Cevimeline and pilocarpine are muscarinic agonists used clinically to treat dry mouth. In this study, we explored fluid secretion from mouse submandibular glands to determine the mechanism of cevimeline, pilocarpine, and an experimentally used agent carbachol. Cevimeline evoked almost the same amount of secretion at concentrations from 30 μM to 1 mM. Pilocarpine also induced secretion at a concentration as low as 1 μM and was the most powerful secretagogue at 10 μM. Secretion was induced by carbachol at 0.1 μM, with maximum secretion at 1.0 μM. Cevimeline induced monophasic secretion at all concentrations tested, whereas higher concentrations of pilocarpine and carbachol induced secretion with variable kinetics, i.e., an initial transient high flow rate, followed by decreased secretion after 2 to 3 min. In the presence of an epithelial Na+ channel blocker, amiloride, neither carbachol nor pilocarpine affected the Na+ level of secreted saliva; however, it significantly increased the Na+ content of cevimeline-induced saliva. The intracellular Ca2+ response of acinar cells was almost identical among all three agents, although recovery after drug removal was slower for cevimeline and pilocarpine. A profound decrease in intracellular pH was observed during pilocarpine and carbachol treatment, whereas intracellular acidification induced by cevimeline was only seen in the presence of a Na+/H+ exchange inhibitor. When external HCO3− was removed, cevimeline-induced saliva significantly decreased. These findings suggest that cevimeline specifically activates Na+/H+ exchange and may promote Na+ reabsorption by stabilizing epithelial sodium channel activity.
Salivary gland hypofunction induces dry mouth, which results in loss of not only oral health-related quality of life but also general health. The salivary glands are innervated by both sympathetic and parasympathetic pathways, with parasympathetic muscarinic receptors being the most important for control of salivary fluid secretion (Melvin et al., 2005). Systemic muscarinic drugs for patients with severe dry mouth, such as those with Sjögren's syndrome and those who have undergone irradiation, are used in addition to supportive therapy for dry mouth.
Cevimeline (CVL) is a rigid analog of acetylcholine and is used clinically for the treatment of dry mouth. CVL induces salivation mainly through the activation of muscarinic M3 acetylcholine receptors (AChRs) on salivary acinar cells. In animal studies, CVL evoked salivation in a dose-dependent manner. This secretion was completely inhibited by atropine, a muscarinic AChR antagonist, and 4-diphenylacetoxy-N-methyl-piperidine methiodide, an M3 AChR antagonist (Iwabuchi and Masuhara, 1994). The binding affinity of M3 receptors for CVL is thought to be the same as or higher than affinities for other muscarinic drugs (Iga et al., 1998). Furthermore, CVL has been shown to increase salivary secretion in both healthy subjects and patients being treated for dry mouth, including patients with Sjögren's syndrome and patients after radiotherapy for head and neck cancer (Fife et al., 2002; Chambers et al., 2007; Braga et al., 2009). CVL is reported to be effective with oral administration as well as with gargling before swallowing to directly activate minor and major salivary glands (Takagi et al., 2004).
Another muscarinic agonist used clinically is pilocarpine (PLC), an alkaloid imidazole obtained from the leaves of Pilocarpus jaborandi. PLC activates M3 AChRs and enhances fluid secretion from salivary tissue in a dose-dependent manner (Omori et al., 2003). It is used clinically to induce saliva secretion in patients with Sjögren's syndrome, patients after radiotherapy, and patients with graft-versus-host disease-induced dry mouth (Greenspan and Daniels, 1987; Rieke et al., 1995; Berk, 2008).
