Abstract
Alpha adrenoceptor antagonists have been convincingly shown to be beneficial in reducing both subjective and objective indices of urethral obstruction in benign prostatic hyperplasia. Rec 15/2739 (SB 216469) is a novel alpha-1 adrenoceptor (alpha-1 AR) antagonist currently being developed for benign prostatic hyperplasia. When evaluated in radioligand binding assays with expressed animal or human alpha-1 ARs, Rec 15/2739 shows marked to moderate selectivity for thealpha-1a AR subtype. Its affinity for the recombinantalpha-2 AR subtypes or native dopaminergic D2 receptor was about 100-fold lower than that foralpha-1a AR subtype. In canine tissues, Rec 15/2739 was 20-fold more potent as an inhibitor of [3H]prazosin binding to prostate vis-a-vis aorta. Functional studies in isolated rabbit tissues also confirmed the uroselectivity of Rec 15/2739, with substantially higher affinity (K b = 2–3 nM) being observed in urethra and prostate, compared with ear artery and aorta (K b = 20–100 nM). The in vitro selectivity observed with Rec 15/2739 was confirmedin vivo in the anesthetized dog, comparing potency against norepinephrine- or hypogastric nerve stimulation-induced urethral contraction with its ability to reduce diastolic blood pressure. In this model, Rec 15/2739 had greater selectivity than any other alpha-1 AR antagonist examined, including terazosin and tamsulosin. Based on the low potency of prazosin and some of its structural analogs in the rabbit and dog lower urinary tract tissues, it appears that norepinephrine contracts these tissuesvia activation of the alpha-1L AR. Hence this alpha-1 AR subtype, rather than thealpha-1A AR, may mediate the contraction in vivo.
Benign prostatic hyperplasia is a condition characterized by a nodular enlargement of prostatic tissue which results in obstruction of the proximal urethra (Caine and Perlberg, 1977). The hyperplastic prostate tissue will contract in response to both sympathetic nerve stimulation and alphaadrenoceptor stimulation with exogenous agonists (Caine et al., 1975; Hieble et al., 1985). This response, observed in vitro with tissue strips obtained during prostatectomy, is caused by contraction of the smooth muscle fibers present in prostatic stroma (Shapiro et al., 1992a). Because the ratio of stromal to glandular tissue is increased in hypertrophic prostate (Bartsch et al., 1979; Shapiro et al., 1992b), the adrenoceptor-mediated contraction of the stromal smooth muscle is likely to make an important contribution to the pressure that the enlarged gland exerts on the urethra. It has been estimated that this dynamic component contributes about 40% of the total tone that the hyperplastic prostate exerts on the urethra (Caine, 1986). Although BPH is currently treated primarily via surgical techniques, both clinical and experimental evidence suggest that pharmacological management of this disease is possible and can delay or prevent the necessity of surgery in a significant percentage of patients.
The alpha adrenoceptor antagonists have been shown to be beneficial in reducing both subjective and objective indices of urethral obstruction in BPH (Caine et al., 1976). Almost all clinical trials using alpha-1 AR antagonists showed a statistically significant improvement in objective parameters indicative of urinary obstruction (e.g., urinary flow rate, peak prostatic pressure), as well as in subjective symptom assessment by the patient and physician (Monda and Oesterling, 1993; Lepor, 1993). However, because of poor organ selectivity, side effects can limit the therapeutic usefulness of this class of drugs. Events such as dizziness, hypotension and asthenia are common, in addition to a “first-dose phenomenon,” which can give rise to pronounced orthostatic hypotension. In practice, side effects necessitate dose titration and limit the maximum tolerated dose. Such a practice also leads to underdosing; therefore, the full potential of completealpha-1 AR antagonism at the level of the bladder neck and prostate has not, as yet, been realized. A drug with selectivity foralpha-1 ARs in the lower urinary tract, with little or no effect on systemic blood pressure, would represent a major advance in the treatment of BPH. Such an agent would be expected to offer immediate and quantitatively greater improvement in terms of both objective measures and symptom scores, and the need for dose titration would be abolished.
Recent studies indicated that the alpha-1 AR that mediates the contraction of human prostate smooth muscle has the pharmacological properties of the alpha-1a (formerly designated asalpha-1c; Hieble et al., 1995a) subtype (Marshallet al., 1992, 1994; Forray et al., 1994a; Nobleet al., 1994). Hence, alpha-1 AR antagonists selective for this subtype were postulated as more efficacious and better tolerated agents for the treatment of symptomatic BPH.
Rec 15/2739 is a new alpha-1 AR antagonist recently synthesized in Recordati Laboratories (Leonardi et al., 1992), within a project aiming to discover new, more selective antagonists of the alpha-1 ARs to be used in the symptomatic treatment of obstructive disorders of the lower urinary tract. This compound is also referred to as SB 216469 and is now undergoing Clinical Phase II for BPH treatment. Its high affinity (pKi= 9.0–9.4) and selectivity for the alpha-1a AR subtype has been previously demonstrated by use of native and cloned animal and human alpha-1 AR subtypes (Testa et al., 1995).
