Cannabinoid receptors in submandibular acinar cells: Functional coupling between saliva fluid and electrolytes secretion and Ca²⁺ signalling

Cannabinoids, the terpenophenolic compounds derived from the plant Cannabis sativa, have been used for centuries as psychotropic drugs and medicinal agents with diverse systemic effects. The actions of cannabinoids in mammalian tissues are primarily mediated by two main types of cannabinoid receptor (CBR): CB₁ and CB₂ (Kano et al., 2009; Matsuda et al., 1990; Matsuda, 1997). CB₁ receptors (CB₁Rs) are expressed at high levels in the central and peripheral nervous systems, where their main role is to modulate neurotransmitter release (Freund et al., 2003). CB₂ receptors (CB₂Rs) are found in immune cells as well as within the cardiovascular system and gastrointestinal tract (Pertwee, 2001; Sanger, 2007; Wright et al., 2008). CB₂Rs are thought to regulate abnormal gut motility, intestinal inflammation, visceral sensitivity and pain (Duncan et al., 2008; Kikuchi et al., 2008; Wright et al., 2008); however, their physiological role remains controversial.

Although Cannabis gains access to the systemic circulation within minutes of penetration, the oral cavity and gastrointestinal tract are points of first contact and, therefore, represent sites of considerable impact. Cannabinoids reduce enteric nerve activity and intestinal motility (Izzo et al., 2000; Izzo et al., 2001; Makwana et al., 2010; Mathison et al., 2004), inhibit saliva secretion (McConnell et al., 1978; Prestifilippo et al., 2006) and fluid and gastric acid secretion (Adami et al., 2002; Coruzzi et al., 2006; Hornby and Prouty, 2004; Izzo and Coutts, 2005) and prevent stimulated ion transport across the mucosa of the intestine, thus reducing water accumulation (Izzo et al., 2003). All these effects have been largely attributed to the activation of CB₁Rs. Although it was demonstrated that cannabinoids affect salivary gland function (Prestifilippo et al., 2006), the precise mechanisms of action remain unknown. Expression of both CB₁Rs and CB₂Rs in the submandibular gland is suggested from preliminary immunohistochemical studies, in which CB₁Rs were found mainly in the ductal system, whereas CB₂Rs were found in the acini (Prestifilippo et al., 2006). The high-affinity, endogenous CBR agonist arachidonyl ethanolamide (anandamide, AEA) markedly inhibited stimulated saliva secretion when injected into the submandibular gland (Prestifilippo et al., 2006) or into the lateral ventricle of the rat brain (Fernandez-Solari et al., 2009), acting either directly on the receptors in the gland or through presynaptic inhibition of neurotransmitter release. In contrast to the effects on submandibular gland, AEA induced amylase release in parotid glands that correlated with increased cAMP content, and also inhibited Na⁺-K⁺-ATPase activity (Busch et al., 2004).

Among other salivary glands, the submandibular gland provides a main source for secretion of saliva fluid and electrolytes, continual secretion of which is required for the moistening of the oral cavity (Ambudkar, 2000; Melvin et al., 2005). Salivary fluid and electrolyte secretion is a two-stage process: saliva is initially formed in the acinar lumen (primary fluid) and then is modified in the salivary ducts by removal of Na⁺ and Cl⁻ and addition of K⁺ and HCO₃⁻ to produce the final saliva that flows out the ducts (Turner and Sugiya, 2002). In acinar cells, the secretion of fluid and electrolytes is triggered by a complex cytosolic Ca²⁺ ([Ca²⁺]_cyt) signal, originating from Ca²⁺ release from the endoplasmic reticulum (ER) with subsequent activation of store-operated Ca²⁺ entry (SOCE), which is required for synchronised activation of spatially separated Ca²⁺-dependent Cl⁻ and K⁺ channels (Ambudkar, 2000; Melvin et al., 2005). In endothelial cells, AEA was shown to initiate [Ca²⁺]_cyt signalling via CB₂R-mediated activation of phospholipase C, 1,4,5-inositol trisphosphate receptor (IP₃R)-mediated Ca²⁺ release from the ER and accompanied Ca²⁺ influx (Waldeck-Weiermair et al., 2008; Zoratti et al., 2003). The CBR agonists, however, inhibited glucose-induced [Ca²⁺]_cyt increase and [Ca²⁺]_cyt oscillations in pancreatic β-cells (Nakata and Yada, 2008) and suppressed Ca²⁺ mobilisation in smooth muscle cells (Chataigneau et al., 1998; Zygmunt et al., 1997). Despite the functional significance of CBR-mediated regulation of salivation, their signalling pathways in acinar cells as the primary sites for saliva secretion have not been elucidated.

In the present study, we initially aimed to explore in vivo the effects of different CBR agonists on outflow and content of saliva produced by the rat submandibular gland. In addition, fast CBR-induced extracellular Na⁺ ([Na⁺]_e) transients were recorded in the lumen of acini in the submandibular gland. Ca²⁺ imaging techniques were then used to characterise CB₁R- and CB₂R-induced [Ca²⁺]_cyt signalling in isolated submandibular acinar cells.

Because the submandibular gland provides mainly a secretion of saliva fluid and electrolytes, we initially determined whether cannabinoids effect submandibular gland salivation. For this, we analysed in vivo the main parameters of salivation, such as saliva flow rate, total protein and electrolyte content of saliva secreted by the gland, before and after administration of cannabinoids. The potent, broad-spectrum CBR agonist WIN55,212-2 (5 μM), but not saline or vehicle, produced a marked decrease of saliva flow rate (43±6% decrease; n=8, P<0.001) at 5 minutes after an intraglandular injection when compared with vehicle. At the same time, no significant changes were observed either in total Ca²⁺ or in protein concentration in secreted saliva (Fig. 1A). Similarly, (R)-(+)-methanandamide (5 μM), a metabolically stable AEA analogue and selective agonist of the CB₁Rs, produced a significant decrease of saliva flow rate (24±4% decrease; n=7, P<0.001) and did not significantly affect total Ca²⁺ and protein concentrations (Fig. 1A). Another endogenous cannabinoid virodhamine (Porter et al., 2002) (30 μM) displayed an inhibitory effect on saliva flow rate similar to that induced by (R)-(+)-methanandamide (22±2% decrease; n=7, P<0.001) but also decreased total Ca²⁺ (14±1% decrease; n=7, P<0.05) and protein concentration (32±2% decrease; n=7, P<0.01; Fig. 1A). Because of the complex pharmacology of virodhamine (Porter et al., 2002; Ryberg et al., 2007; Kozlowska et al., 2006), we also tested JWH015, a selective agonist of the CB₂Rs (Pertwee, 1999). At a dose of 100 nM, JWH015 produced a decrease of saliva flow rate (34±5% decrease; n=12, P<0.001) with no significant effect on either total Ca²⁺ or protein concentrations (Fig. 1A). Thus, both CB₁Rs and CB₂Rs are likely to be involved in the downregulation of salivation, markedly reducing saliva outflow from the submandibular gland. The fact that the effect of the broad-spectrum CBR agonist WIN55,212-2 is greater than that of selective CB₁R or CB₂R agonists alone suggests a potential additive effect in CBR-mediated downregulation of salivation.

