cells. The release of ATP further enhances the [Ca2+]i increase by activating P2X and P2Y receptors via autocrine or paracrine mechanisms.


INTRODUCTION
The urothelium of the bladder has been previously considered a passive barrier. However, in recent years, a large number of reports have shown that the urothelium plays an important role in bladder mechanosensory transduction (de Groat, 2004;Birder, 2005). In response to mechanical stimuli, two critical concomitant intracellular events occur in urothelial cells: an increase in intracellular Ca 2+ concentration ([Ca 2+ ] i ) and the release of signaling molecules such as NO, ATP, and acetylcholine (ACh). The release of ATP from urothelial cells directly activates purinergic receptors on bladder afferent terminals to increase their excitability, thereby enhancing sensory function of the bladder (Yu and . Although several studies and a few reviews have been conducted on mechanosensing in urothelial cells (Merrill et al., 2016;Janssen et al., 2017), the exact cellular mechanisms by which urothelial cells sense mechanical stimuli are not fully understood.
In vivo, changes in intravesical pressure or bladder wall tension are physiologically relevant mechanical stimuli for the urothelium. To simulate these stimuli in vitro, common strategies include stretching of urothelial cells cultured on coverslips using various customized devices (Miyamoto et al., 2014;Guan et al., 2018) or exposing urothelial cells to high hydrostatic pressure (Olsen et al., 2011). For example, custom designed chambers, attached to an extension device, were used to stretch urothelial cells in one study (Miyamoto et al., 2014). In another study, urothelial cells cultured on plastic dishes were exposed to sustained hydrostatic pressure (5-20 cm H 2 O; Olsen et al., 2011).
The above approaches of delivering mechanical stimuli are valuable in revealing the underlying mechanisms of urothelial mechanotransduction. However, they require technically complicated setups. In other reports in the literature, exposure of urothelial cells to low-osmolality solutions to induce cell swelling was also a frequent strategy to alter cell membrane stress (Haeberle et al., 2008;Wu et al., 2011). Delivering hypoosmotic shock to cells is simple. However, this modality might not present a physiologically relevant stimulus. Although low-osmolarity urine is occasionally present in the bladder in vivo, it may not lead to the swelling of urothelial cells. This is because of the presence of specialized lipid molecules and uroplakin proteins in the apical membrane of umbrella cells, which reduce the permeability of the cells to water.
Several studies have reported that cells can be mechanically stimulated by creating an air-liquid interface (ALI). The ALI is produced by reducing the volume of liquid covering cells, and a thin film of liquid on the cells is drawn down to such proximity that tension forces at the ALI cause cell deformation (Patrick et al., 2001;Ramsingh et al., 2011). In the study by Ramsingh et al. (2011), lung epithelial cells were exposed to ALI forces by briefly introducing an air bubble into a closed perfusion chamber or by tilting the culture dish containing the cells. Such stimuli produced reversible cell deformation, a transient increase in [Ca 2+ ] i , and ATP release. In another study, a microfluidic device was developed to create an ALI by liquid evaporation (Huang et al., 2016), to investigate ALI-induced intracellular Ca 2+ response in a human leukemic cell line .
In the present study, we developed another method to create an ALI in a perfusion cell chamber, by simply stopping influx, but maintaining efflux of the perfusion, until a thin layer of liquid was created on the urothelial cells. We found that the ALI can induce a remarkable increase in [Ca 2+ ] i and ATP release from urothelial cells cultured on glass coverslips, and the responses were repeatable and did not affect cell viability. We aimed to investigate the underlying cellular mechanisms of mechanical stimulus-induced increase in [Ca 2+ ] i with this easily-operated model. Our main focus was on the potential involving channels or receptors located either on the cell membrane or those associated with intracellular Ca 2+ stores. We found that both Ca 2+ influx via membrane-bound TRPV4 or pannexin 1 channels, and Ca 2+ release from intracellular Ca 2+ stores via IP3 and ryanodine receptors contribute to an ALI-induced [Ca 2+ ] i increase. Furthermore, ATP released from urothelial cells also contribute to a [Ca 2+ ] i increase via autocrine or paracrine mechanisms.

