Contribution of β-adrenoceptor subtypes to relaxation of colon and oesophagus and pacemaker activity of ureter in wildtype and β₃-adrenoceptor knockout mice

Abstract

1. The smooth muscle relaxant responses to the mixed β₃-, putative β₄-adrenoceptor agonist, (−)-CGP 12177 in rat colon are partially resistant to blockade by the β₃-adrenoceptor antagonist SR59230A suggesting involvement of β₃- and putative β₄-adrenoceptors. We now investigated the function of the putative β₄-adrenoceptor and other β-adrenoceptor subtypes in the colon, oesophagus and ureter of wildtype (WT) and β₃-adrenoceptor knockout (β₃KO) mice.

2. (−)-Noradrenaline and (−)-adrenaline relaxed KCl (30 mM)-precontracted colon mostly through β₁-and β₃-adrenoceptors to a similar extent and to a minor extent through β₂-adrenoceptors. In colon from β₃KO mice, (−)-noradrenaline was as potent as in WT mice but the effects were mediated entirely through β₁-adrenoceptors. (−)-CGP 12177 relaxed colon from β₃KO mice with 2 fold greater potency than in WT mice. The maintenance of potency for (−)-noradrenaline and increase for (−)-CGP 12177 indicate compensatory increases in β₁- and putative β₄-adrenoceptor function in β₃KO mice.

3. In oesophagi precontracted with 1 μM carbachol, (−)-noradrenaline caused relaxation mainly through β₁-and β₃-adrenoceptors. (−)-CGP 12177 (2 μM) relaxed oesophagi from WT by 61.4±5.1% and β₃KO by 67.3±10.1% of the (−)-isoprenaline-evoked relaxation, consistent with mediation through putative β₄-adrenoceptors.

4. In ureter, (−)-CGP 12177 (2 μM) reduced pacemaker activity by 31.1±2.3% in WT and 31.3±7.5% in β₃KO, consistent with mediation through putative β₄-adrenoceptors.

5. Relaxation of mouse colon and oesophagus by catecholamines are mediated through β₁- and β₃-adrenoceptors in WT. The putative β₄-adrenoceptor, which presumably is an atypical state of the β₁-adrenoceptor, mediates the effects of (−)-CGP 12177 in colon, oesophagus and ureter.

Introduction

β₃-Adrenoceptors mediate relaxation of gastrointestinal smooth muscle (reviewed by Arch & Kaumann, 1993). In rat colon, the endogenous catecholamine, (−)-noradrenaline, selective β₃-adrenoceptor agonists as well as (−)-CGP 12177 and (±)-cyanopindolol caused relaxant responses which were resistant to blockade by the β₁- and β₂-adrenoceptor antagonist (−)-propranolol (Kaumann & Molenaar, 1996). Both (−)-CGP 12177 and (±)-cyanopindolol were partial agonists which blocked the relaxant effects of the β₃-adrenoceptor selective agonists and therefore appeared to cause their effects through a common receptor. We used the selective β₃-adrenoceptor antagonist SR 59230A (Manara et al., 1995) and showed that it competitively blocked the effects of (−)-noradrenaline, β₃-selective agonists and (−)-CGP 12177 (Kaumann & Molenaar, 1996). However, SR 59230A blocked the relaxant responses of (−)-CGP 12177 considerably less than those to (−)-noradrenaline and β₃-selective agonists, suggesting involvement of a receptor distinct from β₁-, β₂- and β₃-adrenoceptors, conceivably the putative β₄-adrenoceptor. The pharmacology of the putative β₄-adrenoceptor has been described in the heart of several species (Kaumann 1989; 1997; Lowe et al., 1998; 1999), including man (Kaumann 1996; Kaumann & Molenaar, 1997; Molenaar et al., 1997b) and mouse (Kaumann et al., 1998), and in murine and human adipocytes (Galitzky et al., 1997; Preitner et al., 1998). In cardiac muscle, stimulant effects are observed to a series of compounds described as non-conventional partial agonists which include (−)-CGP 12177. Non-conventional partial agonists cause blockade of responses mediated by stimulation of β₁- and β₂-adrenoceptors but higher concentrations cause stimulant effects which are resistant to blockade by antagonists such as (−)-propranolol, but can be blocked with moderate affinity by (−)-bupranolol. In this respect the pharmacology of the putative β₄-adrenoceptor is similar to that of the β₃-adrenoceptor, but the receptors are distinct because putative β₄-adrenoceptor function is preserved in atrium (Kaumann et al., 1998) and fat cells of mice (Preitner et al., 1998) with targeted disruption of the β₃-adrenoceptor gene.

The purpose of this study was to investigate the role of putative β₄-adrenoceptors in smooth muscle and compare it to the function of β₁-, β₂- and β₃-adrenoceptors. We used the mouse colon, oesophagus and ureter for which the relative role of β-adrenoceptor subtypes is unknown. β₃-adrenoceptors and to a lesser extent β₂-adrenoceptors mediate relaxation of rat oesophagus (de Boer et al., 1993). In canine ureter, CGP 12177 caused relaxation, an effect attributed to stimulation of β₃-adrenoceptors (Tomiyama et al., 1998), however the contribution of the putative β₄-adrenoceptor to CGP 12177 mediated relaxation in canine ureter was not investigated.

β₁- and β₃-adrenoceptors were activated with selective agonists and antagonists. To identify functional putative β₄-adrenoceptors we used (−)-CGP 12177 but it can also stimulate β₃-adrenoceptors (Kaumann & Molenaar, 1996; Molenaar et al., 1997a; Sennitt et al., 1998). To provide unambiguous evidence for the existence of a (−)-CGP 12177 activated receptor that is distinct from the β₃-adrenoceptor, genetically engineered mice which had targeted disruption of the β₃-adrenoceptor were used. This approach was previously used to show that the receptor activated by (−)-CGP 12177 (previously called putative β₄-adrenoceptor) was pharmacologically distinct from the β₃-adrenoceptor and remained functional in cardiac (Kaumann et al., 1998) and brown adipose tissues (Preitner et al., 1998).

