Macrolide antibiotics such as erythromycin (EM) and azithromycin (AZM) are beneficial in the treatment of mucus hypersecretion in inflammatory pulmonary diseases. Several indirect and direct mechanisms of action have been proposed. This study investigates the direct effect of macrolides on secretory function of isolated submucosal mucous gland cells (SMGCs). We hypothesize that macrolides inhibit the calcium influx necessary for evoked mucus secretion. To test this, we quantified mucin protein release using enzyme-linked immunosorbent assay, calcium-activated K+ (KCa), and calcium-activated Cl− (ClCa) currents. We measured nonselective cation current (NSCC) using whole-cell patch-clamp techniques; intracellular calcium concentration ([Ca2+]i) was measured using fura-2 Ca2+ imaging. We found that both EM and AZM are agonists at muscarinic receptors. EM (10 μM) evoked a small but significant increase in mucin release but inhibited the mucin release induced by subsequent acetylcholine (ACh) treatment. Both EM and AZM (10 μM) evoked KCa and ClCa whole-cell currents, which were blocked by atropine. EM and AZM also accelerated the decay of inositol trisphosphate-induced KCa and ClCa currents without changing the peak current amplitudes. Likewise, internal application of AZM (10 μM) enhanced the decay rate of ACh-induced KCa and ClCa currents. EM (1–10 μM) and AZM (0.1–10 μM) slowly (over 25–30 min) inhibited thapsigargin (TG)-induced Ca2+ entry when applied during the plateau phase of Ca2+ entry but blunted TG-induced Ca2+ entry by 70% after a 5-min pretreatment before initiating calcium entry. EM blocked TG-induced NSCC. We conclude that macrolide antibiotics are partial agonists at muscarinic receptors but inhibit stimulated mucus release by inhibiting calcium entry in SMGCs.
Mucus secretion plays a key role in the protection of airways from particles and pathogens. Submucosal gland cells produce approximately 95% of the mucus in normal airways (Reid, 1960). However, excessive mucus secretion is a common pathophysiological characteristic in chronic inflammatory respiratory diseases such as asthma, chronic obstructive pulmonary disease, and cystic fibrosis. Excessive mucus secretion obstructs airways and impairs gas exchange. Thus, reduction in mucus secretion and clearing of excess mucus from the airways are beneficial in chronic inflammatory lung diseases (Knowles and Boucher, 2002).
The main macromolecular component of mucus is mucin, a highly glycosylated protein. MUC5AC and MUC5B are major mucin subtypes secreted in the airways (Hewson et al., 2004). Mucin proteins link to each other, forming high molecular weight chains that are packaged within vesicles in mucous gland cells and released by exocytosis upon activation of the cells by a stimulus (Evans and Koo, 2009). The process of mucin release has been shown to be similar to neurotransmitter release at a nerve synapse (Evans and Koo, 2009). In a review of therapeutic strategies to treat mucus hypersecretion, Rogers and Barnes (2006) note that our understanding of the process of mucus release has provided numerous targets for therapeutic agents to decrease release. One class of drugs listed is the macrolide antibiotics, commonly used to treat respiratory infections and known to have beneficial effects of decreasing inflammation and mucus secretion (Tamaoki et al., 2004).
Macrolide antibiotics, such as erythromycin (EM) and azithromycin (AZM), in addition to their antimicrobial action, are known to be anti-inflammatory and to moderate mucus secretion in chronic inflammatory lung diseases (Gorrini et al., 2001). Macrolide antibiotics have also been shown to directly reduce the secretion of respiratory glycoconjugate induced by histamine and methacholine from human airway cells in vitro (Goswami et al., 1990). Another macrolide, clarithromycin, has been reported to inhibit MUC5AC production by NCI-H292 cells stimulated with ovalbumin and lipopolysaccharide (Shimizu et al., 2003), and clarithromycin has been shown to inhibit mucin secretion in a murine model of diffuse panbronchiolitis (Kaneko et al., 2003). Although these studies have shown that macrolide antibiotics reduce mucus release directly, the mechanisms of the direct action are unclear. In this study, we focus on the direct action of macrolide antibiotics on SMGCs in response to secretagogues, specifically through the mechanism of inhibition of Ca2+ entry, thereby inhibiting increases in [Ca2+]i and Ca2+-mediated responses.
