During the storage phase of the micturition cycle, the urinary bladder must accommodate increasing amounts of urine without major elevation of intravesical pressure. This enhancement of bladder compliance requires relaxation of smooth muscle cells of the detrusor, which is controlled by reflex pathways involving an efferent activity of the sympathetic nervous system, particularly the hypogastric nerve originating from spinal cord segments Th12–L2 (Michel and Peters, 2004). Norepinephrine released from the hypogastric nerves primarily acts upon β-adrenergic receptors in the urinary bladder to promote relaxation during the storage phase. Therefore, β-adrenergic receptor activation is considered to be the most important physiological mechanism mediating urinary bladder relaxation during the filling/storage phase of the micturition cycle (Yamaguchi, 2002; Andersson, 2004).

Several recent reports have investigated the β-adrenergic receptor subtypes mediating urinary bladder relaxation in several species. Atypical β-adrenergic receptors, i.e., β₃ and/or other non-β₁-non-β₂ subtypes, seem to be important for bladder relaxation in all species, but β₂-and possibly even β₁-adrenergic receptors can additionally contribute in some species (Yamaguchi, 2002). Much less attention has been devoted to identifying intracellular signaling pathways mediating the bladder relaxation upon β-adrenergic receptor stimulation. Stimulation of adenylyl cyclase, and hence formation of cAMP, is the prototypical signaling pathway of β-adrenergic receptors (Bylund et al., 1994). Until recently, the smooth muscle relaxant effect of β-adrenergic stimulation was considered to occur via cAMP, partly because the adenylyl cyclase stimulator forskolin, which also increases cellular cAMP content in the bladder (Kories et al., 2003), frequently mimicked the effects of β-adrenergic stimulation. However, more recent data have implicated other signaling pathways, including certain potassium channels (for review, see Peters and Michel, 2003). Specifically, it has been shown in guinea pig urinary bladder that a β-adrenergic receptor agonist can activate potassium channels and that inhibition of such channels attenuates relaxation (Kobayashi et al., 2000).

Therefore, we have investigated the role of cAMP-dependent and -independent pathways in the relaxation of rat urinary bladder by the β-adrenergic agonist isoproterenol. Since previous studies have failed to unequivocally identify the β-adrenergic receptor subtype mediating bladder relaxation in the rat, we have based our present study on isoproterenol, an agonist that similarly acts on all known β-adrenergic receptor subtypes (Hoffmann et al., 2004). Previous studies on bladder relaxation by β-adrenergic agonists have employed two different approaches: relaxation against passive tension (Lecci et al., 1998; Yamazaki et al., 1998; Igawa et al., 1999) or relaxation against active tone, induced, e.g., by depolarizing concentrations of KCl (Oshita et al., 1997; Longhurst and Levendusky, 1999; Kobayashi et al., 2000; Yamanishi et al., 2002). Since we expected this methodological difference to affect the relative role of potassium channels in bladder relaxation, all our experiments were performed under both conditions.

Materials and Methods

Tissue Preparation. Adult male Wistar rats (260–280 g) were purchased from Charles River (Margate, Kent, UK). Animals were anesthetized using pentobarbital (75 mg/kg i.p.) and sacrificed by decapitation. The bladders were harvested, adipose and soft connective tissues were removed, and the middle parts were cut longitudinally into four strips (1 mm in diameter, 18 ± 1 mm in length, 9.4 ± 0.7 mg; n = 70). All experimental procedures were in line with National Institutes of Health guidelines for the use of laboratory animals and approved by the Animal Care Committee of Academisch Medisch Centrum.