Experimental Animals
Female Sprague Dawley rats (virgins, 2-3 months old, 180-250 g; obtained from Pengyue Animal Co., Jinan, China) were used in this study. Rats were allowed free movement, and had free access to food and water. Care and handling of the animals were in accordance with the Shandong University Animal Care and Use Committee. The study was approved by the Ethics Committee of the Second Hospital, Cheeloo College of Medicine, Shandong University [KYLL-2016(GJ)A-0027].

Urothelial Cell Culture
Urothelial cell cultures were performed as previously described (Kullmann et al., 2009). Rats were subjected to isoflurane anesthesia, and bladders were resected and placed in cold minimal essential medium (MEM, Invitrogen, Carlsbad, CA, United States) supplemented with HEPES (2.5 g/L, Sigma, St. Louis, MO, United States) and penicillin/streptomycin/fungizone (PSF, 1%, Sigma). The bladder urothelium was incubated in Dispase (2.5 mg/ml, Worthington Biochemical Lakewood, NJ, United States) overnight at 4°C. Urothelial cells were gently scraped, placed in trypsin (0.25% wt/vol, Sigma) for 10-15 min at 37°C, and dissociated by trituration. Cells were suspended in MEM containing 10% FBS and centrifuged at 416 g for 5 min. The cells were suspended in urothelial cell medium (UCM, ScienCell, San Diego, CA, United States) with 1% PSF, centrifuged again, and resuspended in fresh media. Cells were plated on poly-L-lysine-coated glass coverslips and used for Ca 2+ imaging, 48-96 h after dissociation. Creation of the Air-Liquid Interface The approach to produce an ALI on urothelial cells is illustrated in Figure 1. The cell chamber was continuously perfused with Hank's balanced salt solution (HBSS) containing (in mM): 138 NaCl; 5 KCl; 0.3 KH 2 PO 4 ; 4 NaHCO 3 ; 2 CaCl 2 ; 1 MgCl 2 ; 10 HEPES; and 5.6 glucose. The pH was adjusted with Tris base to 7.4, and osmolality was adjusted with sucrose to 320-325 mOsm. The influx of perfusion was driven by gravity, and the efflux was driven by a vacuum pump. Monolayer urothelial cells were immersed in HBSS when the influx and efflux flow rates were equally maintained, and no ALI was formed on the cells ( Figure 1A). When the influx was stopped but the efflux remained (vacuum was kept running), the amount of HBSS solution in the chamber was reduced to a level, at which a thin liquid layer was formed covering the cells. Cells were mechanically deformed by the tension produced at the ALI, and fluorescence signals at 340 and 380 nm were altered (inset of Figure 1B). The above process took about 10-15 s. The influx of perfusion was resumed immediately after noticing a change in the fluorescence signal. A simplified situation of two hemispherical cells covered by a thin liquid film, where T is the surface tension of the liquid and ΔT is a net component of T, acting perpendicularly to the cell surface to cause cell deformation ( Figure 1C).

Ca 2+ Imaging
Urothelial cells were loaded with Fura 2-AM (2 μM, Dojindo, Japan) for 30-40 min at 37°C in an atmosphere of 95% O 2 and 5% CO 2 . Fura 2-AM was dissolved in HBSS. Ca 2+ imaging was performed as described previously (Zhang et al., 2011). Briefly, coverslips were placed under a fluorescence microscope (Nikon Eclipse Ti) and continuously perfused with HBSS (1.5-2 ml/min). Fura 2-AM was alternately excited with ultraviolet light at 340 and 380 nm. The fluorescence emission was detected at 510 nm using a computer-controlled monochromator. Wavelength selection, timing of excitation, and the acquisition of images were controlled using the MetaFluor software (Molecular Devices, Sunnyvale, CA, United States). Digital images were stored for offline analysis. The ratio of fluorescence signals measured at 340 nm divided by the fluorescence signals measured at 380 nm was used to determine the increase in intracellular Ca 2+ .