An increase in [Ca2+]i is required for ion channel activities and mucin release stimulated by acetylcholine (ACh), histamine, and ATP (Hamada et al., 1993; Choi et al., 2005; Liu and Farley, 2005, 2007). In most nonexcitable cells such as SMGCs, membrane receptor activation induces Ca2+ release from internal stores and Ca2+ influx through store-operated Ca2+ channels (SOCs) or receptor-operated channels (ROCs) (Takemura and Ohshika, 1991; Vig and Kinet, 2007). It has been shown that in SMGCs, Ca2+ influx is important for mucus release to occur (Nagaki et al., 1994), although the pathway is not well described. In this study, we examine the effects of macrolide antibiotics on mucus secretion, ion channel activity, and Ca2+ influx in swine airway SMGCs.
Materials and Methods
SMGCs were prepared according to methods established previouosly in our lab (Liu and Farley, 2007). Swine tracheas were obtained from a local abattoir (Wilson Meat Packing, Crystal Springs, MS). Tracheas were quickly removed after exsanguination of the swine and placed in ice-cold physiological solution for transportation to the laboratory. The transport solution contained 100 U/ml penicillin, 100 μg/ml streptomycin, and 50 μg/ml kanamycin. The epithelium was scraped with a cell scraper (Corning Life Sciences, Lowell, MA), stripped off the cartilage rings, and then digested with 1 mg/ml collagenase IV (CLS-4; Worthington Biochemicals, Freehold, NJ), 0.5 mg/ml dithiothreitol, and 0.5 mg/ml bovine albumin (Sigma-Aldrich, St. Louis, MO) in physiological solution for four 90- to 100-min digestion cycles at 37°C; new enzyme solution was used at each cycle. Only cells collected from digest cycles 3 and 4 were used for experiments. The isolated cells were then mixed with 5 ml of fetal bovine serum to stop the enzymatic reaction. The cell mixtures were centrifuged through a discontinuous Percoll gradient (10, 20, 30, 40, and 60%) at 500g for 10 min. SMGCs were concentrated at the interface between the 40 and 60% layers, removed, washed, and resuspended in complete PC-1 medium containing 2 mM GlutaMAX (Invitrogen, Carlsbad, CA), serum substitutes (Lonza Walkersville, Inc., Walkersville, MD), and antibiotics. Cells were plated on glass coverslips (VWR, West Chester, PA) coated with poly-l-lysine (P-4707; Sigma-Aldrich) for patch-clamp and Ca2+ measurement studies. Cells stored at 4°C in complete PC-1 medium overnight showed no differences in electrophysiological responses to those of freshly isolated cells.
Whole-Cell K+ and Cl− Current Measurement.
Patch-clamp experiments were performed at 24 to 25°C in the whole-cell recording configuration (Liu and Farley, 2007). Pipettes had resistances of 3 to 5 MΩ when filled with internal solution. For whole-cell K+ and Cl− current recording, voltage ramps were applied every 2 s from inter-ramp holding potentials (−40 mV) to cover the range of −100 to +40 mV over 100 ms. Normal external solution consisted of 140 mM NaCl, 6 mM KCl, 1.13 mM MgCl2, 10 mM glucose, 1 mM CaCl2, and 10 mM HEPES, pH 7.4. In Ca2+-free external solution, 1 mM CaCl2 was replaced with 1 mM MgCl2 and 0.5 mM EGTA. Normal internal solution consisted of 140 mM KCl, 8 mM NaCl, 1.13 mM MgCl2, 0.05 mM EGTA, 10 mM HEPES, 5 mM MgATP, and 1 mM Na3GTP, pH 7.2. K+ and Cl− currents were measured at 0 and −80 mV, respectively.
Whole-Cell Nonselective Cation Current Measurement.
External solution contained 140 mM NaCl, 6 mM CsCl, 1 mM CaCl2, 1 mM MgCl2, 10 mM glucose, and 10 mM HEPES, pH 7.4, by NaOH. Ca2+-free external solution was used in which CaCl2 was replaced with EGTA, and divalent-free external solution was also used in which CaCl2 and MgCl2 were replaced with 2 mM hydroxyethyl ethylenediamine triacetic acid to support large monovalent Na+ currents. Internal solution contained 140 mM CS-methane-sulfonic acid, 8 mM NaCl, 1 mM MgCl2, 10 mM EGTA, 10 mM HEPES, and 1 mM Na2ATP, pH 7.2, by CsOH; in some experiments, NaCl was replaced with 5 mM CaCl2. The calculated free Ca2+ concentration in this solution was estimated to be 140 nM using MAXCHELATOR software (http://www.stanford.edu/∼cpatton/maxc.html). After forming the whole-cell configuration, 50-ms voltage ramps (from −100 to +60 mV) were applied every 2 s from 0-mV holding potential. The calculated equilibrium potential for Cl− is −69 mV, the potential at which NSCC current was recorded.