Clinical and animal studies comparing these two muscarinic drugs have revealed that although CVL requires a higher dose, it has minimal adverse effects and a longer-lasting salivation effect than PLC (Masunaga et al., 1997; Braga et al., 2009). However, most reports on CVL lack mechanistic details, and the differences between these agonists have not been well characterized at the cellular or glandular level. At the cellular level, intracellular free Ca2+ plays a critical role in salivation. Within 1 s of stimulation, intracellular Ca2+ ([Ca2+]i) activates the opening of Cl− channels located at the luminal surface of salivary acinar cells, resulting in the rapid loss of Cl− and water. Concomitant with the loss of Cl− is the loss of HCO3−, which results in intracellular acidification and the consequent extrusion of H+ by the Na+/H+ exchanger (NHE) to compensate for the pH decrease (Nauntofte, 1992). It has been reported that pHi affects epithelial Na+ channel (ENaC) activity, which regulates Na+ reabsorption (Reddy et al., 2008).
In the present study, we focused on the effects of CVL and PLC, as well as those of carbachol (CCh), a muscarinic agonist used experimentally, on the secretion and ion composition of saliva, the alteration of intracellular pH, and the mechanism of cellular signaling.
Materials and Methods
Ex Vivo Mouse Submandibular Gland Analysis.
All experiments were approved by the Animal Committee of Kyushu Dental College. C57BL/6J mice aged 7 to 9 weeks, purchased from the Kyushu Animal Laboratory, were habituated, and experiments were performed with the mice at 8 to 10 weeks of age. In the mouse facility, mice were exposed to a 12-h light/dark cycle and fed ad libitum. To eliminate neurological input, we used an ex vivo vascular perfusion technique for glandular fluid secretion analysis. The surgical procedure was described previously (Romanenko et al., 2007; Nakamoto et al., 2008). In brief, mice were anesthetized with chloral hydrate (400 mg/kg b.wt.), and the submandibular glands were dissected out under a dissecting microscope (SZX7; Olympus, Tokyo, Japan). The main artery was cannulated with a 31- to 32-gauge blunt-end cannula, and perfusion solution was applied with a peristaltic pump (IPC4; Ismatec, Glattbrugg, Switzerland) at a flow rate of 1 ml/min. The perfusion solution was composed of 120 mM NaCl, 4.3 mM KCl, 25 mM NaHCO3, 1.0 mM MgCl2, 1.0 mM CaCl2, 5 mM glucose, and 10 mM HEPES and was equilibrated with 95% O2-5% CO2. NaHCO3 was replaced with the same concentration of NaCl to prepare a HCO3− removal solution and was equilibrated with 100% O2. Unlike CCh, CVL and PLC stimulated very little saliva secretion at 25°C (room temperature) in our preliminary experiments. Thus, the solutions and the glands were warmed at 37 ± 0.5°C with a custom-made water jacket heating system. Fluid secretion was stimulated by adding each agonist at the indicated concentrations. Secreted saliva was collected by a precalibrated glass capillary tube, and the saliva flow inside the capillary was collected every 30 s for the first 3 min, and then every 1 min for the duration; the collected saliva was stored in 500-μl tubes at −80°C until analyzed. The Na+ and Cl− concentrations in the saliva were analyzed with an electrode chip using DRI-CHEM 7000 (Fuji Film Medical, Tokyo, Japan).
Intracellular Signaling Measurements: Intracellular Ca2+ and pH.
For cellular signaling, [Ca2+]i and intracellular pH (pHi) were measured. Submandibular glands were removed from anesthetized mice and placed in Krebs-Henseleit-Ringer solution (103 mM NaCl, 4.7 mM KCl, 25 mM NaHCO3, 2.8 mM glucose, 12.5 mM HEPES, 1.2 mM NaH2PO4, 4.9 mM sodium pyruvate, 2.7 mM sodium fumarate, 2.7 mM sodium glutamate, 2.6 mM CaCl2, and 1.1 mM MgCl2 supplemented with 1 mg/ml bovine serum albumin), minced approximately 100 times with fine scissors, and then digested with 520 U/ml collagenase L (Nitta Gelatin, Osaka, Japan) for 20 min at 37°C with continuous shaking at 100 cycles/min. The digested tissue was washed three times and dispersed into 5 ml of Ringer's solution. The cells were incubated for 20 min with 1 to 2 μM Fura-2 AM to detect [Ca2+]i or 1 μM 3′-O-acetyl-2′,7′-bis(carboxyethyl)-4(5)-carboxyfluorescein (BCECF), diacetoxymethyl ester to detect pHi. Fluorescence was detected under a microscope (IX71; Olympus) equipped with a fluorescence analysis system (Argus/Aquacosmos; Hamamatsu Photonics, Hamamatsu, Japan), with excitation and emission wavelengths of 340 or 380 and 510 nm, respectively, for Fura-2 and 440 or 495 and 535 nm, respectively, for BCECF. The [Ca2+]i was determined as the ratio of the emission at 380-nm excitation and emission at 340-nm excitation; the pHi was determined similarly for emissions at 495- and 440-nm excitation.
