Functional characterization of rat submaxillary gland muscarinic receptors using microphysiometry

Abstract

1. Muscarinic cholinoceptors (MChR) in freshly dispersed rat salivary gland (RSG) cells were characterized using microphysiometry to measure changes in acidification rates. Several non-selective and selective muscarinic antagonists were used to elucidate the nature of the subtypes mediating the response to carbachol.

2. The effects of carbachol (pEC₅₀=5.74±0.02 s.e.mean; n=53) were highly reproducible and most antagonists acted in a surmountable, reversible fashion. The following antagonist rank order, with apparent affinity constants in parentheses, was noted: 4-DAMP (8.9)=atropine (8.9)>tolterodine (8.5)>oxybutynin (7.9)>S-secoverine (7.2)>pirenzepine (6.9)>himbacine (6.8)>AQ-RA 741 (6.6)>methoctramine (5.9).

3. These studies validate the use of primary isolated RSG cells in microphysiometry for pharmacological analysis. These data are consistent with, and extend, previous studies using alternative functional methods, which reported a lack of differential receptor pharmacology between bladder and salivary gland tissue.

4. The antagonist affinity profile significantly correlated with the profile at human recombinant muscarinic M₃ and M₅ receptors. Given a lack of antagonists that discriminate between M₃ and M₅, definitive conclusion of which subtype(s) is present within RSG cells cannot be determined.

Introduction

Muscarinic cholinoceptor (MChR) antagonists have therapeutic utility in overactive bladder. However, several side effects associated with these therapies, including dry mouth and blurred vision, limit their applicability (Owens & Karram, 1998; Yarker et al., 1995; Norton et al., 1994). Consequently, the identification of novel compounds that are selective for MChR subtype(s) in urinary bladder over those mediating salivation may provide a significant advance over current therapies.

There are five recognized muscarinic receptor subtypes (M₁ to M₅), each possessing different distributions and functions (Caulfield, 1993). M₁, M₃, and M₅ receptors positively couple to phospholipase C, while muscarinic M₂ and M₄ receptors negatively couple to adenylyl cyclase. Although physiological functions have been proposed for M₁ to M₄, they are ill defined for the M₅ subtype (Eglen & Nahorski, 2000). Of the five muscarinic receptors, M₂ and M₃ are highly abundant in smooth muscle tissue (Eglen et al., 1997). Previous pharmacological studies, functional and binding, in salivary gland suggested only a role for muscarinic M₃ receptors in the regulation of salivation (Dai et al., 1991; Laniyonu et al., 1990; Moriya et al., 1999).

Rat submaxillary gland has frequently been studied for pharmacological characterization of the MChRs involved in secretion by examining the following assays: ⁸⁶rubidium efflux (Bovell et al., 1989), calcium mobilization (Hurley & Brinck, 1992; Valdez & Turner, 1991), inositol phosphates accumulation (Laniyonu et al., 1990; Abdallah et al., 1992), and inhibition of adenylyl cyclase activity (Dai et al., 1991; Laniyonu et al., 1990; Abdallah et al., 1992). While these approaches offer some insight into cellular function in rat submaxillary gland, they generally lack a robust method to determine antagonist affinity and thus characterize the muscarinic subtype. An alternative approach is to employ a functional response in intact cells by using microphysiometry. In the current studies, the Cytosensor® microphysiometer (Molecular Devices Corp.) was utilized to operationally characterize the MChR subtype(s) responsible for the changes in the total metabolic activity, in vitro, of freshly dispersed submaxillary gland cells from the rat.

Preliminary reports of these findings have been presented to Experimental Biology 98 (Meloy et al., 1998).

Methods

Microphysiometer running medium

Modified RPMI 1640 medium lacking sodium bicarbonate (Molecular Devices Corp., Sunnyvale, CA, U.S.A.) was used in these studies. Supplied powder was dissolved in Nanopure water and titrated up to pH 7.40 with 0.1 M and 1 M NaOH. The medium was sterile-filtered using Nalgene tissue culture units (1 L, 0.2 μM; Nalge Nunc International, Rochester, NY, U.S.A.).

