Introduction

Top Abstract Introduction Methods Results Discussion References

Muscarinic receptors mediate cholinergic excitation in the distal smooth muscle esophagus. The overall contribution of this excitatory mechanism to normal esophageal peristalsis differs between species. Thus, atropine potently blocks swallow-induced peristalsis in the cat, the monkey and the human esophagus, but in the opossum, it has a more modest effect (Goyal and Paterson, 1989; Diamant, 1989). These findings are supported by in vitro studies in which nerve-mediated responses evoked by EFS of muscle strips from the cat or the human esophagus also are inhibited by atropine (Behar et al., 1989; Preiksaitis et al., 1994), whereas in the opossum, a significant atropine-resistant, noncholinergic, nonadrenergic contraction is seen (Crist et al., 1984).

Five subtypes of the muscarinic receptor (m₁-m₅) have been identified by molecular cloning techniques (Bonner et al., 1987; Peralta et al., 1987; Dorje et al., 1991). To date muscarinic receptor agonists or antagonist with sufficient selectivity to distinguish any one subtype from all the others have not been developed. However, four subtypes of the receptor (M₁-M₄) can be differentiated pharmacologically based on the pattern of selectivity for several muscarinic antagonists (Hulme et al., 1990; Caulfield, 1993; Eglen et al., 1996a; Dorje et al., 1991). Previous in vivo studies to characterize muscarinic receptor subtypes in the cat (Blank et al., 1989) and the opossum esophagus (Gilbert and Dodds, 1986) concluded that cholinergic activation accompanying peristalsis occurred mainly via a M₂-mediated mechanism. This conclusion was based in part on the lack of effect of pirenzepine, a M₁ receptor antagonist, vs. the greater selectivity of 4-DAMP, originally designated as a M₂-selective antagonist. 4-DAMP is now known to be more selective for the M₃ receptor (Eglen et al., 1996a). Hence, it is more likely that the M₃ receptor mediates peristalsis in vivo in the cat and the opossum (Goyal, 1989).

However, in a recent in vitro study on isolated smooth muscle cells from the circular layer of the cat esophagus, Sohn et al. (1993) demonstrated that methoctramine, a selective M₂ antagonist, was more effective in antagonizing acetylcholine-stimulated cell shortening than p-HHSiD, a selective M₃ antagonist, and concluded that the response was therefore M₂-mediated. In other regions of the gastrointestinal tract, receptor binding studies have shown that smooth muscle expresses both M₂ and M₃ receptors; however, the contractile response in most gastrointestinal smooth muscles is mediated mainly by the M₃ subtype (Eglen et al., 1996a). This holds true in several species and preparations, including guinea pig terminal ileum (Eglen et al., 1992b; Barocelli et al., 1994; Michel and Whiting, 1990b; Michel and Whiting, 1988b; Giraldo et al., 1988), rat terminal ileum (Lazareno and Roberts, 1989), canine terminal ileum (Shi and Sarna, 1997) and human colon (Kerr et al., 1995; Gomez et al., 1992). Thus, the studies of Sohn et al. (1993), which show that the functional response of the circular muscle from the esophageal body of the cat is mediated by the M₂ receptor, represent a noteworthy exception for gastrointestinal smooth muscles. This observation holds significant clinical importance because it implies that anticholinergic agents that have sufficient selectivity for the M₂ receptor could be used to target esophageal motor disorders, whereas undesirable M₃-mediated systemic effects could be minimized.

In the present study, we address this controversy by characterizing the muscarinic receptor subtype(s) that mediates cholinergic responses in the cat esophagus using smooth muscle strips studied in vitro. We consider separately the longitudinal and circular layers since significant differences in the distribution of muscarinic receptors between muscle layers may exist (Preiksaitis et al., 1996). Additionally, we examine the relative effectiveness of selective receptor antagonists on cholinergic nerve-mediated responses.