To investigate more persistent effects of cannabinoids on outflow and content of saliva, we performed repetitive applications of agonist to the gland (every 5 minutes), mimicking the conditions of single or persistent cannabis intake. Consistent with the previous results (Fig. 1A), we observed a significant reduction of saliva flow rate after a single application of WIN55,212-2 (5 μM), but not saline, which developed over 5 minutes (42±11% of the corresponding time point in the saline-treated group; n=5, P<0.05), reached a peak level at 10 minutes (45±9% decrease; n=8, P<0.001) and recovered to basal level within 15–20 minutes (Fig. 1Ci). No significant changes of total Ca²⁺ and proteins concentration in saliva were observed after single administration of WIN55,212-2 (Fig. 1Cii–iii). By contrast, repetitive applications of WIN55,212-2 at the same dose (5 μM) produced marked and constant inhibition of saliva flow rate, which began at 5 minutes (39±7% of the corresponding time point in the saline-treated group; n=7, P<0.05) and lasted for at least 30 minutes (59±11% of the saline-treated group at 20 minutes; n=10, P<0.001; Fig. 1Ci). Repetitive applications of WIN55,212-2 (5 μM) also induced an increase of both Ca²⁺ and protein concentrations in saliva. An increase of total protein concentration began at 10 minutes (52±9% of corresponding time point in the saline-treated group; n=10, P<0.05) and was maintained for the whole period tested (72±9% of the saline-treated group at 15 minutes; n=9, P<0.01; Fig. 1Cii). A significant increase of Ca²⁺ concentration was detected at 20 minutes (47±5% of the saline-treated group at 20 minutes; n=8, P<0.05) and lasted for 30 minutes during a repetitive WIN55,212-2 application (Fig. 1Ciii). As expected, saline did not affect salivation after either single or repetitive application (Fig. 1Ci–iii). On further analysis of the electrolyte content of saliva, we found that repetitive application of WIN55,212-2 did not significantly change K⁺ and P²⁺ concentrations in the final saliva during the period tested (P>0.05; Table 1). Taken together, our results suggest that persistent activation of CB₁Rs and CB₂Rs in the submandibular gland causes a significant decrease in saliva flow rate, a concomitant increase in total protein content and Ca²⁺ concentration with no changes in other electrolytes tested in final saliva.

To confirm that the effects of cannabinoid ligands are CBR specific, we investigated the effects of the CB₁R- or CB₂R-selective antagonists AM251 and AM630, respectively. The inhibitory effect of either (R)-(+)-methanandamide (5 μM) or JWH015 (100 nM) on the parameters of salivation was completely eliminated by AM251 (1 μM) or AM630 (1 μM), injected intraglandularly 20-30 minutes prior to the agonist administration (Fig. 1B). Interestingly, AM251 alone stimulated the saliva flow rate (48±7% increase; n=5, P<0.05 compared with vehicle) and slightly and insignificantly increased total Ca²⁺ and protein concentrations in secreted saliva (Fig. 1B). Selective inhibition of CB₂Rs by AM630 (1 μM) did not produce any significant changes in total protein and Ca²⁺ concentrations in saliva and did not alter saliva flow rate (Fig. 1B). To elucidate the possible physiological role of CB₁Rs and CB₂Rs in regulation of salivation, we studied the time dependence of the effects of their selective antagonists. At a concentration of 1 μM, AM251 significantly potentiated saliva outflow when injected every 5 minutes for all periods tested. An increase of saliva flow rate started within 10–15 minutes of AM251 administration (49±10% increase, P<0.05 compared with vehicle [time point ‘0’)] and was maintained over 30 minutes (Fig. 1Di). Neither total protein nor Ca²⁺ concentration was significantly changed during the period of observation (Fig. 1Dii–iii). Selective blockade of CB₂Rs by AM630 (1 μM) did not significantly alter the main parameters of salivation when compared with vehicle (time point ‘0’; Fig. 1Di–iii).

Taken together, our data indicate that cannabinoids cause a significant decrease in saliva flow rate and modify the content of final saliva through activation of CB₁Rs and CB₂Rs in the submandibular gland.

As cannabinoids do not affect K⁺ and P²⁺ concentrations in the final saliva, we tested next whether cannabinoids alter electrolytes secreted in the primary saliva fluid. For this, we utilised a microelectrode technique for recording fast secretory responses in submandibular gland slices in vitro (Fedirko et al., 2006a). Using concentric Na⁺-selective microelectrodes, we measured the changes of [Na⁺]_e in the acinar lumen in response to an agonist applied closely to the basal pole of acinus (Fig. 2A). Acetylcholine (ACh, 5 μM), a principal neurotransmitter that stimulates secretion of fluid and electrolytes, evoked a substantial increase of [Na⁺]_e in the lumen of the acinus (Fig. 2B). This increase of [Na⁺]_e reflects Na⁺ leakage into the lumen and indicates an increased secretory response. The typical [Na⁺]_e response to ACh consisted of a fast initial transient rise in [Na⁺]_e (half-rise time was 201±25 mseconds; n=14; Fig. 2B,E) followed by a much slower (within several seconds) recovery to the baseline. The average amplitude of the ACh-induced [Na⁺]_e transient was 120±12 μM (n=14). By contrast, WIN55,212-2 evoked a transient decrease in [Na⁺]_e when applied to the basal pole of acinus at different concentrations (Fig. 2C). The EC for the decrease in [Na⁺ 50]_e induced by WIN55,212-2 was 2.8×10⁻⁷ M (Fig. 2F). The average amplitudes of WIN55,212-2-induced [Na⁺]_e decreases were –266±32 μM for 1 μM (n=5) and –396±16 μM for 10 μM WIN55,212-2 (n=14) (P<0.01); the half-time of decrease in [Na⁺]_e was 103±26 mseconds (n=5) and 134±12 mseconds (n=14) for 1 μM and 10 μM WIN55,212-2, respectively (Fig. 2E). These data indicate that CBRs inhibit Na⁺ transport in the acinar lumen of the submandibular gland and alter electrolyte content in primary saliva fluid.