The in vitro and in vivo animal pharmacology reported in the present paper demonstrates that this compound fulfills the therapeutic need for organ selectivity, as expressed above.
The experiments described in this paper were performed in comparison with drugs now clinically used for the therapy of BPH, namely prazosin, terazosin, alfuzosin and tamsulosin (Monda and Oesterling, 1993; Lepor, 1993). The data obtained with four other new compounds, chemically related to Rec 15/2739 (Leonardi et al., 1993), are also reported, as well as the results obtained with one new prazosin-like compound (see table 1; Leonardi et al., 1995), with thealpha-1a selective antagonists 5-methylurapidil (Grosset al., 1989), with the 1,4-dihydropyridine derivative SNAP 5089 (Wetzel et al., 1995) and with thealpha-1N-selective antagonist HV 723 (Muramatsu et al., 1990). Some of the data on Rec 15/2739 and the reference standards have already been presented in abstract form (Testa et al., 1994a, 1994b).
Materials and Methods
Male Sprague Dawley rats (200–250 g), New Zealand White rabbits (2.5–3.5 kg) and mongrel and Beagle dogs (10–12 kg) were used in these experiments. Animals were housed with free access to food and water and maintained on a forced 12-hr light-dark cycle at 22–24°C for at least 1 week before the experiments were carried out.
Affinity for the recombinant alpha-1 andalpha-2 AR subtypes, and native D2and 5-HT1A receptors.
The affinity of the tested compounds for the recombinant rat alpha-1d (previously alpha-1A/D), hamster alpha-1b and bovine alpha-1a (previously alpha-1c) ARs in COS-7 cells, as well as for the recombinant human subtypes, expressed in CHO cells, was evaluated as described previously (Testa et al., 1995). Recombinant animal alpha-1 AR subtypes expressed on COS-7 cell membranes were provided by Dr. Susanna Cotecchia, University of Lausanne, Switzerland. The affinity for the recombinant human alpha-2a, alpha-2b andalpha-2c adrenoceptors in CHO cells was evaluated as described previously (Hieble et al., 1995b). The affinity for the dopaminergic D2 (rat striatum) and serotoninergic 5-HT1A (rat hippocampus) receptors was evaluated according to the methods reported by Meltzer et al. (1989) and Hoyeret al. (1985), respectively.
Displacement of specific [3H]prazosin binding from canine prostate and aorta membranes.
Male mongrel dogs were sacrificed by intravenous injection of 5 ml of pentobarbital (324 mg/ml) into the cephalic vein. The aorta was removed and placed into ice-cold sucrose MOPS buffer, pH 7.2. The prostate was removed and placed in cold Krebs’ solution. The aorta was cleaned of adhering tissue, frozen in liquid nitrogen and stored at −70°C until used for membrane preparation. Microsomal membranes were prepared as reported previously (Shi et al., 1990). The aortae were minced in 10 volumes of ice-cold sucrose-MOPS buffer and homogenized with a Polytron homogenizer until a homogeneous suspension was obtained. The crude tissue homogenate was centrifuged at 2500 rpm for 10 min. The supernatant was filtered through two layers of nylon gauze and centrifuged at 9,000 rpm for 10 min. The resulting supernatant was centrifuged at 105,000 × g for 30 min. The microsomal pellet was suspended in sucrose-MOPS buffer. The prostatic tissue, cleaned of adhering tissue and the urethra, was minced in 10 volumes of ice-cold TRIS-HCl buffer, pH 7.4, and homogenized with a Polytron until a homogeneous suspension was obtained. The crude homogenate was filtered as reported above, and centrifuged at 2500 rpm for 10 min. The supernatant was stored on ice, and the pellet was suspended in buffer and homogenized and centrifuged. The supernatants were pooled and centrifuged at 9000 rpm for 10 min. The pellet was discarded, and the supernatant was centrifuged at 105,000 × g for 30 min. The crude pellets of canine aorta and prostate were resupended in assay buffer (50 mM MOPS with 10 mM MgCl2, pH 7.4, and 50 mM TRIS-HCl, pH 7.4, respectively), giving a final protein concentration of about 0.12 mg/ml and 0.26 mg/ml, respectively.
In saturation binding studies, the pellets were incubated with different concentrations of [3H]prazosin, ranging from 0.03 to 0.5 nM for aorta, and 0.02 to 10.4 nM for prostate. In displacement binding studies, they were incubated in the presence or not of different concentrations of the compounds tested, as well as a fixed concentration (near K d value) of [3H]prazosin, at 25°C for 45 min.
Nonspecific binding was determined in the presence of 10 μM phentolamine. The reaction was terminated by adding 5 ml of ice-cold Mg MOPS buffer and filtration through Whatman GF/B glass filter with a Brandell M-24 automatic cell harvester. The filters were dried, placed in 10 ml of scintillation solution, allowed to stand overnight at room temperature and counted in a scintillation counter.
Functional in vitro alpha-1 antagonistic activity.
The functional alpha-1 antagonistic activity of Rec 15/2739 and reference compounds was evaluated by studying the effects on NE-induced contractions of rabbit vascular (aorta and ear artery) and lower urinary tract tissues (urethra and prostate).