Our results show that (−)-noradrenaline-evoked relaxation of mouse colon and oesophagus is mediated through β₁- and β₃-adrenoceptors. In the colon from β₃KO mice, β₁-adrenoceptors compensate for the lack of β₃-adrenoceptors by mediating increased relaxation. Relaxation of colon, oesophagus and reduction of ureter pacemaker activity by (−)-CGP 12177 is mediated through putative β₄-adrenoceptors.

Methods

129 Sv mice (male, 2–4 months) were obtained from the Animal Resources Centre, Canning Vale, Western Australia. Balb/c mice (either sex, 7–10 weeks) were bred at the Babraham Institute. 129 Sv×C57BL/6 wildtype (WT) and β₃KO mice were off-springs of the previously described founders (Revelli et al., 1997). 4–5 month old female homozygous WT and β₃KO mice were used.

Southern blot genotyping

The identification of the WT or homozygous β₃KO mice was performed as previously described (Revelli et al., 1997; Kaumann et al., 1998).

Northern blots

Total RNA was isolated from colon fragments using RNeasy Mini Kit (QIAGEN, Basel, Switzerland). Twelve μg of total RNA was electrophoresed on a 1% agarose gel containing formaldehyde according to standard protocols and transferred to Electran Nylon blotting membranes (BDH Laboratory Supplies, Poole, U.K.) by vacuum blotting. Hybridizations were performed with radiolabelled probes using Quikhyb™ hybridization solution (Stratagene, La Jolla, CA, U.S.A.) as previously described (Revelli et al., 1997). The β₁-adrenoceptor probe (size: 756 bp, between positions 297 and 1053 on the mouse sequence GenBank Accession L10084) was obtained by reverse transcription-polymerase chain reaction (RT-PCR) on mouse heart polyA+ RNA. The β₂-adrenoceptor probe (size: 488 bp, between positions 641 and 1129 on the rat sequence GenBank Accession J03024) was obtained by polymerase chain reaction (PCR) on a rat β₂-adrenoceptor cDNA. RNA size estimates for the RNA species were established by comparison with an RNA ladder (Gibco, BRL). Films were exposed for 3–4 days at −80°C. RNA levels were quantified by scanning photodensitometry of the autoradiograms. Subsequent hybridization of the blots with a γ−³²P-ATP labelled synthetic oligonucleotide specific for the 18S rRNA subunit was used to correct for the differences in the amounts of RNA loaded onto the gels.

Longitudinal muscle of the colon

A partially depolarized colon preparation (30 mM K⁺) essentially as described previously (Kaumann & Molenaar, 1996) was used. Sv 129, WT and β₃KO mice (either sex, aged 6–12 weeks) were killed either by dislocation of the neck or rapid exsanguination from the carotid arteries. One ‘whole' 2.5 cm (approximately) segment of colon was dissected free and placed immediately into continuously oxygenated (95% O₂/5% CO₂) modified Krebs' solution (mM): Na⁺ 125, K⁺ 5, Ca²⁺ 2.25, Mg²⁺ 0.5, Cl⁻ 98.5, SO₄²⁻ 0.5, HCO₃⁻ 32, HPO₄²⁻ 1, EDTA 0.04, ascorbic acid 0.2 at room temperature. Tissues were mounted longitudinally in pairs in 50 ml organ baths, with care taken not to occlude the lumen, attached to Swema 4–45 strain gauge transducers and given 10 mN force. The incubation medium was exchanged with modified Krebs solution containing in addition (mM): Na⁺ 15, fumarate 5, pyruvate 5, L-glutamate 5 and glucose 10, together with 30 μM corticosterone (to block extraneuronal uptake of amines), 3 μM cocaine (to block neuronal uptake of amines) and 1 μM phentolamine (to block α-adrenoceptors) at 37°C. Tissues were allowed to equilibrate for 30 min, some in the presence of β-adrenoceptor antagonists before the addition of 30 mM KCl. Tone was maintained by KCl for 30 min, the tissues washed twice and allowed to equilibrate for a further 60 min before the re-addition of KCl. Antagonists were re-added after each wash. Agonists were added, either cumulatively in half log increments or as a bolus at maximally effective concentration. To terminate the experiment a maximally effective concentration of (−)-isoprenaline (200–600 μM) was added. Relaxant responses were expressed as a percentage of the effect of (−)-isoprenaline.

Longitudinal muscle of the oesophagus

Mice (Balb/c, WT, β₃KO) were killed by dislocation of the neck (Balb/c) in accordance with Home Office (U.K.) procedures or rapid exsanguination (WT, β₃KO) and the oesophagus immediately taken out and placed in equilibrated oxygenated (95% O₂/5% CO₂) solution at room temperature containing (mM): Na⁺ 146, K⁺ 5.2, Ca²⁺ 1.8, Mg²⁺ 1, Cl⁻ 122.6, HCO₃⁻ 25, H₂PO₄²⁻ 1.2, glucose 10, pyruvate 2, EDTA 0.04 and ascorbic acid 0.2. Oesophagi were cut open longitudinally and set up in pairs at 37°C in a 50 ml organ bath containing the above solution. Tissues were attached to Swema 4–45 strain gauge transducers and force was recorded on a 12 channel Watanabe polygraph. Tissues had slight spontaneous activity and were stretched to 2 mN force. Tissues were incubated with 3 μM cocaine, 10 μM corticosterone and 1 μM phentolamine.

Before commencement of experiments, strips were washed once and allowed to equilibrate for 30 min. After obtaining a steady state contraction to carbachol (1 μM), cumulative concentration-relaxation curves to (−)-noradrenaline, CL 316,243 and (−)-CGP 12177 were constructed. Antagonists were present for 60 min (30 min before and 30 min during stabilization of the carbachol response) prior to the beginning of each curve. At the end of each concentration-effect curve (±)-isoprenaline (600 μM) was added to obtain maximal relaxation.