In order to establish the contribution of the different CBR subtypes in mediating this effect we used the selective CB₁R antagonist AM251 (10 μM). The decrease in [Na⁺]_e induced by WIN55,212-2 (10 μM) was strongly inhibited in the presence of AM251 (95±2% decrease; n=7, P<0.001; Fig. 2D). This suggests that it is CB₁Rs that predominantly contribute to the inhibition of Na⁺ transport into the acinar lumen.

The Na⁺-K⁺-ATPase present in the basolateral membrane of polarised acinar cells represents one of the core regulatory systems for fluid secretion driven by an osmotic gradient. To elucidate a possible mechanism for reduced fluid secretion upon activation of CBRs in the submandibular gland, we studied the effect of CB₁R and CB₂R agonists on Na⁺-K⁺-ATPase activity. A microsomal fraction was prepared from the submandibular gland tissue. We found that an agonist of either CB₁Rs or CB₂Rs markedly decreased Na⁺-K⁺-ATPase activity in gland tissue (Fig. 3A). In particular, the decrease in Na⁺-K⁺-ATPase activity was 47±10% (n=6, P<0.05) after 10 minutes incubation with (R)-(+)-methanandamide (5 μM) compared with the vehicle, and 50±8% (n=7, P<0.05) after incubation with JWH015 (100 nM). Furthermore, this effect was time dependent: the decrease in Na⁺-K⁺-ATPase activity started at 1 minute of agonist presence [37±2% decrease, n=5, P<0.05 for (R)-(+)-methanandamide; 42±10% decrease, n=6, P<0.05 for JWH015], reached a maximum at 5 minutes [59±5% decrease, n=6, P<0.01 for (R)-(+)-methanandamide; 60±11% decrease, n=6, P<0.01 for JWH015], and lasted for all periods tested (Fig. 3B). The rate of inhibition of Na⁺-K⁺-ATPase activity, calculated as the difference in Na⁺-K⁺-ATPase activity between vehicle and CBR agonist for the period of incubation, was the highest at 1 minute and was on average 13.6±7.2 μM P_i/mg protein/minute for (R)-(+)-methanandamide and 18.3±8.4 μM P_i/mg protein/minute for JWH015. The rate of inhibition then decelerated over 15 minutes of incubation either with CB₁R or CB₂R agonists.

Thus, CB₁R and CB₂R agonists inhibit Na⁺-K⁺-ATPase activity in submandibular gland tissue. However, despite the fact that acinar cells are a major cell type in salivary glands, a microsomal fraction obtained from the gland tissue also includes the components from other cells (e.g. duct cells and blood vessels). We therefore obtained membrane microsomes derived from isolated acinar cells to determine whether the effect of cannabinoids on Na⁺-K⁺-ATPase activity is observed in this specific cell type. In acinar cells, the inhibitory effect of WIN55,212-2 (5 μM) on Na⁺-K⁺-ATPase activity (58±8% decrease; n=14, P<0.001; Fig. 3D) was similar to the effect of cannabinoids in submandibular gland tissue.

To estimate the contribution of CBR subtypes in mediating such inhibition in acinar cells, we evaluated the effects of the selective CB₁R and CB₂R antagonists AM251 and AM630, respectively, on the broad-spectrum CBR agonist WIN55,212-2. The inhibitory effect of WIN55,212-2 (5 μM) on Na⁺-K⁺-ATPase activity did not significantly alter in the presence of AM251 (1 μM; applied for 15 minutes prior to the agonist administration) (60±8% decrease, n=8, P<0.01 compared with the vehicle versus 58±8% decrease, n=14, P<0.001 in the absence of AM251; P>0.8; Fig. 3D), indicating a significant contribution of CB₂Rs to the WIN55,212-2-induced decrease of Na⁺-K⁺-ATPase activity. This effect was similar to that induced by the selective CB₂R agonist JWH015 in gland tissue, indicating a substantial contribution of CB₂Rs to decreased Na⁺-K⁺-ATPase activity in acinar cells. In the presence of AM630 (1 μM), the inhibitory effect of WIN55,212-2 (5 μM) on Na⁺-K⁺-ATPase activity in acinar cells was reduced to 34±13% decrease (n=6; Fig. 3D) compared with the vehicle. This residual effect is likely to represent a contribution of CB₁Rs to decreased Na⁺-K⁺-ATPase activity, which is normally masked by CB₂R activation. Treatment with AM630 alone did not change significantly Na⁺-K⁺-ATPase activity in acinar cells (n=8, P>0.3).

Altogether, our results indicate that activation of either CB₁Rs or CB₂Rs leads to an inhibition of Na⁺-K⁺-ATPase activity in acinar cells of submandibular salivary gland.

In many cell types, including secretory cells, CBRs mediate their effect via modulation of [Ca²⁺]_cyt signalling machinery (Chaudhry et al., 1988; Nakata and Yada, 2008). To determine whether CBRs could trigger [Ca²⁺]_cyt signalling in salivary cells, we measured intracellular Ca²⁺ dynamics in isolated acinar cells. Bath administration of WIN55,212-2 evoked an increase of [Ca²⁺]_cyt, which, in contrast to the fast ACh-induced [Ca²⁺]_cyt transient, was characterised by a slow rising phase and [Ca²⁺]_cyt remained elevated for up to 30 minutes after agonist removal (Fig. 4A,B). The EC₅₀ for initiating [Ca²⁺]_cyt signalling by WIN55,212-2 was 1.75×10⁻⁷ M (Fig. 4C). The [Ca²⁺]_cyt response to 200 nM WIN55212-2 consisted of fast [Ca²⁺]_cyt oscillations with a gradual background [Ca²⁺]_cyt elevation (Fig. 4A). A supramaximal concentration of WIN55,212-2 (5 μM) produced a much faster increase in [Ca²⁺]_cyt without [Ca²⁺]_cyt oscillations (Fig. 4B); the half-rise time of gradual [Ca²⁺]_cyt increase induced by WIN55,212-2 (5 μM) was 353±29 seconds (n=5).