ATP Assay
The ATP concentration was measured by using luciferin-luciferase bioluminescence. Samples (100 μl) of perfusate were collected 2 min before and immediately after ALI stimulation. A mixture of 100 μl luciferin-luciferase was added to a 100 μl sample, according to the manufacturer's instructions using the Promega CellTiter-Glo Luminescent Cell Viability Assay Kit (Promega). ATP detection was evaluated using the GloMax 20/20 Luminometer (Promega). Sample bioluminescence was compared to that of standard amounts of ATP used within the same concentration range. All samples were run in duplicate.

Immunofluorescence Staining
Immunofluorescence staining was performed to identify the expression of TRPV4 and pannexin 1 in urothelial cells. Hank's balanced salt solution (HBSS), for which influx was driven by gravity and efflux was driven by a vacuum pump. Monolayer urothelial cells were immersed in HBSS, and no ALI was formed when influx and efflux were balanced. Fluorescence signals at 340 or 380 nm were monitored simultaneously, and signals showed no change (inset). (B) To produce the ALI, influx was stopped but efflux was continued. Once the amount of HBSS solution in the chamber was reduced to a specific level at which a thin liquid film was formed on the cells, they were mechanically deformed by the tension force at the ALI, which altered the fluorescence signals (inset). The above process usually took about 10-15 s. Perfusion was resumed immediately after noticing a change in fluorescence signals. (C) Simplified situation of two hemispherical cells covered by a thin liquid film, where T is the surface tension of the liquid and ΔT is the net component of T, acting perpendicularly to the cell surface to cause cell deformation.

Statistical Analysis
All data were expressed as the mean ± SEM. Statistical significance was evaluated using the Student's t-test or paired t-test, or one-way or two-way ANOVA with Bonferroni post hoc test for multiple comparisons, as appropriate. When the test for normality failed, the Mann-Whitney rank sum test was applied. Data were considered statistically significant when p < 0.05.

ALI Evokes an Intracellular Ca 2+ Increase and ATP Release in Urothelial Cells
Several forms of mechanical stimuli, such as stretching (Miyamoto et al., 2014;Guan et al., 2018) and osmotic swelling (Haeberle et al., 2008;Wu et al., 2011) can trigger Ca 2+ influx in urothelial cells. In the current study, ALI stimulation also evoked a remarkable increase in [Ca 2+ ] i in the cultured urothelial cells (Figure 2A). Moreover, this response was reversible (Figure 2A), and the second ALI-evoked increase in [Ca 2+ ] i showed no significant differences from the first stimulation (amplitude, 0.579 ± 0.017 for the first transient, vs. 0.546 ± 0.016 for the second transient, n = 420 cells, p > 0.05, Figure 2C). For the same coverslip, the application of ATP (100 μM, Sigma) at the end of the experiment evoked a significant increase in Ca 2+ (Figure 2A), indicating that cells retained good viability after formation of the ALI. A significant increase in the ATP concentration of perfusates from five coverslips was observed after ALI stimulation (p < 0.001, Figure 2E). A model of two successive ALI stimulations with a 10-15 min interval (Figure 2A) was used in the following experiments to test the underlying mechanisms of the [Ca 2+ ] i increase.

Mechanosensitive Channels Involved in the ALI-Evoked [Ca 2+ ] i Increase
The MSCs have been shown to be involved in tension-induced Ca 2+ influx (Olsen et al., 2011). To determine the potential involvement of MSCs in the ALI-evoked [Ca 2+ ] i increase, GdCl 3 (Sigma), an unspecific MSC blocker, was applied between the two ALI. The GdCl 3 (50 μM) treatment almost abolished the second ALI-evoked [Ca 2+ ] i increase (amplitude reduced by 94.30%, p < 0.001, Figures 2B,D). The inhibition produced by 50 μM GdCl 3 was long lasting, and required 20-30 min washout for recovery of the ALI-induced response ( Figure 2B). The ALI-evoked ATP release was also significantly reduced in the presence of GdCl 3 (p < 0.001, Figure 2E).