Concentration-effect curves to (−)-noradrenaline were constructed in the absence or presence of ICI 118,551 (fold selectivity: 125 β₂ vs β₁, Bilski et al., 1983; >6000 β₂ vs β₃, putative β₄, Kaumann & Molenaar, 1996); CGP 20712A (fold selectivity: 16,000 β₁ vs β₂, Lemoine & Kaumann, 1991; >12,000 β₁ vs β₃, Kaumann & Molenaar, 1996; 1600 β₁ vs putative β₄, Kaumann & Molenaar, 1996); SR 59230A (fold selectivity: 125 β₃ vs putative β₄, Kaumann & Molenaar 1996) and (−)-propranolol (fold selectivity: >100 β₁, β₂ vs β₃, Kaumann & Molenaar 1996; >600 β₁, β₂ vs putative β₄, Kaumann & Molenaar 1996) either in combination or alone.

Curves to CL 316,243 and (−)-CGP 12177 were all constructed in the presence of 200 nM (−)-propranolol in the absence or presence of 10 μM SR 59230A. In some experiments carbachol was omitted and (−)-CGP 12177 added as a bolus concentration (2 μM).

Concentration-effect curves of CL 316,243 and (−)-CGP 12177 were expressed as a percentage reduction of the (1 μM) carbachol-induced contraction. Curves to (−)-noradrenaline were expressed as absolute values in mN relaxation to clearly show differences in the extent of contraction caused by carbachol and to show the extent of relaxation.

Ureter

β₃KO and WT mice were rapidly exsanguinated as above, the left and right ureters excised and placed in oxygenated modified Krebs solution (composition as for colon experiments) at room temperature. Two preparations (2 mm in length) from the middle portion of each ureter were mounted on 40 μm stainless steel wires in Mulvany-style small vessel myographs (J.P. Trading, Aarhus, Denmark) for isometric recording of circular smooth muscle activity. The incubation medium was exchanged with modified Krebs solution containing in addition (mM): Na⁺ 15, fumarate 5, pyruvate 5, L-glutamate 5 and glucose 10 together with 30 μM corticosterone, 3 μM cocaine, 1 μM phentolamine and 1 μM prazosin. After a 30 min equilibration period at 37°C, the preparations were stretched to a passive force of 2.5 mN and allowed to equilibrate for a further 30 min. During this time baseline tension gradually stabilized at approximately 1 mN. Spontaneous beating was initiated in these otherwise quiescent preparations by increasing the KCl concentration to 25 mM in the modified Krebs solution and by adding the thromboxane A₂ mimetic U46619 (100–300 nM). After beating commenced, preparations were allowed to stabilize for 10 min before agonists were applied. In other experiments using tissues from Balb/c mice, time control experiments demonstrated that these preparations maintained a steady rate of beating for up to 40 min.

Statistics and calculations

All data are expressed as mean±standard error. Differences between groups were assessed and tested for significance with Student's t-test at P⩽0.05.

The concentration ratio (CR) caused by an antagonist B, was determined by use of log forms, pD₂ - pD_2,B where (−log EC₅₀=pD₂, −log EC₅₀ in the presence of antagonist=pD_2,B).

The error of log CR was calculated from (s.e.mean²pD₂+s.e.mean²pD_2,B)^1/2 (Kaumann, 1990).

The blockade by SR 59230A of the effects of CL 316,243 was analysed with a Schild-plot (Arunlakshana & Schild, 1959).

Drugs used

(−)-CGP 12177 was a gift from Dr Jonathan Arch (Smith Kline Beecham Pharmaceuticals, Harlow, Essex, U.K.), (−)-bupranolol was a gift from Dr Klaus Sandrock (Sanol-Schwarz, Monheim, Germany), SR 59230A was a gift from Dr Luciano Manara (Sanofi, Milan, Italy), CL 316,243 was a gift from Dr Kurt Steiner, (Wyeth-Ayerst Research Princeton, NJ, U.S.A.), ICI 118,551 from Zeneca (Wilmslow, Cheshire, U.K.), CGP 20712A from Novartis (Basel, Switzerland), U46619 from Sapphire Bioscience (Sydney, Australia), (−)-propranolol hydrochloride; (−)-isoprenaline bitartrate, (±)-isoprenaline bitartrate, corticosterone, phentolamine methanesulphonate, carbachol (carbamylcholine chloride) from Sigma (St Louis, MO, U.S.A.), cocaine HCl from SmithKline Beecham Pharmaceuticals (Harlow, U.K.) or Victorian Hospitals Association (Mulgrave, Australia), prazosin HCl from RBI (Nattick, Massachusetts, U.S.A.), (−)-noradrenaline bitartrate from RBI (Nattick, Massachusetts, U.S.A.) or Sigma (St Louis, MO, U.S.A.).

Results

Contribution of β₁-, β₂-, β₃- and putative β₄-adrenoceptors to relaxation of mouse (SV 129) colon

(−)-Noradrenaline caused concentration-dependent relaxation of the longitudinal muscle of 30 mM KCl-precontracted mouse colon (Figure 1a,b, Table 1). The effects were blocked by the β₁-adrenoceptor-selective antagonist CGP 20712A (300 nM) but not by the β₂-adrenoceptor-selective antagonist ICI 118,551 (50 nM, P=0.07) (Figure 1a, Table 1). However, the shift of the (−)-noradrenaline concentration-effect curve by CGP 20712A was 2-log units less than expected from its corresponding affinity for β₁-adrenoceptors (Table 1). The simultaneous addition of CGP 20712A and ICI 118,551 did not cause significantly greater blockade than CGP 20712A alone. (−)-Propranolol (200 nM) caused similar blockade to CGP 20712A but again it was less than expected from its affinity for β₁- and β₂-adrenoceptors. The selective β₃-adrenoceptor antagonist SR 59230A at a concentration of 3 μM caused a 1.0-log unit rightward shift of the concentration-effect curve to (−)-noradrenaline (Figure 1b, Table 1). When combined with 200 nM (−)-propranolol to block β₁- and β₂-adrenoceptors, 3 μM SR 59230A caused an additional rightward shift of the concentration-effect curve (Figure 1b, Table 1).