Drugs and Statistical Analysis.
Pilocarpine was purchased from Nacalai Tesque (Kyoto, Japan). Collagenase L was from Nitta Gelatin. Fura-2 AM and BCECF, diacetoxymethyl ester were from Dojindo (Kumamoto, Japan). All other chemicals were purchased from Sigma Japan (Tokyo, Japan). Cevimeline was a gift from Nippon Kayaku (Tokyo, Japan).
The collected data are shown as mean ± S.E. Student's t test was used for a comparison between two group means. For multiple comparisons, one-way ANOVA and then the Tukey post hoc test was applied to detect statistically significant differences using SPSS (version 14.0J; SPSS, Inc., Chicago, IL). Results were deemed significant at p ≤ 0.05.
Cevimeline- and Pilocarpine-Induced Salivation in Mouse Submandibular Glands Ex Vivo.
CVL and PLC are commonly used to induce fluid secretion in vivo, but few studies have examined the function of these agents ex vivo or in vitro. We examined the effects of these agonists in an in vitro system. At concentrations as low as 1 μM, CVL was able to induce salivation in perfused mouse submandibular glands ex vivo. Stable salivation was observed using 10 μM CVL [as well as 30 μM (126.5 ± 2.5 μl/10 min, mean ± S.E.)], and the kinetics of secretion during agonist stimulation was unchanged with up to nearly 1 mM CVL (Fig. 1A). PLC at a concentration as low as 0.1 μM induced salivation (1.5 ± 0.8 μl/10 min), and stable secretion was observed with approximately 3 μM PLC (170.5 ± 7.5 μl/10 min) (Fig. 1B). In addition, PLC was the most effective secretagogue during the 10 min of stimulation tested in this study (187.9 ± 13.7 μl/10 min at 10 μM). CCh, a muscarinic agonist used experimentally, showed moderate secretion induction at 0.1 μM, and stable secretion was observed at 0.3 μM CCh (148.4 ± 3.2 μl/10 min) (Fig. 1C). Higher concentrations of PLC induced salivation for a longer time period after removal of the agonist at 10 min, giving a greater total amount of secretion over a collection time of 20 min (Fig. 1E) compared with CVL (Fig. 1D) or CCh (Fig. 1F).
The kinetics of fluid secreted in response to PLC and CCh varied depending on concentration. At higher concentrations, secretion was biphasic, i.e., an initial rapid phase, followed by a lower plateau or gradual decrease (Fig. 1, B and C). However, CVL lacked the initial peak phase (Fig. 1A). Both PLC and CCh induced a rapid high flow rate in less than 1 min with strong stimulation at most concentrations tested, whereas CVL required approximately 3 min to reach the highest flow rate at all concentrations tested. Higher concentrations of PLC and CVL required 3 to 5 min to fully wash out their effects (Fig. 1, D and E).
The analysis of Na+ and Cl− concentrations in secreted saliva revealed clear differences among the agonists. Compared with CVL, both PLC and CCh induced significantly higher Na+ concentrations in saliva (CVL versus PLC: p < 0.05 at 10 and 30 μM, p < 0.01 at 100 μM; CVL versus CCh: p < 0.05 at 10 μM) and showed significantly higher Cl− concentrations (CVL versus PLC: p < 0.01 at 10 μM) (Fig. 1, G–I). Whereas Na+ and Cl− contents changed depending on the PLC and CCh agonist concentration, they were unchanged in CVL saliva.