Relaxation Experiments. The bladder strips were mounted under a resting tension of 10 mN in organ baths containing 7 ml of Krebs-Henseleit buffer of the following composition: 118.5 mM NaCl, 4.7 mM KCl, 1.2 mM MgSO₄, 0.025 mM Na₄EDTA, 2.5 mM CaCl₂, 1.2 mM KH₂PO₄, 25 mM NaHCO₃, and 5.5 mM glucose at 37°C, yielding a total potassium concentration of 5.9 mM. The organ baths were continually gassed with 95% O₂/5% CO₂ to maintain a pH of 7.4. The bladder strips were equilibrated for 75 min, during which the buffer solution was refreshed every 15 min. Following the equilibration, the tissues were challenged with 50 mM KCl for 6 min. Thereafter, they were again equilibrated with normal buffer, and readjustment of passive tension to 10 mN every 10 min until stabilization had occurred, usually within 45 min. Thereafter, inhibitors of signal transduction or their vehicles were added. The further experimental design is depicted in Fig. 1. After 10 min, some strips were precontracted with 50 mM KCl. After 30 min (20 min after KCl), when the KCl-exposed strips had reached a stable tension, cumulative isoproterenol concentration-response curves (0.3 nM–30 μM, with concentration increases every 3 min) were started (Fig. 1). At the end of each experiment, 10 μM forskolin was added to each preparation to assess receptor-independent relaxation. To avoid desensitization, only a single relaxation curve was generated in each bladder strip. To avoid false negative findings, we have generally used high inhibitor concentrations corresponding to those used in previous studies (Czyborra et al., 2002).

Preliminary experiments had shown that a passive tension of 10 mN and a KCl concentration of 50 mM yielded more stable contractions than various other combinations of passive tension and KCl concentration (data not shown). α-Adrenergic receptor antagonists (e.g., phentolamine or phenoxybenzamine) or uptake blockers (e.g., desipramine, hydrocortisone, or phenoxybenzamine) were not used because isoproterenol is unlikely to stimulate α-adrenergic receptors and because uptake blockers did not even affect relaxation by norepinephrine, a much better uptake substrate than isoproterenol (data not shown). Moreover, a high concentration of desipramine (10 μM) markedly reduced contraction responses to both carbachol and KCl, 50 μM phenoxybenzamine markedly reduced the contraction to carbachol, and phentolamine increased myogenic activity of the preparation (data not shown).

Chemicals. Pentobarbital was obtained from OPG Groothandel B.V. (Utrecht, The Netherlands). Hydrocortisone was acquired from Bufa B.V. (Uitgeest, The Netherlands). All other chemicals were purchased from Sigma Chemie (Deisenhofen, Germany). Stock solutions of the various experimental compounds were made as follows: 30 mM ODQ in dimethylsulphoxide; 1mM glibenclamide and 10 mM H89 in ethanol; 3 mM apamin in 5% acetic acid; and 10mM H7, 30 mM Rp-cAMPS, 300 μM tetraethylammonium chloride, 10 mM l-NNA, 3 μM iberiotoxin, 3 μM charybdotoxin, and 1 mM SQ 22,536 in deionized water.

Data Analysis. Nonlinear regression was used to fit sigmoidal curves to the isoproterenol concentration-response curves to determine agonist potency (pEC₅₀) and maximum effects (E_max). The force of contraction immediately prior to addition of the first isoproterenol concentration within a given experiment was defined as 0% relaxation, and a force of contraction of 0 mN was defined as 100% relaxation.

All data are expressed as mean ± S.E.M. of n experiments. Statistical significance of inhibitor effects on the E_max or pEC₅₀ of isoproterenol was assessed by two-tailed t tests if a single inhibitor was compared with a given vehicle or by one-way analysis of variance if multiple inhibitors were compared with a given vehicle. If the latter indicated that variance between groups was significantly greater than within groups, individual inhibitors were compared with vehicle with Dunnett's post hoc tests. Similar calculations were made for the forskolin effects. All statistical analysis was calculated using the Prism program (GraphPad software, San Diego, CA), and a P < 0.05 was considered statistically significant.

Results

Signal Transduction Underlying Relaxation against Passive Tension. In the absence of inhibitors, resting tension immediately prior to the addition of isoproterenol was approximately 10 mN. None of the vehicles or inhibitors substantially altered resting tension (data not shown). Moreover, resting tension declined by less than 5% over the course of an experiment if no isoproterenol was added (Fig. 2). Therefore, no correction for such spontaneous decline was made in the analysis of the isoproterenol concentration-response curves.

Neither ethanol, acetic acid, nor dimethylsulphoxide significantly affected the relaxing response to isoproterenol relative to that in the presence of water (Table 1), and even triple vehicle had no effect (data not shown). When data for all vehicles were pooled, isoproterenol relaxed bladder tone with a pEC₅₀ of 8.16 ± 0.04 and maximum reduction of 33 ± 2% (n = 35, Fig. 2).