TRPV4 Activation Is Involved in the [Ca 2+ ] i Increase
The epithelial Na + /degenerin (ENaC/DEG) channel (Du et al., 2007); Piezo 1 channel (Miyamoto et al., 2014); Piezo 2 channel (Marshall et al., 2020); connexin channel (Shimura et al., 2018), and TRPV4 channel (Mochizuki et al., 2009) are all MSCs that are reportedly present on urothelial cells. We examined the contribution of these MSCs to the ALI-induced [Ca 2+ ] i increase. The application of amiloride (100 μM, Sigma), the ENaC channel antagonist, or GsMTX4 (GsX4, 10 μM, MCE, China), one commonly used Piezo channel antagonist (Lee et al., 2014;Alcaino et al., 2017;Velasco-Estevez et al., 2020), or carbenoxolone (100 μM), a connexin channel blocker between the two successive inductions of the ALI and during the second ALI showed no significant inhibition (Figures 3Aa-Ac). No significant differences were observed in the amplitude between the first and the second ALI-induced [Ca 2+ ] i increases (p > 0.05, Figures 3Ad-Af), suggesting no involvement of the ENaC, Piezo, or connexin channels. However, application of ruthenium red (RR, 10 μM, Sigma), a nonspecific inhibitor of many cation channels, including the TRP channels, significantly reduced the ALI-induced [Ca 2+ ] i increase (Figure 3Ba), and the amplitude of the second ALI-induced [Ca 2+ ] i increase was reduced by 35.7% (n = 183, p < 0.001, Figure 3Bc). The TRPV4 channels are highly expressed in the urothelium (Mochizuki et al., 2009). To evaluate the contribution of TRPV4 to the ALI-induced [Ca 2+ ] i increase, HC0674 (10 μM, Abcam), the specific TRPV4 antagonist, was applied between the two inductions of the ALI and during the second ALI. The amplitude of the second ALI-induced [Ca 2+ ] i increase was reduced by 38.8% (n = 184, p < 0.001, Figures 3Bb,Bd), compared with that of the first ALI-induced [Ca 2+ ] i increase.

Source of the ALI-Evoked [Ca 2+ ] i Increase
To examine the contribution of Ca 2+ influx from the extracellular space, cells were perfused with Ca 2+ -free HBSS, which was modified from normal HBSS through the omission of CaCl 2 , the addition of 0.1 mM EGTA, and an increase in the concentration of MgCl 2 to 3.1 mM. The osmolality of Ca 2+ -free HBSS was adjusted with sucrose to 320-325 mOsm. In the presence of Ca 2+ -free HBSS, the ALI-induced [Ca 2+ ] i increase was significantly reduced by 22%, compared with that in the presence of normal HBSS (0.936 ± 0.018 vs. 0.731 ± 0.016, n = 496, p < 0.01 Figures 4Aa,Ac). To determine the involvement of Ca 2+ release from the endoplasmic reticulum (ER) calcium  Figures 4Ab,Ad). Notably, thapsigargin alone evoked a remarkable [Ca 2+ ] i increase (Figure 4Ab).

Activation of IP3 and Ryanodine Receptors on the ER Contribute to the [Ca 2+ ] i Increase
The ryanodine receptor (RyR) and inositol 1,4,5-trisphosphate receptor (IP3R; Galione and Churchill, 2002) are the two major Ca 2+ -release receptor channels of the ER. To determine whether these two receptors contribute to the ALI-induced [Ca 2+ ] i increase, xestospongin-C (Xest-C; Mijares et al., 2020;Sigma), the IP3R specific antagonist, or ryanodine (MCE, China), the RyR antagonist, were applied between the two ALI stimuli and during the second ALI stimulus. Both Xest-C (5 μM) and ryanodine (100 μM) significantly reduced the second ALI-induced [Ca 2+ ] i increase (Figures 4Ba,Bb), and the amplitude was decreased by 71.6 and 67.7%, respectively (Figures 4Bd,Be).
To further examine the contribution of intracellular Ca 2+ release, BAPTA-AM (2 μM), a membrane-permeable form of BAPTA was applied between the two ALI stimuli and during the second stimulus to clamp intracellular Ca 2+ levels (Birder et al., 2003). Second ALI induced [Ca 2+ ] i was significantly reduced (Figures 4Bc,Bf).