	Methods

Top Abstract Introduction Methods Results Discussion References

Animals and tissue retrieval. The experimental protocol was in accordance with the ethical guidelines of the Canadian Council of Animal Care and approved by the University of Western Ontario Animal Care Committee. Thirty-six cats of either sex weighing between 3.1 and 6.2 kg were euthanized with a lethal dose of phenobarbitol (100 mg/kg intraperitoneally). The abdomen and chest were opened and the esophagus was excised en bloc from the aortic arch distally to include a 2- to 3-cm portion of the proximal stomach and was placed in room-temperature Kreb's solution equilibrated with 5% CO₂ and 95% O₂. The Kreb's solution had the following composition (in mM): sodium 143, potassium 5.0, calcium 2.5, magnesium 1.2, chloride 128, phosphate 2.2, bicarbonate 24.9, sulfate 1.2, and glucose 10. In some experiments, a small full thickness biopsy (~1 by 3 mm) taken from the mid-distal third of the esophageal body muscle was obtained immediately on opening the chest. The mucosa was removed and the muscle was frozen on dry ice and stored at 70°C for subsequent RNA extraction. Samples of terminal ileum were obtained similarly, immediately after opening the abdomen. An equivalent sized portion of cardiac tissue, which included all layers, was obtained from the ventricular muscle immediately on opening the chest and used for RNA extraction as detailed below.

Tissue bath studies. The esophagus was freed of surrounding fascia, opened lengthwise and pinned to its approximate in situ dimensions. After removal of the mucosa by sharp dissection, longitudinally and circularly oriented strips of approximately the same size (0.2 × 1.0 cm) were prepared, with the aid of a magnifying glass, from the same region of the esophageal body, 1 to 3 cm above the lower esophageal sphincter muscular ring. Care was taken to ensure that the long axis of each strip followed the direction of the muscle fibers. Strips were mounted vertically in 10-ml jacketed organ baths containing Kreb's solution held at 37°C and continuously bubbled with 5% CO₂ and 95% O₂. One end of each strip was fixed to an electrode holder, and the other end was fastened by a silk tie to a Grass FT03 isometric force transducer coupled to a Grass 79E chart recorder (Grass Instruments, Quincy, MA).

8

3

RT-PCR and molecular cloning of cat m₂ and m₃ receptors. Total RNA was isolated from cat esophagus using the method of Chomczynski and Sacchi (1987). RNA samples were run out on agarose gels to verify integrity. One microgram of total RNA from each sample was reverse transcribed for 90 min at 42°C using random hexamers and Superscript RNase H- (GIBCO BRL, Gaithersburg, MD). The cDNA was diluted 2.5 times, and 5 µl was used in each 50 µl PCR reaction. PCR reactions were carried out for 35 cycles with 2.5 mM MgCl₂, 0.2 mM deoxynucleotide triphosphates, 2 µM of m₂ or m₃ primers and 0.2 µl of Taq DNA polymerase (ID Labs Biotechnology) in the reaction mixture. The timing of each cycle was 0.5 min at 94°C, 0.5 min at 58°C and 1 min at 72°C, followed by a final 7-min extension at 72°C. PCR products (13 µl) were analyzed by electrophoresis on 2% agarose gels and visualized by ethidium bromide staining. Primers were selected based on the known sequences for the human m₂ and human m₃ genes, and were purchased from GIBCO BRL. The respective upstream and downstream primers for m₂ were 5'-GGTCAGCAATGCCTCAGTTA-3' and 5'-CTTGGTGCCAATTCTGATGC-3', and for m₃ were 5'-TGATGATCGGTCTGGCTTGG-3' and 5'-TGCTGCTGTGGTCTTGGTCC-3'. The predicted PCR product sizes were 676 base pairs for m₂ and 441 base pairs for m₃. PCR products were purified using PCRapid (ID Labs Biotechnology), subcloned into the pGEM-T Vector (Promega, Madison, WI), and transformed into JM109 competent cells (Promega). Plasmid DNA was purified using the RPM kit (BIO 101). Clones containing the PCR inserts were identified by restriction digest and sequenced by the Queen's University Core Facility for Protein/DNA Chemistry (Kingston, Ontario, Canada).