Figure 1.

Relaxation of 129 Sv mouse colon by (−)-noradrenaline (a,b) and (−)-adrenaline (c,d) in the absence or presence of 50 nM ICI 118,551, 300 nM CGP 20712A, 50 nM ICI 118,551+300 nM CGP 20712A, 3 μM SR 59230A, 200 nM (−)-propranolol or 3 μM SR 59230A+200 nM (−)-propranolol. Values shown are mean±s.e.mean (vertical lines, where larger than symbol size), n=4–6 tissues for each curve.

Table 1.

Effects of antagonists on the relaxant effects of (−)-noradrenaline, (−)-adrenaline, CL 316,243 and (−)-CGP 12177 in colon of 129 Sv mice

(−)-Adrenaline caused concentration dependent relaxant responses in colon (Figure 1c,d, Table 1). As for (−)-noradrenaline, the relaxant effects of (−)-adrenaline were partially resistant to blockade with CGP 20712A (Figure 1c, Table 1). ICI 118,551 (50 nM) caused a greater rightward shift of the concentration effect curve to (−)-adrenaline (0.56-log units, P=0.03) than to (−)-noradrenaline (0.30-log units, not significant). (−)-Propranolol at a concentration of 200 nM and SR 59230A (3 μM) each caused 1-log unit rightward shifts of the concentration-effect curves to (−)-adrenaline (Figure 1d, Table 1). The combination of (−)-propranolol and SR 59230A caused a further shift to the right (Figure 1d, Table 1).

The presence of the β₃-adrenoceptor in colon was confirmed with the selective β₃-adrenoceptor agonist CL 316,243 which was a partial agonist with respect to (−)-isoprenaline. In the presence of 200 nM (−)-propranolol, the concentration-effect curve to CL 316,243 was shifted to the right in a concentration dependent manner by SR 59230A (Figure 2a, Table 1). The pK_B value for SR 59230A, (determined from the Schild plot with slope constrained to 1), was 6.81 (Figure 2b).

Figure 2.

Relaxant responses in 129 Sv mouse colon to CL 316,243 and blockade by SR 59230A. Responses are to CL 316,243 in the absence or presence of SR 59230A. All experiments were carried out in the presence of 200 nM (−)-propranolol. Values shown are mean±s.e.mean (vertical lines, where larger than symbol size for experiments with n⩾4), n=2–8 tissues for each curve. The corresponding Schild-plot is shown in (b). The slope of the Schild-plot was not significantly different from 1.0 (slope 0.90±0.16) and was therefore constrained to unity.

In contrast, 3.3 μM SR 59230A did not cause a rightward shift of the concentration-response curve to (−)-CGP 12177 in the presence of 200 nM (−)-propranolol. (−)-Bupranolol (1 μM) however, caused an 0.77-log unit rightward shift (Figure 3, Table 1).

Figure 3.

Relaxant responses in 129 Sv mouse colon of (−)-CGP 12177 in the absence or presence of 3 μM SR 59230A or 1 μM (−)-bupranolol. All experiments were carried out in the presence of 200 nM (−)-propranolol. Values shown are mean±s.e.mean (vertical lines, where larger than symbol size), n=6–9 tissues for each curve.

Contribution of β₁-, β₂- and putative β₄-adrenoceptors to relaxation in β₃KO mouse colon

There was no difference in the 30 mM KCl induced increase in tone in colon from WT (6.9±0.3 mN, n=156) and β₃KO mice (6.0±0.6 mN, n=20, P=0.3). (−)-Noradrenaline relaxed colon from WT and β₃KO mice with similar potencies (Figure 4a, Table 2). CGP 20712A (300 nM) caused a greater shift of the concentration-effect curve to (−)-noradrenaline in β₃KO mice compared to WT mice (Figure 4a, Table 2). The addition of 50 nM ICI 118,551 to 300 nM CGP 20712A did not cause further blockade in WT or β₃KO mice (Figure 4a, Table 2).

Figure 4.

(a) Relaxant responses to (−)-noradrenaline in WT and β₃KO mouse colon in the absence or presence of 300 nM CGP 20712A or 300 nM CGP 20712A+50 nM ICI 118,551. (b) Relaxant responses to (−)-CGP 12177 in the presence of 200 nM (−)-propranolol in β₃KO and Sv 129 (from Figure 3 for comparison). Values shown are mean±s.e.mean (vertical lines, where larger than symbol size), n=4–8 tissues for each curve. Note that for WT, concentration–effect curves to (−)-noradrenaline in the presence of CGP 20712A alone or CGP 20712A+ICI 118,551 were nearly superimposable.

Table 2.

Effects of antagonists on the relaxant effects of (−)-noradrenaline in colon from wildtype (WT) and β₃KO mice

In the presence of 200 nM (−)-propranolol, a single concentration of CL 316,243, 2 μM caused 81.1±1.9% relaxation with respect to 600 μM (−)-isoprenaline (n=8) of colon from WT mice but had no effect in colon from β₃KO mice, n=4 (Figure 5). In the presence of CL 316,243, the addition of 2 μM (−)-CGP 12177 caused 21.9±0.3% relaxation with respect to 600 μM (−)-isoprenaline in β₃KO mice (n=4). A representative experiment is shown in Figure 5. In other colon from β₃KO mice, (−)-CGP 12177 caused concentration dependent relaxation (pEC₅₀, 7.19±0.04, n=4, Figure 4b).