The role of cAMP- and cGMP-dependent pathways on bladder relaxation was assessed using the adenylyl cyclase inhibitor SQ 22,536 (1 μM); the protein kinase A (PKA) inhibitors H7 (10 μM), H89 (1 μM), and Rp-cAMPS (30 μM); and the guanylyl cyclase inhibitor ODQ (3 μM). All five inhibitors caused only minor attenuation of the isoproterenol effects, and such inhibition reached statistical significance only for Rp-cAMPS (Figs. 3 and 4; Table 1). l-NNA (100 μM), an inhibitor of NO synthase, also failed to affect the E_max of isoproterenol but significantly enhanced its potency (Fig. 4; Table 1); however, the meaning of this alteration is difficult to judge due to a change in the shape of the concentration-response curve in the presence of l-NNA.

Six different inhibitors were used to study the role of various types of potassium channels in the isoproterenol-induced bladder relaxation: BaCl₂ (10 μM), tetraethylammonium chloride (3 μM), glibenclamide (10 μM), apamin (300 nM), charybdotoxin (30 nM), and iberiotoxin (30 nM). None of the potassium channel blockers significantly altered the E_max or pEC₅₀ of isoproterenol (Fig. 5; Table 1).

Finally, we tested the effects of the combined inhibition of the signaling pathways. A combination of SQ 22,536 and ODQ significantly inhibited maximum isoproterenol responses without altering its potency (Fig. 6; Table 1). The addition of charybdotoxin to SQ 22,536, ODQ or its combination, however, had no additional inhibitory effect (Fig. 6; Table 1).

Forskolin (10 μM, added after the highest isoproterenol concentration) yielded only little additional relaxation (Fig. 1), i.e., caused a total relaxation of 38 ± 2% (pooled data for all vehicles). All of the above inhibitors had similar effects against forskolin (Table 2) as against the maximum effect of isoproterenol; i.e., only a combination of SQ 22,536 and ODQ significantly attenuated the forskolin responses (Table 1).

Signal Transduction Underlying Relaxation of KCl-Induced Tension. From a passive tension of 10 mN, the addition of 50 mM KCl initially increased tension to approximately 35 mN; thereafter, tension slowly declined to a relatively stable plateau of approximately 25 mN (Fig. 1). This plateau was not markedly affected by any of the vehicles of inhibitors (data not shown) and declined by less than 5% over the course of an experiment if no isoproterenol was added (Fig. 2).

Neither ethanol, acetic acid, nor dimethylsulphoxide significantly affected the relaxing response to isoproterenol relative to that in the presence of water (Table 1). When data for all vehicles were pooled, isoproterenol relaxed bladder tone with a pEC₅₀ of 7.43 ± 0.04 and an E_max of 53 ± 1% (n = 34, Fig. 2). None of the inhibitors of cyclic nucleotide- or nitric oxide-dependent pathways caused statistically significant alterations of isoproterenol-induced relaxation of precontracted bladder strips (Figs. 3 and 4; Table 1).

The BK_Ca channel inhibitors charybdotoxin and iberiotoxin significantly reduced the relaxing potency of isoproterenol, and charybdotoxin additionally significantly reduced its E_max in precontracted bladder (Fig. 5; Table 1). In contrast, BaCl₂, tetraethylammonium chloride, glibenclamide, or apamin did not significantly affect the isoproterenol responses (Table 1).

A combination of SQ 22,536 and ODQ did not inhibit isoproterenol-induced relaxation of rat bladder (Table 1). Moreover, SQ 22,536 and ODQ alone or in combination did not enhance the inhibitory effect of charybdotoxin (Table 1).

Relaxation responses to 10 μM forskolin in the absence of signaling inhibitors but in the presence of precontraction were 69 ± 2% (n = 34, pooled data from all vehicles). The forskolin response was significantly inhibited by charybdotoxin or iberiotoxin (Table 2). A significant inhibition of forskolin responses was also seen with a combination of SQ 22,536 and ODQ, but neither of these alone or in combination enhanced the inhibitory effect of charybdotoxin (Table 2).