Data analysis and statistics. Carbachol responses and the effects of muscarinic antagonists were analyzed using the curve-fitting facilities of Prism V 2.0 (GraphPAD Software, San Diego, CA). Using this program, EC₅₀ and IC₅₀ values were obtained from the sigmoidal dose-response relationship generated from the experimental data by nonlinear regression. The method of Arunlakshana and Schild (1959) was used to determine pA₂ values. Straight lines were fitted by least-squares linear regression. The slope of a straight line was considered to be not different from unity if the 95% confidence interval for the slope included 1. Results are reported as mean ± S.E. The number of cats studied is indicated by n. Statistical comparisons were made with the Student's t test. P < 0.05 was considered significant.

Drugs and materials. Carbachol (carbamyl choline), atropine sulfate and L-NNA were obtained from Sigma Chemical (St. Louis, MO). Methoctramine, 4-DAMP, pirenzepine and p-HHSiD were obtained from Research Biochemicals (Natick, MA). AF-DX-116 (11-[[[2-diethylamino-0-methyl]-1-piperidinyl]acetyl]-5,11-dihydrol-6H-pyridol[2,3,-b][1,4]benzodiazepine-6-one) was generously provided by Boehringer Ingleheim, Germany. Zamifenacin was a gift from Pfizer Central Research, Sandwich, UK. All drugs were prepared as concentrated stock solutions and diluted into Kreb's buffer just before use such that drugs were added to the 10 ml tissue bath in a volume of 10 to 100 µl.

	Results

Top Abstract Introduction Methods Results Discussion References

Antagonist effects on carbachol-mediated responses in circular and longitudinal muscle. Carbachol caused a concentration-dependent increase in tension in both the longitudinal and circular muscles. The mean log EC₅₀ values for the circular muscle was 6.17 ± 0.06 and 6.88 ± 0.07 for longitudinal muscle indicating a slightly greater agonist potency in the longitudinal layer (P < .001, n = 10). Since carbachol may also act at nicotinic and muscarinic receptors on parasympathetic ganglia, the net effect of this drug could be influenced by additional activation of inhibitory or excitatory nerve pathways. Neither 100 µM L-NNA (which inhibits nitric oxide synthase, thus blocking inhibitory nerve effects) nor 10 µM hexamethonium, alone or in combination, had a significant effect on the maximum tension response or the EC₅₀ for carbachol in longitudinal or circular muscle strips (table 1).

View larger version (18K): [in this window] [in a new window]   Fig. 1.   Comparison of the effects of six muscarinic receptor antagonists on carbachol-mediated contractions in circular and longitudinal muscle of the esophagus. a and b, demonstrate the effects of increasing concentrations of six muscarinic antagonists (n = 4-11) on carbachol (1 µM) contraction responses in circular and longitudinal esophageal muscle, respectively. c, Comparison of the IC50 for each antagonist in longitudinal and circular muscle layers. The resulting line has a slope of 0.90 ± 0.05 (r2 = .98), which is not significantly different from unity. Results for the following antagonists are shown: 4-DAMP (), zamifenacin (), p-F-HHSiD (), pirenzepine (×), AF-DX-116 () and methoctramine ().

View larger version (30K): [in this window] [in a new window]   Fig. 2.   Antagonism of carbachol-mediated contractions in circular (a-c) and longitudinal (d-f) muscles from the cat esophagus by methoctramine (a, d), pirenzepine (b, e) and 4-DAMP (c, f). Symbols represent the following antagonist concentrations: () none, (×) 109 M, () 108 M, () 107 M, () 106 M, () 105 M and () 104 M. Values shown represent the means for 3 to 6 muscle strips obtained from 9 cats.

View larger version (18K): [in this window] [in a new window]   Fig. 3.   Schild plots of methoctramine (, , n = 5), pirenzepine (, , n = 4) and 4-DAMP (, , n = 4) antagonism of carbachol-mediated responses in longitudinal (open symbols) and circular (closed symbols) smooth muscles from the cat esophagus. Lines shown are the best fit by least-squares linear regression. The resulting slopes and x-intercepts (pA2 values) are summarized in table 3.