Adult rabbits were sacrificed by intravenous injection of pentobarbital into the marginal ear vein or by cervical dislocation. The central ear artery, aorta, urethra and prostate were removed, placed in Krebs-Henseleit buffer and dissected free of adhering tissue. Rings (3–4 mm wide) were cut from the aorta and ear artery, and two tungsten wire hooks (0.008 inches in diameter) were passed through the lumen of each ring. The prostate was bisected, and 5–0 surgical silk was tied to each end of both sections. Two strips were prepared from the prostatic urethra (1- to 2-cm-long specimens, starting from the trigone). The vascular and prostatic tissues were suspended between an isometric transducer and a stationary tissue holder, in organ baths containing buffer equilibrated with 95% O2:5% CO2 and maintained at 38°C. A resting tension of 1 g was applied to the tissues and allowed to equilibrate under these conditions for 45 to 60 min before testing. Cocaine (6 μM) was present in the buffer through the experiments to block neuronal uptake. The strip preparations from urethra were attached to isotonic transducers with 1 g of resting tension and suspended in Krebs’ solution also containing 0.1 μM desmethylimipramine and 1 μM corticosterone to block neuronal and extraneuronal uptake of NE, 1 μM (±)-propranolol to block beta adrenoceptors and 0.1 μM yohimbine to block alpha-2 adrenoceptors. Concentration-response curves were constructed in all tissues by stepwise cumulative addition of NE until no further increase of contractile response could be obtained. After washout of NE and reequilibration of the tissue (45 min), the antagonist to be tested was added to the bath, and after 30 min incubation, a second NE cumulative concentration-response curve was generated.
In vivo selectivity for the lower urinary tract.
The selectivity of Rec 15/2739 and reference compounds for the lower urinary tract tissues versus the cardiovascular system was evaluated by assessing the inhibition of the increase of UP induced by NE injection or hypogastric nerve stimulation, and its effects on DBP in anesthetized dogs. These experiments were performed according to the method of Imagawa et al. (1989), with substantial modifications.
Adult dogs were anesthetized with pentobarbital sodium (30 mg/kg i.v. and 2 mg/kg/hr i.v.), intubated and spontaneously ventilated with room air. To monitor the blood pressure, a PE catheter was introduced into the aortic arch through the right common carotid artery. A collateral of the left femoral vein was cannulated for infusion of the anesthetic, and the right femoral vein was cannulated for administration of the compounds. For intraarterial injection of NE, a PE catheter was introduced into the lower portion of abdominal aorta via the right external iliac artery. Through such procedure, NE is selectively distributed to the lower urinary tract. Via a midline laparotomy, the urinary bladder and proximal urethra were exposed. To prevent filling of the bladder, the two ureters were cannulated and urine was led outside. In order to record the prostatic UP, a Mikro-tip catheter (6 F) was introduced into the bladder via the external urethral meatus, and withdrawn until the pressure transducer was positioned in the prostatic urethra. A ligature was secured between the neck of the bladder and urethra to isolate the response of the latter and avoid any interaction with the bladder. Another ligature was put around the Mikro-tip catheter at the external urethral meatus to secure the catheter itself. After a stabilizing period following the surgical procedure (30 min), in which arterial and prostatic UPs were monitored continuously, five to seven intraarterial administrations of NE were made at 10-min intervals, and the mean of the urethral responses considered as basal value. The dose of NE used produced an increase of at least 100% in UP (usually 1.0–1.5 μg/kg). The test compounds were then administered i.v. in a cumulative manner with intervals of 15 min between administrations. Intraarterial injections of NE were repeated approximately 5 min after every dosing of test compound.
The same protocol was used in experiments in which urethral contractions were induced by hypogastric nerve stimulation. The hypogastric nerve was freed from surrounding tissue and cut 1 cm distal from the inferior mesenteric ganglion. The distal end of right or left branch of the nerve was placed on a bipolar electrode. The nerve stimulation was made with a train of rectangular pulses of 10 to 15 V, 10 to 30 Hz, width 5 msec, 8 sec duration. The compounds were administered i.v. as reported above and nerve stimulation was repeated 5 min after each dose.
In this model, the effects of some compounds were evaluated also after intraduodenal administration. After the administration of a single dose of the compounds (in a volume of 1 ml/kg), stimulations of the urethra were performed from 15 to 270 min, with intervals of 15 min in the first hour, and intervals of 30 min in the following hours.
Compounds and solutions.
The following compounds were used: [3H]prazosin (7-methoxy-3H), [3H]rauwolscine, [3H]spiperone and [3H]8-OH-DPAT (NEN Life Science Products, Cologno Monzese, Milano, Italy); norepinephrine tartrate, prazosin-HCl, phentolamine-HCl (Sigma-Aldrich, Milano, Italy); 5-methylurapidil (RBI); compounds in table 1, alfuzosin-HCl, SNAP 5089-HCl, terazosin-HCl and tamsulosin-HCl (all synthesized in Recordati Laboratories). HV 723 fumarate was kindly given by Prof. Muramatsu (Research Biochemicals International, Natick, MA); (Dept. of Pharmacology, Fukui Medical School, Matsooka, Fukui, Japan).