Activation of the Pannexin 1 Channel Is Involved in the ALI-Induced [Ca 2+ ] i Increase and ATP Release
Pannexin 1 channels play essential roles in urothelial mechanotransduction (Negoro et al., 2014). Besides its mechanical sensitivity, a nonvesicular mechanism of ATP release, mediated by pannexin channels in urothelial cells, has been identified in recent years (Penuela et al., 2013). To determine whether pannexin 1 channels are involved in the ALI-induced [Ca 2+ ] i increase and release of ATP, 10 panx (Burma et al., 2017), the pannexin 1 channel peptide blocker, was applied between the

Role of Autocrine ATP in the ALI-Induced [Ca 2+ ] i Increase
In urothelial cells, ATP can induce a [Ca 2+ ] i increase by activating P2X or P2Y receptors (Locovei et al., 2007;von Kugelgen and Hoffmann, 2016). The autocrine ATP signaling in urothelial cells contributes to a major part of the mechanical-mediated [Ca 2+ ] i increase (Guan et al., 2018). To investigate whether the ALI-induced ATP release could lead to a [Ca 2+ ] i increase in an autocrine or paracrine manner, we first applied the ATP-diphosphohydrolase (apyrase, from Sigma) to degrade extracellular ATP by catalyzing its breakdown to ADP. As shown in Figure 5B, the application of apyrase ( Figures 5Bc,Bf), and these effects were reversible.

TRPV4 And Pannexin-1 Expression in the Urothelium
Immunofluorescence staining for pannexin 1 and TRPV4 channels revealed strong staining in cells adjacent to the lumen of the bladder (Figures 6B,F). Colocalization studies with AE1/ AE3, a molecular marker of urothelial cells, confirmed that pannexin 1 and TRPV4 are both expressed in the urothelium (Figures 6C,G). Negative control experiments with normal rabbit IgG showed no staining (Figures 6D,H).

