Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Benign prostatic hyperplasia is a hormone-dependent disease in which progressive enlargement of the prostate leads to bladder outlet obstruction and disturbance in urinary outflow. Alpha adrenoceptor-mediated contraction of prostate tissue has been demonstrated in humans (Caine et al., 1976) as well as in several other species, including dogs and rabbits (Hieble et al., 1986; Honda et al., 1985). The use of selective alpha-1 antagonists in the therapy of BPH is therefore widely accepted (Monda and Oesterling, 1993; Lepor, 1993). The alpha-1 blockers currently marketed for BPH show, however, poor selectivity for the lower urinary tract and their systemic side effects, primarily attributable to blockade of vascular alpha-1 ARs, limit their acceptance and clinical utility. Therefore, there is a widespread interest in the design of alpha-1 AR antagonists having prostate versus vascular selectivity.

One approach to the design of prostate selective antagonists is based on the cloning of three distinct alpha-1 AR proteins, corresponding to three subtypes identified in native tissues. These subtypes have been designated as alpha-1A (alpha-1a), alpha-1B (alpha-1b) and alpha-1D (alpha-1d), with lower case subscripts being used for recombinant receptors and upper case subscripts for receptors in native tissues (Hieble et al., 1995). Distribution of mRNA for these alpha-1 AR subtypes shows that the message for the alpha-1a AR is predominant in human prostate (Price et al., 1993), and autoradiographic studies in human prostate show that the alpha-1A AR is localized to prostatic stroma (Lepor et al., 1995). These studies suggest a role for the alpha-1A AR in mediating the prostatic contractile response to alpha-1 AR activation, and are consistent with the ability of potent and selective alpha-1A AR antagonists to produce functional blockade of NE-induced contraction of human prostate and with the good general correlation of functional antagonist potency in human prostate with affinity for the recombinant alpha-1a AR (Forray et al., 1994a, b; Testa et al., 1996a). However, compounds have been identified, such as RS 17053 and abanoquil, which have high affinity for the alpha-1a AR and yet are very weak functional antagonists of NE-induced contraction of human prostate (Ford et al., 1995, 1996a; Marshall et al., 1992).

There is also evidence, based both on radioligand binding (Takeda et al., 1993; Maruyama et al., 1996) and functional experiments (Muramatsu et al., 1994) for an additional alpha-1 AR population, distinct from either alpha-1A, alpha-1B or alpha-1D ARs, in human prostate. These alpha-1 ARs, designated as alpha-1L ARs based on their relatively low sensitivity to blockade by prazosin, have been characterized functionally in several blood vessels (Muramatsu et al., 1990, 1995). A series of novel alpha-1 AR antagonists was identified by us (Leonardi et al., 1992); the best studied was Rec 15/2739 whose selectivity for the alpha-1a AR subtype as well as its uroselectivity have been reported previously (Testa et al., 1994a, b,1995; Blue et al., 1996; Kenny et al., 1996b). In part I of this paper (Leonardi et al., 1997, companion paper) we hypothesized that the potency in inhibiting NE-induced contractions of tissues of the lower urinary tract of several compounds, including Rec 15/2739, could be related to their affinity for the alpha-1L AR subtype rather than for the alpha-1a subtype. The in vitro potency of prazosin on rabbit urethra and prostate, in fact, was low, giving pK_b values of 8.11 and 7.8, respectively. On these tissues, the potency of SNAP 5089 and Rec 15/2627, two alpha-1a subtype potent and selective compounds, was also lower than that expected on the basis of their binding affinity for the alpha-1a subtype. On the other hand, the in vivo potency of the tested compounds against NE-induced increase of pressure in dog urethra was shown to give a reasonable correlation with the in vitro functional potency observed in rabbit urethra, but not with the binding affinity for the alpha-1a AR subtype.