In binding studies, the compounds were dissolved in absolute ethanol. For the isolated organ preparations, HV 723, prazosin, terazosin, alfuzosin and phentolamine were dissolved in distilled water. Rec 15/2739 was dissolved in distilled water containing 0.05 Eq of methanesulfonic acid; 5-methylurapidil, Rec 15/2627, Rec 15/2869, Rec 15/2841, Rec 15/2802 and Rec 15/2636 were dissolved in distilled water containing 3% dimethylformamide and 3% Tween 80; tamsulosin was dissolved in dimethyl sulfoxide and water (1:1). All these stock solutions (10− 3 M) were further diluted with distilled water.
Statistical analysis.
The displacement curves of the antagonists on the receptor studied were analyzed by nonlinear curve fitting of the logistic equation according to the method reported by De Lean et al. (1978), by use of the ALLFIT program (from the National Institutes of Health). The IC50 values and pseudo-Hill slope coefficients were estimated by the program. The value for the inhibition constant, K i, was calculated by use of the Cheng and Prusoff equation (Cheng and Prusoff, 1973):K i = IC50/(1 +L/K D), where L is the concentration of the 3H-ligand used.
In the in vitro functional studies, the dissociation constant (K b) at the alpha-1 AR in the different tissues studied was estimated by the technique ofArunlakshana and Schild (1959). Dose ratios (i.e., the ratio between the concentrations of NE required to produce half-maximal response in the presence and in the absence of the antagonist tested) were calculated at each concentration of the compounds. The logarithm of these dose ratios −1 was plotted against the logarithm of the compound concentrations (Schild plot). The slope of the resulting regression line will not differ significantly from unity if the antagonism is competitive and the intercept on the x-axis is the pA2 value. From a Schild plot with slope constrained to unity the intercept can be considered to be a representation of the negative logarithm of the K b. When only two concentrations of the tested compounds were used, theK b value was calculated with the formula:K b = [B]/(dose ratio − 1), where [B] is the antagonist concentration. If theK b values obtained at both concentrations were similar, the antagonism was assumed to be competitive. For the different tissues studied, the concentrations (nM) used to evaluate theK b values are shown in table 2.
In the in vivo studies, dose-response curves were constructed by computing the percent inhibition of the increase in UP, the percent fall in DBP or the percent inhibition of the increase in systolic blood pressure produced by the test compound. ED25for DBP (dose inducing 25% decrease) and ED50 (dose inducing 50% inhibition of increase in UP and systolic blood pressure) values were computed by means of linear regression analysis.
Results
Affinity for recombinant alpha-1 andalpha-2 adrenoceptor subtypes and native D2 and 5-HT1Areceptors.
Table 3 shows the affinity of a series of antagonists for human recombinant alpha-1 AR subtypes and for recombinant alpha-1 ARs from cow (alpha-1a), hamster (alpha-1b) and rat (alpha-1d), respectively. Compounds tested included Rec 15/2739 and several structural analogs, prazosin and other quinazolines, SNAP 5089, a niguldipine analog reported to be highly selective for thealpha-1a AR and tamsulosin, a potent and partially selectivealpha-1a AR antagonist marketed for BPH. The classicalalpha-1a AR antagonist 5-methylurapidil and HV 723 were also included. A good correlation between the affinities for human and animal alpha-1 subtypes was observed (R2 values were .782, .790 and .871 for the alpha-1a,alpha-1b and alpha-1d subtypes, respectively), although the affinity for human recombinant alpha-1b andalpha-1d subtypes was generally higher than that evaluated on the animal clones, a finding also reported by other authors (e.g., Saussy et al., 1996).
Rec 15/2739 shows selectivity for the alpha-1a AR, although the magnitude of this selectivity is less than observed with the 1,4-dihydropyridine SNAP 5089. Rec 15/2739 shows 9- and 26-fold greater selectivity for the alpha-1a AR than the alpha-1d and alpha-1b subtypes, respectively, when affinities for human clones are compared. Data from animal receptors show greateralpha-1a selectivity, with ratios of 85 and 178 versus alpha-1d and alpha-1b subtypes, reflecting the trend reported above.
The selectivity profile of the structural analogs of Rec 15/2739 is similar (with the exception of Rec 15/2636) to all the flavones showing selectivity for the alpha-1a AR subtype. The affinity profile of the quinazolines is consistent with expectations from previously published data on members of this structural series. Based on data with human alpha-1 ARs, none of these compounds, nor HV 723, show significant subtype selectivity. Likewise, the profile of tamsulosin on human receptors is consistent with previous findings (Shibata et al., 1995), which show that this compound has high affinity for both alpha-1a and alpha-1d ARs, with a slight preference for the alpha-1a subtype. The affinity of Rec 15/2739 and its structural analogs for the recombinantalpha-2 adrenoceptors and dopaminergic D2receptor was dramatically lower than that for the alpha-1a subtype (table 4). Affinity for the 5-HT1Aserotoninergic subtype was similar to that of tamsulosin and lower than that of 5-methylurapidil.