Discussion

Smooth muscle tone in the urinary bladder is largely regulated by a balance between contraction elicited via muscarinic acetylcholine receptors and relaxation elicited via β-adrenergic receptors. Recently, we have intensively characterized the signaling pathways underlying muscarinic receptor-mediated contraction of rat and human urinary bladder (Fleichman et al., 2004; Schneider et al., 2004a,b). In the present study, we have investigated the signal transduction underlying rat urinary bladder relaxation by the β-adrenergic agonist isoproterenol. Mechanisms involved in the relaxant effects of forskolin, added after the highest isoproterenol concentration, were preliminarily studied in comparison.

Methodological Considerations. Previous studies have employed two different approaches to measure urinary bladder relaxation by β-adrenergic receptor agonists, i.e., relaxation against passive tension (Lecci et al., 1998; Yamazaki et al., 1998; Igawa et al., 1999) or relaxation against active tone, induced, e.g., by depolarizing concentrations of KCl (Oshita et al., 1997; Longhurst and Levendusky, 1999; Kobayashi et al., 2000; Yamanishi et al., 2002, 2003). Since we hypothesized that the extracellular potassium concentration may affect the relative role of potassium channels, we have used both approaches in parallel. Our data show that the presence of KCl-induced precontraction indeed has important implications for the quantification of the isoproterenol-induced bladder relaxation. Thus, measurements in the presence of KCl yielded approximately one log unit lower isoproterenol potency but somewhat greater maximum effects. This is in line with indirect comparisons in the published literature, where an pEC₅₀ for isoproterenol of 8.3 (Yamazaki et al., 1998) versus 7.2 (Longhurst and Levendusky, 1999) and of 9.1 (Yamazaki et al., 1998) versus 7.3 (Oshita et al., 1997) were reported in rats and rabbits, respectively, for passive tension versus precontraction. Physiological urinary potassium concentrations range between 1 and 50 mM, but due to the presence of the urothelium, it is not fully clear whether and/or how much of this variation is sensed by bladder smooth muscle cells. Moreover, increasing bladder filling itself alters bladder distension. Therefore, we cannot definitively determine which of the two experimental conditions approximate the physiological situation more closely. Irrespective of such considerations, these differences will impact estimates of relative selectivity of an agonist for bladder versus, e.g., cardiac β-adrenergic receptors and hence may be important in defining uroselectivity of such agents.

Signaling Underlying Bladder Relaxation. Cyclic AMP is the classic second messenger of β-adrenergic receptors, and the adenylyl cyclase activator forskolin mimics the effect of isoproterenol on bladder tone in the present and many previous studies. Therefore, we have evaluated the role of the cAMP/PKA pathway in rat bladder relaxation using the adenylyl cyclase inhibitor SQ 22,536 and the three PKA inhibitors H7, H89, and Rp-cAMPS. All four inhibitors caused only weak inhibition against either isoproterenol or forskolin, and only inhibition by Rp-cAMPS of isoproterenol or forskolin responses against passive tension reached statistical significance. Although we cannot rule out that some of these negative results were affected by an insufficient tissue penetration of the inhibitors, it should be noticed that the chosen inhibitor concentrations were rather high and similar to those used by many other investigators in studies on vascular or bronchial smooth muscle. Therefore, our data do not support a major role for cAMP and/or PKA in isoproterenol or forskolin-induced bladder relaxation. They are in line with those of several other tissues where smooth muscle relaxation by β-adrenergic agonists was insensitive or only partially sensitive to inhibitors of the cAMP/PKA pathway (for review, see Peters and Michel, 2003). A rather minor role for the cAMP/PKA pathway may be particularly relevant for atypical β-adrenergic receptors since, e.g., in rabbit corpus cavernosum, relaxation via β₂-adrenergic receptors involved adenylyl cyclase activation, whereas that via atypical β-adrenergic receptors did not (Teixeira et al., 2004). If isoproterenol acts cAMP-independently, this could be a direct effect of the receptor or its linked G-protein to a different effector such as a potassium channel. Direct coupling of β-adrenergic receptors to at least some types of potassium channels has indeed been demonstrated (Kathöfer et al., 2003). On the other hand, it remained surprising that the cAMP/PKA pathway also explained only a minor part of the relaxation by the adenylyl cyclase activator forskolin. Although forskolin is a direct adenylyl cyclase activator, its effect upon intracellular cAMP content may be partly G_s-dependent (Laurenza et al., 1989). A future study with full concentration-response curves for forskolin in the absence of isoproterenol may be necessary to fully define the signaling underlying forskolin-induced bladder relaxation.