Whole-cell currents were acquired using an Axopatch 200B amplifier, and the data acquisition software used for all patch-clamp experiments was Axon pClamp, version 10 (Axon Instruments, Sunnyvale, CA). Thapsigargin (2 μM) and other compounds were applied in the external solution via an MPS-2 multichannel perfusion system (WPI, Sarasota, FL).
Ratiometric fura-2/acetoxymethyl ester dye was used to measure calcium fluorescence signals in a single SMGC. For Ca2+ measurement, excitation illumination with wavelengths of 340 and 380 nm was produced by a Xenon 75 W lamp (Chiu Technical Corporation, Kings Park, NY) with a dichromatic reflection lens. Intracellular Ca2+ was measured as the ratio of fura-2 fluorescence emitted during alternating excitation at 340- and 380-nm wavelengths using a Lambda 10-2 filter wheel (Sutter Instrument Company, Novato, CA) outfitted with bandpass filters for each wavelength (Semrock, Inc., Rochester, NY). Average intensities at 340- and 380-nm excitation for regions of interest in each cell were measured at a frequency of 1 to 5 Hz. Images were corrected for background fluorescence.
The cells plated on the No. 0 glass coverslips were washed with standard external solution and loaded with 2 μM fura-2/acetoxymethyl ester for 30 min at room temperature. Coverslips were mounted in a chamber on the stage of an inverted microscope (Nikon Diaphot; Tokyo, Japan). A Nikon 60× UV oil-immersion objective lens was used to visualize the cells. Fluorescence was captured using an intensifier (MTI GENIISYS; Dage-MTI, Michigan City, IN) and a CCD camera (MTI CCD72). The shutter and gain of the intensifier were set to fixed values giving a response (e.g., 1 and 200, respectively). Images obtained at 340 and 380 nm were recorded on a personal computer and acquired at one to five frames per second through a Data Translation 3155 board (Data Translation, Inc., Marlboro, MA). Data were analyzed using Imaging Workbench 5 (INDEC BioSystems, Santa Clara, CA). Averages of fluorescence intensity in regions of interest defined in each image provided an estimate of intracellular calcium concentration.
Mucus Collection and Enzyme-Linked Immunosorbent Assay.
Mucus sample collecting was similar to the previous protocol used in our lab (Dwyer and Farley, 1997). Five million SMGCs were loaded onto Durapore membrane filters (5-μm pore size; Millipore Corporation, Billerica, MA) housed in 25-mm Swinney filter holders (Millipore). The filter holders were continuously perfused with physiological HEPES saline (standard external solution) that contained 140 mM NaCl, 6 mM KCl, 1 mM MgCl2, 10 mM glucose, 1 mM CaCl2, and 10 mM HEPES, pH 7.4, for 30 min for stabilization. Baseline control and treatment samples were obtained after this stabilization period. Samples were collected every 5 min. All solutions were maintained at 37°C and continuously equilibrated with 100% O2.
Enzyme-linked immunosorbent assay was performed as described previously by Phillips et al. (2006). Porcine stomach mucin (Sigma-Aldrich) was used as a standard starting from a concentration of 100 μg/ml and then serially diluted to 1:1024. Samples were added to 96-well microtiter plates (Corning Life Sciences) in duplicate (100 μl in every well). The plates were incubated at room temperature for 90 min and then washed three times with phosphate-buffered saline containing 0.1% Tween 20 (Acros Organics, Fairlawn, NJ). Sea Block blocking buffer (Pierce, Rockford, IL) was then added, and the plates were incubated for 60 min (200 μl in every well), followed by three washes as described above. Next, 45M1 (MUC5AC) antibody (Sigma-Aldrich) at a 1:100 dilution in Sea Block blocking buffer was added, and the plates were incubated for 30 min (100 μl in every well) and washed again. Afterward, 100 μl of polyclonal goat anti-mouse IgG (diluted 1:5000 in blocking buffer; Zymed Laboratories, South San Francisco, CA) was added to each well for a 30-min incubation followed by three washes. After the plates were incubated for 30 min with streptavidin-horseradish peroxidase conjugate (Pierce) and three washes, 100 μl of tetramethylbenzidine reagent with peroxide (Pierce) was added to each well and incubated for 15 min. The reaction was stopped by the addition of 50 μl of 2 M sulfuric acid to each well. The optical density was read at a wavelength of 450 nm using a SpectraMax 190 microplate reader (Molecular Devices, Sunnyvale, CA). Mucin release is expressed as picograms of protein per minute per cell, calculated by using a standard curve for each 96-well plate. Stimulated mucin release was normalized for comparison by calculating the ratio of stimulated release for cells on each filter to the average basal release for the 5-min collection period immediately preceding stimulation of the same cells with secretagogue.
