|Home | About | Journals | Submit | Contact Us | Français|
Coordination of Ca2+ signaling among cells contributes to synchronization of salivary gland cell function. However, mechanisms that underlie this signaling remain elusive. Here, intercellular Ca2+ waves (ICW) in submandibular gland cells were investigated using fura-2 fluorescence imaging. Mechanical stimulation of single cells induced ICW propagation from the stimulated cells through ~7 layers of cells or ~120 μm. Our findings indicate that an extracellular ATP-dependent pathway is involved because the purinergic receptor antagonist suramin and the ATP hydrolyzing enzyme apyrase blocked ICW propagation. However, the gap junction uncoupler oleamide had no effect. ATP is released from mechanically stimulated cells possibly through opening of mechanosensitive maxi-anion channels, and does not appear to be directly linked to cytosolic Ca2+. The ICW is propagated by diffusing ATP, which activates purinergic receptors in neighboring cells. This purinergic signaling induces a Ca2+ transient that is dependent on Ca2+ release via IP3 receptors in the ER and store operated Ca2+ entry (SOCE). Finally, inhibition of mitochondrial Ca2+ uptake modified ICW indicating an important role of these organelles in this phenomenon. These studies increase our understanding of purinergic receptor signaling in salivary gland cells, and its role as a coordination mechanism of Ca2+ signals induced by mechanical stimulation.
Intracellular Ca2+ signaling has a central role in the regulation of salivary gland cell function . Once increased in the cytosol, Ca2+ activates various ion channels and transporters such as the Ca2+-activated K+ or Cl− channels, and Na+-K+-2Cl− transporters that are involved in saliva production and modification [1–3]. In a multi-cellular system, coordination of this Ca2+ signal between cells is important for synchronized and effective tissue function.
Intercellular Ca2+ wave (ICW) propagation has been reported in many different cell types such as respiratory tract epithelial , glial , aortic epithelial , liver epithelial , osteoblastic , and renin secreting juxtaglomerular  cells. Two main pathways are currently proposed as underling mechanisms of ICW propagation. One pathway is extracellular while the other is mediated through gap junction intercellular communication (GJIC). Two apposing connexin hemichannels form gap junctions between cells, which allow direct intercellular communication through passage of small signaling molecules such as Ca2+ and inositol 1,4,5-triphosphate (IP3) in the GJIC pathway. Previous studies in salivary gland cells indicate that GJIC is involved in the synchronization or propagation of Ca2+ signals. Muscarinic stimulation induced a synchronized Ca2+ signal among individual acinar cells that was disrupted by pretreatment with the gap junction uncoupler octanol in rat submandibular glands . In blowfly salivary glands, IP3 microinjection induced propagation of a Ca2+ wave through gap junctions .
ATP and purinergic (P2) receptor dependent signaling form a common extracellular pathway that, like GJIC, also contributes to ICW propagation. Various stimuli like muscarinic receptor activation, mechanical stress, and hypoxic conditions cause a release of ATP via exocytosis or ion channels . Two different subtypes of P2 receptors are involved in extracellular ATP-dependent signaling. P2X subtype receptors are Ca2+ permeable ion channels and the cytosolic Ca2+ increase upon P2X receptors activation is dependent on extracellular Ca2+. P2Y subtype receptors are G-protein coupled receptors that activate phospholipase C (PLC) which generates IP3. This IP3 induces a Ca2+ release via IP3 receptors in the endoplasmic reticulum (ER). Depletion of the ER Ca2+ store by IP3 receptor activation can further induce Ca2+ influx from the extracellular fluid through store operated Ca2+ channels (SOCC). Subtypes of both the P2X (P2X4, P2X7) and P2Y (P2Y1, P2Y2) receptors have been identified in different salivary gland cells .
Salivary glands experience repetitive mechanical stress during mastication. Myoepithelial cells, which contain myosin, contract to generate direct mechanical stimulation of salivary gland cells. Furthermore, mechanical stimulation, e.g. in chewing gum, has been suggested as an alternative treatment for xerostomia patients to increase saliva production [14, 15]. However, the detailed mechanisms underlying this cell signaling induced by mechanical stimulation are not clearly understood in salivary gland cells.