In the absence of extracellular Ca²⁺, WIN552,12-2 (1 μM) evoked only small and slow [Ca²⁺]_cyt transients with an average amplitude of 0.33±0.04 (n=14; Fig. 4D,E), which reflects Ca²⁺ release from the intracellular stores, mainly from the ER. By contrast, the amplitude of [Ca²⁺]_cyt signal induced by 1 μM of WIN 55,212-2 substantially increased (71±9% increase; n=12, P<0.0001) in Ca²⁺-containing extracellular medium (Fig. 4D,E), indicating a major contribution of SOCE to [Ca²⁺]_cyt rise.

Although both CB₁Rs and CB₂Rs are thought to be expressed in the submandibular gland, only CB₂Rs have been detected immunohistochemically in the acini of the gland (Prestifilippo et al., 2006). We recorded [Ca²⁺]_cyt transients in isolated acinar cells induced by a selective agonist of CB₂Rs, JWH015. At 100 nM, JWH015 evoked a small [Ca²⁺]_cyt rise with an average amplitude of 0.052±0.005 (n=33; Fig. 5A,B), the half-rise time of the JWH015-induced [Ca²⁺]_cyt transient was 9.2±0.9 seconds (n=26; Fig. 5B). The stimulatory effect of JWH015 on [Ca²⁺]_cyt was completely abolished in the presence of the CB₂R-selective antagonist AM630 (1 μM), indicating specificity of the effect due to the activation of CB₂Rs subtype (Fig. 5C). Thus, CB₂Rs in acinar cells are functional and their activation triggers [Ca²⁺]_cyt signalling.

Our findings that CB₁Rs predominantly decrease Na⁺ transport into the acinar lumen and inhibit Na⁺-K⁺-ATPase activity in acinar cells also indicate the presence of functional CB₁Rs in acinar cells. Moreover, in isolated acinar cells, the CB₁R selective agonists (R)-(+)-methanandamide and arachidonylcyclopropylamide (ACPA) also triggered [Ca²⁺]_cyt signalling. Similarly to WIN55,212-2, (R)-(+)-methanandamide (10 μM) induced an increase of [Ca²⁺]_cyt which was characterised by an initial fast transient followed by a gradual [Ca²⁺]_cyt rise (Fig. 6A). In contrast to WIN55212-2, (R)-(+)-methanandamide induced a [Ca²⁺]_cyt rise with a much smaller amplitude (average amplitude was 0.20±0.02, n=11; half-rise time of fast [Ca²⁺]_cyt transient was 7.34±0.85 seconds, n=11; Fig. 6C). The increase of [Ca²⁺]_cyt was dose dependent with an EC₅₀ of 5.4×10⁻⁶ M (Fig. 6B). Another selective CB₁R agonist, ACPA (50 nM), produced similar changes in [Ca²⁺]_cyt in acinar cells (Fig. 6D). In particular, in Ca²⁺-containing extracellular medium, ACPA induced a [Ca²⁺]_cyt rise with an amplitude of 0.34±0.14 (n=14) and half-rise time of fast [Ca²⁺]_cyt spike of 2.49±1.42 seconds (n=6; Fig. 6E). It should be noted that the half-rise time of the Ach-induced [Ca²⁺]_cyt transient was 2.04±0.14 seconds (n=10), which is similar to that induced by ACPA. Interestingly, the amplitude of the ACPA-induced [Ca²⁺]_cyt rise in Ca²⁺-free extracellular medium was similar to that induced by WIN55,212-2 in Ca²⁺-free medium (0.28±0.03, n=4 versus 0.33±0.04, n=14 for ACPA and WIN55,212-2, respectively). This similarity indicates a predominant participation of CB₁Rs to WIN55,212-2-induced Ca²⁺ release from the ER. Upon re-addition of extracellular Ca²⁺, an additional [Ca²⁺]_cyt increase was observed as a result of SOCE (Fig. 6D). The SOCE-mediated component of the ACPA-induced [Ca²⁺]_cyt signal was much smaller than observed with WIN55,212-2 (18±2%, n=14 versus 71±9%, n=12, respectively, P<0.01; Fig. 4D, Fig. 6E) indicating that SOCE only contributes slightly to CB₁R-mediated [Ca²⁺]_cyt signalling. Comparison of the effects of the selective CB₁R agonist ACPA and those of the broad spectrum CBR agonist WIN55,212-2 suggests that CB₂Rs might contribute to WIN55,212-2-induced SOCE.

Because prolonging inhibition of CB₁Rs affects salivation, we tested the effect of CB₁Rs antagonist per se on the changes in [Ca²⁺]_cyt in isolated acinar cells. In contrast to CB₂R blockade [AM630 (1 μM) did not produce any significant [Ca²⁺]_cyt rise in acinar cells over 10 minutes of application (Fig. 5C)], inhibition of CB₁Rs with AM251 (1 μM) produced a significant rise in [Ca²⁺]_cyt that was characterised by slow, gradual kinetics for the whole period of antagonist application (10–15 minutes) (Fig. 6F).

Thus, our findings demonstrate that CB₁Rs and CB₂Rs are present in submandibular acinar cells and are functionally coupled to [Ca²⁺]_cyt signalling cascades.

Here, we demonstrated that in vivo and in vitro activation of CBRs in the submandibular salivary gland caused a substantial reduction of saliva outflow and significant alterations of saliva content in both the primary and final saliva secreted by the gland. At the same time, CBRs in acinar cells inhibited Na⁺-K⁺-ATPase activity and triggered [Ca²⁺]_cyt signalling by different receptor subtype-specific signalling cascades.

Salivary glands permanently produce saliva ensuring resting (or basal) salivation, which occurs at a low rate and increases up to tenfold during the tasting and chewing of food (agonist-evoked saliva secretion) (Ambudkar, 2000; Proctor and Carpenter, 2007). Resting saliva is an electrolyte-enriched fluid necessary for the moistening of the oral cavity between food intakes. Decreased saliva secretion is linked to pathological disorders and can result in mouth dryness (xerostomia), increased risk of periodontal diseases, caries, oral Candida and gland swelling (Ben-Aryeh et al., 1993).

High levels of endocannabinoids have been reported in gastric tissues (Di Marzo et al., 2008; Izzo et al., 2003; Izzo and Sharkey, 2010), which makes them potential candidates for intrinsic regulators of physiological functions. Moreover, accumulating evidence suggests that salivary glands are able to synthesise endocannabinoids; their stimulation with arachidonic acid increased the level of 2-arachidonoylglycerol (2-AG) and caused the formation of AEA (Fezza et al., 2003). The ability of the endocannabinoid AEA to modify agonist-stimulated saliva secretion has been shown in both submandibular and parotid glands (Busch et al., 2004; Fernandez-Solari et al., 2009; Prestifilippo et al., 2006). Despite the importance of unstimulated salivation for the physiology of the oral cavity, the role of CBRs in the regulation of salivation has not been fully defined.