Cell preparation and microphysiometry procedure

Male Sprague-Dawley rats (250 – 600 g) from Charles River Laboratories (Wilmington, MA, U.S.A.) were asphyxiated with CO₂ and decapitated. Submaxillary (submandibular) glandular tissue was dissected and placed on a McIlwain Tissue Chopper set at 350 μm, turned 90°, chopped, turned 90°, and chopped for a third time. Two glands per experiment were digested with 1 mg ml⁻¹ collagenase (Sigma Chemical Co., St. Louis, MO, U.S.A.) in 50 ml Ham's F-12 medium (GibcoBRL, Gaithersburg, MD, U.S.A.) at 37°C for 1 h with frequent shaking. The cells were filtered through a Cellector tissue sieve (40 mesh screen), washed twice with F-12 medium, centrifuged at 1200 r.p.m. between washes, and resuspended in 20 ml F-12 medium. The cells were then counted using a hemacytometer. Approximately 2×10⁶ cells were centrifuged and resuspended in 100 μl low-buffered RPMI (running medium), pH 7.4 plus 50 μl of melted agarose cell entrapment medium (Molecular Devices Corp., Sunnyvale, CA, U.S.A.). A 7 μl aliquot of the cell/agarose suspension was dotted onto the centre of a capsule cup (pre-chilled at 4°C for >15 min) in a transwell capsule plate (3 μm pore size; Corning Costar Corp., Cambridge, MA, U.S.A.) and allowed to air dry 10 min before addition of the spacer and RPMI buffer. For treatment with test compounds in the microphysiometer, a transwell spacer and insert (Molecular Devices Corp., Sunnyvale, CA, U.S.A.) were added to each transwell capsule, and the capsules placed in sensor chambers (one capsule per chamber). Eight sensor chambers were used per experiment. The sensor chambers were loaded onto the microphysiometer and running medium was pumped through the transwell capsules at a rate of 100 μl min⁻¹ (microphysiometer setting=50%) with a peristaltic pump. Test compound pump cycle, exposure times, recovery intervals, and acceptable baseline pH accumulation rates were empirically determined for the rat submaxillary gland cells. Total pump cycle time was 2 min, which included 53 s pump-off period, during which the acidification rate was measured for 45 s (i.e. the Get Rate). RSG cells were exposed to carbachol for 65 s beginning 55 s after pump-on began (12 s before pump-off began), and then allowed to recover during a 28 min 55 s wash interval before the next exposure to compound. RSG cells were first allowed to equilibrate with running medium on the microphysiometer for approximately 2 h, then with antagonist or running medium for an additional 17 min before exposure to carbachol.

Experimental protocol and measurement of cellular responses

Cell chambers were exposed to increasing concentrations of agonist in the presence or absence of antagonist at a fixed concentration. Agonist additions to the cells were made with valve switches during perfusion. The maximum metabolic acidification rate at each concentration was recorded and expressed as per cent change over the basal acidification rate. All measurements were made using Cytosoft (Molecular Devices Corp., Sunnyvale, CA, U.S.A.) software provided with the Cytosensor® microphysiometer (McConnell et al., 1992). Briefly, a silicon chip, located at the bottom of the sensor chamber, functioned as an extremely sensitive detector of hydrogen ions. A light emitting diode (LED) positioned below the sensor chamber allowed pH readings in the centre of the cell capsule to be made continuously throughout the pump cycle (see above), and pH readings recorded and displayed on the computer as ‘Raw Data' in millivolt (mV) units. The data collected when the pumps were off were fitted to a straight line and the slope of the line was plotted as the acidification rate (‘Rate Data' in μV s⁻¹ units). All data were recorded throughout the experiment using the Cytosoft software package. After data collection, selected segments of the acidification rates were normalized using Cytosoft.

Changes in cellular metabolic activity in response to muscarinic agonist (carbachol), in the presence or absence of antagonist, were expressed as Acidification Response (% above basal). Normalized Rate Data were exported to a spreadsheet program (Microsoft Excel) and analysed using the following equation: Acidification Response [% above basal]=(CP_rate−Baseline_rate)×(Baseline_rate×100)⁻¹. CP_rate indicates the normalized peak metabolic Rate Data point after exposure to test compound (Chosen Point) and Baseline_rate indicates the normalized average of the three Rate Data points immediately preceding exposure to test compound.