Here, we investigated the mechanism of ICW induced by mechanical stimulation in a monolayer of human submandibular gland (HSG) cells and in freshly isolated submandibular gland tissues using fluorescence Ca2+ imaging. The results below demonstrate that the propagation of Ca2+ waves from the mechanically stimulated cells to the neighboring cells relies on extracellular ATP-dependent signaling. Pharmacological characterization revealed that P2Y2 subtype receptors are involved in ICW. Our data also indicate that mechanosensitive maxi-anion channels are likely candidates for the ATP-release pathway in mechanically stimulated cells. Both intracellular Ca2+ release from the ER and Ca2+ influx from the extracellular medium contribute to ICW. Finally, mitochondria were found to play an important role by actively regulating the Ca2+ mobilization pathway.
HSG cells were grown in MEM (minimum essential medium Eagle, Mediatec Inc.) containing 2 mM glutamine and supplemented with 10% fetal bovine serum (FBS), 1% penicillin and streptomycin. Cells were maintained at 37°C in a humidified 5% CO2 incubator and passed twice a week.
Native submandibular gland cells were freshly prepared as previously described [16, 17] with some modifications. In brief, Sprague-Dawley rats were sedated by 100% CO2 and decapitated using a guillotine. Submandibular glands were surgically removed from the ventral cervical area and finely minced in MEM supplemented with 20 mM HEPES. The minced tissues were digested for 20 min with a collagenase (100 Unit/mL CLS type II, Wathington Biochemical Co.). The remaining collagenase was washed out twice and followed by gentle agitations for further dissociation of the tissues. Animals were carefully handled and sacrificed in accordance with the guide lines of National Institutes of Health (NIH) and University Committee on Animal Resource in New York University.
HSG cells were seeded on 25 mm No.1 glass coverslips (Fisher Scientific) at least 1 day before the experiment. Coverslips were transferred to Rose chambers and washed with imaging buffer containing (in mM): 145 NaCl, 4.5 KCl, 1.8 CaCl2, 0.5 MgCl2, 5 HEPES, 5.5 glucose, and pH 7.4 was adjusted with NaOH. 5 μM Fura-2 acetoxymethyl ester (Fura-2 AM, Invitrogen), a cell permeable fluorescence Ca2+ indicator, was loaded for 20 min at 37°C. 0.02% Pluronic F-127 (Invitrogen) was used to help dispersion of Fura-2 AM. After dye loading, the remaining Fura-2 AM in the extracellular medium was removed by washing twice with the same imaging buffer and incubated another 20 min. The cell chamber with its custom-built regulated heating block was mounted on an inverted microscope (Eclipse TE2000-E, Nikon). The freshly isolated submandibular gland cells were preloaded with 5 μM Fura-2 AM and kept on ice. After transferring the native tissues to the same cell chamber, cells were equilibrated in the imaging buffer for 30 min before starting experiments. Cells were observed with a 20x objective (S fluor, numerical aperture, 0.75; Nikon) and fluorescence micrographic images were captured at 510 nm using a digital CCD camera (CoolSNAP HQ2, Photometrics) and an imaging software (NIS-Element AR 3.0, Nikon) after alternating excitations at 340 and 380 nm. Fura-2 ratio (Fratio 340/380) images were displayed and the Fratio values from the regions of interest (ROIs) drawn on individual cells were monitored during the experiments and analyzed later offline with NIS-Element AR 3.0 software. The ICW propagation was measured by counting cell layers that showed peak amplitudes at least 0.1 Fratio values increase (ΔFratio). To minimize UV light exposure, 4×4 binning function was used. All experiments were performed at 37±1°C.
Cells were loaded with 5 μM Rhod-2AM (Invitrogen) for 20 min at room temperature, washed with imaging buffer, and incubated another 20 min at 37°C. For the co-localization study, mitochondria were visualized by transiently expressing the mitochondria matrix targeted pAc-GFP1 (mtGFP, a generous gift of E. Pavlov of University of Calgary) using a modified polyethylenimine (PEI, Sigma) transfection method originally described by Boussif et al . Briefly, 2 μg of plasmid were mixed with 10 μL of PEI stock (1g/L in distilled water, pH 7.4) in 500 μL culture medium without FBS and incubated for 15 min at room temperature to polymerize PEI/DNA. Cells were incubated with this mixture for 2 hours at 37°C and washed with normal culture medium containing FBS. After 24 hours, mtGFP expression was checked on a fluorescence microscope. Both Rhod-2 and mtGFP fluorescence images were deconvoluted using a built-in deconvolution module in NIS-Element AR 3.0 software.