Figure 5.

Original traces obtained from WT and β₃KO mouse colonic tissues which were mounted in the same tissue bath. Tissues were pre-contracted with 30 mM KCl and then exposed to 2 μM CL 316,243 (CL), followed by 2 μM (−)-CGP 12177 (CGP) and finally 600 μM (−)-isoprenaline (ISO). Arrows indicate administration of drug. (−)-CGP 12177 caused relaxation in colon from β₃KO mouse. Note the lack of effect of CL 316,243 in the same colon. Horizontal bar=5 min.

Quantitative determination of β₁- and β₂-adrenoceptor mRNA in WT and β₃KO colon

β₁- and β₂-adrenoceptor mRNA was quantified in WT and β₃KO mouse colon. β₁- and β₂-adrenoceptor transcripts had sizes of 3.1 kb and 2.3 kb respectively (Figure 6) which corresponded to those observed in white and brown adipose tissue (Revelli et al., 1991; 1997). The targeted disruption of the β₃-adrenoceptor did not change the level of β₁- or β₂-adrenoceptor mRNA.

Figure 6.

Photodensitometric comparisons of autoradiogram signals obtained from WT (+/+) or β₃-adrenoceptor deficient (−/−) 2–3-month-old female mouse colon total RNA hybridized with β₁- and β₂-adrenoceptor (AR) probes. Representative autoradiogram signals obtained in colon with β₁- and β₂-ARs and 18S rRNA (18S) probes are shown in (a). The results are expressed as percentage±s.e. of the mean value in +/+ mice taken as 100%; n=20 experiments (b). The signals were quantified by scanning photodensitometry and normalized using the corresponding 18S rRNA values. Only signals obtained on the same Northern blot were compared. The sizes of the RNAs were 3.1 kb for the β₁-AR and 2.3 kb for the β₂-AR.

Contribution of β₁-, β₂-, β₃- and putative β₄-adrenoceptors to relaxation of mouse (Balb/c) oesophagus

The mouse oesophagus had slight spontaneous activity (Figure 7). Carbachol (1 μM) contracted the oesophagus (Figure 7a) and the contraction was enhanced by pre-incubation with ICI 118,551 (50 nM) (P=0.005) (Figure 8b) or (−)-propranolol (200 nM) (P=0.0009) (Figure 8d) but not CGP 20712A (300 nM) (P=0.26) (Figures 7a and 8a) or SR 59230A (10 μM) (P=0.19) (Figures 7a and 8c). The combination of both CGP 20712A (300 nM) and SR 59230A (10 μM) enhanced carbachol-evoked contractions (P=0.04, Figure 8e). (−)-Propranolol (200 nM) also caused an increase in basal tone (Figure 7b), as did a combination of both CGP 20712A (300 nM) and ICI 118,551 (50 nM) (Figure 7a), while CGP 20712A (Figure 7a), SR 59230A (Figure 7a) or ICI 118,551 (data not shown) alone had no effect.

Figure 7.

Representative recordings of mouse oesophageal muscle. In (a) concentration response curves to (−)-noradrenaline were constructed in the absence or presence of CGP 20712A (CGP 300 nM), a combination of both CGP 20712A (CGP 300 nM) and ICI 118,551 (ICI 50 nM) and SR 59230A (SR 10 μM) respectively. Half log unit increments in (−)-noradrenaline concentration were added. Maximal relaxation was tested by the addition of (±)-isoprenaline (ISO) (600 μM). In (b) a single concentration of (−)-CGP12177 ((−)-CGP 2 μM) caused a decrease of basal tone, either in the presence of (−)-propranolol or a combination of both (−)-propranolol and SR 59230A (10 μM). Maximal relaxation was tested by the addition of (±)-isoprenaline (ISO 600 μM). In (c) oesophagi from WT and β₃KO were mounted in the same bath, incubated with 200 nM (−)-propranolol, exposed to (−)-CGP 12177 ((−)-CGP 2 μM) and then 600 μM (−)-isoprenaline (ISO). Note doubling of gain and readjustment of pen position in (b) and after equilibrium of carbachol response in (a).

Figure 8.

Relaxant responses to (−)-noradrenaline in mouse oesophageal muscle, in the absence or presence of (a) CGP 20712A 300 nM, (b) ICI 118,551 50 nM, (c) SR 59230A 10 μM, (d) (−)-propranolol 200 nM and a combination of both CGP 20712A 300 nM and ICI 118,551 50 nM, (e) a combination of both CGP 20712A 300 nM and SR 59230A 10 μM, (f) a combination of ICI 118,551 50 nM and SR 59230A 10 μM and (g) a combination of CGP 20712A 300 nM, ICI 118,551 50 nM and SR 59230A 10 μM. n=8−17 tissues for each group (see Table 3).

(−)-Noradrenaline caused concentration-dependent relaxation of the mouse oesophagus (Figure 7a and 8). The effects of (−)-noradrenaline were blocked surmountably by CGP 20712A (300 nM), SR 59230A (10 μM) or (−)-propranolol (200 nM), but not ICI 118,551 (50 nM) (Figure 8a–d, Table 3). The combination of both CGP 20712A (300 nM) and ICI 118,551 (50 nM) caused a greater shift of the concentration effect curve of (−)-noradrenaline compared to CGP 20712A alone (Figures 7a and 8a,d), as did the combination of both CGP 20712A (300 nM) and SR 59230A (10 μM) (Figure 8e). ICI 118,551 (50 nM) caused no additional blockade when combined with SR 59230A (10 μM) (Figures 8c,f, Table 3) or with the combination of CGP 20712A (300 nM) and SR 59230A (10 μM) (Figure 8g, Table 3). When shifts of concentration-effect curves were analysed at the 0.35 mN level of tone, ICI 118,551 alone or in combination with other antagonists caused small right-ward shifts (Table 3).