Effect of β-Agonist and Contribution of Epithelial Na+ Channels in CVL-, PLC-, and CCh-Induced Fluid Secretion.
The Na+ and Cl− concentrations in mouse submandibular saliva are lower than those in parotid saliva, owing mainly to epithelial Na+ channels in the apical membrane of ductal cells (Romanenko et al., 2008). The cystic fibrosis transmembrane conductance regulator Cl− channel is critical for Cl− reabsorption in ductal cells (Catalán et al., 2010). Both mechanisms are up-regulated by phosphodiesterase, which can be pharmacologically activated by β-stimulation (Romanenko et al., 2008). Here, when 10 μM isoproterenol (IPR) was added in the presence of a muscarinic agonist (30 μM CVL, 3 μM PLC, or 0.3 μM CCh), the flow rate decreased slightly (Fig. 2, A, C, and E), and the Na+ and Cl− concentrations decreased by ∼50% for all agonists (Fig. 2, B, D, and F).
We next examined whether the CVL-induced decrease in Na+ in saliva was attributable to specific activation of ENaC by using low concentrations of the inhibitor amiloride to specifically block ENaC. To ensure thorough inhibition of ENaCs, the glands were perfused with 10 μM amiloride for 30 min before muscarinic stimulation (Catalán et al., 2010). Only the saliva of CVL-treated glands showed significantly increased Na+ levels in the presence of amiloride. Amiloride also blocked the IPR-induced increase in Na+ and Cl− absorption observed in the presence of all muscarinic agonists.
Effect of Muscarinic Agonists on [Ca2+]i and pHi.
Because an increase in intracellular Ca2+ is central to regulating fluid secretion, we hypothesized that differences in activities, especially initial secretion, among the agonists stem from differences in the ability to modulate cellular signaling. The induction of fluid secretion by CVL was much slower than that for the other agonists, but increases in Ca2+ in the sustained phase did not differ significantly among the agonists (Fig. 3A). Both CVL and PLC showed a slower decrease in [Ca2+]i after removal of the agonist, and this was completely blocked by a muscarinic antagonist (0.5 μM atropine) (Fig. 3B).
With regard to pHi, both PLC and CCh induced a sustained decrease in pHi (Fig. 4, B and C), whereas CVL at 30 μM and higher concentrations (1 mM) produced only a minor change in pHi (Fig. 4A). CVL showed a statistically significant, but minor, decrease in pH at maximal concentration (CVL versus PLC: p < 0.05; CVL versus CCh: p < 0.01).
Melvin et al. (1988) revealed that stimulation with CCh acidified salivary cells and then the resting pH level recovers, which was accomplished by the Na+/H+ exchanger. We hypothesized that CVL produces only a minor change in pHi because it specifically activates this exchanger. Several types of NHEs are expressed in salivary glands. NHE1 is the primary form, and it controls proton concentration in exocrine glands in both acinar and ductal cells (Evans et al., 1999; Brown et al., 2003). Lower doses of 5-(N-ethyl-N-isopropyl) amiloride (EIPA) have been shown to specifically inhibit NHE1 (Orlowski and Kandasamy, 1996; Praetorius et al., 2000). Perfusion of glands with 5 μM EIPA for 2 min before muscarinic stimulation in the present study unmasked intracellular acidification, confirming that the reduced effect of CVL stimulation was due to its specific activation of NHE1 (Fig. 5A). With the maximum drop in pH in the presence of EIPA, the differences among the agonists disappeared (Fig. 5, A–C).
Anion accumulation from basolateral membranes and anion release at apical membranes are two main initiating components of saliva secretion. For continuous saliva flow, both the Na+-K+-2Cl− transporter and Cl−/HCO3− exchanger [anion exchanger (AE)] work together, coupled with NHE at the basolateral membrane (Evans et al., 2000; Melvin et al., 2005). We confirmed the effect of NHE on fluid secretion by removing HCO3− from the external solution, which was accomplished by replacing NaHCO3 with the same concentration of NaCl. The fluid in response to CVL decreased significantly (p < 0.01); however, that in response to PLC and CCh was not significantly different (Fig. 6).