Micropipettes of ~1 μm diameter were fabricated using a preconfigured program in a horizontal pipette puller (Model P-87, Sutter Instrument Co.). The tips of micropipettes were sealed and polished with a microforge (MF-83, Narishige). These micropipettes were used to briefly touch the top of target cells in a monolayer of HSG cells using a micromanipulator (MO-303, Narishige) to provide the initial mechanical stimulation and initiate the ICW. The freshly isolated submandibular gland cells were mechanically stimulated with a side wall of the micropipettes to avoid rotation and movement of tissues.
HSG cells were microinjected with 10 mM Lucifier Yellow CH lithium salt (Invitrogen) using a micropipette connected to a pressure microinjector (IM-30, Narishige). Dye coupling was confirmed up to 30 min after microinjection.
ATP, UTP, suramin, apyrase, SK&F96365, thapsigargin, 2-aminoethoxydiphenyl borate (2-APB), p-trifluoromethoxy carbonyl cyanide phenyl hydrazone (FCCP), oligomycin, GdCl3, LaCl3, 5-nitro-2-(phenylpropylamino)-benzoate (NPPB), carbenoxolone, glibenclamide, verapamil, and carbamylcholine chloride (carbachol) were purchased from Sigma. U73122 and its inactive analog U73343 were purchased from Biomol International. Brilliant Blue G was purchased from Alfa Aesar. 2-[(4-Fluorobenzoyl)amino]-benzoic acid methyl ester (Exo1) was purchased from Tocris Bioscience. Concentrated stock solutions were prepared in either distilled water or dimethyl sulfoxide (DMSO) and kept at −20°C. The concentration of DMSO in the final solution was less than 0.1%. DMSO alone tested up to 0.5% had no effect on ICW or ATP-induced Ca2+ responses.
Data were presented as mean±S.D. and number of cells tested (n). For statistical analysis, independent two-sample Student’s t-test were applied with an equal variance assumption to calculate p-values using an Origin 8.0 software (OriginLab Corp.). Two independent groups were considered statistically different if p-values <0.05.
A monolayer of HSG cells was used to mimic the microenvironment of a salivary gland during mastication and investigate the intercellular Ca2+ wave (ICW) induced by mechanical stimulation. The cell clusters were 300–400 μm in diameter and contained 150–300 cells. Focal mechanical stimulation of single cells in the center of cell clusters was achieved by gently touching the target cell for less than 1 sec with a sealed micropipette. The induced ICW was monitored by Fura-2 fluorescence ratio imaging. At the time of focal mechanical stimulation, there was an initial small increase in cytosolic Ca2+ concentration in the target cell. The amplitude of this Ca2+ signal reached a peak value 2–3 sec after mechanical stimulation. Without perfusion of the extracellular medium, this Ca2+ signal propagated from the target cell to the surrounding 7.2±2.2 layers of cells (~120 μm, n=21) with a speed of ~15 μm/sec in all directions (Fig. 1A and a movie in Suppl. Fig. 1). Therefore, about 200 cells showed synchronous Ca2+ signals in less than 10 seconds by focal mechanical stimulation of a single cell. The Ca2+ waves lasted up to ~2 minutes before returning to resting cytosolic Ca2+ level.
The ICW were propagated through an extracellular pathway. This Ca2+ wave was propagated to neighboring cells that did not have direct physical contact (Fig. 1B). Interestingly, the direction of Ca2+ wave propagation was affected by the flow of extracellular medium. With a flow rate of ~6mL/min, the Ca2+ wave was propagated more than 200 μm in the direction of perfusion whereas propagation against the flow was limited to less than 20 μm (Fig. 1C). In order to evaluate any contribution of GJIC pathway to propagation, cells were pretreated with a gap junction uncoupler oleamide (100 μM); this treatment had no effect on ICW (Fig. 1D). Furthermore, HSG cells showed poor dye coupling between cells when microinjected with Lucifer yellow (Suppl. Fig. 2). These results indicate that an extracellular pathway, but not GJIC, is used to propagate ICW induced by focal mechanical stimulation in HSG cells.