The aim of the present work was therefore to extend the in vitro and in vivo studies by adding other subtype-selective alpha-1 antagonists, and by evaluating also their functional affinity for the putative alpha-1L AR subtype expressed in rabbit aorta pretreated with CEC, as reported by Oshita et al. (1993) and Muramatsu et al. (1996). The structure of the Rec derivatives evaluated in the present study are shown in figure 1; some compounds whose structures are reported in part I have been excluded. Some of the data on Rec 15/2739 and the reference standards have already been submitted (Leonardi et al., 1997, companion paper) or presented in abstract form (Testa et al., 1996b).

View larger version (15K): [in this window] [in a new window]   Fig. 1.   Structure of the new alpha-1 antagonists evaluated in this study.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Male Sprague Dawley rats (180-240 g), New Zealand White rabbits (2.5-3.5 kg) 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.

In the present study, the following compounds were evaluated: BMY 7378, RS 17053, Rec 15/2731, Rec 15/2784, Rec 15/3011, Rec 15/2615 and Rec 0/0232. The data of the other compounds cited were previously obtained with the same methods (Leonardi et al., 1997, companion paper), with the exception of the functional affinity for the alpha-1L AR subtype that was evaluated for all the compounds discussed in this paper.

Affinity for the native alpha-1A AR subtype. The affinity of tested compounds for the native alpha-1A AR subtype was evaluated by use of rat hippocampal membranes pretreated with CEC, according to the method reported previously (Testa et al., 1993).

Affinity for the recombinant animal and human alpha-1a, alpha-1b and alpha-1d AR subtypes. The affinity of the tested compounds for the recombinant rat alpha-1d, hamster alpha-1b and bovine alpha-1a ARs expressed in CHO cells and human alpha-1 AR subtypes expressed in CHO cells was evaluated as described previously (Testa et al., 1995).

Affinity for canine prostate alpha-1 ARs. The affinity of the tested compounds for the canine prostate alpha-1 ARs was evaluated according to the method reported previously (Leonardi et al., 1997, companion paper).

Functional in vitro alpha-1 antagonistic activity. The functional alpha-1 antagonistic activity of Rec 15/2739 and the other tested compounds against NE-induced contractions of rabbit urethra and prostate was evaluated according to a previously published method (Testa et al., 1993).

Functional affinity for the alpha-1L AR subtype. The functional affinity of the tested compounds for the alpha-1L AR was evaluated as antagonism against NE-induced contraction of rabbit aorta pretreated with CEC (Oshita et al., 1993) according to the following method: Adult rabbits were sacrificed by cervical dislocation. The aorta was removed, placed in Krebs-Henseleit buffer and dissected free of adhering tissue. Rings were prepared from each artery (8 rings per aorta, about 4-5 mm wide) and suspended in 20-ml organ baths containing Krebs' bicarbonate buffer of the following composition (mM): NaCl, 112.0; KCl, 5.0; CaCl₂, 2.5; KH₂PO₄, 1.0; MgSO₄, 1.2; NaHCO₃, 12.0; and glucose, 11.1, equilibrated at 37°C with 95% O₂:5% CO₂. Desmethylimipramine (0.1 µM) and corticosterone (1 µM) to block neuronal and extraneuronal uptake of NE, (±)-propranolol (1 µM) to block beta adrenoceptors and yohimbine (0.1 µM) to block alpha-2 adrenoceptors were added to the buffer. Tissues were subjected to a passive load of 2 g, and determination of the developed tension was measured by an isometric transducer (Basile 7003). The preparations were allowed to equilibrate for 60 min and then challenged every 30 min with 10 µM NE (three times). After washing, the tissues were incubated with the alkylating agent (5 × 10⁵ M) for 30 min and then extensively washed three times (in 0.5 hr) before constructing the NE concentration-response curve. After washout of NE and re-equilibration of the tissue (45 min), the drug to be tested was added; after 30 min, a second NE cumulative concentration-response curve was constructed. Proper control preparations treated with CEC and following vehicle incubation were prepared to identify any time-dependent change of NE response. Each antagonist concentration was tested with use of two to three aortic rings from different rabbits.

In vivo selectivity for the lower urinary tract. The selectivity of Rec 15/2739 and the other tested compounds for the lower urinary tract tissues versus the cardiovascular system was evaluated by assessing the inhibition of urethral pressure elevated by i.a. NE injection and the effects on DBP in anesthetized dogs, according to the method reported previously (Leonardi et al., 1997, companion paper).