By testing the effects of cannabinoids on salivation produced by the submandibular gland in vivo, we found that virodhamine, an agonist of CB₂R with in vivo antagonistic activity at CB₁R (Porter et al., 2002), inhibited saliva flow rate similarly to the CB₁R agonist (R)-(+)-methanandamide (~30% decrease, Fig. 1A), indicating involvement of both CB₁R and CB₂R subtypes in the downregulation of saliva outflow. Taking into account the complex pharmacology of virodhamine, the effects on other receptors cannot be excluded (Ho and Hiley, 2004; Ryberg et al., 2007; Kozlowska et al., 2006). Therefore, to clarify the role of CB₂Rs in saliva secretion, we also tested a selective agonist of CB₂R, JWH015. The effect of JWH015 was similar to that observed for (R)-(+)-methanandamide, suggesting an equal contribution of CB₁R and CB₂R in the downregulation of salivation. In support of these observations, highly potent, broad-spectrum CBR agonist, WIN55,212-2, demonstrated a pronounced decrease in saliva flow rate, indicating an additive action of CB₁R and CB₂R in reduced saliva outflow. The inhibition of salivation observed upon cannabinoid administration is mediated directly by CBR activation as the inhibitory effect of either the CB₁R or the CB₂R agonist was completely eliminated by their selective antagonists. When administered alone, the CB₁R antagonist had effects on saliva outflow, which might indicate a regulatory role for CBRs in continuous secretion of saliva fluid under unstimulated (basal) conditions. Interestingly, the effect of CB₁R blockade was opposite to those produced by the receptor agonists: a prolonged inhibition of CB₁R leads to increased salivation, whereas activation of CB₁R markedly decreases salivation. A similar effect was not clearly observed with the CB₂R antagonist, which did not show any significant changes in saliva flow rate. Consistent with our findings, an antagonist of CB₁R was reported to augment electrically evoked acetylcholine release in ileum and to evoke contraction and peristalsis (Coutts and Pertwee, 1997; Izzo et al., 2000), whereas the CB₂R-selective antagonist AM630 alone potentiated electrically stimulated relaxation of the rat fundus (Storr et al., 2002). By contrast, CB₁R and CB₂R antagonists alone did not alter agonist-induced fluid secretion in the submandibular gland (Fernandez-Solari et al., 2009), but when injected into the submandibular gland simultaneously (inhibiting both CB₁R and CB₂R) they increased salivation induced by the low dose of an agonist (Prestifilippo et al., 2006).

Our present findings support the current clinical observations of reduced saliva flow rate in people addicted to marijuana (Verstraete, 2005). By mimicking the conditions of single or repetitive cannabis intake, we found that a reduction of saliva outflow reached its maximum level within 10 minutes and recovered briefly after a single administration of cannabinoid but was reduced for prolonged periods if the agonist was applied repetitively (Fig. 1C).

Our data are consistent with the results demonstrating an inhibitory effect of CB₁R on water accumulation in gastrointestinal tract (Izzo et al., 2003; Sanger, 2007). Unstimulated (basal) salivation is maintained by low doses of a neurotransmitter released from the nerves innervating the gland (Ambudkar, 2000), CB₁Rs might reduce excitatory cholinergic neurotransmission in the intestine (Hinds et al., 2006; Storr et al., 2004) by presynaptic inhibition of evoked and spontaneous ACh release from myenteric nerves (Coutts and Pertwee, 1997), resulting in reduced peristalsis and decreased gastrointestinal motility and transit in vivo (Izzo et al., 2001). CBR-induced decrease of electrically stimulated salivary flow is also mediated by diminishing ACh release (McConnell et al., 1978). Reduced autonomic neurotransmission to the submandibular gland (Fernandez-Solari et al., 2009) might reduce the blood supply of the gland (McConnell et al., 1978) and lead to decreased water flow to the acini. However, evidence of CB₂R-mediated inhibition of salivation is still elusive because of limited functional evidence for a role of CB₂Rs in saliva secretion. The fact that CB₂Rs are thought to be expressed in submandibular acinar cells points to a potential role of this receptor subtype in saliva secretion. Indeed, our findings demonstrated a reduced saliva outflow upon activation of CB₂Rs that might be caused by CB₂R-mediated regulation of saliva release from the acinar cells to the salivary ducts due to their predominant location peripherally to the acini (Prestifilippo et al., 2006).

The submandibular salivary gland is composed of acinar cells, which secrete proteins and electrolytes (primary fluid), and ductal cells, which modify the electrolyte content producing final saliva (Ambudkar, 2000; Turner and Sugiya, 2002). We did not find any significant changes in K⁺ and P⁺ concentrations in the final saliva, collected from the gland ducts, in response to repetitive administration of cannabinoid in vivo. However, we found a significant increase of total protein concentration, which started 10–15 minutes after WIN55,212 administration (time point of maximal inhibition of saliva outflow) and was maintained for the entire period if the agonist was applied repetitively (Fig. 1Cii). Although we only measured the total fraction of saliva proteins, our data are consistent with previous reports demonstrating CB₁R-induced amylase release from rat parotid glands (Busch et al., 2004) and CB₁R-mediated stimulation of insulin and glucagon secretion by the pancreas (Bermúdez-Silva et al., 2008). In salivary gland cells, two major signal transduction pathways are implicated in protein secretion: cAMP generation and the Ca²⁺-phosphoinositide messenger system (Rubin and Adolf, 1994). The enhancement of protein secretion induced by CBR activation might be mediated through cAMP-induced Ca²⁺ signalling. Indeed, the cannabinoid-induced increase of protein concentration was associated with an increase of total Ca²⁺ concentration in the final saliva (Fig. 1Cii–iii), suggesting an extrusion of Ca²⁺ during exocytosis of secretory vesicles (Gerasimenko et al., 1996) and supporting the hypothesis for a role of Ca²⁺ in CBR-induced secretion of salivary proteins. In agreement with this assumption, in the case of single administration of cannabinoids, any significant changes in both total protein and Ca²⁺ concentrations were found in final saliva after 10 minutes of drug presence (Fig. 1A). Also, any effects of the antagonists of CB₁Rs and CB₂Rs alone on either basal total protein or Ca²⁺ concentrations were observed in final saliva during all periods tested (30 minutes). This is consistent with the observations of others showing an absence of an effect of antagonists per se, used at the same doses, on basal level of amylase release from parotid glands (Busch et al., 2004) as well as basal or even KCl-stimulated amylase release from acinar cells of rat pancreatic glands (Linary et al., 2009).