Data analysis

Rate Data were analysed to construct concentration-effect (E [A]⁻¹) curves by iterative curve fitting to a four-parameter logistic equation to determine minimum, maximum, pEC₅₀ and Hill slope (n_H) values. Kaleidagraph software was used to fit data to the general logistic function: E=basal+[E_m×A^nH]×[(A^nH+[A]₅₀^nH)]⁻¹ for agonist stimulation curves. Schild plots were constructed to study the nature of antagonism over a wide range of concentrations. pA₂ values were estimated from x-intercept, and pK_B values, where appropriate, were estimated by constraining regression line slope to one (assuming Schild slope was not significantly different from one; Jenkinson et al., 1995).

The correlation coefficient, r value (ranging from −1 to 1), reflects the degree of linear relationship between two variables. The larger the absolute r value, the stronger the linear association between the two variables. Σ (x−y)² is the sum of squares of differences in affinity estimates for each plot, and describes, in relative terms, to what degree two data sets differ.

Statistical analysis

Data are expressed as mean and standard error of the mean (s.e.mean) for each compound tested (3 – 6 independent experiments; 5 – 16 replicates).

Compounds used

Carbachol, atropine and oxybutynin were purchased from Sigma Chemical Co. (St. Louis, MO, U.S.A.). 4-DAMP (4-diphenylacetoxy-N-methylpiperidine), methoctramine, pirenzepine and p-F-HHSiD (para-fluoro-hexahydro-sila-difenidol) were purchased from RBI (Natick, MA, U.S.A.). Himbacine was kindly provided by Professor W.C. Taylor (University of Sydney, New South Wales, Australia). AQ-RA 741 (structure) was a gift from Boehringer Ingelheim Pharmaceuticals Inc., U.S.A.. Darifenacin and zamifenacin were provided by Pfizer Central Research (Kent, U.K.). Remaining compounds were synthesized in the Medicinal Chemistry Group, Neurobiology Unit, Roche Bioscience (Palo Alto, CA, U.S.A.). All stock solutions of compounds were dissolved in running medium, with the exceptions of darifenacin and zamifenacin, which were dissolved in DMSO (100% in 10⁻² M stock solution). All serial dilutions were performed using running medium.

Results

Measurement of responses

Muscarinic receptor activation in response to carbachol was assessed as changes in the acidification rate during the pump-off period of the experiment. Figure 1a – e illustrates the progression from Raw Data to Rate Data obtained in a single channel. Exposure of the cells to agonist occurred 12 s before the peristaltic pump was turned off during a single pump cycle (Figure 1a). Data were gathered during the pump-off period, when the metabolic activity of the cells acidified the extracellular medium. The drop in pH was measured as a decrease in μV over time, where 1 μV s⁻¹ was approximately 0.001 pH min⁻¹ (Figure 1b) (McConnell et al., 1992). The data gathered during the pump-off period were fitted to a straight line using a least-squares method, with the resulting slope of the line reported as the acidification rate (Figure 1c). Acidification rate data for exposure to carbachol were normalized to baseline rate data (Figure 1d) at the conclusion of the experiment, and subsequently exported to a graphing program to generate concentration-effect (E [A]⁻¹) curves (Figure 1e).

Figure 1.

Representative experiment performed using carbachol-stimulated rat submaxillary gland (RSG) cells in microphysiometer. Sample of pump cycle (a), raw data (b), rate data (c), normalized segments (d), and resulting curve (e) generated from a single sensor chamber exposed to increasing concentrations of carbachol (10⁻⁹ to 10⁻⁵ M). (f) Mean of seven individual carbachol curves from a single experiment (pEC₅₀=5.99±0.10 (s.d.), max=107±15% (s.d.)).

Figure 1f shows the mean E [A]⁻¹ curve generated with data from seven separate sensor chambers in one experiment. Carbachol produced concentration-dependent changes in cellular acidification with a potency (pEC₅₀) of 5.99±0.03 and a maximum of 107±5%.