Data are shown as means ± S.E.M. Differences were examined for statistical significance using one-way ANOVA for more than two groups plus a post hoc test (Fisher's least significant difference) and a Student's t test between two groups. A paired t test or a one-way repeated-measures ANOVA was used to compare different treatments in the same group. Cells from at least three animals per experimental protocol were used. Statistical significance was defined as p < 0.05.
EM Induced Mucin Release but Inhibited ACh-Induced Release from SMGCs.
As described under Materials and Methods, mucin release from freshly isolated SMGCs was measured from cells continuously superfused with oxygenated solutions at 37°C. Baseline samples were collected only after a 30-min equilibration period to stabilize basal mucin release. Mucin release is expressed as picograms of protein per minute per cell, calculated based on a standard curve. ACh (1 and 10 μM) induced a significant increase in mucin release above the baseline (0.42 ± 0.03 and 0.38 ± 0.03 pg/min per cell compared with basal release of 0.07 ± 0.018 and 0.05 ± 0.003 pg/min per cell) (Fig. 1A). This represents 6.4- ± 0.9-fold and a 7.4- ± 0.1-fold increases relative to baseline secretion, respectively. EM at 10 μM, but not at 1 μM, induced a significant increase in mucin release. In the presence of 1 μM EM, the release was 0.1 ± 0.03 pg/min per cell; after exposure to 10 μM EM, the release was 0.18 ± 0.06 pg/min per cell (Fig. 1B). When expressed as fold changes above baseline, the release values are 1.9- ± 0.8-fold (p > 0.05) and 2.5- ± 0.7-fold (p < 0.05), respectively (Fig. 1C). After 10 min of pretreatment with EM (1 or 10 μM), ACh did not induce significant release above baseline (0.11 ± 0.02 and 0.11 ± 0.03 pg/min per cell, respectively) (Fig. 1B). The bars for each treatment group in Fig. 1 represent mucin release during five consecutive minicollection periods. We then further examined the agonist action by recording ionic currents from isolated SMGCs using the patch clamp.
EM and AZM Activated Muscarinic Receptors in SMGCs.
Upon whole-cell formation, voltage-ramp protocols as described in previous studies in our lab (Liu and Farley, 2007) were used to detect ACh-, EM-, or AZM-induced whole-cell KCa and ClCa currents. KCa and ClCa currents were recorded at 0 and −80 mV during each ramp (equilibrium potentials for ClCa and KCa, respectively) at an inter-ramp holding potential of −40 mV. Similar to what Liu and Farley (2007) found, ACh elicits both KCa and ClCa, as shown in Fig. 2A, left that rise to peaks with slowly declining tail. The responses to ACh are blocked by atropine (Fig. 2B, left). As shown in Fig. 2A, 10 μM EM (middle) and 10 μM AZM (right) induced transient peaks of KCa and ClCa currents, followed by slower declining tails. Pretreatment of cells with 10 μM atropine abolished the KCa and ClCa currents induced by both EM (Fig. 2B, middle) and AZM (Fig. 2B, right). The results in Fig. 2 are representative of three to six cells in each group.
To compare the time course of the currents induced by EM, AZM, and ACh, half-times of decay from the initial peak response were compared between ACh-, EM-, and AZM-induced KCa and ClCa currents (Fig. 3A). As shown in Fig. 3C, left, the half-times of the decay of IK induced by EM (10 μM, t1/2 = 16 ± 4 s) or AZM (10 μM, t1/2 = 16 ± 10 s) are significantly faster than for IK induced by ACh (10 μM, t1/2 = 44.3 ± 7 s; p < 0.05). Likewise, Fig. 3C, right shows that the half-times for decay of peak ICl induced by EM (10 μM, t1/2 = 16 ± 4 s) or AZM (10 μM, t1/2 = 13 ± 8 s) are significantly faster than for ICl induced by ACh (10 μM, t1/2 = 85 ± 23 s; p < 0.05). The rates of decay were estimated to be two to seven times faster for IK and ICl induced by macrolides compared with similar amplitude currents induced by ACh (p > 0.05; Fig. 3D).