Compounds and solutions. The following compounds were used: [³H]prazosin (7-methoxy-³H), (NEN Life Science Products, Cologno Monzese, Milano, Italy); norepinephrine tartrate, prazosin-HCl, phentolamine-HCl (all Sigma-Aldrich, Milano, Italy); compounds of figure 1, alfuzosin-HCl, SNAP 5089-HCl, terazosin-HCl and tamsulosin-HCl (all synthesized in Recordati Laboratories). RS 17053-HCl was kindly given by Syntex (USA) Inc., Palo Alto, CA. For the other Rec derivatives, see part I of this paper (Leonardi et al., 1997).

Statistical analysis. The displacement curves of the antagonists in the binding studies were analyzed by nonlinear curve fitting of the logistic equation according to the method reported by De Lean et al. (1978), with 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).

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Affinity for the native alpha-1A AR subtype and recombinant animal and human alpha-1a, alpha-1b and alpha-1d AR subtypes. The affinity of Rec 15/2739 and the other tested compounds for the cited alpha-1 AR subtypes is listed in table 1. Rec 15/2739 and some of its related compounds (namely Rec 15/2841, Rec 15/3011 and Rec 15/2869) were selective for the alpha-1a subtype. Among the other N-arylpiperazine derivatives, BMY 7378 was selective for the alpha-1d subtype, whereas the remaining compounds were nonselective compounds. The reference quinazoline derivatives prazosin, terazosin, alfuzosin, Rec 15/2627 and the abanoquil analog Rec 0/0232 were also nonselective. On the other hand, the newly synthesized quinazoline derivative Rec 15/2615 was alpha-1b subtype selective. The alpha-1a subtype selectivity of the other reference antagonists tested, in particular SNAP 5089 and RS 17053, was confirmed.

View this table: [in this window] [in a new window]   TABLE 1 Affinity of Rec 15/2739, related compounds and reference drugs for the native alpha-1A AR and for recombinant alpha-1 AR subtypes Data represent the mean pKi (log10 Ki) values (at least two duplicate determinations) for inhibition of [3H]prazosin binding to membrane homogenates.

Affinity for alpha-1 ARs in different animal tissues. The binding affinities of the tested compounds for the canine prostate alpha-1 ARs, as well as their functional affinities (antagonism against NE-induced contraction) for the alpha-1 ARs of rabbit prostate and urethra are listed in table 2. The functional affinities of the considered compounds for the alpha-1L AR subtype present in rabbit aorta after CEC incubation (Muramatsu et al., 1996) are also listed in table 2. NE caused concentration-dependent contractions of rabbit aorta. After CEC incubation (5 × 10⁵ and 1 × 10⁴ M), the cumulative concentration-response curves of NE were slightly shifted to the right (the EC₅₀ value for NE before CEC incubation was 0.24 µM, and EC₅₀ values after CEC incubation were 1.27 and 1.43 µM, respectively) with no depression of the maximal tension attainable, which indicated that part of the contraction induced by NE is caused by a CEC-resistant alpha-1 AR subtype. These results are in full agreement with those recently reported by Auguet et al. (1995). Most of the compounds tested in this model behaved as competitive antagonists; the slope of Schild plot was not different from unity (figs. 2 and 3), with the exception of Rec 0/0232, Rec 15/2627 and RS 17053 which deviate from the unity. An apparent pK_b was evaluated for these compounds.

View larger version (15K): [in this window] [in a new window]   Fig. 2.   Representative Schild analysis of the effects of the N-arylpiperazine listed in table 2 against NE-induced contractions of rabbit aorta (preincubated with CEC). For clarity, the mean of the DR-1 (dose-ratio-1) obtained with different preparations at each concentration tested was plotted. The slopes of the curves were not different from unity and were constrained to unity. Symbols: squares, Rec 15/2739; upper triangles, Rec 15/2869; circles, Rec 15/3011; stars, Rec 15/2841; diamonds, Rec 15/2731; upper semicircle, Rec 15/2874; lower semicircles, Rec 15/2636; lower triangles, Rec 15/2802; hexagons, BMY 7378.