Displacement of specific [3H]prazosin binding from canine prostate and aorta membranes.
Radioligand binding assays demonstrated high-affinity binding of [3H]prazosin to membrane homogenates of canine prostate and aorta. In both tissues, Scatchard analysis showed the presence of only one binding site (data not shown). K dvalues were 1.0 nM and 0.24 nM in the prostate and aorta, respectively.B max values were 83 fmol/mg protein (prostate) and 157 fmol/mg protein (aorta). Rec 15/2739 produced monophasic inhibition of [3H]prazosin binding in both prostate and aorta (fig. 1, upper panel). The K ivalues calculated for this inhibition were 27 nM in the aorta and 0.8 nM in the prostate. In contrast, unlabeled prazosin (fig. 1, lower panel) was a more potent inhibitor in the aorta (K i = 0.5 nM) than in the prostate (K i = 2.8 nM). Corresponding data, summarized in table 5, show that Rec 15/2739 and some of its closest structural analogs, as well as 5-methylurapidil, possess substantial selectivity for the prostatic alpha-1 ARs, whereas quinazoline derivatives were more potent in the aorta. Tamsulosin, phentolamine and SNAP 5089 were not selective.
Although some of the antagonists showed similar affinities at canine prostate adrenoceptors and at the recombinant alpha-1a ARs, others, particularly the 1,4-dihydropyridine SNAP 5089 and the quinazoline derivative Rec 15/2627, were substantially less potent in the canine prostate, leading to poor correlation with human and animalalpha-1a clones. No relationship between binding affinity in the canine prostate and affinity for the recombinantalpha-1b or alpha-1d ARs was observed.
Correlation of the pKi values of the compounds tested for inhibition of [3H]prazosin binding to membranes of canine aorta correlated very well with their affinity for the recombinant human and animal alpha-1b AR (R2 = 0.945 and 0.684, respectively), and to a lesser extent for thealpha-1d AR (R2 = 0.710 and 0.628, respectively). No correlation was found for the alpha-1a AR (both R2 < 0.03).
Functional in vitro alpha-1 antagonistic activity.
Rec 15/2739 produced a potent blockade of NE-induced contraction in isolated rabbit urethra and prostate. The potency in these urogenital tissues was about one order of magnitude greater than in two isolated rabbit blood vessels, the aorta and ear artery.
Data for Rec 15/2739 and other alpha-1 AR antagonists are shown in table 6. Although there was some variation in the potency profile of individual compounds among the four rabbit tissues, the affinities of the compounds for urethra and prostate were usually quite similar (R2 > 0.6); likewise, affinity for the aorta paralleled that for the ear artery.
Selectivity for the urogenital tissues was also observed with the structural analogs of Rec 15/2739, Rec 15/2841, Rec 15/2869 and Rec 15/2802. Interestingly, Rec 15/2636 proved not to be selective, and 5-methylurapidil was poorly selective. The quinazolines were generally more potent in the vascular tissues. Tamsulosin and phentolamine showed no significant selectivity in these functional assays. SNAP 5089 had very low potency in the rabbit urethra and rabbit aorta, and showed substantially higher affinity in the rabbit ear artery and prostate. HV 723 was potent in rabbit aorta.
Potency in the urogenital tissues could not be clearly related to binding affinity at any of the alpha-1 AR subtypes. In particular, the functional affinity of the quinazolines in both urethra and prostate was substantially lower than their radioligand binding affinity for the recombinant alpha-1a ARs. A good correlation was obtained between pKb values against NE-induced contraction in the rabbit aorta and the pKi for inhibition of [3H]prazosin binding to the recombinant animal and human alpha-1d AR (R2 = 0.609 and 0.711, respectively), although this tissue was formerly reported as containing a mixed population of alpha-1B andalpha-1L AR subtypes (Muramatsu et al., 1990). The functional potency of these 12 antagonists in the rabbit ear artery could not be related to binding affinity at any of thealpha-1 AR subtypes (R2 < 0.50), particularly the alpha-1a subtype (R2 < 0.10).
In vivo selectivity for the lower urinary tract.
In anesthetized dogs, urethral contractions could be elicited by intraarterial injection of NE into the lower portion of abdominal aorta. Via this procedure, NE was selectively distributed to the lower urinary tract and produced only minor effects on systemic blood pressure. Reproducible increases in UP in the range of 10 to 15 mm Hg were obtained, not differing significantly between experimental groups. Cumulative i.v. administration of Rec 15/2739 and prazosin produced dose-dependent inhibition of the urethral contractions induced by intraarterial NE without influencing basal UP. Drug solvent had no significant effect. As an inhibitor of this response, Rec 15/2739 and prazosin had equal potency. On the other hand, prazosin produced a 25% decrease in DBP at doses only slightly higher than those inhibiting the NE-induced contractions of the urethra, whereas Rec 15/2739 reduced DBP only at doses much higher than those active on the urethra (fig.2).