DISCUSSION
In the current study, the cellular mechanisms underlying ALI-induced increase in [Ca 2+ ] i of cultured urothelial cells were Frontiers in Physiology | www.frontiersin.org investigated using various pharmacological blockers. Our main findings were as follows: (1)  In vitro, commonly used approaches to deliver mechanical stimulation to urothelial cells include: stretching the cells with various custom designed devices (Miyamoto et al., 2014;Guan et al., 2018); exposing the cells to high hydrostatic pressure; or simply subjecting the cells to a hypoosmotic shock (Olsen et al., 2011). Using these approaches, valuable information has been obtained regarding urothelial cell mechanotransduction. For example, TRPV4 or Piezo1 channels have been revealed as the MSCs responsible for sensing mechanical stimuli (Mochizuki et al., 2009;Miyamoto et al., 2014). In addition, the activation of pannexin 1 channels has been considered an important mechanism of ATP release (Penuela et al., 2013). However, the above approaches usually require technically complicated setups.
In the current study, another form of mechanical stimulation was established, i.e., creation of an ALI on urothelial cells. The ALI evoked similar cellular responses in urothelial cells as the mechanical modalities mentioned above, i.e., [Ca 2+ ] i increase and ATP release. Moreover, the whole procedure did not cause cell lysis, which was evidenced by the remarkable [Ca 2+ ] i increase in response to ATP after inductions of ALI. Thus, an ALI can induce the [Ca 2+ ] i increase and ATP release, but does not necessarily cause damage to the plasma membrane. Although the chances that the bladder urothelium is exposed to an ALI are not common under physiological conditions, this may occur during some invasive procedures, like bladder surgeries under laparoscopy or during cystoscopy.
An ALI may also be frequently encountered in urothelial cell culture experiments. For example, when the culture medium is changed. In the current study, we did not observe to what extent cell deformation was evoked by the ALI. When the cell is exposed to an ALI, surface tension forces up to 320 Newtons/in 2 could be exerted on the cell membrane (Ludwig et al., 2008), and the pressure generated by surface tension is large enough to cause significant cell deformation. Thus, we established a simple procedure to deliver mechanical stimulation to cultured urothelial cells, which can be used as Tension-induced Ca 2+ entry into cells is reportedly achieved through the activation of MSCs (Olsen et al., 2011;Leybaert and Sanderson, 2012;Jafarnejad et al., 2015). In the current study, the involvement of MSC channels in the ALI-induced [Ca 2+ ] i increase could be supported by the observation of the almost complete blockade of the ALI-induced [Ca 2+ ] i increase by GdCl 3 . As a common nonspecific MSC blocker, Gd 3+ has been shown to block many MSC channels, including several types of TRP channels (White et al., 2016), pannexin 1 channels (Jiang and Penuela, 2016), as well as mechanoinsensitive channels like those associated with purinergic receptors (Khakh et al., 2001).
In previous studies, TRPV4 (Mochizuki et al., 2009), epithelial Na + (ENa; Du et al., 2007), Piezo1 (Miyamoto et al., 2014), Piezo 2 (Marshall et al., 2020), connexin (Shimura et al., 2018), and pannexin 1 channels (Timoteo et al., 2014;McLatchie and Fry, 2015) have been implicated in mechanotransduction of urothelial cells. In agreement with these previous reports, our observations of the significant reduction of the [Ca 2+ ] i increase by RR, as well as the highly TRPV4-selective antagonist, suggest the involvement of TRPV4 channels. The expression of TRPV4 in the urothelium further supports the contribution of TRPV4. However, our observations that the ENaC antagonist (amiloride), the Piezo channel antagonist (GsMTx4), and the connexin channel antagonist (carbenoxolone) could not block the ALI-induced [Ca 2+ ] i increase invalidates the involvement of these MSCs.
These results are in contrast with those of previous reports, which state that Piezo1, Piezo2 (Marshall et al., 2020), ENaC, or connexin channels (Shimura et al., 2018) contribute to mechanosensory transduction of the bladder urothelium (Du et al., 2007;Miyamoto et al., 2014). These inconsistent results may be attributed to the different ways in which the cells were stimulated. In the study Miyamoto et al. (2014), mouse urothelial cells were horizontally elongated by extending the cell chamber. In the current study, the apical surface of urothelial cells were exposed to the ALI, such that any curved cell surface was subjected to a net component of tension forces that acted perpendicularly to the cell membrane ( Figure 1B). Thus, the inconsistency in results may imply that urothelial cells use different mechanosensing channels in response to various forms of tension forces (horizontal vs. perpendicular).