Contrary to observations for the final saliva, cannabinoids dramatically altered the ionic content of the primary saliva. Using a fast concentric microelectrode technique, we measured [Na⁺]_e in the acinar lumen close to the acinar cell apical membrane. WIN55,212-2 application to the basal pole of acini resulted in a transient drop of [Na⁺]_e in the acinar lumen reflecting inhibited transport of Na⁺ ions, which might, in turn, lead to the suppressed saliva flow rate observed in vivo. The fact that the WIN55,212-2-induced drop of [Na⁺]_e was significantly abolished by the selective inhibition of CB₁R (Fig. 2D) indicates not only CB₁Rs functioning in acinar cells, but also their significant role in regulation of Na⁺ transport. The inhibited Na⁺ transport is likely to be mediated by decreased Na⁺-K⁺-ATPase activity, significant inhibition of which by cannabinoids has been directly demonstrated either in whole gland or in isolated acinar cells by ourselves (Fig. 3) and by others (Busch et al., 2004). The effect of cannabinoids on Na⁺-K⁺-ATPase activity is mediated by CBR activation as it was eliminated by the selective antagonists, demonstrating an involvement of CB₁Rs and CB₂Rs in the inhibition of Na⁺-K⁺-ATPases in acinar cells. The reduced ion transport might also be mediated by a suppression of K⁺ channels located in the basolateral membrane of acinar cells. The ability of CBR agonists to suppress different types of voltage-dependent K⁺ currents has been shown in many cell types (Fan and Yazulla, 2003; Schweitzer, 2000; Hampson et al., 2000); such suppression was significant (~45%) and was eliminated by CB₁R antagonists. Thus, inhibited Na⁺ transport by cannabinoids might be due to suppressed K⁺ channel conductance and inhibited Na⁺-K⁺-ATPase activity in the basolateral membrane of acinar cells, which prevents efflux of K⁺ ions and extrusion of Na⁺ into the interstitium in order to maintain the electrical gradient necessary to cause a driving force for movement of Na⁺ into the lumen across the tight junctions of the acinar cells.

The fact that CBR-induced changes of ion concentrations were observed only in primary saliva but not in the final saliva, collected from the gland ducts, indicates that the effect of cannabinoids is acinar cell specific. This is also supported by the altered concentration of proteins, secretion of which is produced only by acinar cells. At the same time, any effects of cannabinoids on ductal cells, which modify electrolyte content in final saliva, cannot be excluded, particularly considering the expression of CB₁Rs in the ductal system (Prestifilippo et al., 2006).

Secretion of fluid and electrolytes is governed by the [Ca²⁺]_cyt signal initiated by IP₃R-mediated Ca²⁺ release from the ER and subsequent SOCE (Ambudkar, 2000; Melvin et al., 2005). The [Ca²⁺]_cyt signal also functions as a cellular messenger in regulation of salivary protein secretion (Rubin and Adolf, 1994). Previous studies have shown that CB₂Rs induce [Ca²⁺]_cyt mobilisation from the intracellular stores in endothelial cells (Waldeck-Weiermair et al., 2008; Zoratti et al., 2003) whereas CB₁R stimulation evokes capacitative Ca²⁺ entry (Filipeanu et al., 1997). Our findings that cannabinoids increased total Ca²⁺ concentration in secreted saliva simultaneously with increased protein concentration (Fig. 1C) suggest that altered basal salivation might be the physiological consequence of [Ca²⁺]_cyt signalling triggered by activation of CBRs in acinar cells. Indeed, the broad CB₁R and CB₂R agonist WIN55212-2 induced a dose-dependent increase of [Ca²⁺]_cyt in isolated acinar cells (Fig. 4). In contrast to epithelial (Nakata and Yada, 2008; Sugiura et al., 2000) and endothelial (Waldeck-Weiermair et al., 2008; Zoratti et al., 2003) cells, in which CBRs evoked a rapid transient increase in [Ca²⁺]_cyt, in acinar salivary cells we observed an initial fast transient [Ca²⁺]_cyt rise and a subsequent slow [Ca²⁺]_cyt increase, which persists even after the removal of the agonist. A comparison of the kinetics of the [Ca²⁺]_cyt rise induced by a CBR agonist with those produced by ACh (mediated by activation of muscarinic G-protein-coupled receptors, GPCRs), showed a similarity of the half-rise time of fast [Ca²⁺]_cyt spikes induced by JWH015 or (R)-(+)-methanandamide and ACPA to those induced by ACh, suggesting that the responses were due to a typical GPCR-mediated InsP₃-evoked increase in [Ca²⁺]_cyt.

Testing the effects of both CB₁R (AM251) and CB₂R (AM630) antagonists on the agonist-induced changes in [Ca²⁺]_cyt as well as the effects of antagonists alone, we found that, in the case of CB₂Rs, AM630 did not produce significant change in [Ca²⁺]_cyt when applied alone, and completely prevented [Ca²⁺]_cyt transient induced by the selective CB₂Rs agonist JWH015, thus confirming that JWH015-induced [Ca²⁺]_cyt responses were due to CB₂Rs activation. In the case of CB₁Rs, we were unable to antagonise this receptor subtype without observing a run-up of resting [Ca²⁺]_cyt. This run-up of resting [Ca²⁺]_cyt upon CB₁R blockage was similar to that observed following CB₁R activation and might be mediated partially by non-specific effects, such as increased permeability of the plasma membrane due to cannabinoid adhesiveness and/or SERCA inhibition followed by ER depletion and the consequent increase in SOCE, that could be promoted in the conditions of prolonged bath application of an agonist. The great sensitivity of CBR-induced [Ca²⁺]_cyt increase to extracellular Ca²⁺ indicates a substantial contribution of Ca²⁺ influx to generation of the [Ca²⁺]_cyt signal. In submandibular acinar cells, Ca²⁺ influx represents SOCE through the TRP1 channel (Liu et al., 2003) and a strong correlation between SOCE, TRP1 and fluid secretion has been established (Singh et al., 2000). Taking this into account, possible cross-talk between CBRs and TRP1 might exist, although an effect of cannabinoids on TRP1 has also been demonstrated (Maccarrone et al., 2008; Singh et al., 2000).