Effects of EM and AZM on IP3-Induced KCa and ClCa Currents.
To examine actions of macrolides on SMGC responses independent of muscarinic receptor activation, we blocked muscarinic receptors with atropine and determined the effect of EM and AZM on the currents elicited by 30 μM IP3 applied intracellularly to SMGC via the patch pipette. As shown in Fig. 4A, upon whole-cell formation, IP3 induced transient increases in both KCa and ClCa currents that peaked within 4 s (two samples) and then declined from the peak to or near baseline levels over the next ∼200 s. The decay of KCa and ClCa currents in all treatment groups was fit to a single exponential decay model by using the function y = A1 × exp(−x/t) + y0. Neither the peak currents (p > 0.05; Fig. 4B) nor the time course of the currents (Fig. 4C) were altered by the exposure of cells to 10 μM atropine for 5 min before and during the whole-cell formation (atropine: IK, t = 88.3 ± 6.4 s ICl, t = 116.6 ± 17.6 s; control: IK, t = 82.6 ± 7.4 s and ICl, t = 89.9 ± 10.5 s) (p > 0.05; Fig. 4C). As shown in Fig. 4C, there is good correspondence in the rates of decay of the currents with or without atropine exposure. In other groups of cells, extracellular 10 μM EM or 10 μM AZM was applied in the presence of 10 μM atropine for 5 min before whole-cell formation, and the effects on IP3-induced currents were examined. Peak current amplitudes of both KCa and ClCa induced by IP3 were not reduced by AZM or EM in the presence of atropine (p > 0.05; Fig. 4B). However, as shown in Fig. 4C, the decay phases of IK were accelerated in the presence of EM (t = 29.5 ± 1.2 s) and AZM (t = 23.3 ± 1.1 s) (p < 0.05, compared with atropine alone), as were those of ICl in the presence of EM (t = 38.4 ± 2.1 s) and AZM (t = 53.6 ± 6.4 s) (p < 0.05, compared with atropine alone). To further examine nonmuscarinic receptor actions of macrolides, we circumvented receptor stimulation by directly applying 10 μM AZM intracellularly via the patch pipette.
Effects of Internal Application of AZM on ACh-Induced KCa and ClCa Currents.
The effects of AZM on Ca2+-dependent ACh-induced currents were measured using the voltage-ramp protocol as described above. SMGCs were dialyzed with normal internal solution or internal solution containing 10 μM AZM (Fig. 5A) for 5 min before application of 10 μM ACh. As shown in Fig. 5A, left, with normal internal solution in the pipette, ACh induces a transient increase in IK and ICl that peaks within a few seconds and then decays to near baseline levels in ∼200 s, as noted in Fig. 2 (inter-ramp holding potential of −40 mV). In the presence of 10 μM AZM applied intracellularly, there is no significant difference in ACh-induced peak KCa and ClCa currents from control (Fig. 5B; p > 0.05). As shown in Fig. 5C, the half-times of decay of both KCa (Fig. 5C, left) and ClCa (Fig. 5C, right) currents were significantly accelerated by 10 μM AZM. The half-times for decay were as follows: IK, t1/2 = 7.6 ± 0.8 s and ICl, t1/2 = 22.6 ± 10 for ACh-elicited currents in the presence of 10 μM AZM; and IK, T1/2 = 45.5 ± 9.2 s and ICl, t1/2 = 77 ± 16.1 s (p < 0.05) for currents evoked by 10 μM ACh alone. Calcium entry is known to be important during the decay phase in response to ACh (Liu and Farley, 2007). We further examined the effect of EM and AZM on TG-induced calcium entry.
EM and AZM Inhibit Thapsigargin-Induced Ca2+ Entry.