The possible involvement of signaling through P2 receptors was then examined to further characterize this extracellular pathway. Addition of the P2 receptor inhibitor suramin (Fig. 1E) or the ATP hydrolyzing enzyme apyrase (Fig. 1F) significantly decreased the ICW propagation induced by mechanical stimulation to 2.5±1.1 (n=10, p<0.001) and 1.2±0.9 (n=10, p<0.001) layers of cells, respectively. In summary, these findings indicate that ATP and P2 receptors are involved in the propagation of the ICW through an extracellular pathway.
Once the involvement of an extracellular ATP-dependent pathway was revealed, further pharmacological characterization was performed to identify the subtype of the P2 receptors involved in ICW induced by mechanical stimulation. Phospholipase C (PLC) and IP3 dependent signaling are activated by metabotropic P2Y receptors, but not by ionotropic P2X receptors. Pretreatment with 10 μM U73122, a PLC inhibitor, completely abolished ICW induced by mechanical stimulation (Fig. 2B). However, its inactive analog U73343 had no effect on ICW (Fig. 2B). This finding suggests that the effect of U73122 is not due to nonspecific effects but is related to PLC inhibition. These findings indicate that P2Y receptors are involved.
The affinity for purines and pyrimidines was used to identify the P2 receptor subtype responsible for these ICW. Addition of ATP to the extracellular solution induced dose-dependent Ca2+ increases with a half effective concentration (EC50) of 1.3±0.05 μM and maximal responses at ~10 μM ATP (Fig. 2A). The Ca2+ responses were not sustained at >100 μM ATP, but decreased with time. Once the cytosolic Ca2+ concentration returned to basal level still in the presence of 100 μM ATP, further addition of up to 500 μM ATP failed to induce subsequent Ca2+ responses. UTP activates only P2Y2 and P2Y4 receptors . P2Y2 has equal potency and efficacy with UTP and ATP, whereas P2Y4 is preferentially activated by UTP . Addition of UTP induced dose-dependent Ca2+ responses similar to ATP (Fig. 2A). The EC50 for UTP was 0.91±0.06 μM and maximal responses were at ~10 μM UTP. Higher concentrations of UTP also showed the same desensitization of the Ca2+ response as ATP. After desensitization by 200 μM UTP, mechanical stimulation did not induce further ICW (Fig. 2B). Previously, benzoyl (Bz)ATP sensitive P2z, or P2X7, receptors as well as UTP sensitive P2u, or P2Y2, receptors were identified in HSG cells . However, a specific inhibitor P2X7 receptor, brilliant blue G, had no effect on ICW induced by mechanical stimulation (Fig. 3B), suggesting that P2X7 receptors are not involved. Taken together, these results indicate that UTP-sensitive and PLC-linked metabotropic P2 receptors, likely P2Y2, are involved in propagation of ICW induced by mechanical stimulation.
What is the signal that triggers ATP release from the mechanically stimulated cells? Previous reports showed that mechanical stimulation or cell swelling induces ATP release from different cell types [9, 20, 21]. Mechanosensitive channels such as maxi-anion channels, connexin hemichannels, pannexin1 channels, P2X7 receptors, and anion transporters were proposed as ATP-release mechanisms. Thus, we tested whether the inhibitors of these pathways affect the mechanical stimulation induced ICW in our model. Among various inhibitors, we found that Gd3+ and a broad spectrum anion channel inhibitor NPPB effectively inhibited ICW (Fig. 3A and 3B). Since Gd3+ and NPPB are also known to block connexin hemichannels and pannexin1 channels as well as maxi-anion channels, the effect of carbenoxolone, which blocks both connexin hemichannels and pannexin1 channels, was tested. However, carbenoxolone did not inhibit ICW (Fig. 3B), suggesting that the effect of Gd3+ is not due to the inhibition of connexin hemichannels or pannexin1 channels. Furthermore, inhibitor of P2X7 receptors brilliant blue G, cystic fibrosis transmembrane regulator (CFTR) regulated Cl- channel inhibitor glibenclamide, multidrug resistant (MDR1) P-glyocoprotein regulated Cl- channel inhibitor verapamil, and inhibitor of exocytosis exo1 had no effect on ICW (Fig. 3B). These data suggest that mechanical stimulation itself may be the signal for ATP release from the target cells, and mechanosensitive maxi-anion channels are good candidates to provide a route for ATP release in our model.