Measurement of apparent antagonist affinities

Carbachol concentration-effect curves were constructed in the presence of increasing concentrations of muscarinic antagonists. In this manner, rightward, concentration-dependent shifts of the curves allowed for the determination of pA₂ and pK_B values (Table 1).

Table 1.

Antagonist affinity estimates (pK_B, pA₂) for muscarinic antagonists in rat submaxillary gland using microphysiometry

Figure 2 contains representative antagonist curve shifts using himbacine (0.3 – 5 μM). In the presence of antagonist, there was no reduction in basal rate seen during the antagonist equilibration phase, suggesting a lack of inverse agonism or the presence of endogenous agonist in these preparations. The presence of antagonist resulted, in most cases, in the rightward, parallel shift of the concentration-effect curves. Deviations from this occurred with darifenacin, p-F-HHSiD, and zamifenacin.

Figure 2.

Acidification response in freshly dispersed RSG cells to carbachol in the absence and presence of himbacine. Data shown are from a single representative experiment performed with single point determinations. Each curve represents a concentration-effect curve generated from a single sensor chamber (one individual determination). Individual determinations (n=5 – 16) were repeated in at least three independent experiments.

Figure 3 contains Schild plot analysis for the 12 antagonists tested (4-DAMP, AQ-RA 741, atropine, darifenacin, himbacine, methoctramine, oxybutynin, p-F-HHSiD, pirenzepine, S-secoverine, tolterodine, and zamifenacin; n=5 – 16). pK_B determinations were made for the nine compounds in which the Schild slopes did not differ significantly from one (4-DAMP, AQ-RA 741, atropine, himbacine, methoctramine, oxybutynin, pirenzepine, S-secoverine, and tolterodine). However, for darifenacin, p-F-HHSiD, and zamifenacin, the slopes were 0.47±0.21, 1.19±0.04 and 1.40±0.13, respectively, and pK_B values could not be calculated. In addition, there was significant suppression of the curve maxima in the presence of darifenacin, precluding rigorous Schild analysis. Table 1 contains the pK_B, pA₂, Schild slope and r values, where appropriate.

Figure 3.

Schild analysis for (a) 4-DAMP, (b) AQ-RA 741, (c) atropine, (d) himbacine, (e) methoctramine, (f) oxybutynin, (g) pirenzepine, (h) S-secoverine, (i) tolterodine, (j) darifenacin, (k) p-F-HHSiD, and (l) zamifenacin. Data shown are single point determinations from individual curve shifts (single chamber, n=5 – 16). The data shown on the Schild plot were fit to a straight line using a least-squares method (linear regression). The average of the individual determinations±s.e.mean are reported in Table 1.

Correlation plots (Figure 4) were constructed using affinity estimates from Table 2 to demonstrate the relationship between binding affinity estimates (pK_i) generated at human recombinant muscarinic receptors (M₁ to M₅) expressed in CHO K1 cells, and functional affinity estimates (pK_B) generated in the present study. Each plot contains the line of identity (y=x) indicated by the solid line and the SSQ value. A correlation between affinity estimates for muscarinic receptor(s) present in RSG and cloned M₁, M₃ and M₅ receptors was shown in these plots.

Figure 4.

Correlation plots showing relationship of binding affinity data (pK_i) generated at human recombinant muscarinic (a) M₁, (b) M₂, (c) M₃, (d) M₄ and (e) M₅ expressed in CHO K1 cells with functional affinity estimates (pK_B) generated in the RSG. In each plot, a solid line indicates the line of identity (y=x). SSQ=sum of squares of differences between affinity estimates (Σ(x−y)²).

Table 2.

Antagonist affinity estimates (pK_B) derived from microphysiometry studies in RSG compared with radioligand binding data (pK_i) at cloned human muscarinic M₁ – M₅ receptors

Discussion

Muscarinic receptor antagonists effectively control symptoms of overactive bladder, but their side effects, particularly dry mouth, lead to increased fluid intake and subsequent exacerbation of bladder overactivity. Elucidation of potential differences between MChR subtypes in bladder and salivary gland may enable the discovery of novel, selective therapies with reduced side effects. Although several methods have been applied in the search for such tissue differences, none have yielded conclusive results. In the present study, we have developed an in vitro assay using the Cytosensor® microphysiometer to characterize the MChR(s) in freshly dispersed rat submaxillary (submandibular) gland (RSG) cells.