Thapsigargin (TG), an irreversible sarco/endoplasmic reticulum Ca2+-ATPase pump blocker, was used to deplete intracellular Ca2+ stores and evoke what has been termed store-operated Ca2+ entry. Freshly dissociated cells were loaded with fura-2 as described. The cells were then treated with 1 μM TG for 5 min in Ca2+-free external solution to deplete internal Ca2+ stores. A sustained increase in intracellular Ca2+ influx was evoked when external Ca2+ was restored to 2 mM (Fig. 6A). The increase was due to Ca2+ influx because the removal of calcium causes a rapid, reversible decline in intracellular Ca2+ to near baseline levels within 2 to 3 min after Ca2+ removal (Fig. 6A). EM (1 and 10 μM; Fig. 6B) and AZM (0.1, 1, and 10 μM; Fig. 6C) inhibited TG-induced Ca2+ influx in a concentration-dependent manner. After 28 min of exposure to EM (1 and 10 μM, Fig. 6B), [Ca2+]i is reduced to 76.5 ± 7 and 57.1 ± 5%, respectively (Fig. 6D; p < 0.05; compared with pretreatment status). After 28 min of exposure to AZM (0.1, 1, and 10 μM, Fig. 6C), [Ca2+]i is reduced to 80 ± 5, 65.6 ± 3, and 65.1 ± 6%, respectively (Fig. 6E; p < 0.05; compared with pretreatment status). The rates of decline in the intracellular Ca2+ are clearly much slower than those observed for the removal of extracellular calcium.
In another series of experiments, TG in Ca2+-free solution was applied to SMGCs and replaced 5 min later by Ca2+-free solution containing 10 μM atropine alone or 10 μM atropine plus 10 μM EM or AZM. Atropine was included to eliminate muscarinic receptor activation by the macrolides. Five minutes later, 2 mM calcium was introduced into the above solutions, and the magnitude of TG-induced calcium entry was measured as described above. As shown in Fig. 7, A and D, atropine had no effect on the magnitude of the increase in calcium induced by TG when calcium was restored compared with control data [atropine group: ΔF340/380 = 0.83 ± 0.08, p > 0.05; control group (see Fig. 6A): ΔF340/380 = 0.82 ± 0.14]. Five-minute pre-exposure to either 10 μM EM (Fig. 7B) or 10 μM AZM (Fig. 7C) significantly reduced the increase in intracellular calcium induced upon reintroduction of extracellular calcium (ΔF340/380 = 0.28 ± 0.03 and ΔF340/380 = 0.21 ± 0.04, respectively; p < 0.05) compared with atropine alone (control), as summarized in Fig. 7D.
EM Inhibits Store Depletion-Activated Current.
The effects of EM on store depletion-induced nonselective cation current were also measured. TG (2 μM) was applied in the external solution and EGTA (10 mM), and 140 nM free Ca2+ was included in internal solution to deplete calcium stores during whole-cell voltage-clamp experiments to induce NSCC. The magnitude of the NSCC current was 49 ± 11 pA when divalent-free external solution was used (Fig. 8A; n = 5). EM (10 μM) applied in the divalent-free external solution causes a 32 ± 4% decrease in NSCC (p < 0.05, n = 5; Fig. 8B). Reintroduction of 1 mM Ca2+ and 1 mM Mg2+ into the external solution reduced this current to near baseline levels (Fig. 8A). The effects of the NSCC blockers 1-[β-(3-(4-methoxyphenyl)propoxy)-4-methoxyphenethyl]-1H-imidazole hydrochloride (SKF96365), 2-aminoethoxydiphenyl borate (2-APB), Gd3+, and La3+ on TG-induced current are shown in Fig. 8B. All NSCC blockers tested significantly blocked the current induced by TG to the same extent (p < 0.05 compared with the current amplitude before drug exposure, one-way ANOVA).
Macrolide antibiotics reduce mucus hypersecretion and improve lung function in the treatment of cystic fibrosis and panbrochiolitis (Crosbie and Woodhead, 2009) by an effect unrelated to their antibiotic properties. Several mechanisms have been proposed for inhibition of mucus secretion, including 1) inhibition of actions meditaed by neutrophil elastase (Gorrini et al., 2001), a known mucus secretagogue (Dwyer and Farley, 2000); 2) anti-inflammatory actions of macrolides (Targowski and Jahnz-Rózyk, 2008); and 3) inhibition of release directly at the gland cell (Kaneko et al., 2003; Imamura et al., 2004). The purpose of this study was to elucidate a mechanism for the direct effects of macrolides on mucus secretion and ion channels from swine SMGCs. We hypothesized that macrolides interfere with calcium influx into cells to reduce secretion.