We further examined whether the Ca2+ increase causes a secondary ATP release in the neighboring cells during ICW propagation. Carbachol is known to activate M3 muscarinic receptors and increase IP3 and cytosolic Ca2+ in salivary gland cells . Thus, we examined whether suramin, a P2 receptor inhibitor, affected the dose response of cabachol-induced Ca2+ responses. If cabachol-induced Ca2+ increase causes ATP release, then, it was expected that the P2 receptor would have some contribution to the calcium increase. However, the carbachol induced Ca2+ responses were not affected by 200 μM suramin (Suppl. Fig. 3). These data suggest that Ca2+ may not be the direct triggering signal for ATP release.
Activation of P2Y2 receptors induces Ca2+ release via IP3 receptors in the ER. Thus, it was expected that this ICW still could be induced by mechanical stimulation without Ca2+ influx from the extracellular medium. As expected, mechanical stimulation still induced ICW in the absence of extracellular Ca2+ (no added Ca2+ plus 1 mM EGTA), although the peak amplitudes of these waves were decreased (Fig. 4A). The effect of the store operated Ca2+ channel (SOCC) inhibitor, SK&F96365, on ICW was determined to further evaluate the Ca2+ influx pathway. In the presence of normal extracellular Ca2+, SK&F96365 (30 μM) also decreased the peak amplitude of the Ca2+ wave without inhibiting ICW (Fig. 4B). These findings indicate that intracellular Ca2+ release alone is enough to propagate ICW. However, Ca2+ influx through SOCC may also be involved in the regulation of the amplitude of the ICW.
In order to determine the contribution of the ER calcium reservoir to ICW, the effect of pretreatment with thapsigargin was determined. Thapsigargin is an ER Ca2+ pump inhibitor that effectively empties the ER store of Ca2+. The intracellular Ca2+ release after removal of extracellular Ca2+ is dependent on the ER Ca2+ store because thapsigargin (2 μM) completely abolished the ICW induced by mechanical stimulation (Fig. 4C). Consistent with the effect of the PLC inhibitor U73122 (Fig. 2B), pretreatment with an IP3 receptor inhibitor 2-aminoethoxydiphenyl borate (2-APB, 50 μM) in the absence of extracellular Ca2+ also completely abolished ICW (Fig. 4D). The involvement of IP3 receptors and SOCC in this Ca2+ mobilization further supports a mechanism reliant on P2Y2 receptor signaling during ICW.
Unlike the Ca2+ signals in the neighboring cells, the Ca2+ signal in the target cells was not completely inhibited by suramin and apyrase (Fig. 1B and 1C). These findings suggest that there might be some additional Ca2+ mobilization mechanisms in the target cells other than autocrine activation of its own P2 receptors. Removal of extracellular Ca2+ markedly decreased the Ca2+ signal in the target cells (Fig. 4), indicating that a Ca2+ influx pathway is activated by mechanical stimulation. Thus, we examined the effect of 500 μM Gd3+, an inhibitor of mechanosensitive cation channels, on the Ca2+ signal in the target cells. Above, we showed that 30 μM Gd3+ effectively inhibited the ICW propagation, but did not inhibit the Ca2+ increase in the target cells (Fig. 3A). However, 500 μM Gd3+ significantly decreased the Ca2+ signal in the target cells (Fig. 4E). These results indicate that both autocrine activation of P2Y2 and mechanosensitive cation channels contributed to the Ca2+ signal in the target cells.
Mitochondria buffer intracellular Ca2+ , which precipitated our investigation of the role of mitochondrial Ca2+ uptake in ICW. HSG cells were loaded with a mitochondrial Ca2+ indicator Rhod-2AM in order to determine if mechanical stimulation induces these organelles to take up Ca2+. Mechanical stimulation induced an increase of Rhod-2 fluorescence signal in a punctuate pattern around nucleus (Fig. 5A). This pattern perfectly co-localized with mitochondria targeted GFP (mtGFP) fluorescence (Fig. 5A). This mitochondrial Ca2+ uptake during ICW propagation was completely inhibited by treatment with the uncoupler FCCP (2.5 μM) and ATP synthase inhibitor oligomycin (2.5 μg/mL, not shown). Next, the impact of eliminating mitochondrial Ca2+ buffering with FCCP and oligomycin on the ICW was determined. The peak amplitude of the ICW induced by mechanical stimulation and the Ca2+ response to extracellular ATP additions were both significantly reduced when mitochondrial Ca2+ uptake was inhibited (Fig. 5B and C). However, oligomycin alone did not affect these Ca2+ waves or responses (not shown). These results indicate mitochondria play an important role in regulation of these Ca2+ transients.