Table 3.

Effects of antagonists on relaxation of mouse oesophagus by (−)-noradrenaline

Both CL 316,243 and (−)-CGP 12177 caused concentration dependent relaxation of the mouse oesophagus in the presence of (−)-propranolol (200 nM). SR 59230A (10 μM) caused a 1.36-log unit rightward shift of the relaxant effects of CL 316,243 (Figure 9a, Table 4) but did not cause significant blockade of the relaxant effects of (−)-CGP 12177 (p=0.07, Figure 9b, Table 4).

Figure 9.

Relaxant responses in mouse oesophageal muscle of (a) CL 316,243 and (b) (−)-CGP 12177 in the absence or presence of 10 μM SR 59230A. All experiments were carried out in the presence of 200 nM (−)-propranolol. Values shown are mean±s.e.mean (vertical lines, where larger than symbol size), n=8–11 tissues for each curve.

Table 4.

Effects of SR 59230A (10 μM) on relaxation of mouse oesophagus by CL 316,243 and (−)-CGP 12177

In the presence of 200 nM (−)-propranolol and absence of carbachol, a single concentration of (−)-CGP12177 (2 μM) reduced basal tone by 67±6% (relaxation expressed as a percentage of the effect of 600 μM (±)-isoprenaline, n=6) which was unaffected by 10 μM SR 59230A (68±4%, Figure 7b).

(−)-CGP 12177 mediates relaxation of oesophagus from WT and β₃KO mice

In the presence of 200 nM (−)-propranolol, 2 μM (−)-CGP 12177 caused similar relaxation (expressed as a percentage of the relaxation to 600 μM (−)-isoprenaline) of oesophagi from WT (61.4±5.1%, n=4) and β₃KO (67.3±10.1%, n=4) mice (Figure 7c).

Reduction in pacemaker activity of ureter from WT and β₃KO mice

CL 316,243 (2 μM) reduced KCl- and U46619-induced pacemaker activity in ureter from WT (per cent of control 80.5±2.8, n=4, P=0.006) but not β₃KO (per cent of control 97.4±2.1, n=3, P=0.3) mice. In the presence of CL 316,243, 2 μM (−)-CGP 12177 caused a reduction of pacemaker activity in ureter from both WT (per cent of control 68.9±2.3, n=4, P=0.0009) and β₃KO (per cent of control 68.7±7.5, n=3, P=0.05) mice (Figure 10).

Figure 10.

Original traces showing the effects of 2 μM CL 316,243 (CL) and 2 μM (−)-CGP 12177 (CGP) on KCl- and U46619-induced spontaneous contractions in WT and β₃KO ureter. In this experiment the frequency of spontaneous contractions was lower in the ureter from the β₃KO compared to the WT mouse. CL 316,243 reduced the frequency of spontaneous contractions in WT but not β₃KO ureters. (−)-CGP 12177 reduced the frequency of contractions in β₃KO. Time bar=1 mm

Discussion

The longitudinal smooth muscle of mouse colon and oesophagus is relaxed by stimulation of both β₁- and β₃-adrenoceptors. The effects of (−)-CGP 12177 were presumably mediated through the ‘putative β₄-adrenoceptor' in colon, oesophagus and ureter, previously identified in heart (Kaumann, 1989; 1997; Kaumann & Molenaar, 1997; Molenaar et al., 1997b; Kaumann et al., 1998; Lowe et al., 1998; 1999) and adipose tissue (Galitzky et al., 1997; Preitner et al., 1998).

β₁- and β₃-adrenoceptors mediate relaxation of smooth muscle

The relaxant effects of (−)-noradrenaline in colon and oesophagus were partially blocked by selective β₁- (CGP 20712A) and β₃- (SR 59230A) but not β₂-adrenoceptor antagonists (ICI 118,551) indicating relaxation mediated by both β₁- and β₃-adrenoceptors. Assessment of the blockade caused by CGP 20712A and SR 59230A alone showed that both β₁- and β₃-adrenoceptors share to around the same extent the mediation of relaxation by both (−)-adrenaline and (−)-noradrenaline. In colon of β₃KO mice, the 2.6-log unit rightward shift of the concentration-effect curve to (−)-noradrenaline by 300 nM CGP 20712A showed that β₁-adrenoceptors were mostly responsible for relaxation. Previously we observed that 200 nM (−)-propranolol, the concentration used in this study, did not affect the potency of (−)-noradrenaline in rat colon indicating that (−)-noradrenaline mediates its effects essentially by stimulation of β₃-adrenoceptors and that this concentration of (−)-propranolol does not block β₃-adrenoceptors (Kaumann & Molenaar, 1996). However 200 nM (−)-propranolol caused a 0.8-log unit rightward shift in the concentration-effect curve to (−)-noradrenaline in mouse colon indicating a greater relative involvement of β₁-adrenoceptors in mouse compared to rat colon.

The partial resistance of blockade by β₁- and β₂-adrenoceptors with (−)-propranolol or with the combination of CGP 20712A and ICI 118,551 together with blockade by the β₃-selective antagonist SR 59230A of relaxant responses of the catecholamines in colon and oesophagus revealed the involvement of the β₃-adrenoceptor. The presence of the β₃-adrenoceptor in mouse colon and oesophagus was confirmed by the use of the selective β₃-adrenoceptor agonist CL 316,243 and blockade of its relaxant effects by the antagonist SR 59230A. In mouse colon the affinity of SR 59230A, pK_B=6.8 was similar to that observed in rat colon against the β₃-adrenoceptor agonists ZD 2079 (7.0), SR 586118 (6.9) and BRL 37344 (6.9) at β₃-adrenoceptors (Kaumann & Molenaar, 1996).