EM was shown to be a motilin receptor agonist, a property responsible for gastrointestinal side effects (Yurkowski and Calis, 1992). Irokawa et al. (1999) demonstrate that EM is a muscarinic receptor agonist; we confirm here that EM acts as an agonist at muscarinic receptors, eliciting IK, ICl, and mucin release. However, the sensitivity of swine gland cells to EM muscarinic actions is ∼1000 times greater than estimated for gland cells isolated from the cat (Irokawa et al., 1999). The reason for this difference is unknown, but it may be related to the species or the specific cell type used. EM (10 μM) increased mucin release modestly but induced peak KCa and ClCa currents similar in amplitude to those induced by a maximal concentration of ACh. In addition, EM inhibited ACh-induced mucin release. Therefore, we suggest EM is a partial agonist with respect to mucus release but a full agonist with respect to induction of peak currents.
Inhibition of Mucus Release.
EM slightly stimulated mucin release but inhibited mucin secretion induced by ACh (Fig. 1) at EM concentrations consistent with the estimated micromolar therapeutic tissue levels of the drugs (Jain and Danziger, 2004). These concentrations are similar to or lower than the concentrations used by others to block mucin release in vitro (Goswami et al., 1990; Kaneko et al., 2003). If EM and AZM are partial agonists at muscarinic receptors, this may explain, in part, the inhibition of ACh-induced mucus release by competitive antagonism or desensitization. However, the actions of EM and AZM are unlikely to be solely through direct receptor effects alone because mucus release induced by ACh (Fig. 1), lipopolysaccharide, ovalbumin, RV14, and N-(3-oxododecanoyl)homoserine lactone has been shown to be inhibited by macrolides (Shimizu et al., 2003; Imamura et al., 2004; Inoue et al., 2008). The general nature of the inhibition suggests that there are effects of macrolides on common pathways involved in mucus release. Our data suggest a common pathway involved is inhibition of calcium influx. Secretagogue-induced mucin release occurs through exocytosis of preformed vesicles, a process requiring calcium influx (Evans and Koo, 2009).
Nonreceptor Inhibition of SMGC Function.
Both ClCa and KCa currents were used as reporters of intracellular calcium levels as previously shown to be valid (Liu and Farley, 2007). When muscarinic receptor activation by macrolides is circumvented by blocking muscarinic receptors with atropine or applying the macrolides intracellularly, EM and AZM accelerate the decay phase of KCa and ClCa currents without altering the peak responses to IP3 (atropine protocol; Fig. 4) or ACh (internal application protocol; Fig. 5). The decay phase of the currents evoked by muscarinic receptor activation was shown to partially depend on the influx of calcium (Liu and Farley, 2007). In addition, the falling phase of currents evoked by AZM and EM was significantly faster than ACh-induced currents of similar amplitude, a condition that might be predicted to occur if the macrolides both activate muscarinic receptors and modify a downstream-signaling event in the transduction process. We propose that this signaling event downstream of receptor activation that is inhibited is calcium influx. Using A549 cells (a human epithelial-like cell line), Zhao et al. (2000) conclude that EM blocked ATP-induced calcium influx by inhibiting P2X receptor channels. Kondo et al. (1998) show that 100 μM EM applied before ATP inhibited ATP-induced steady-state increases in intracellular calcium in bovine epithelial cells but not the initial IP3-induced calcium transient, a result similar to our findings. Thus we suggest that EM and AZM inhibit calcium influx linked to muscarinic receptor activation and/or release of calcium from intracellular stores in addition to being muscarinic agonists.
Calcium Influx Pathway.
Secretory epithelia in general are nonexcitable and do not express voltage-gated calcium channels (Petersen, 2003). However, exocrine glands such as pancreatic acinar cells (Petersen, 2003), salivary gland cells (Ambudkar, 2000), and sweat glands (Ko et al., 1999) express calcium channels that are activated by second messengers or by depletion of intracellular calcium stores. We report here that calcium entry occurs after muscarinic receptor activation, intracellular IP3 application, or TG-mediated intracellular calcium store depletion. Liu and Farley (2007) demonstrate that intracellular application of IP3 induced an influx of calcium, presumably linked to release of calcium from internal stores. Kondo et al. (1998) report a sustained calcium influx in bovine airway epithelial cells induced by P2u receptor activation. We propose that calcium entry pathways in swine SMGCs are activated by muscarinic receptors via protein kinase C activation and/or the subsequent IP3-mediated calcium release from internal stores. Canonical transient receptor potential channels (TRPCs) and calcium release-activated calcium channels (e.g., Orai1) have been implicated in store-operated calcium entry (SOC) (Lee et al., 2010) and via store-independent processes linked to G protein-coupled processes (ROCs) (Nishida et al., 2009). Both pathways may possibly be operative in SMGCs because receptor activation and store depletion using TG can activate calcium influx.