Inhibition of mitochondrial Ca2+ uptake modified store operated calcium entry (SOCE) but not IP3-dependent release of intracellular Ca2+ from the ER. In the absence of extracellular Ca2+ (no added Ca2+ plus 1 mM EGTA), IP3-dependent Ca2+ release was induced by the addition of 10 μM ATP to the medium. The peak amplitude of this intracellular Ca2+ release was not affected by the inhibition of mitochondrial Ca2+ uptake (Fig. 6). However, both peak amplitude and rate of SOCE after re-introduction of normal extracellular Ca2+ were significantly decreased (Fig. 6C). These results indicate that mitochondria actively regulate SOCE, which affects repetitive Ca2+ responses elicited by mechanical stimulation and ATP additions.
Finally, we examined the possibility that the P2Y2 receptor dependent ICW can be induced by mechanical stimulation in freshly isolated rat submandibular gland ductal cells. Previously, the existence of P2Y2, or P2u, receptors was reported in native submandibular gland cells [16, 17, 23]. Thus, we first checked whether P2Y2 receptor agonist UTP can induce any Ca2+ signaling in freshly isolated rat submandibular gland cells. Similar to previous reports, a small increase of cytosolic Ca2+ was induced by 100 μM UTP (Fig. 7A), which was ~ 30% of the amplitude observed in HSG cells. Furthermore, mechanical stimulation also induced ICW in native rat submandibular gland cells (Fig. 7B). These results suggest that UTP sensitive P2u, or P2Y2, receptors can generate ICW induced by mechanical stimulation in native submandibular gland tissues as well as the HSG cell line.
In the present study, the mechanism of intercellular Ca2+ wave (ICW) propagation induced by mechanical stimulation in HSG cells and freshly isolated submandibular gland cells was investigated (proposed model shown in Fig. 8). Extracellular ATP and P2Y2 receptor dependent signaling is central to the resulting Ca2+ wave propagation (Fig. 1 and Fig. 2). Our data suggest that a mechanosensitive maxi anion channel is a good candidate for the ATP release mechanism (Fig. 3). We found that both IP3-dependent Ca2+ release from the ER and SOCE contribute to the cytosolic Ca2+ signal during ICW propagation (Fig. 4). Interestingly, mitochondria are actively involved in the regulation of SOCE, which contributes to fast store refilling during repetitive mechanical or P2 receptor stimulation (Fig. 5 and Fig. 6).
The role of extracellular ATP and P2 receptor signaling in salivary glands is not completely understood. It seems that this signaling is involved in production and modification of saliva by modulating various ion transporters through cytosolic Ca2+ or protein kinase C (PKC) dependent signaling [12, 13, 17]. In our experimental model, we found that ATP and P2Y2 receptors can mediate the propagation of ICW induced by mechanical stimulation. However, extracellular ATP-dependent signaling will be more complex in vivo because multiple subtypes of P2 receptors such as P2Y1, P2Y2, P2X4, P2X7 have been reported in different salivary gland cells . P2Y2 and P2X7 receptors were previously described by others in HSG cells . P2Y2 receptors are known to be activated by low micromolar range of ATP, whereas P2X7 receptors are activated by a higher (>1 mM) concentration of ATP in the presence of normal extracellular Mg2+ . High affinity receptors such as P2Y2 and P2X4 are probably required for this intercellular communication because of the high activity of ectonucleotidases at the cell surface and in saliva under in vivo conditions . However, depending on the amount of ATP released, other lower affinity P2 receptors may also be involved in vivo. Recently, it was reported that P2X7 receptors are involved in nucleotide induced fluid secretion in mouse submandibular gland . However, excessive amounts of ATP in the extracellular medium are also known to cause apoptosis or necrosis by forming a large pore constituted by this low affinity P2 receptor .