Utility of microphysiometry in muscarinic RSG receptor characterization

Microphysiometry is an attractive alternative to previous methods, and we report here the first use of this technique in receptor characterization of freshly dispersed RSG cells. While the quantitative characterization of muscarinic subtypes in vitro using bladder smooth muscle is feasible, such analysis of salivary gland function has proven more challenging. In these studies, we show that microphysiometry provides a quantitative profile of cellular response to MChR antagonists and, with few exceptions, is robust and practical. Most studies of MChR characterization of salivary gland have involved measurements of such endpoints as ⁸⁶Rb efflux, calcium influx, inositol phosphates accumulation, and inhibition of adenylyl cyclase activity upon exposure of the cells to known MChR antagonists. None of these methods are either sufficiently robust or relevant to offer unambiguous pharmacological characterization that is reflective of total metabolic activity of the target cells. For example, ⁸⁶Rb efflux and calcium influx determinations measure a downstream effect of a compound on target cells, which has been proposed to reflect saliva secretion. However, ⁸⁶Rb efflux suffers from poor data fidelity due to a small signal to noise ratio, and thus it is difficult to quantitatively differentiate antagonist affinities (Bovell et al., 1989). Calcium influx-based approaches measure the endpoint of only one metabolic event in response to an effector, rather than measuring a global metabolic response. Inositol phosphates accumulation and inhibition of adenylyl cyclase activity are assays that only measure a single second-messenger, thus they represent biased functional response signals that are receptor-specific.

All techniques relying on functional analogy have drawbacks, and microphysiometry is no exception. Some difficulties relate to the physiochemical properties of the antagonist as well as its known pharmacological properties. We found that compounds such as darifenacin, p-F-HHSiD, and zamifenacin had Schild slopes significantly different from one. This finding could be due to compound accumulation within the chamber, to compound sticking to the peristaltic tubing, or to pseudo-irreversible antagonism. With darifenacin, we found a decreased E_max in agonist curves with increasing concentration of this compound. Further studies with multiple exposures to agonist after a single exposure to darifenacin (data not shown) revealed that the drug washed out of the system very slowly, displaying typical pseudo-irreversible antagonism. Thus, for these compounds, pK_B values could not be estimated and their utility in our screen was limited.

Our studies with microphysiometry also identified some technical limitations of this method that were not universal and affected a minority of responses. For example, we were not able to add agonist too close to the scheduled pump-off time because peristaltic pump speeds varied slightly among chambers. However, in studies where a resolution time of greater than 5 s can be tolerated, this effect becomes less significant. Some drawbacks could also be attributed to the software used in our analyses. For example, the optimization software used (Cytosoft) did not enable extraction of all experimental data for the entire duration of the experiment. In addition, the software did not allow us to select the greatest change in acidification during the Get Rate (i.e. greatest slope), reporting only the value of the slope for the entire Get Rate period.

Another area of note was the rate of data capture that the microphysiometer used, which could limit the quantity of usable data. For example, when carbachol (a rapidly acting agonist) was used at the higher concentrations (Figure 1e), there was a characteristic fall-off in response because the equipment was not able to capture data quickly enough. Since the response recorded for these higher concentrations was not an accurate reflection of the true response to the agonist, these data points were omitted from further analysis. Although we describe this observation with RSG cells, this type of response has been reported by others to occur with other targets, including MChRs transfected into CHO K1 cells (Baxter et al., 1994). Thus, we were compelled to select a sub-maximal stimulus level to collect linear data. This decision represents a need for compromise in data acquisition through microphysiometry, because with some agonists the Get Rate may be found to begin too soon at low concentrations of agonist and too late at high concentrations due to physical limitations in delivering compound to the chamber. Despite these noted limitations, we found microphysiometry to be a worthwhile and advantageous method in studies of this type.