View larger version (14K): [in this window] [in a new window]   Fig. 3.   Representative Schild analysis of the effects of the reference compounds listed in table 2 against NE-induced contractions of rabbit aorta (preincubated with CEC). For clarity, the mean of the DR-1 (dose-ratio-1) obtained with different preparations at each concentration tested was plotted. The slopes of the curves were not different from unity and were constrained to unity. Symbols: squares, tamsulosin; upper triangles, phentolamine; circles, SNAP 5089; stars, prazosin; diamonds, terazosin; upper semicircle, alfuzosin; lower semicircles, Rec 15/2615.

In vivo selectivity for the lower urinary tract. The previously published results (Leonardi et al., 1997, companion paper) obtained by evaluating the compounds in the dog model, and the additional results obtained during the present investigation for other reference compounds, as well as other compounds structurally related to Rec 15/2739, are listed in table 3. Among the molecules structurally related to Rec 15/2739, Rec 15/3011, Rec 15/2869 and Rec 15/2841 also showed relevant selectivity (evaluated as the ratio between the effective dose lowering blood pressure and the potency at the urethral level), clearly higher than that of the reference compounds studied. Rec 15/2802 proved as uroselective as SNAP 5089, but it was clearly more potent at the urethral level. On the other hand, other structurally related N-arylpiperazines, namely Rec 15/2731 and Rec 15/2784, proved poorly selective or, like Rec 15/2636, nonselective. Prazosin and the other quinazoline derivatives (terazosin, Rec 15/2615, Rec 15/2627 and the analog of abanoquil Rec 0/0232) were not selective. In this class of compounds, only alfuzosin showed some degree of selectivity. The reference alpha-1a selective compounds used and belonging to different chemical classes, RS 17053, tamsulosin and phentolamine, were similar to alfuzosin, whereas SNAP 5089 proved relatively selective, albeit weakly potent on the urethra.

Correlation analysis. By linear regression analysis, we evaluated the correlation between the binding affinity for the animal and human alpha-1 AR subtypes of the compounds tested and their affinity for the alpha-1 ARs present in the rabbit prostate and urethra (functional) or dog prostate (binding). The results of these correlation analyses are shown in table 4.

View larger version (15K): [in this window] [in a new window]   Fig. 4.   Correlation between the binding affinity (pKi) of the different compounds tested for the recombinant human alpha-1a AR subtype (left) or their functional affinity (pKb) for the alpha-1L AR subtype (right), and their functional affinity (pKb) for the alpha-1 ARs of rabbit urethra. The continuous line represents the calculated equation, and the dotted line represents the line of identity. Outliers (filled symbols) were not included in the regression.

View larger version (15K): [in this window] [in a new window]   Fig. 5.   Correlation between the binding affinity (pKi) of the different compounds tested for the recombinant human alpha-1a AR subtype (left) or their functional affinity (pKb) for the alpha-1L AR subtype (right), and their functional affinity (pKb) for the alpha-1 ARs of rabbit prostate. The continuous line represents the calculated equation, and the dotted line represents the line of identity. Outliers (filled symbols) were not included in the regression.

View larger version (15K): [in this window] [in a new window]   Fig. 6.   Correlation between the binding affinity (pKi) of the different compounds tested for the recombinant human alpha-1a AR subtype (left) or their functional affinity (pKb) for the alpha-1L AR subtype (right), and their binding affinity (pKi) for the alpha-1 ARs of dog prostate. The continuous line represents the calculated equation, and the dotted line represents the line of identity. Outliers (filled symbols) were not included in the regression.

View larger version (13K): [in this window] [in a new window]   Fig. 7.   Correlation between the binding affinity (pKi) of the different compounds tested for the recombinant human alpha-1a AR subtype (left) or their functional affinity (pKb) for the alpha-1L AR subtype (right), and their in vivo potency (log ED50 values) as antagonists of NE-induced contractions of dog urethra. Outliers (filled symbols) were not included in the regression.