Antagonist effects on EFS responses in circular muscle. The inhibition of nitric oxide synthase by 100 µM L-NNA was used to enhance and stabilize on-contractions of circular muscle as illustrated in figure 4 and previously described for human esophageal smooth muscle (Preiksaitis et al., 1994). Without L-NNA, EFS of circular muscle strips produced typical off-contractions, which followed EFS after a brief latency (fig. 4B). These contractions remained stable with repetitive stimulation and were inhibited by 10 µM atropine. A small (<5%) residual, atropine-insensitive component of the off-contraction was noted in some preparations. After the addition of L-NNA, off-contractions were suppressed and replaced by on-contractions (fig. 4A) which also were stable with repetitive stimulation and were always completely suppressed by 10 µM atropine (fig. 4B).

View larger version (23K): [in this window] [in a new window]   Fig. 4.   Effect of L-NNA and atropine on electrical field stimulation (EFS) responses in circular esophageal smooth muscle. Each tracing is from a single smooth muscle strip, with the heavy bar just below the tracing indicating the time of application of the EFS train. a, L-NNA (100 µM) caused the suppression of the off-contraction (which follows the cessation of the EFS train) and the emergence of the intra-stimulus on-type contraction. b, Repetitive EFS with 3-sec trains applied every 180 sec caused reproducible off-contractions (upper trace of b) and on-contractions in the presence of 100 µM L-NNA (lower trace of b). Both on- and off-contractions were antagonized completely by 10 µM atropine.

View larger version (21K): [in this window] [in a new window]   Fig. 5.   Effects of increasing concentrations of muscarinic antagonists on EFS-mediated responses in circular esophageal muscle. a, Effect on off-contractions (no L-NNA present). b, Effect on on-contractions (in the presence of 100 µM L-NNA). c, Comparison of IC50's for effects of muscarinic antagonists on EFS-mediated on-contractions (closed symbols and ×) and off-contractions (open symbols and +) to carbachol-mediated responses in circular esophageal muscle. Both comparisons yielded straight lines with slopes of 1.2 ± 0.2 (r2 = .89) for off-contractions and 1.3 ± 0.1 (r2 = .97) for on-contractions, neither of which differs significantly from unity. Results for the following antagonists are shown: 4-DAMP (, ), zamifenacin (, ), p-F-HHSiD (, ), pirenzepine (×, +), AF-DX-116 (, ) and methoctramine (, ). Each data point represents the mean of n = 3-8 cats.

Identification of m₂ and m₃ receptor mRNA by RT-PCR. To further explore the potential roles of M₂ and M₃ receptor subtypes in the cat esophageal smooth muscle, RT-PCR was used to identify mRNA species encoding these two receptor types in RNA isolated from esophageal muscle samples, which included both longitudinal and circular layers. Since the gene sequences for the cat muscarinic receptor subtypes were previously unknown, PCR primers based on the known human gene sequences were used. Messenger RNA for the m₃ receptor could be readily identified in 3 of 3 cats, as well as terminal ileum which served as control (fig. 6). The level of expression of m₂ mRNA was not detected in the same 3 cats despite using optimum conditions to amplify the m₂ sequence as determined for cat heart tissue, which served as the positive control for the m₂ receptor. The identity of the PCR products for both receptor subtypes was verified by sequence analysis using the m₂ product from myocardial tissue and the m₃ product from esophageal body (fig. 7). Comparison with the known sequences for the human m₂ and m₃ genes demonstrated a high degree of nucleotide sequence homology for the amplified portions cloned from cat tissues (93% and 89%, respectively). The upstream and downstream PCR primer sequences selected were identical for the cat and the human genes (fig. 7). Figure 8 shows the comparison of the corresponding amino acid sequences for the human and cat m₂ and m₃ receptors.

View larger version (74K): [in this window] [in a new window]   Fig. 6.   Identification of M2 and M3 receptor messenger RNA in esophageal smooth muscle, heart and terminal ileum by RT-PCR. Messenger RNA coding for the M3 receptor was readily identified in esophageal tissues from 3 cats and terminal ileum which served as a positive control. In contrast, mRNA for the M2 receptor was not detected in any esophageal specimen, but was readily detected in heart tissue which served as a positive control. Control lanes shown are the products of the PCR reaction without cDNA present. The ladder shows bands corresponding (top to bottom) to 2000, 1500, 1000, 700, 500, 400 and 300 kb.