In isolated acinar cells, (R)-(+)-methanandamine and ACPA (CB₁R-selective agonist) induced a significant increase of [Ca²⁺]_cyt, confirming the presence of functional CB₁Rs coupled with Ca²⁺ signalling cascades in submandibular acinar cells. In addition, the activation of CB₁Rs or CB₂Rs alone or in combination resulted in different amplitudes of SOCE-mediated [Ca²⁺]_cyt rise, suggesting that these receptors activate distinct signalling pathways.

Overall, our data suggest that functional CB₁Rs and CB₂Rs are expressed in submandibular acinar cells and their activation inhibits Na⁺-K⁺-ATPase activity and triggers [Ca²⁺]_cyt signalling, which might underlie their effects on salivation and saliva content. Although the overall effects of CB₁R and CB₂R activation on salivary gland function are similar, there are differences in their effects at the cellular level. CB₂Rs appear to be more important for regulation of Na⁺-K⁺-ATPase activity, whereas CB₁Rs markedly effect Ca²⁺ signalling by distinct intracellular mechanisms. CB₁Rs trigger intracellular Ca²⁺ mobilisation from the ER, whereas CB₂Rs might initially induce Ca²⁺ release but preferentially activate SOCE. Although the precise nature of CB₁R- and CB₂R-mediated [Ca²⁺]_cyt signalling cascades in submandibular acinar cells are still unclear, the altered [Ca²⁺]_cyt level upon prolonged CBR inhibition suggests a physiological role for functional CBRs in [Ca²⁺]_cyt signalling and, together with our in vivo data, give a preliminary indication of the physiological role played by endogenous cannabinoids in the regulation of salivary gland functions. The substantial role of CB₁Rs and CB₂Rs in modulation of salivation and saliva content indicates that the endocannabinoid system could represent a novel therapeutic target in salivary glands dysfunction.

Male Wistar rats (150-200 g) were housed in cages on a standard 12:12 hours light/dark cycle. Water and food were available ad libitum until rats were transported to the laboratory before experiments. The animals were used in accordance with protocols that were approved by the Animal Care and Use Committee at the Bogomoletz Institute of Physiology and Ivan Franko National University. All efforts were made to minimise animal suffering and to reduce the number of animals used.

Ketamine was purchased from CuraMed Pharma (Karlsruhe, Germany), lysthenon from Nycomed (Austria), collagenase (type II) and agarose (type VIIA) from Sigma-Aldrich (Chemie, Taufkirchen, Germany). Fura-2 was obtained from Invitrogen (Carlsbad, CA, USA). WIN55,212-2 [(R)-(+)-[2,3-dihydro-5-methyl-3-(4-morpholinylmethyl)pyrrolo-[1,2,3-de]-1,4-benzoxazin-6-yl]-1-naphthalenylmethanone mesylate], (R)-(+)-methanandamide [(R)-(+)-arachidonoyl-1′-hydroxy-2′-propylamide], ACPA (arachidonylcyclopropylamide), AM251 [N-(piperidin-1-yl)-5-(4-iodophenyl)-1-(2,4-dichlorophenyl)-4-methyl-1H-pyrazole-3-carboxamide], AM630 (6-iodopravadoline) and JWH015 [(2-Methyl-1-propyl-1H-indol-3-yl)-1-naphthalenylmethanone] were from Tocris Bioscience. Cannabinoids were dissolved in dimethylsulfoxide (DMSO).

The rats were anesthetised by intraperitoneal injection of a mixture of ketamine (100 mg/kg body weight) and lysthenon (0.05 ml/kg) and then fixed in the supine position. To collect whole saliva secreted by the submandibular gland, two main ducts of both right and left glands were cannulated with tightly closed glass cannulae. Extra care was taken to prevent any damage of the submandibular duct papilla as well as saliva mixing of other types of saliva. To achieve this, the tapered hole of a glass cannula was put to cover the submandibular duct papilla, providing an accurate collection of the secreted saliva, which flows out into a tube connected to the other end of cannula via a variable speed peristaltic pump. To avoid any pressure on the duct papilla during pumping and prevent any stimulus for saliva flow, the pump was used only when the cannula was filled with secreted saliva to draw the fluid into a tube.

Saliva secreted over a period of 5–10 minutes was collected in a tube to evaluate salivation under naïve (unstimulated) conditions and upon pharmacological treatment. After that, the cannabinoids were applied either once or every 5 minutes as described in the text; their effect on salivation was expressed as a value normalised to that under naïve (unstimulated) conditions in a corresponding animal. Saline (0.9%, 0.05 ml) injections were used as control. In a separate group of experiments, the effects of vehicles were studied, including distilled water and DMSO (final concentration 1:1000).

Salivation was evaluated by the following parameters: saliva flow rate, concentration of total proteins and concentration of electrolytes (Ca²⁺, P²⁺, K⁺) in saliva collected from the ducts before and after drug administration. Saliva flow rate was calculated as a volume of saliva secreted per 5 or 10 minutes normalised to a weight of the animal and over a time course of 1 hour (Fedirko et al., 2006b). The concentration of proteins was determined using the Lowry method; electrolyte concentrations were measured as follows: Ca²⁺, using the colourimetric method; K⁺, by atomic-absorption spectroscopy; and P²⁺, using the chemistry analyser ‘Humalyzer 2000’.

WIN55,212-2 (5 μM), (R)-(+)-methanandamide (5 μM), virodhamine (30 μM), JWH015 (100 nM), AM251 (1 and 10 μM), AM630 (1 μM) or vehicle (either saline or DMSO with water) were given intraglandularly into both right and left submandibular glands either by injection (50 μl/rat) or through surgical skin section (50 μl/rat). Prior to cannabinoid administration, salivation was evaluated under naïve conditions in each animal for 5 or 10 minutes. After that, compounds at the doses indicated above were applied either once or every 5 minutes as described in the text; their effect on the parameters of salivation was expressed as a value of a given parameter normalised to a corresponding value obtained in the same animal under naïve (non-stimulated) conditions. Such normalisation is extremely useful to avoid any technical non-specific effects on salivation caused by, for example, cannulation using a pump, or anaesthesia with ketamine. The minimum period between two administrations of drugs was 5 minutes. To study the effect of antagonists on salivation per se, AM251 or AM630 at dose of 1 μM was injected intraglandularly every 5 minutes over a time course of 30 minutes; saliva was collected every 5–10 minutes between the injections. Additionally, the effects of the agonists were studied when CB₁Rs or CB₂Rs were blocked. Antagonists were applied over a period of 25 minutes as described above, then (R)-(+)-methanandamide or JWH015 was injected for 5 minutes.