View larger version (12K): [in this window] [in a new window]   Fig. 8.   Correlation between the binding affinity (pKi) of the different compounds tested for the recombinant animal (left) or human (right) alpha-1d AR subtype, and their in vivo hypotensive effects (log ED25 values) in the dog model.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

In part I of this paper (Leonardi et al., 1997, companion paper) we reported that Rec 15/2739, and other N-phenylpiperazine derivatives structurally related to it (Rec 15/2841 and Rec 15/2869), were markedly more selective for the lower urinary tract tissues than all the reference compounds studied. These compounds, in fact, were clearly more potent in antagonizing the increase in UP induced by NE than in lowering DBP in dogs. The studies described in this paper show that at least another phenylpiperazine derivative (Rec 15/3011) was markedly uroselective, whereas others (Rec 15/2731 and Rec 15/2784) were not. In addition, other compounds included in the present paper such as the alpha-1a subtype selective compound RS 17053 (Ford et al., 1995, 1996a), the abanoquil analog Rec 0/0232 (fig. 1), the alpha-1b subtype selective quinazoline Rec 15/2615 (fig. 1) and the alpha-1d subtype selective BMY 7378 (Saussy et al., 1996) proved nonselective in our dog model. This study therefore extends previous reports (Poirier et al., 1988; Breslin et al., 1993; Testa et al., 1994a; Kenny et al., 1994, 1996a, 1996b) showing that alpha-1 AR antagonists decrease urethral resistance in in vivo animal models, consistent with clinical reports describing improvements in urinary flow with these agents (Wilde et al., 1993). Changes in intraurethral pressure measured in the prostatic portion in these models are, in fact, reported to be caused by contraction of both the prostate gland and urethral smooth muscle (Brune et al., 1995).

To clarify whether or not the uroselectivity observed for some compounds could be related to a different affinity profile for the alpha-1 AR subtypes, we evaluated their binding affinity for the recombinant animal and human alpha-1 AR subtypes, as well as their in vitro affinity (binding or functional) for the alpha-1 ARs of different isolated tissues of the lower urinary tract. A point of criticism for these comparisons could be the correlation of data from binding and functional experiments, without a direct comparison of these two methods in a single tissue. On the other hand, we have previously reported (Leonardi et al., 1997, companion paper) that binding and functional estimates of affinity for Rec 15/2739 and Rec 15/2627 on dog prostate were the same. In the present study, we extended this comparison. We found that the binding affinities (pK_i) of Rec 15/2615 and RS 17053 on dog prostate (table 2) were very close to their functional affinities (evaluated as antagonism to NE-induced contraction) evaluated on the same tissue (pK_b values were 7.13 and 7.58, respectively; data not shown).

The functional affinity for the putative alpha-1L AR subtype was also evaluated as antagonism of NE-induced contraction in rabbit aorta pretreated with CEC (Muramatsu et al., 1996). Although several studies, in fact, point to the involvement of the alpha-1a AR in lower urinary tract tissues and human prostate contraction (Forray et al., 1994a,b; Price et al., 1993; Faure et al., 1994; Goetz et al., 1994), there is also evidence to support a functional role of the alpha-1L AR (Hiraoka et al., 1995; Muramatsu et al., 1994). The alpha-1L AR is defined by its relatively low affinity (pK_b < 9.0) for prazosin (Muramatsu et al., 1990), in contrast with the high affinity (pK_b > 9.0) of this compound for the other alpha-1 AR subtypes. Genes encoding the alpha-1 AR possess introns, and recently, splice variants of the alpha-1A AR have been described (Hirasawa et al., 1995). Moreover, the inhibition of NE-stimulated accumulation of [³H]inositol phosphates in CHO cells, which stably express the human alpha-1A AR (Ford et al., 1996b), shows that, in this assay, the human alpha-1a clone displays pharmacological recognition properties similar to the putative alpha-1L AR. Whether these characteristics explain the alpha-1L AR pharmacology observed in many native tissues requires further investigation.