We explored the mechanistic differences among two clinically used muscarinic agonists, cevimeline and pilocarpine, and the experimentally used muscarinic agonist, carbachol (Fig. 1). Both carbachol and pilocarpine showed different properties, depending on concentrations. At higher concentrations, they show a biphasic response, i.e., an initial transient high flow rate followed by decreased secretion. In comparison, cevimeline demonstrated slower induction of secretion at all concentrations tested and achieved stable secretion levels at higher doses. During the 10-min drug application, CVL and PLC induced sustained secretion over a very broad range of concentrations. CCh was capable of inducing fluid secretion over only a very narrow concentration range (0.3–1.0 μM). We identified the lowest concentration at which we could detect stable secretion in response to 10 min of stimulation and confirmed stable secretion in response to several 10-min pulse stimulations (data not shown). On the basis of these findings, we used 30 μM CVL, 3 μM PLC, and 0.3 μM CCh for further experiments.
Glands continued to secrete for several minutes after stimulation with CVL or PLC (and with higher concentrations of CCh), suggesting that these agents bind strongly to muscarinic receptors. This suggestion was confirmed by measuring [Ca2+]i in vitro (Fig. 3), which gave results consistent with those for ex vivo fluid secretion. The slow decay in [Ca2+]i levels that was observed after removal of CVL or PLC was completely inhibited by atropine, producing a rapid return to resting [Ca2+]i (Fig. 3B). Furthermore, perfusion with 0.5 μM atropine on muscarinic stimulation completely blocked ex vivo saliva secretion, indicating that CVL, PLC, and CCh induce salivation by activating muscarinic receptors (data not shown).
Compared with PLC and CCh, CVL treatment resulted in a significantly lower Na+ level in ex vivo saliva. There are at least three possible explanations for the significant Na+ decrease in ex vivo saliva in CVL-treated glands. First, Na+ absorption is flow rate-dependent; thus, the lower saliva flow rate observed with CVL relative to PLC and CCh may account for this difference. However, this possibility was eliminated by our tests on the effects of perfusion flow rate (data not shown). Decreasing the perfusion flow rate to 0.5 ml/min during exposure to 0.3 μM CCh decreased the fluid secretion rate to approximately 20% less than that observed during perfusion with 30 μM CVL at 1 ml/min, yet the Na+ concentration remained significantly lower in the CVL-induced saliva (42.6 ± 2.1 mM Na+ for CVL versus 53.5 ± 1.8 mM Na+ for CCh; p = 0.002, Student's t test). Second, CVL may activate a Na+ reabsorption pathway, presumably NHE3, which is highly expressed on the apical membrane of salivary duct cells (Park et al., 2001). However, no change was observed in the Na+ levels of ex vivo parotid (Park et al., 2001) or in vivo submandibular (Catalán et al., 2010) saliva from NHE3-gene disrupted mice, suggesting that NHE3-mediated Na+ absorption is probably not important at these flow rates. Finally, CVL may specifically activate ENaCs. We examined this possibility using amiloride, which inhibits ENaCs (Lingueglia et al., 1996), but it also has other broad effects on membrane proteins in epithelial cells. We used amiloride at lower concentration (10 μM) because it can inhibit not only ENaC but also NHEs at higher concentrations (Orlowski and Kandasamy, 1996; Praetorius et al., 2000). Furthermore, the Na+-dependent Ca2+ transport mechanism (presumably the Na+/Ca2+ exchanger) was shown to be functionally expressed in the rat submandibular gland (Gallacher and Morris, 1987; Morris et al., 1987) and may be affected by amiloride. However, replacement of external Na+ with Li+ to inhibit Na+/Ca2+ exchanger activity (Blaustein and Santiago, 1977; Laskowski and Medler, 2009) failed to show any significant change in [Ca2+]i in response to CCh, suggesting that it may be present in mouse submandibular gland, but it is not functionally important during muscarinic stimulation (data not shown). In glands infused with amiloride starting 30 min before and during muscarinic stimulation, Na+ absorption was blocked significantly, and the differences among the agonists disappeared, indicating that ENaC activity increased in CVL-stimulated glands. The interaction between cystic fibrosis transmembrane conductance regulator and ENaC, which is discussed extensively elsewhere (Reddy and Quinton, 2003), was made evident by treatment with a β-agonist (Fig. 2, B, D, and F), but the Na+ concentration was still significantly lowest in CVL, and ENaC inhibition abolished the differences in ion compositions among agonists. This finding suggests that the decreased Na+ concentration in CVL resulted from activation of ENaC, which was not associated with mechanisms involving β-agonists.