View larger version (87K): [in this window] [in a new window]   Fig. 7.   Partial sequences for the cat m2 and cat m3 receptor genes.

View larger version (55K): [in this window] [in a new window]   Fig. 8.   Comparison of the known amino acid sequences for the human m2 (hm2) and m3 (hm3) receptors and the portions of the cat m2 (cm2) and m3 (cm3) receptors amplified by RT-PCR in the present study. The approximate positions of the seven transmembrane-spanning regions (MI-MVII) are shown in bold type above the amino acid sequences. The portions of the human sequences corresponding to the RT-PCR primer sets used are shown in bold type.

	Discussion

Top Abstract Introduction Methods Results Discussion References

Despite the large number of muscarinic receptor agonists and antagonists available, none has sufficient selectivity for one receptor subtype over all others to allow the unequivocal pairing of receptor type and pharmacological response. However the order and pattern of selectivities of a group of antagonists for the receptor can provide sufficient grounds for the classification of a given response (Eglen et al., 1996a; Caulfield, 1993; Hulme et al., 1990). By applying this strategy, we were able to characterize the muscarinic receptor subtype in cat esophageal smooth muscle. Our findings can be summarized as follows: (1) Carbachol-induced contraction of smooth muscle in both the longitudinal and circular layer of the esophageal body demonstrates a pattern of antagonist selectivity best represented by the M₃ receptor subtype. (2) The pattern of antagonist selectivity for inhibiting both on- and off-type intrinsic nerve-mediated contractions also is characteristic of the M₃ receptor subtype. Thus, both nerve-mediated and carbachol-induced contractions of the cat esophageal body muscles result through activation of M₃ receptors. (3) These observations are further corroborated by RT-PCR data, which show that mRNA encoding the m₃ receptor is readily detected in the cat esophageal body smooth muscle, whereas mRNA for the m₂ receptor is not.

Both the longitudinal and circular muscles of the esophagus showed similar relative antagonist selectivity for carbachol-mediated contractions. The high selectivity of 4-DAMP, intermediate selectivity of pirenzepine, and low selectivity of methoctramine in antagonizing carbachol-mediated responses in both muscle layers constitute compelling evidence that contraction is mediated by M₃ receptors. The pA₂ values obtained are similar to those found in functional studies of the guinea pig terminal ileum (Eltze and Figala, 1988; Eglen et al., 1992a; Eglen and Harris, 1993), a typical M₃-mediated response, and for M₃-mediated contractile responses in other smooth muscle types (Shi and Sarna, 1997; Eglen et al., 1996a; Caulfield, 1993). Furthermore, the pA₂ values for these three antagonists correspond to the log K_i determined by radioligand binding studies on the M₃ receptor in native tissues (Hulme et al., 1990) and using heterologous expression of the human M₃ receptor gene product (Eglen et al., 1996a). In both muscle layers, these three drugs showed competitive antagonism of the carbachol responses, producing parallel rightward shifts in the dose response curves, and linear Schild plots with slopes not significantly different from unity. Since 4-DAMP has poor selectivity for M₃ over M₁ receptors, and only a ~9- to 10-fold selectivity ratio of M₃ over M₂ receptors (Dorje et al., 1991), our conclusion depends substantially on the observation that methoctramine (M₂/M₄ selective) poorly antagonized carbachol responses, whereas pirenzepine (M₁ selective) had an intermediate effect. The pA₂ value we found for methoctramine is too low to consider either a M₂- or M₄-mediated response (Eglen et al., 1996a). Schild slopes for methoctramine that deviate significantly from unity have been observed in some preparations, suggesting inadequate tissue equilibration time (Barocelli et al., 1993) or a noncompetitive, allosteric effect of this antagonist (Eglen et al., 1988). In the present study, exposure to methoctramine for 2 hr did not result in any difference in antagonist effectiveness compared to 30 min. Thus, we found no evidence for disequilibrium or allosteric effects of methoctramine in the present study.