As shown in Fig. 8, the nonselective SOC and ROC inhibitors SKF96365, 2-APB, Gd3+, and La3+ (Leung et al., 1996; Bootman et al., 2002) all inhibit TG-induced nonselective cation current. Liu and Farley (2007) also demonstrate that SKF96365 and 2-APB inhibit short-circuit currents induced by ACh across SMGC monolayers. These findings are consistent with receptor-operated and/or store-operated calcium entry in swine SMGCs but do not permit further channel identification.
Mechanism of EM and AZM Inhibition of Calcium Entry.
Kondo et al. (1998) propose that the 14,15-membered ring macrolides (e.g., EM, AZM) may act like another macrolide, FK506 (tacrolimus), an immunosuppressant drug known to inhibit calcineurin, a phosphatase, by binding to FK506-binding protein (FKBP, immunophyllin). This inhibition ultimately blocks activation of T lymphocytes (Morita et al., 2006). It has also been reported that FKBP12, one subtype of FKBP, associates with TRPC3, TRPC6, and TRPC7; moreover, FKBP52 associates with TRPC1, TRPC4, and TRPC5 (Sinkins et al., 2004). Sinkins et al. (2004) demonstrate that FKBP is involved in TRPC6 activation. It is intriguing to speculate that EM and AZM may interfere with calcium entry by interacting with FKBP. TG-induced calcium entry, however, has generally been linked to store depletion-induced STIM1 (an endoplasmic reticulum protein) accumulation near the plasma membrane and activation of Orai1 channels (Putney, 2007). Our finding that EM and AZM inhibit TG-induced calcium entry is interesting and suggests possible modulation of the STIM1-Orai1 activation pathway. However, Zhao et al. (2000) found that EM did not alter TG-induced calcium entry in A549 cells after a 20-min pretreatment with 100 μM EM. This difference suggests that the pathway for depletion-activated calcium entry may differ depending on the cell type. It is now known that Orai1 and TRPC can form heteromeric channels (Ambudkar et al., 2007). Although our data do not permit us to select between activation mechanisms, the concept that macrolides inhibit one pathway to account for both receptor- and depletion-operated calcium influx is appealing.
It is interesting that EM and AZM inhibit TG-induced calcium entry slowly after the calcium entry has developed, whereas inhibition of responses to TG, ACh, and IP3 require no more than 5 min of pre-exposure to EM or AZM to be inhibited. In addition, the currents induced by EM and AZM as a result of activation of muscarinic receptors decline more rapidly than do ACh-induced currents of similar amplitude. This is possibly due to EM and AZM interfering with calcium influx during activation, or it may represent differences in receptor-binding kinetics between macrolides and ACh. These findings are consistent with AZM and EM, inhibiting the activation of the calcium influx pathway more readily than reversing the calcium influx once it has been activated. Further research into the activation pathway for calcium entry into SMGCs needs to be performed to understand these issues.
EM and AZM are partial agonists at muscarinic receptors with respect to mucin release, inducing modest release but inhibiting mucin release induced by ACh. Both macrolides inhibit calcium entry induced by muscarinic receptor activation and store depletion. Understanding the calcium entry pathway and the molecular mechanism of macrolide inhibition of this pathway should provide targets for the development of therapeutics to moderate excessive mucus secretion.
Participated in research design: Lu, Liu, and Farley.
Conducted experiments: Lu and Liu.
Performed data analysis: Lu, Liu, and Farley.
Wrote or contributed to the writing of the manuscript: Lu, Liu, and Farley.
Other: Farley acquired funding for the research.
We thank Terry Dwyer for his critical reading of the manuscript.
This work was supported by the American Heart Association [Grant 0655086B] (to J.M.F.).
Article, publication date, and citation information can be found at http://jpet.aspetjournals.org.
- intracellular calcium concentration
- submucosal mucous gland cell
- store-operated Ca2+ channel
- receptor-operated Ca2+ channel
- nonselective cation channel
- analysis of variance
- calcium-activated K+
- Ca2+-activated Cl− channel
- inositol 1,4,5-trisphosphate
- 1-[β-(3-(4-methoxyphenyl) propoxy)-4-methoxyphenethyl]-1H-imidazole hydrochloride
- 2-aminoethoxydiphenyl borate
- FK506 binding protein
- transient receptor potential channel.
- Received June 17, 2010.
- Accepted September 24, 2010.
- Copyright © 2011 by The American Society for Pharmacology and Experimental Therapeutics