Unlike PLC and CCh, CVL produced minor intracellular acidification. Intracellular acidification during muscarinic receptor stimulation has been linked to up-regulation of NHE1, which is thought to be coupled with enhanced anion exchanger activity and increased fluid secretion. To determine whether CVL activated NHE, acinar cells were treated with EIPA, a specific inhibitor of NHE, in particular NHE1 (Orlowski and Kandasamy, 1996; Praetorius et al., 2000). This maneuver revealed that in the presence of EIPA CVL produced a drop in pHi to a level similar to that induced by PLC and CCh. This result suggests that CVL rapidly enhanced the activity of NHE (presumably NHE1) during fluid secretion.
Removing the HCO3− from the external solution did not affect PLC and CCh ex vivo saliva, which was consistent with previously reported results in rats (Martinez, 1987). On the contrary, HCO3− removal decreased CVL saliva significantly (Fig. 6), suggesting that a HCO3−-dependent Cl−/HCO3− exchange mechanism, which was coupled with NHE transport, was more active in the presence of CVL. Putative mechanisms discussed here are summarized in Fig. 7. There are many reports on NHE activation, and an increase in [Ca2+]i activates NHE exchange in salivary gland (Manganel and Turner, 1990, 1991). Later the Ca2+-calmodulin complex was proposed to activate NHE, because no additional phosphorylation occurred in response to muscarinic stimulation in salivary gland (Robertson et al., 1997). In addition, intracellular Na+ concentration sensing mechanisms have been discussed (Ishibashi et al., 1999). In this study, it is likely that [Ca2+]i alone was not involved in NHE up-regulation because we observed almost identical increases at both 30 μM CVL and 3 μM PLC (Figs. 3⇑–5). In addition, we show that HCO3− may be involved in NHE activation mechanisms (Fig. 6), consistent with previous reports (Yao et al., 1999). However, HCO3−-involved mechanisms are extremely complex. HCO3− can pass through anion channels, it is transported by AE and Na+-HCO3− cotransporters, and it can be produced by cellular metabolism by oxygen consumption and carbonic anhydrase (CA) (Li et al., 2006). In addition, HCO3− transport is regulated as part of a so-called metabolon, in which NHE, CA, and AE interact (Gonzalez-Begne et al., 2007). The activation mechanism by which CVL regulates the HCO3−-dependent pathway will be addressed in future studies.
Participated in research design: Kondo, Nakamoto, Masaki, and Hosokawa.
Conducted experiments: Kondo, Nakamoto, Mukaibo, and Kidokoro.
Performed data analysis: Kondo, Mukaibo, and Kidokoro.
Wrote or contributed to the writing of the manuscript: Kondo and Nakamoto.
We thank Dr. James Edward Melvin for constructive discussion of the manuscript and Nippon Kayaku for kindly providing cevimeline.
This work was supported by KAKENHI [Grant 20890203] from the Japan Society for the Promotion of Science.
Article, publication date, and citation information can be found at http://jpet.aspetjournals.org.
- acetylcholine receptor
- Na+/H+ exchanger
- epithelial Na+ channel
- intracellular pH
- analysis of variance
- 5-(N-ethyl-N-isopropyl) amiloride
- Cl−/HCO3− exchanger (anion exchanger)
- carbonic anhydrase.
- Received September 8, 2010.
- Accepted January 13, 2011.
- Copyright © 2011 by The American Society for Pharmacology and Experimental Therapeutics