The pA₂ value we report for pirenzepine in the longitudinal muscle (7.26) is significantly greater than that found in the circular muscle (6.79) and is slightly greater than might be expected based on other M₃-mediated responses (6.7-7.1, Caufield, 1993). Nonetheless, our conclusion that the M₃ receptor mediates this response in the longitudinal muscle is valid for several reasons: (1) The pA₂ value we calculate for all three antagonists fits the pattern expected for a M₃-mediated response better than for any other subtype (Eglen et al., 1996a; Caulfield, 1993). (2) A pA₂ value for pirenzepine in the range of 8.1 to 8.5 would be expected for a M₁-mediated response (Caulfield, 1993). (3) A similar pA₂ value of 7.23 was reported for the M₃-mediated contraction of human colonic smooth muscle (Kerr et al., 1995). (4) Finally, M₁ receptors are present on enteric ganglia, but have not been localized to gastrointestinal smooth muscle cells (Goyal, 1989; Eglen et al., 1996a). Investigation of the possible mechanisms underlying the difference in the effectiveness of pirenzepine or the potency of carbachol in longitudinal and circular muscles is beyond the scope of this study.

The good linear correlation between the log IC₅₀ values for the six muscarinic antagonists against 1 µM carbachol-mediated contractions and both on- and off-type nerve-mediated contractions in the circular muscle layer leads us to conclude that nerve-mediated responses in this tissue also are mediated predominantly by M₃ receptors. On- and off-contractions occur via different mechanisms, the latter being dependent on the action of noncholinergic, nonadrenergic inhibitory nerves, the main mediator being nitric oxide (Preiksaitis et al., 1994; Murray et al., 1991). The mechanism by which cholinergic and noncholinergic nerves interact to bring about the off-contraction is unknown. Differences in the mechanism of on- and off-contractions might include the amounts of acetylcholine released. We speculate that these factors could explain why the correlation between antagonist effects on carbachol-responses and both types of nerve-mediated contractions results in distinct parallel lines, with slopes not different from unity. In both cases, pirenzepine appears to have a greater effect on nerve-mediated responses than carbachol contractions possibly due to a selective interaction of pirenzepine with M₁ receptors present on enteric ganglia, which may additionally modulate the nerve-mediated responses (Gilbert et al., 1984). Our conclusion that nerve-mediated (acetylcholine) contractions and carbachol responses are due mainly to activation of M₃ receptors is based on the assumption that acetylcholine and carbachol display similar selectivity for muscarinic receptor subtypes. In the absence of evidence to the contrary, this assumption can be justified since carbachol is a close analogue of acetylcholine and the minor structural difference does not involve the key site for interaction with the receptor (Hulme et al., 1990).

Zamifenacin and p-F-HHSiD, two additional M₃-selective antagonists, also were similarly effective in inhibiting nerve- and carbachol-mediated contractions. Schild analysis of these antagonists was not carried out in the present study, however both antagonists were less effective inhibitors than could be anticipated based on previous studies on other tissues (Eglen et al., 1990; Watson et al., 1995; Barlow et al., 1995; Feifel et al., 1990). In the present study, the IC₅₀ value obtained for the carbachol response in circular muscle was 100-fold less for p-F-HHSiD compared with 4-DAMP. Low pA₂ values for p-F-HHSiD previously have been noted by others and this appears to be dependent on the preparation used: the pA₂ values for p-F-HHSiD and 4-DAMP differ by ~100-fold in the guinea pig trachea (Eglen et al., 1990). A recent study which demonstrated that carbachol responses in human colonic muscle showed an antagonist profile most consistent with the M₃ receptor, found a 45- to 300-fold difference in the pA₂ values for p-F-HHSiD and 4-DAMP in the longitudinal and circular muscle layers, respectively (Kerr et al., 1995). Similarly, a range of pA₂ values has been observed for zamifenacin antagonism of M₃ responses depending on the smooth muscle type studied (Watson et al., 1995). The pA₂ for zamifenacin in guinea pig terminal ileum was 9.3, similar to the pA₂ for 4-DAMP in the same preparation, but ~50-fold less effective for guinea pig urinary bladder (pA₂ = 7.6). In the present experiments, zamifenacin was ~20-fold less effective than 4-DAMP, indicating a more modest antagonist effect in cat esophageal smooth muscle.