Assessment of the muscarinic receptors in RSG

Concentration-effect curves to carbachol established in the absence and presence of muscarinic antagonists, resulted in monophasic, parallel, surmountable shifts (except darifenacin, p-F-HHSiD, and zamifenacin) consistent with competitive antagonism at a single site. Additionally, regression analysis revealed linear Schild slopes providing no evidence for involvement of more than one receptor subtype. Currently, there are few muscarinic antagonists that are clearly selective for one subtype over the remaining four. Therefore, it is critical to use several selective compounds to determine the subtype(s) present within a tissue. Previous studies in salivary gland tissue have typically used less than five different antagonists to characterize the functional MChR(s) present within the tissue.

In this study, we attempted to distinguish the functional MChR subtype(s) present in RSG cells. The sums of squares (SSQ) of the differences between RSG microphysiometer pK_B and binding pK_i affinity estimates were generated as a mathematical representation of the similarity (or difference) between the two affinity values. The smaller the SSQ value, the smaller the difference between RSG microphysiometer pK_B and binding pK_i data. As shown in the correlation plots, the data in the present study demonstrate the close resemblance that M₁, M₃ and M₅ binding affinity values have to pK_B values obtained for RSG microphysiometry data (Figure 4). However, since key muscarinic antagonists selective for M₁ (methoctramine pK_i=6.7; pirenzepine pK_i=8.0) displayed poor affinity in RSG cells (methoctramine pK_B=5.9, pirenzepine pK_B=6.9), the data suggest the function of either muscarinic M₃ or M₅ receptor.

Schild slopes equivalent to one (for most compounds) and the lack of any inflection point in Schild analysis suggest the presence of only one functional receptor subtype rather than a combination of subtypes. Given the current lack of pharmacological tools able to discriminate further between M₃ and M₅ receptor subtypes, it is not definitively clear which subtype functions predominantly within the rat salivary gland. Our results echoed previously reported literature values, which reported a lack of differential receptor subtype(s) between bladder (Hegde et al., 1997; Longhurst et al., 1995) and salivary gland tissue (Dai et al., 1991; Laniyonu et al., 1990; Moriya et al., 1999) as determined in functional and binding studies.

In the present study the affinity constants for the compounds studies were representative of both subtypes. This may suggest either that the nature of the receptor, and thus its pharmacology, was affected during tissue preparation or that both subtypes were detected by the microphysiometry technique. Numerous lines of data (see Caulfield, 1993, for review) indicate that M₃ receptors are expressed in rat salivary gland. In addition, data from knockout mice for the M₃ receptor (Matsui et al., 2000) show impaired salivary secretion upon stimulation with pilocarpine. Interesting, also, are preliminary data from Yeomans et al. (1999) indicating that mice with a null mutation in the M₅ gene have a reduction in pilocarpine-stimulated salivation. This may suggest a functional role of the M₅ subtype in murine tissue and supports radioligand binding data in rat tissue (Flynn et al., 1997) also demonstrating expression of the subtype.

The pharmacological differentiation between the MChR playing the predominant physiological role in bladder and salivary function remains a challenge and a prerequisite to the discovery of therapies that offer symptomatic control for bladder overactivity with fewer side effects. The current studies indicate that most likely one subtype functions within rat submaxillary gland, and correlation data suggest the MChR subtype is pharmacologically consistent with either M₃ or M₅. Because known muscarinic antagonists have affinity values that differ little between M₃ and M₅ (Buckley et al., 1989; Dörje et al., 1991; Jones et al., 1991; Watson et al., 1999; Eglen & Nahorski, 2000), new compounds that demonstrate improved selectivity for MChR subtypes may prove to be better tools for discrimination and may be optimal candidates for novel therapeutics.

Abbreviations

AQ-RA 741

11-[[4-[4-(diethylamino)butyl]-1-piperidinyl]acetyl]-5,11-dihydro-6H-pyrido(2,3-b)(1,4)-benzodiazepine-6-one

CHO K1

Chinese hamster ovary cells

4-DAMP

4-diphenylacetoxy-N-methylpiperidine

MChR

muscarinic cholinoceptor

RSG

rat submaxillary gland

p-F-HHSiD

para-fluoro-hexahydro-sila-difenidol

SSQ

sum of squares of differences (Σ (x−y)²)