Saline injections (0.9%, 0.05 ml) were used as a control. In a separate group of experiments, the effect of vehicles was studied; vehicles used were distilled water and dimethylsulfoxide (DMSO, final concentration 1:1000, 40 μl/rat).

The submandibular salivary gland slices were prepared from 7- to 9-week-old male rats. After rats were deeply anesthetised with an overdose of isoflurane, the submandibular gland was removed and transferred into 3% agarose in basic extracellular buffer (140 mM NaCl, 10 mM HEPES-NaOH, 4.7 mM KCl, 1.13 mM MgCl₂, 1 mM CaCl₂, 10 mM d-glucose, pH 7.3) at 35°C. Submandibular gland slices (200–300 μm thick) were cut on a vibratome (World Precision Instrument, FL, USA) in basic extracellular buffer at 4–8°C. Slices were maintained at room temperature in basic extracellular buffer.

Concentric high-speed Na⁺-selective microelectrodes with a response time of a few mseconds were made as described previously (Fedirko et al., 2006a). Briefly, the inner micropipette of microelectrodes with a tip diameter of ~1 μm was filled with phosphate-buffered 3 M KCl and inserted within the outer, ion-selective barrel and into the ion exchange column. The outer pipette was filled with a solution of 150 mM NaCl; its tip diameter was 4–6 μm. Recording of the changes in extracellular Na⁺ concentration ([Na⁺]_e) in the submandibular gland slices was carried out as described in previous studies with minor modifications (Fedirko et al., 2006a). Briefly, the acini were visually identified with a video microscopy system using a 40× water-immersion objective (0.95 N.A., Olympus, Japan). A puffing pipette was used for local application of an agonist on the basal pole of acinus (Fig. 2A). Recordings were obtained in the lumen of the acinus, into which primary saliva is secreted. The concentration of [Na⁺]_e was estimated by calibrating the microelectrodes in NaCl solutions over a range of 0.01–10 mM.

Acinar cells were isolated from the submandibular gland using collagenase treatment (0.25 mg/ml) as described previously (Fedirko et al., 2006b; Kopach et al., 2008). The extracellular solution contained 135 mM NaCl, 5 mM KCl, 10 mM HEPES-NaOH, 2 mM MgCl₂, 2 mM CaCl₂, 10 mM glucose, pH 7.35. To obtain a Ca²⁺-free solution, CaCl₂ was omitted, MgCl₂ was increased to 4 mM and 1 mM EGTA was added. All experiments were performed at room temperature (21–23°C).

Isolated acinar cells were loaded with Ca²⁺ dye fura-2/AM (5 μM) for 30 minutes at 35°C. Fura-2 fluorescence was measured by using a 60× water-immersion objective and a 12-bit cooled CCD camera and capturing board (Sensicam, PCO, Germany), as described elsewhere (Fedirko et al., 2006b; Kopach et al., 2008). Briefly, the images were acquired in a fast mode using 2×2 binning with a temporal resolution of ~300 mseconds per frame. The Ca²⁺ probe was excited by PolyChrome IV monochromator (Till Photonics, Germany) at 340/380 nm; Fura-2 emission was measured at >510 nm. The monochromator was attenuated with neutral density filters to avoid bleaching of fura-2. Imaging Workbench software (INDEC System, USA) was used to measure changes in fura-2 fluorescence, which were expressed as changes in the ratio of fura-2 fluorescence at 340 and 380 nm (F₃₄₀/F₃₈₀), which is proportional to [Ca²⁺]_cyt. The background area closest to the areas of interest was subtracted.

Membrane microsomes were obtained from the submandibular gland tissue or from the acinar cells, isolated from the submandibular gland, as described above. Submandibular gland collected from rats (7–9 weeks old; deeply anesthetised with an overdose of isoflurane) or a suspension of acinar cells was homogenised at 4°C in the following solution: 250 mM sucrose, 1 mM EDTA, 10 mM Tris-HCl (pH 7.2) and protease inhibitor complex (1:1000). The microsomal fraction was obtained by differential centrifugation (homogenate was centrifuged for 10 minutes at 1000 g, then supernatant was collected and centrifuged for 20 minutes at 10,000 g and for 60 minutes at 100,000 g) as described (Fedirko et al., 2006b; Busch et al., 2004). The final pellet was then re-suspended in Tris-HCl (10 mM), EDTA (1 mM) buffer containing protease inhibitors, and stored at −40°C until use.

To study the effects of cannabinoids on Na⁺-K⁺-ATPase activity, aliquots of membrane vesicles of the submandibular gland tissue (~2–5 mg/ml of protein) were transferred to the intracellular-like medium containing 50 mM NaCl, 100 mM KCl, 3 mM MgCl₂, 1 mM EGTA, 10 mM HEPES, 3 mM ATP (pH 7.2) and were incubated for 1–15 minutes at 36°C in the presence of selective agonists of CB₁Rs or CB₂Rs. Vehicle (DMSO at a final concentration of 1:1000) was used as a control. After termination of incubation with an agonist, aliquots were transferred to the Na⁺-K⁺-ATPase assay medium (final volume 1 ml) containing 100 mM NaCl, 20 mM KCl, 3 mM MgCl₂, 160 mM Tris-HCl and 1 mM ouabain. The reaction was initiated by the addition of 4 mM ATP at 37°C and was terminated at 5 minutes by addition of 30% ice-cold trichloracetic acid. Samples were centrifuged for 10 minutes at 3000 g and the concentration of inorganic phosphate (P_i) released (total ATPase activity) was measured colourimetrically. The activity of Na⁺-K⁺-ATPase was calculated as the difference between the mean of total ATPase activity and ouabain-insensitive ATPase activity and expressed as μmol P_i/mg protein per hour.

All data are expressed as mean ± s.e.m. Statistical significance was calculated using paired and unpaired two-tailed Student's t-test as appropriate. The results obtained in vivo were analysed by one-way ANOVA followed by Dunnett's post-hoc test or Bonferroni's post-hoc correction and two-way ANOVA with Bonferroni's post-hoc correction where appropriate comparing with the effect of vehicle. A P value less than 0.05 was considered to be statistically significant. Each trace shown is representative of at least four independent experiments.

The authors thank David Brown and Alexei Verkhratsky for their editorial assistance. The authors declare no conflict of interests.