In vivo, pirenzepine has little effect on esophageal peristalsis in animal models (Gilbert and Dodds, 1986; Blank et al., 1989). In contrast, 4-DAMP completely blocks peristalsis in the cat and significantly decreases contraction amplitude in the opossum (Gilbert and Dodds, 1986; Blank et al., 1989). This species difference could be anticipated, since an atropine-resistant contribution to the off-contraction accounts for ~60% in the distal esophagus of the opossum (Crist et al., 1984), while previous studies (Behar et al., 1989; Leander et al., 1982) and our findings indicate that the off-contraction in the cat is highly sensitive to atropine. In the opossum lower esophageal sphincter, 4-DAMP potently inhibited agonist-mediated in vivo contraction (Gilbert et al., 1984). These earlier studies were interpreted to indicate that contractions in circular muscle of the esophagus and lower esophageal sphincter are M₂-mediated. Since it is now recognized that 4-DAMP potently antagonizes the M₃ receptor, it is more likely that the receptor characterized in the above reports is the M₃ subtype as demonstrated here. The present study and those cited above concern the muscularis propria of the esophagus. The M₃ muscarinic receptor subtype also mediates cholinergic responses in the musclularis mucosae of the esophagus in the rat, guinea pig, and rabbit (Hatakeyama et al., 1995; Thomas and Ehlert, 1996; Eglen et al., 1996b).

In studies on smooth muscle cells isolated from the circular layer of the muscularis propria of the cat esophageal body and lower esophageal sphincter, Sohn et al. (1993) concluded that cell shortening was mediated by the M₃ receptor in the sphincter and the M₂ receptor in the esophageal body. They observed a low pA₂ value for p-F-HHSiD (6.78) and an unusually high pA₂ value for methoctramine (9.05) in cells from the esophageal body, and a low pA₂ value for methoctramine (6.53) and a high pA₂ value for p-F-HHSiD (8.61) in cells from the sphincter. Our conclusion that cholinergic contractions in intact muscle strips from the cat esophageal body are mediated by a M₃ receptor mechanism are at odds with the above findings. We were unable to detect mRNA for the M₂ receptor in the cat esophagus, although the significance of this finding must be interpreted cautiously since we have not assayed for the presence of M₂ receptors per se; the mRNA content of a tissue may not consistently reflect receptor expression since other factors such as receptor turnover may be important. In many smooth muscles, the M₂ receptor population dominates, while the less plentiful M₃ receptor mediates the contraction response (Eglen et al., 1996a). Although, our experiments show a dominant role for the M₃ receptor in mediating functional cholinergic responses, we cannot rule out a contribution by the M₂ receptor subtype. As yet there is no explanation for the differing contribution of M₂ and M₃ receptors in intact muscle strips compared with muscle cells isolated by enzymatic digestion (Sohn et al., 1993).

The M₃ receptor preferentially activates phospholipase C, stimulating the production of inositol-1,4,5 trisphosphate, which triggers the release of calcium from intracellular stores (Hulme et al., 1990; Eglen et al., 1996a). In the cat, a significant increase in inositol 1,4,5 trisphosphate in response to acetylcholine stimulation was observed in cells isolated from the circular muscle of the lower esophageal sphincter but not the esophageal body (Sohn et al., 1993). Previous studies on cat and opossum smooth muscle strips showed that the cholinergic contraction is dependent on extracellular calcium (De Carle et al., 1977; Biancani et al., 1987). On the other hand, patch-clamp experiments using cells isolated from the cat esophageal muscle reported by Sims et al. (1990) have demonstrated cholinergic activation of potassium current, indicative of the release of calcium from intracellular stores. Moreover, Kirber and Biancani (1996) recently have found direct evidence for the release of intracellular calcium accompanying acetylcholine-induced contraction of these cells. These latter observations are compatible with a M₃-mediated activation of the phospholipase C pathway. Finally, recent studies on human esophageal muscle, from our laboratory, are consistent with a dominant functional role for the M₃ receptor and a contractile mechanism, which also involves the mobilization of intracellular calcium stores (Sims et al., 1997; Preiksaitis et al., 1996). Although the present data clearly support a major role for the M₃ receptor in the cholinergic response of the cat esophageal smooth muscle, further studies are required to clarify the postreceptor mechanisms involved.