The results obtained in this model with Rec 15/2739, its structural analogs and other reference compounds are expressed as the dose required to produce a 50% inhibition of NE-induced urethral contractions, the dose required to reduce basal DBP by 25% and as a selectivity index obtained from the ratio of these two doses (table7). The selectivity index showed Rec 15/2739, Rec 15/2869 and Rec 15/2841 to be the most uroselective antagonists (selectivity ratios between 50 and 110). A second group, including tamsulosin, alfuzosin and phentolamine, showed moderate uroselectivity (selectivity ratios of about 10). Another structural analog of Rec 15/2739, Rec 15/2802, showed uroselectivity intermediate between these two groups. The other quinazolines showed no pharmacologically significant uroselectivity, with selectivity ratios <3. SNAP 5089 was a very weak antagonist of NE-induced urethral contraction. This compound had almost no effect on DBP, which prevented the quantitation of its uroselectivity. Extrapolation of the dose-response curve for reduction of DBP suggests a selectivity index of about 30 for SNAP 5089.
This canine model can also be used to study the effects of hypogastric nerve stimulation. All drugs tested, including Rec 15/2739, were slightly less potent against nerve stimulation-induced urethral contraction than against the response to exogenous NE, which resulted in lower selectivity ratios (table 8). However, Rec 15/2739 retained its enhanced uroselectivity, compared with tamsulosin, terazosin and prazosin. Rec 15/2739 and two of its structural analogs were also evaluated against stimulation-induced urethral contractionvia the intraduodenal route. All these compounds showed approximately 10-fold greater selectivity than tamsulosin when evaluated under these conditions (table 8).
Discussion
The design of alpha-1 AR antagonists capable of differentiating between alpha-1 ARs of the lower urinary tract and those of the vasculature has attracted the attention of different research groups during the past few years, primarily because of the increased interest in this class of therapeutic agents for the medical management of BPH.
In the anesthetized dog model, comparing the ability to block the increases in UP induced by NE or nerve stimulation with reduction in basal DBP, Rec 15/2739 shows substantially greater uroselectivity than other alpha-1 AR antagonists being marketed or developed for use in BPH. Our data in the dog with reference compounds are in full agreement with previously reported studies with similar procedures (Breslin et al., 1993; Poirier et al., 1988). Recently, the uroselectivity of Rec 15/2739 in the dog has been confirmed (Blue et al., 1996; Katwala et al., 1996). Blue et al. (1996) reported a 30-fold difference between the doses of Rec 15/2739 required to block phenylephrine-induced increases in blood pressure and intraurethral pressure. In agreement with our findings, they did not find that prazosin or terazosin showed uroselectivity. Although tamsulosin showed some uroselectivity in our model (table 7), Blue et al.(1996) did not observe uroselectivity for this agent, in agreement with previous observations in a similar model (Kenny et al., 1994). This uroselectivity is also observed in several in vitro models, evaluating either functional blockade of vascular and urogenital alpha-1 ARs, or receptor affinity as determined in radioligand binding assays. A recent report (Auguetet al., 1995) confirmed the in vitro selectivity of Rec 15/2739 between urethral and vascular responses in rabbit tissues. Interestingly, the potency of Rec 15/2739 and some related compounds on rabbit urethra and prostate found in the present study proved very close to that previously observed in human prostate (Testaet al., 1996).
With the recent identification of alpha-1 AR subtypes, functional and radioligand binding studies have been performed with prostate and other urogenital tissues to characterize thealpha-1 AR subtype responsible for mediating the contractile response of these tissues. Correlations have been reported between the functional dissociation constants for blockade of NE-induced contraction of human prostate and affinity for the recombinantalpha-1 AR subtypes (Forray et al., 1994a,b;Marshall et al., 1994). These studies suggested that contraction was mediated by the alpha-1a AR [designated asalpha-1c at the time; see Hieble et al. (1995a)for currently accepted nomenclature]. This is consistent with previous data which suggest a correlation between functional potency in urogenital tissues from several species and affinity for the nativealpha-1a (Testa et al., 1993) and recombinantalpha-1a AR (Testa et al., 1996). Localization of mRNA for the alpha-1 AR subtypes in human prostate has also demonstrated that the alpha-1a AR represents the predominant subtype (Price et al., 1993; Faure et al., 1994). As a result of these findings, drug discovery efforts in this area have focused on the design of highly selective alpha-1a AR antagonists. Compounds showing 100-fold or greater selectivity for thealpha-1a AR versus the other alphaadrenoceptor subtypes have been identified in several structural series (Huff et al., 1995; Gluchowski et al., 1994). Data are available on a large number of niguldipine analogs, which show up to 1000-fold selectivity for the alpha-1a AR (Wetzelet al., 1995; Gluchowski et al., 1994).
Although it is not yet known whether the high selectivity of these compounds for the alpha-1a subtype will be reflected in selectivity for the tissues of the lower urinary tract, preliminary data suggest that they have reduced orthostatic liability in the rat (Gong et al., 1994).