NO, guanylyl cyclase, and cGMP are not prototypical signaling molecules of β-adrenergic receptors, but recent studies in vascular preparations have shown that inhibition of these pathways can attenuate β-adrenergic relaxation of smooth muscle in vascular (Cardillo et al., 1997; Hutri-Kähönen et al., 1999) and nonvascular preparations (Li et al., 2004; Teixeira et al., 2004). Therefore, we have assessed their potential role in isoproterenol-induced bladder relaxation using the NO synthesis inhibitor l-NNA and the guanylyl cyclase inhibitor ODQ. Whereas ODQ did not significantly attenuate the isoproterenol or forskolin effect, l-NNA lacked inhibition of maximum isoproterenol effects but may have enhanced its potency. These findings are difficult to interpret since the overall role of NO for detrusor function remains controversial (Garcia-Pascual et al., 1996; Liu and Lin-Shiau, 1997).

Interestingly, concomitant inhibition of both adenylyl and guanylyl cyclase inhibited the isoproterenol response in rat urinary bladder against passive tension by about two thirds. Therefore, we propose that adenylyl cyclase and guanylyl cyclase each contributes to isoproterenol-induced bladder relaxation in a minor way only, because the two pathways may compensate for each other. Only combined inhibition reveals their relevance. Nevertheless, it must be stated that even combined inhibition accounted for only two thirds of the overall response against passive tension and did not at all explain relaxation against active tension. A similar synergistic effect was recently demonstrated for vascular smooth muscle relaxation by ceramide, which was only weakly sensitive to inhibition of guanylyl cyclase or BK_Ca channels, whereas their combined inhibition abolished it (Czyborra et al., 2002).

Potassium channels are present in many tissues, including smooth muscle of the lower urinary tract (Gopalakrishnan et al., 1999; Chen et al., 2004). Potassium channel opening, particularly as induced by BK_Ca channel openers, mediates relaxation (Malysz et al., 2004). Some previous studies have linked β-adrenergic receptor stimulation to potassium channel opening (Czyborra et al., 2002; Horinouchi et al., 2003; Kathöfer et al., 2003). Specifically, it has been shown that β-adrenergic receptor stimulation in guinea pig urinary bladder activates BK_Ca channels (Kobayashi et al., 2000). Therefore, we have studied the role of various potassium channels in isoproterenol-induced rat urinary bladder relaxation using a variety of blockers, including the broad-spectrum inhibitors BaCl₂ and tetraethylammonium, the K_ATP channel inhibitor glibenclamide, the Ca²⁺-activated potassium small-conductance channel inhibitor apamin, and the BK_Ca high-conductance calcium-dependent potassium-channel inhibitors charybdotoxin and iberiotoxin. Previous studies in guinea pigs had already demonstrated that iberiotoxin can attenuate isoproterenol-induced bladder relaxation (Kobayashi et al., 2000). In the present study, only inhibitors of calcium-dependent potassium channels, particularly charybdotoxin, inhibited the isoproterenol effects, whereas inhibitors of other potassium channels lacked significant effects. Similar to the attenuation by adenylyl or guanylyl cyclase inhibitors and in agreement with the previously reported guinea pig data (Kobayashi et al., 2000), this was only a partial inhibition. Opposite to the effects of the adenylyl or guanylyl cyclase inhibitors, the inhibition was seen only in precontracted bladder strips and not against passive tension.

Since adenylyl cyclase, guanylyl cyclase, and calcium-dependent potassium channels each contributed to isoproterenol-induced bladder relaxation in a minor way only, we tested combinations of the various inhibitors. Although the combination of the adenylyl and the guanylyl cyclase inhibitor was quite effective against passive tension (see above), the combination of either alone or both with charybdotoxin did not produce any additional effects either against passive tension or in precontracted bladder strips.

In summary, we propose that adenylyl and guanylyl cyclase but not potassium channels contribute to isoproterenol effects against passive tension, whereas calcium-dependent potassium channels but not adenylyl or guanylyl cyclase contribute to effects against KCl-induced precontraction. Since physiological potassium concentrations in urine can vary over a wide range, it remains to be studied in vivo which of these mechanisms is operative under physiological conditions.