β₂-adrenoceptors in colon and oesophagus

β₂-adrenoceptors are present in colon and oesophagus but play a minor role in mediating relaxation. The small degree of blockade caused by 50 nM ICI 118,551 to the relaxant responses of (−)-adrenaline (but not (−)-noradrenaline) revealed a small contribution of β₂-adrenoceptors to relaxation in colon. The presence of β₂-adrenoceptors in colon was confirmed by the detection of β₂-adrenoceptor mRNA. Higher concentrations of (−)-noradrenaline may cause relaxation of oesophagus by stimulation of β₂-adrenoceptors. This was shown by the rightward shift of the concentration-effect curve to (−)-noradrenaline at the 0.35 mN level by the combination of CGP 20712A and ICI 118,551 (2.06-log units) which was 1-log unit greater than in the presence of CGP 20712A alone (0.99-log units) (Table 3).

In the oesophagus (−)-propranolol and a combination of both CGP 20712A and ICI 118,551 increased basal tone. (−)-Propranolol and ICI 118,551 also enhanced the carbachol evoked contraction. Blockade of either postsynaptic β₁- or β₃-adrenoceptors alone by CGP 20712A or SR 59230A respectively did not increase basal tone or the carbachol evoked contraction however the combination of CGP 20712A and SR 59230A enhanced the carbachol evoked contraction. These observations may be explained by the presence of prejunctional sympathetic nerve β₂-adrenoceptors which facilitate release of (−)-noradrenaline in a number of tissues (Majewski, 1983; Göthert & Kollecker, 1986; Münch et al., 1996). In the mouse oesophagus, stretch in the organ bath could release endogenous noradrenaline, possibly facilitated by stimulation of prejunctional β₂-adrenoceptors which in turn stimulates post-synaptic β₁- and β₃-adrenoceptors to reduce muscle tone. The increase of basal tone and carbachol evoked contraction observed after administration of ICI 118,551 and (−)-propranolol could be caused by blockade of pre-junctional β₂-adrenoceptors, leading to a reduction of endogenous (−)-noradrenaline release. Interestingly ICI 118,551 on its own did not increase basal tone but did so in the presence of carbachol. Carbachol and other muscarinic receptor agonists can reduce endogenous (−)-noradrenaline release from sympathetic nerves by stimulating prejunctional muscarinic receptors (Muscholl, 1980) and when combined with ICI 118,551 to block prejunctional β₂-adrenoceptors, a further reduction in transmitter release may occur. Blockade of both post-synaptic β₁- and β₃-adrenoceptors could prevent endogenously released (−)-noradrenaline from reducing muscle tone.

Evidence for the putative β₄-adrenoceptor in smooth muscle

The selective β₃-adrenoceptor antagonist SR 59230A in the presence of (−)-propranolol blocked the relaxant effects of CL 316,243 but not significantly the effects of (−)-CGP 12177 in both colon and oesophagus, indicating a lack of involvement of β₁-, β₂-and β₃-adrenoceptors for (−)-CGP 12177 mediated relaxation. This together with the maintenance of (−)-CGP 12177-evoked relaxation in colon and oesophagus of β₃KO mice, suggested an effect mediated by stimulation of putative β₄-adrenoceptors. In the oesophagus a single concentration of (−)-CGP 12177 also reduced basal tone in the absence of carbachol. (−)-CGP 12177 is also an antagonist at both β₁-and β₂-adrenoceptors (Kaumann, 1983; Nanoff et al., 1987) and would have been expected to increase basal tone as observed for (−)-propranolol if not for its ability to stimulate the putative β₄-adrenoceptor.

In the ureter the selective β₃-adrenoceptor agonist CL 316,243 and (−)-CGP 12177 both mediated a reduction in pacemaker frequency suggesting the involvement of both β₃- and possibly putative β₄-adrenoceptors. The effect of (−)-CGP 12177 but not CL 316,243 was maintained in ureter from β₃KO mice, consistent with mediation through the putative β₄-adrenoceptor.

Compensatory increases in the function of β₁- and putative β₄-adrenoceptors in colon of β₃KO mice

Interestingly, there was no difference in the potency of (−)-noradrenaline in colonic tissues taken from β₃KO (pEC₅₀ 7.1) or WT mice (pEC₅₀ 7.0), suggesting that β₁-adrenoceptors took over the relaxant function of β₃-adrenoceptors. The compensatory hyperfunction of β₁-adrenoceptors in β₃KO is strongly supported by the considerably greater blockade of the relaxant effects of (−)-noradrenaline by CGP 20712A in colon from β₃KO than WT mice. It is not yet known whether the compensatory hyperfunction of β₁-adrenoceptors is due to an increase in receptor density of these receptors but the lack of change β₁-adrenoceptor transcript level in colon from the two groups argues against an increase of receptor density. Instead, β₁-adrenoceptors may be coupled more efficiently to the pathway leading to relaxation. The regulation of β₁-adrenoceptor mRNA is tissue specific and is not necessarily predictable as no change was observed in white adipose tissue of β₃KO mice, but a reduction in brown adipose tissue was observed in the same study (Revelli et al., 1997). (−)-CGP 12177 was 2 fold more potent in colon from β₃KO (pEC₅₀ 7.2) than WT mice (pEC₅₀ 6.9, P=0.005). The markedly enhanced blocking potency of CGP 20712A of the relaxant effects of (−)-noradrenaline in β₃KO colon suggests greater compensatory increases of β₁-adrenoceptor function than putative β₄-adrenoceptor function. Thus, the presence of the β₃-adrenoceptor may have an inhibitory effect on coupling of both β₁- and putative β₄-adrenoceptors with β₁-adrenoceptor>putative β₄-adrenoceptor.