Considering Rec 15/2739 alone, most of its uroselectivity can be explained by its selectivity for alpha-1A versus alpha-1B or alpha-1D ARs. Its binding affinity for thealpha-1 AR site in canine prostate as well as its functional potency against NE-induced contraction of rabbit urethra and prostate is consistent with alpha-1a AR affinity, whereas potency in the dog and rabbit aorta suggests an alpha-1b andalpha-1d AR mediated effect. However, when our data with all the alpha-1 AR antagonists are considered, it becomes clear that uroselectivity involves more than alpha-1A versus alpha-1B or alpha-1D selectivity. If only the three known recombinant alpha-1 ARs are considered, the data would point to the involvement of the alpha-1a subtype in the response of the tissues of the lower urinary tract to adrenoceptor activation (table 9). No significant correlation of these data with alpha-1b or alpha-1d AR affinity was obtained in fact. However, two compounds, the dihydropyridine SNAP 5089 and the quinazoline Rec 15/2627, deviate substantially from the line of identity between functional activity or potency in the isolated urogenital tissues and alpha-1a AR affinity (fig.3). Elimination of these two compounds improves the correlation (table 9).
These differences can not be interpreted in terms of chemical instability and loss of compound (because of high lipophilicity) that may lead to underestimates of functional affinity for, e.g.,the dihydropyridine derivative SNAP 5089. We, in fact, evaluated by radioreceptor assay using recombinant human alpha-1a AR, the amount of this compound in the organ bath at the beginning and the end of the incubation time, during experiments performed to evaluate its antagonistic activity against the NE-induced contraction of rabbit urethra (data not shown). The concentration of SNAP 5089 in the bath was not changed at all after 30 min incubation time. Moreover, Fordet al. (1996a), by use of SNAP 5089 in the same experimental conditions, reported pA2 values of < 6.5 and 9.5 for human prostate and rat caudal artery, respectively, which indicated clearly that the lack of activity of this compound on a tissue can not be related to instability or loss of compound.
Similar results are obtained when the in vivo potency of the antagonists at urethral alpha-1 ARs in the dog, as reflected by their ED50 values against NE-induced increases in UP, are correlated with affinity for the recombinant alpha-1a ARs (table 9). In this case Rec 15/2627 and SNAP 5089 deviate even further from the regression line (fig. 3), and their elimination produces a dramatic increase in the correlation coefficient (table 9). Another example of this disparity is provided by RS 17053, an antagonist which has high affinity (K i < 1 nM) for both bovine and human alpha-1a ARs, high potency in a functional alpha-1A AR assay (perfused rat kidney;K b < 1 nM), but has low functional potency (K b > 30 nM) in isolated human prostate and prostatic urethra (Ford et al., 1995, 1996a). These results suggest that the functional potency on lower urinary tract tissues of most, but not all, alpha-1 AR antagonists correlates with their affinity for the recombinant alpha-1a AR.
Despite comparison of different species, and comparison of an in vitro to an in vivo assay, an excellent correlation is obtained between the potency of the entire series of alpha-1 AR antagonists against NE-induced contraction in isolated rabbit urethra with their ability to block NE-induced urethral contraction when administered intravenously in the anesthetized dog. All compounds, including SNAP 5089 and Rec 15/2627, fall on this regression line, and an excellent correlation (R2 = 0.861) is obtained with all compounds included (fig. 4).
Although some of the evidence cited above points to the involvement of the alpha-1a AR in lower urinary tract tissues and human prostate contraction, there is also evidence to support a functional role of the alpha-1L AR (Muramatsu et al., 1994,1995). The alpha-1L AR is defined by its relatively low affinity for prazosin in functional assays (Flavahan and Vanhoutte, 1986; Muramatsu et al., 1990) and can be distinguished from the alpha-1N subtype by use of the high-affinity selective antagonist HV 723 (Muramatsu et al., 1990). OurK b values in rabbit prostate and urethra are in the range reported for prazosin at the alpha-1L AR in other tissues (Muramatsu et al., 1990, 1995), and the low-potency value of HV 723 in the same tissues suggests the absence of thealpha-1N subtype.