Pannexin 1 channels are characterized by their mechanosensitivity and their ability to form large pores that are permeable to ATP and other molecules (Penuela et al., 2013). Therefore, they are reportedly involved in the nonvesicular release of ATP in the bladder urothelium, in response to mechanical stimulation (McLatchie and Fry, 2015). Consistent with this theory, our observation that 10 panx, the specific pannexin 1 channel antagonist, significantly attenuated both the ALI-induced [Ca 2+ ] i increase (by 39.3%) and ATP release indicates the involvement of pannexin 1 channels. Furthermore, our immunohistochemical analysis demonstrated the expression of the pannexin 1 channel in urothelial cells. We found that the ALI-induced [Ca 2+ ] i increase was reduced by 22% in Ca 2+ -free HBSS and was reduced by 62.4% following the application of thapsigargin, a blocker that irreversibly inhibits the ER Ca 2+ -ATPase pumps. These results suggest that compared with Ca 2+ entry, Ca 2+ release from the ER calcium stores are the dominant source of the ALI-induced [Ca 2+ ] i increase. The contribution of Ca 2+ release from the ER calcium stores can also be supported by the observation that ALI induced [Ca 2+ ] i was significantly reduced by application of BAPTA-AM to clamp intracellular Ca 2+ levels (Figure 4Bc). The IP3 receptor (IP3R) and ryanodine receptor on the ER have been implicated in Ca 2+ release from intracellular calcium stores, and play an important role in Ca 2+ -induced Ca 2+ release (CICR; Galione and Churchill, 2002). In agreement with this idea, when the ER calcium stores were blocked with the antagonist, IP3R (Xest-C), or the ryanodine receptor, the ALI-induced [Ca 2+ ] i increase was reduced by 71.6 and 67.7%, respectively. The CICR was mediated by the activation of ryanodine receptors located on the ER, which can be blocked by high concentrations of ryanodine (100 μM; Bouchard et al., 2003). In the current study, the blocking effects of Xest-C and ryanodine may imply the involvement of the CICR mechanism in the ALI-induced [Ca 2+ ] i increase.
One report on a human urothelial cell line (T24 cells) showed that ATP released upon cell stretching contributes to a major part of Ca 2+ signaling (Guan et al., 2018). In agreement with this finding, we also observed that ALI-evoked ATP release contributed to the ALI-induced [Ca 2+ ] i increase via paracrine or autocrine mechanisms. This conclusion can be supported by the following evidence: (1) the ALI-induced [Ca 2+ ] i increase was significantly reduced by ATP hydrolysis with apyrase and (2) the [Ca 2+ ] i increase was significantly reduced by blocking the P2X and P2Y purinergic receptors with PPADS. Thus, autocrine ATP signaling works like an amplifier to increase the concentration of cytosolic Ca 2+ . We did not examine which type of purinergic receptors is involved with these processes. However, previous studies have shown that the P2X receptor that mediates the increase in [Ca 2+ ] i is dependent on extracellular Ca 2+ . On the other hand, P2Y receptors are G-protein-coupled receptors that can induce the release of Ca 2+ from intracellular stores via the IP3 pathway (von Kugelgen and Hoffmann, 2016). Based on the dominant role of intracellular Ca 2+ release from the ER in the ALI-induced [Ca 2+ ] i increase, also based on our observation that the amplitude of ATP evoked [Ca 2+ ] i increase in zero Ca 2+ bath is not significantly different from that in normal HBSS bath (data not shown), we propose that P2Y receptors may have a greater contribution to the process than P2X receptors.
Based on the current results, a working model has been proposed (Figure 7). The process is likely initiated with the activation of TRPV4 or pannexin 1 channels and Ca 2+ influx, which in turn activates IP3 or ryanodine receptors on the ER to induce Ca 2+ release from intracellular Ca 2+ stores to further increase [Ca 2+ ] i . Activation of the pannexin 1 channels or [Ca 2+ ] i increase leads to ATP release through nonvesicular or vesicular mechanisms, respectively, which further enhances the increase in [Ca 2+ ] i by activating at P2X and P2Y receptors via autocrine or paracrine mechanisms.
Notably, the method used to create the ALI-induced plasma membrane deformation by causing changes in the perfusion volume may alter the equilibrium and concentration of signaling molecules close to the plasma membrane, thereby altering their autocrine/paracrine effects. Typically, this is the case with the ATP concentration; transient reduction of the perfusion fluid volume may increase the concentration of the nucleotide. Compared with other techniques, this is a disadvantage of the ALI method to induce urothelial cell deformation and is a limitation of the current study. Another limitation is the pharmacological approach employed for most of our experiments to determine the associated channels or receptors. Some of the antagonists used in the current study may not be very specific.
FIGURE 7 | A working model of cellular mechanisms for ALI-induced a [Ca 2+ ] i increase. Cell deformation activates membrane-bound TRPV4 or pannexin 1 channels to cause Ca 2+ influx and [Ca 2+ ] i increase, which in turn activates IP3 or ryanodine receptors on the endoplasmic reticulum (ER) to induce Ca 2+ release from intracellular Ca 2+ stores to further increase the [Ca 2+ ] i levels. The activation of pannexin 1 channels or the [Ca 2+ ] i increase can lead to ATP release, through nonvesicular or vesicular mechanisms, respectively. ATP further enhances the [Ca 2+ ] i increase by acting on P2Y receptors to increase IP3 level and Ca 2+ release from Ca 2+ stores.

CONCLUSION
In summary, we revealed the cellular mechanisms underlying membrane tension force induced [Ca 2+ ] i increase in urothelial cells. Changes of any one of the associated channels or receptors (Figure 7) may affect bladder mechanotransduction and sensory function. Although the chances that the bladder urothelium exposed to an ALI is uncommon under physiological conditions, detrusor overactivity that occurs in overactive bladder conditions can increase deformation in the urothelium, which could trigger an increase in [Ca 2+ ] i and the release of ATP, independent of urine volume.

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.

ETHICS STATEMENT
The animal study was reviewed and approved by the Ethics Committee of the Second Hospital, Cheeloo College of Medicine, Shandong University [KYLL-2016(GJ)A-0027].