In native rat submandibular gland cells, specific localizations of P2 receptor subtypes were identified. UTP sensitive P2u or P2Y2, receptors exist on the basolateral side whereas BzATP sensitive P2z or P2X7, receptors exist on the luminal side of acinar and ductal cells [16, 17]. The amplitude of Ca2+ responses and Cl− currents after P2u receptor stimulation was ~1/3 of P2z receptor stimulation, suggesting that a significant level of P2u receptors are expressed in native tissues . Consistent with these observations, 100 μM UTP induced a small increase in cytosolic Ca2+ (Fig. 7A). Mechanical stimulation also induced ICW in freshly isolated rat submandibular gland cells (Fig. 7B). However, the Ca2+ response and the propagation of ICW were much smaller than that observed in HSG cells, which is possibly due to a lower expression level of P2Y2. In contrast, mechanical stimulation in submandibular gland cells from mouse did not elicit ICW. The differences in Ca2+ responses between rat and mouse may be due to different P2Y2 expression level, isolation procedures, or their sensitivity to mechanical stimulation. Interestingly, the expression level of P2Y2 receptors in acinar and intercalated ductal cells of salivary glands are up-regulated during tissue injury  and in a mouse model of Sjögren’s syndrome . Thus, P2Y2 based ICW may play a larger role under these pathophysiological conditions.
P2Y2 receptor up-regulation is apparently part of a cell survival mechanism. P2Y2 receptor activation induces ERK, and PI3K/Akt dependent cell survival signaling . Ca2+ signaling induced by muscarinic receptor activation is the major stimulus for saliva secretion and is significantly decreased in certain diseases like Sjögren’s syndrome [24, 25]. Hence, P2Y2 receptors may provide an alternative target to alleviate complications due to dry mouth or xerostomia while helping cells to survive under these pathological conditions. However, the experiments in these pathological conditions are beyond the scope of present study.
In our experimental model, unlike previous reports [10, 11], gap junction intercellular communication (GJIC) does not contribute to the propagation of ICW. However, there are still possibilities that gap junctions as well as P2 receptors are involved in synchronization or coupling of local Ca2+ signaling under in vivo conditions. Connexin 32 and 43 were previously identified in acinar and myoepithelial cells . However, connexin hemi-channels, which did not form gap junctions, can also contribute to extracellular ATP and P2 receptor signaling by providing an ATP-release pathway . Additionally, the gap junction inhibitors such as octanol, heptanol, and glycyrrhetinic acid, which are often used to prove the involvement of GJIC, are not specific and also may directly interfere with intracellular Ca2+ signaling . In the present study, oleamide, a more specific inhibitor of gap junctions than other inhibitors, did not affect ICW (Fig. 1D) and untreated cells did not spread Lucifer yellow staining to neighboring cells (Suppl. Fig. 2).
Mechanosensitive maxi-anion channels are possible candidates for the ATP release mechanism in our model (Fig. 8). The activity of maxi-anion channels was previously recorded in wide variety of cell types including cardiomyocytes , macular densa cells , astrocytes , and mammary fibroblastic C127 cells. However, the molecular identity of maxi anion channel is not yet clearly known . A voltage dependent anion-selective channel (VDAC) resembling mitochondrial VDAC was considered as a candidate for maxi-anion channel because maxi-anion channels have biophysical properties similar to mitochondrial VDAC. However, a recent knockout study indicates that none of the mitochondrial VDAC subtypes are responsible for the maxi-anion channel activity in the plasma membrane . The pore-sizing study using a non-electrolyte exclusion method indicates that the pore diameter of the maxi-anion channel (~2.6 nm) is large enough for the passage of ATP, whose diameter is ~1.2 nm .
Several other pathways were also proposed for ATP release mechanisms in other cell types which include connexin hemichannel, pannexin1 channel, P2X7 receptor, CFTR regulated Cl− channel, p-glycoprotein regulated Cl− channel, and exocytosis. Among these pathways, connexin hemichannel, pannexin1 channel, P2X7 receptor as well as the maxi-anion channel are known to be sensitive to mechanical stimulation or swelling. Gd3+ and some anion channel inhibitors such as NPPB, DIDS, and SITS, which inhibit maxi-anion channel, effectively block ATP release induced by cell swelling [20, 35]. Even if these anion channel inhibitors also inhibit connexin hemichannel, pannexin1 channel, and P2X7 receptors, the absence of an effect of carbenoxolone, an inhibitor of both connexin and pannexin1 hamichannnels, or brilliant blue G, a P2X7 receptor inhibitor on ICW, suggests that the maxi-anion channel, rather than connexin hemichannels, pannexin1 or P2X7 receptors, is a candidate for ATP release mechanism in our model (Fig. 3). Vesicular pathway was also examined (Fig. 3B), but its’ role seems to be minor at least in ductal cells. We hope more specific inhibitors or molecular approaches become available for maxi-anion channels in the near future.