Although affinity for the alpha-1L AR can only be quantitated in functional assays, because the receptor has not yet been cloned and reliable radioligand binding assays are not available, it appears that many alpha-1 AR antagonists have equivalent affinity for alpha-1a and alpha-1L ARs (Muramatsuet al., 1996). If contraction of the smooth muscle of the lower urinary tract is indeed mediated by the alpha-1L AR, antagonist potency at this site might correlate well withalpha-1a affinity for most compounds, with the only compounds deviating from the regression line being those with substantially lower affinity for the alpha-1Lvis-a-vis alpha-1a AR. Such compounds might include SNAP 5089, Rec 15/2627 and RS 17053. Although the alpha-1L AR was defined based on its low affinity for prazosin, these three compounds may differentiate alpha-1L and alpha-1a ARs more clearly than prazosin. The radioligand binding results would also be consistent with the presence of an alpha-1L AR and absence of an alpha-1N AR in canine prostate. Although most of the antagonists have similar potencies against [3H]prazosin binding in canine prostate and recombinant human alpha-1a AR, in agreement with the data reported by Goetz et al.(1994), some of the antagonists are substantially weaker in the prostate, Rec 15/2627 and SNAP 5089 being the compounds showing the greatest affinity difference. We evaluated the functional affinity in dog prostate (substantially with the same method used to evaluate affinity in rabbit prostate) of Rec 15/2739 and Rec 15/2627 (data not shown). The pKb values obtained (9.22 and 7.92, respectively) were very close to the pKi values obtained in radioligand binding studies performed on the same organ (9.1 and 8.1, respectively). Moreover, our binding data in canine prostate of tamsulosin, prazosin, terazosin, alfuzosin, 5-methylurapidil and phentolamine strictly correlate with the functional data (pA2) recently reported by Buckner et al. (1996)in the same organ (R2 = 0.880), which suggests no discrepancies between binding and functional assay on this tissue. TheK D for [3H]prazosin at thealpha-1 ARs of canine prostate is also about 4-fold higher than that observed in the canine aorta. These results suggest that, although a monophasic association of [3H]prazosin with the alpha-1 ARs of canine prostate is observed, the binding may represent labeling of both alpha-1A andalpha-1L ARs. The difficulty in separating the effects onalpha-1A and alpha-1L ARs could be partly explained by the recent finding that the alpha-1L AR may be a different conformer of the alpha-1A subtype, as recently suggested by Ford et al. (1996b) with regard to its functional activity. In human prostate, Ford et al. (1995)observed a biphasic displacement of [3H]prazosin binding by RS 17053. We tested this compound in our assay in canine prostate and did not consistently observe biphasic inhibition curves. However, its high ratio between affinity in canine prostate and affinity for the human recombinant alpha-1a AR (pKi values = 7.41 and 9.24, respectively; data not shown) again suggests that [3H]prazosin may be labeling an alpha-1L AR in this tissue.
Regarding the vascular tissues, our data show that the ability of the tested series of alpha-1 AR antagonists to inhibit [3H]prazosin binding to membranes of canine aorta and their functional potency at rabbit aortic alpha-1 ARs gives the best correlation with alpha-1b and alpha-1d AR affinity, respectively. Several authors have reported and Vargas and Gorman (1995) have recently reviewed that rabbit and canine aorta functionally express three alpha-1 ARs: alpha-1A,alpha-1B and alpha-1L subtypes. The presence of these three subtypes could interfere with the interpretation of correlation analysis. It has been suggested that the identification of a single vascular alpha-1 AR in vitro may vary with the experimental conditions and design, e.g., the number and range of antagonist concentrations used in competition binding studies or in generating a Schild plot and the use of a limited number of subtype selective antagonists (Vargas and Gorman, 1995). Nevertheless, the correlation coefficients obtained by us comparing the binding affinity for the recombinant alpha-1 AR subtypes and the binding or functional affinity in these vascular tissues seem to exclude the presence or the functional relevance of thealpha-1A subtype (R2 values always < 0.06), whereas the presence of the other subtypes can not be ruled out. Further studies are needed to clarify this point.
In conclusion, our results indicate that NE-induced contraction of tissues of the rabbit urogenital tract is probably mediatedvia the alpha-1L rather than thealpha-1A AR. The alpha-1L AR also appears to be labeled by [3H]prazosin in binding assays to canine prostate, but not aorta, and is responsible for the urethral contraction induced by intravenous NE administration to the anesthetized dog. Antagonists such as Rec 15/2739, which have selectivity for the alpha-1L AR relative to thealpha-1B and alpha-1D ARs, show uroselectivity in these in vitro and in vivo models. The apparent correlation of antagonistic activity in tissues of the lower urinary tract with affinity for the recombinant alpha-1a AR is probably a consequence of the fact that most alpha-1 AR antagonists cannot discriminate between alpha-1A andalpha-1L ARs. Although the uroselective alpha-1 AR antagonists thus far identified, including Rec 15/2739, have high affinity for the alpha-1A AR, the sole involvement of thealpha-1A AR in the contraction of urogenital tissues is inconsistent with the low potency on these tissues of antagonists such as the 1,4-dihydropyridine SNAP 5089 and the quinazoline analog Rec 15/2627. This point will warrant further research. Whatever the subtypes involved, our results with Rec 15/2739 and some of its analogs substantiate the possibility of obtaining molecules endowed with high functional selectivity for the alpha-1 ARs of the lower urinary tract.
Acknowledgments
The authors thank Dr. D. Colombo, P. Angelico, M. Ibba and C. Taddei of Recordati S.p.A, for their collaboration in obtaining the experimental data included in this paper.
Footnotes
-
Send reprint requests to: Rodolfo Testa, Pharmaceutical R&D Division, RECORDATI S.p.A., Via Civitali 1, 20148, Milano, Italy.
- Abbreviations:
- BPH
- benign prostatic hyperplasia
- α1-AR
- alpha-1 adrenoceptor
- NE
- norepinephrine
- UP
- urethral pressure
- DBP
- diastolic blood pressure
- MOPS
- 4-morpholinepropanesulfonic acid
- TRIS
- tris(hydroxymethyl)methylamine
- PE
- polyethylene
- COS-7 cells
- CV-1 monkey kidney epithelial cells, SV 40
- CHO cells
- Chinese hamster ovary cells
- Received April 29, 1996.
- Accepted February 5, 1997.
- The American Society for Pharmacology and Experimental Therapeutics