The putative β₄-adrenoceptor as an atypical state of the β₁-adrenoceptor

As with other non-conventional partial agonists (Kaumann, 1989; 1997), (−)-CGP 12177 shows dissociation of antagonist (high affinity for β₁- and β₂-adrenoceptors) and agonist (low cardio-stimulant potency) properties. Interestingly, Pak & Fishman (1996) also observed a similar pattern of agonist and antagonist activity for (±)-CGP 12177 on rat and human recombinant β₁- and β₂-adrenoceptors expressed at high non-physiological densities in several cell lines. As found with native receptors (Sarsero et al., 1998b; 1999), Pak & Fishman (1996) detected a high and low affinity population for (−)-[³H]-CGP 12177. The effects of (±)-CGP 12177 were attributed to stimulation of a guanine nucleotide sensitive low-affinity form of the β₁-adrenoceptor which comprised approximately 10% of the total population of β₁-adrenoceptors (Pak & Fishman, 1996).

But can the effects ascribed to the putative β₄-adrenoceptor be explained in terms of a low affinity conformation of the β₁-adrenoceptor? In mouse (Kaumann et al., 1998), rat (Sarsero et al., 1998b; 1999) and human cardiac tissues (Sarsero et al., 1998a), (−)-³H-CGP 12177 labels a population of high affinity receptors (pK∼9) corresponding to β₁- and β₂-adrenoceptors and another distinct low affinity population (pK∼7) attributed to the putative β₄-adrenoceptor. This population is not affected by guanine nucleotides unlike rat and human recombinant and transfected β₁-adrenoceptors whose affinity for (−)-[³H-CGP 12177 is reduced (Sarsero et al., 1998b). Furthermore, the population attributed to putative β₄-adrenoceptors in rat heart can be subdivided into a high-affinity form (β₄H) and a low-affinity form (β₄L, pK∼5) for several non-conventional partial agonists but not for (−)-CGP 12177 which has the same affinity for both forms (Sarsero et al., 1999). Because the affinity of several non-conventional partial agonists for the β₄H form agrees with their corresponding inotropic and chronotropic potencies it has been suggested to mediate the positive inotropic and chronotropic effects of these ligands (Sarsero et al., 1999). The suggestion is also in line with the low binding affinity of (−)-propranolol for the putative β₄-adrenoceptor (Sarsero et al., 1998b) which correlates with its failure to block the cardiostimulant effects of (−)-CGP 12177 and other non-conventional partial agonists (Kaumann, 1996; Kaumann & Molenaar, 1996). In contrast, (−)-propranolol is a potent antagonist of the cardiostimulant effects of catecholamines mediated through β₁-adrenoceptors, suggesting that the putative β₄-adrenoceptor is distinct from the β₁-adrenoceptor or is a (−)-propranolol-resistant form of the β₁-adrenoceptor.

The hypothesis that some agonist effects of (−)-CGP 12177 are mediated through an atypical state of the β₁-adrenoceptor has recently received conclusive support. Cardiostimulation by (−)-CGP 12177 was observed in atria from mice lacking β₂-adrenoceptors (β₂-AR knockout) but not in atria from mice lacking both β₁- and β₂-adrenoceptors (double β₁-/β₂-AR knockout) (Kaumann et al., 2000). Furthermore, CGP 12177 stimulates adenylyl cyclase of mouse brown fat through a high and low affinity receptor population; the high-affinity and low-affinity effects were not observed in β₁-AR knockout and β₃-AR knockout, respectively (Konkar et al., 2000). Taken together, this evidence indicates that β₁-adrenoceptors have an obligatory role for the mediation of at least part of the cardiostimulant effects and agonist effects of CGP 12177 in brown fat. It is therefore plausible, if not likely, that the smooth muscle relaxant effects of (−)-CGP 12177 are also mediated through an atypical state of the β₁-adrenoceptor. In line with this contention is our observation that both the relaxant effects of (−)-noradrenaline, mediated through β₁-adrenoceptors, and of (−)-CGP 12177 were found to be enhanced in smooth muscle from β₃-AR knockout compared to wild-type mice.

(−)-CGP 12177 has recently been found to be 40 times more potent than (−)-isoprenaline in eliciting arrhythmic calcium transients in mouse ventricular myocytes (Freestone et al., 1999). This effect of (−)-CGP 12177 was also (−)-propranolol-resistant and its potency subnanomolar, in agreement with its high-affinity for β₁-adrenoceptor, suggesting the existence of a high-affinity β₁-adrenoceptor state which is distinct from the low affinity state that presumably mediates positive chronotropic, inotropic and lusitropic effects (Kaumann et al., 1998; Sarsero et al., 1999) and relaxation of smooth muscle (this work).

Conclusions

In mouse colon and oesophagus, activation of β₁-and β₃-adrenoceptors mediate relaxation. In β₃KO, β₁- and putative β₄-adrenoceptors compensate for the lack of β₃-adrenoceptor contribution to relaxation. In comparison, β₂-adrenoceptors have a minor role. As recently demonstrated for cardiac and adipose tissues, the putative β₄-adrenoceptor in smooth muscle may correspond to an atypical state of the β₁-adrenoceptor.

Abbreviations

β₃KO

β₃-adrenoceptor knockout

Carbachol

carbamylcholine chloride

(−)-CGP 12177

(−)-4-(3-tertiarybutylamino-2-hydroxypropoxy) benzimidazol-2-one

CL 316,243

disodium (R,R)-5-[2-[[2-(3-chlorophenyl)-2-hydroxyethyl-amino]propyl]-1,3-benzodioxole-2,2-dicarboxylate]

CGP 20712A

2-hydroxy-5(2-((2-hydroxy-3-(4-((1-methyl-4-trifluoromethyl) 1H-imidazole-2-yl) -phenoxy) propyl) amino) ethoxy)-benzamide monomethane sulfonate

CR

concentration-ratio

ICI 118,551

(erythro-DL-1(7-methylindan-4-yloxy)-3-isopropylamino-butano-2-ol)

SR 59230A

3-(2-ethylphenoxy)-1-[(1S)-1,2,3,4-tetrahydronaphth-1-ylamino]-2S-2-propanol oxalate

WT

wildtype