The cytosolic Ca2+ increase upon mechanical stimulation in the target cells is contributed by at least two different mechanisms. One is possibly through direct activation of Ca2+ permeable mechanosensitive nonselective cation channels  and the other is an autocrine activation of its own P2Y2 receptor secondary to ATP release. We showed that Gd3+ decreased the Ca2+ signal induced by mechanical stimultion in the target cells at 500 μM (Fig. 4E), whereas 30 μM was enough to effectively inhibit ICW (Fig. 3A). We speculate that the binding affinity of Gd3+ to mechanosensitive cation channels and maxi-anion channels are different; the maxi-anion channels may have high affinity Gd3+ binding sites.
For subsequent ICW in the surrounding cells, both IP3-dependent Ca2+ release from the ER and SOCE contribute to cytosolic Ca2+ increase after P2Y2 receptor activation (Fig. 4). Since the ER Ca2+ store is limited, further Ca2+ influx from the extracellular medium is necessary for a more sustained or repetitive cytosolic Ca2+ signals. ER-store depletion is known to activate Ca2+ release activated Ca2+ channel (CRAC)/or SOCC in the plasma membrane. Recently, it was found that the specific interaction between the ER Ca2+ sensor stromal interaction molecule 1 (STIM1) and SOCC (ORAI1/TRPC1 channels) underlies the activation of SOCE [37, 38]. The contribution of extracellular Ca2+ influx and intracellular Ca2+ release to cytosolic Ca2+ signaling will vary depending on which P2 receptor signaling pathway underlies the ICW in different cell types.
Mitochondria, located next to ER and plasma membrane, regulate local Ca2+ concentrations as well as metabolic state. IP3 receptors and SOCC are known to be regulated by local Ca2+ and ATP concentration . Mitochondrial regulation of these Ca2+ mobilization pathways seems to be cell type specific. In rat basophilic leukemia cells, inhibition of mitochondrial Ca2+ uptake specifically caused a decrease in SOCE without affecting IP3-dependent Ca2+ release . However, in smooth muscle cells, Ca2+ release from IP3 receptor is decreased by mitochondrial depolarization . Here, we found that inhibition of mitochondrial Ca2+ uptake caused a decrease in the amplitude of ICW induced by repetitive mechanical stimulations (Fig. 5B). This decrease of Ca2+ responses in the FCCP/Oligomycin treated group is not simply due to a decrease in ATP release but may be due to the regulation of Ca2+ mobilization pathway by mitochondria. Importantly, additions of the same amount of extracellular ATP still showed decreased Ca2+ responses in FCCP/Oligomycin treated group. Furthermore, inhibition of mitochondrial Ca2+ uptake specifically decreased and slowed SOCE without affecting Ca2+ release from IP3 receptor in HSG cells (Fig. 6). Since oligomycin alone did not affect the repetitive Ca2+ signaling, mitochondrial regulation of SOCC is probably due to changes in the local concentration of Ca2+ rather than ATP. We speculate that the local Ca2+ concentration just next to SOCC is dynamically regulated by mitochondrial Ca2+ uptake, which in turn regulates Ca2+ dependent inactivation of SOCC (Fig. 8) . It is noteworthy that the basal cytosolic Ca2+ levels after mechanical stimulation or ATP addition remained slightly elevated in a FCCP/oligomycin treated group compared to that of control (Fig. 5B). This residual cytosolic Ca2+ might also help to keep SOCC inactivated and hence inhibit ER store refilling in between mechanical or P2 receptor stimulations.
The present study enhances our understanding about how the extracellular ATP and P2Y2 receptor dependent signaling pathway mediates intercellular Ca2+ signaling, which synchronizes salivary gland cell function. Mitochondria were found to have a specific role in regulating the Ca2+ mobilization mechanism in salivary gland cells. Finally, this information identifies P2Y2 receptors as an alternative therapeutic target to alleviate dry mouth symptoms in many different oral pathologies related to salivary hypo-function.
We thank Dan Malamud for providing HSG cells, Evgeny Pavlov for providing mtGFP, and Laurent Dejean and Oscar Teijido for critical discussions. This work was supported by NIH grants, RO1 DE14756 to David I. Yule and GM57249 to Kathleen W. Kinnally.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.