Our K_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). In our previous paper (Leonardi et al., 1997, companion paper), moreover, we showed that 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. No significant correlation of these data with alpha-1b or alpha-1d AR affinity was in fact obtained. However, two compounds, the dihydropyridine SNAP 5089 and the quinazoline Rec 15/2627, deviated substantially from the line of identity between functional activity or potency in the isolated urogenital tissues and alpha-1a AR affinity. Elimination of these two compounds improved the correlation.

The present paper, extended to other alpha-1 AR antagonists, including the alpha-1a AR subtype selective compound RS 17053 (Ford et al., 1995, 1996a), confirms our previous observation that the affinity for cloned alpha-1a ARs does not entirely explain the antagonist potency on the alpha-1 ARs mediating contractile responses to NE in the lower urinary tract tissues both in vitro and in vivo. The potencies of RS 17053 and the quinazoline derivative Rec 15/2615 on dog and rabbit prostate and urethra were lower than expected on the basis of their affinity for the alpha-1a AR subtype, as was observed with SNAP 5089 and Rec 15/2627. The low potencies of RS 17053 and SNAP 5089, relative to their alpha-1a affinity, as antagonists of NE-induced contraction of human prostate were recently confirmed by other authors (Marshall et al., 1996; Kenny et al., 1996b).

The relatively low affinity exhibited by prazosin and the other compounds cited above against alpha-1-mediated contractile responses in the tissues of the lower urinary tract, in contrast to their affinity for recombinant alpha-1a ARs, suggests the presence of a distinct receptor which has the characteristics of the alpha-1L AR. We therefore evaluated the functional affinity of these compounds for the alpha-1L AR by use of rabbit aorta pretreated with CEC, according to the suggestions of Muramatsu et al. (1996). In our experience, this vessel gives results very similar to those obtained with canine femoral artery or vein, other tissues which functionally express the alpha-1L AR (Kohno et al., 1994). The pK_b values of prazosin, 5-methylurapidil and WB 4101 obtained by us in rabbit aorta after CEC alkylation (pK_b values = 7.7, 7.9 and 8.3, respectively) were in fact close to those reported by Kohno et al. (1994) in dog femoral vein for the same compounds (pK_b values = 8.3, 8.0 and 8.4, respectively).

Despite comparison of different species, and comparison of an in vitro with 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 CEC-pretreated rabbit aorta, and their binding affinity for the alpha-1 ARs of dog prostate (fig. 6), ability to block NE-induced contraction of isolated rabbit prostate and urethra (Fig. 4 and 5), or potency against urethral contractions in the anesthetized dog (fig. 7). All compounds fall on the regression line, without the need for discarding outliers. The correlations between alpha-1a AR and alpha-1L AR affinity and in vivo potency on UP in the dog are very similar to those that might be extrapolated by use of the data of Blue et al. (1996) with a different set of alpha-1 AR antagonists, including Rec 15/2739, prazosin, tamsulosin, terazosin and RS 17053. Their data on the ID₅₀ (UP), representing the intravenous dose required to produce 50% blockade of nerve stimulation-induced increase of UP, were poorly related with the affinity for the recombinant animal alpha-1a AR subtype (R² < 0.2; n = 9), but correlated very well with the functional affinity for the alpha-1L AR subtype (R² > 0.8; n = 9). 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, our data indicate that many alpha-1 AR antagonists have equivalent affinity for alpha-1a and alpha-1L ARs. 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 with alpha-1a affinity for most compounds, with the only compounds deviating from the regression line being those with substantially lower affinity for the alpha-1L vis-a-vis alpha-1a AR. Such compounds might include SNAP 5089, Rec 15/2627, Rec 15/2615 and RS 17053. Again, the good correlations obtained between the affinity for the alpha-1L AR subtype and those for the dog and rabbit prostate and urethra, as well as between these tissues (table 6) confirm the assumption that NE-induced contraction of these tissues is mediated primarily by this subtype (Muramatsu et al., 1995; Leonardi et al., 1997, companion paper).

In the context of uroselectivity, the findings reported in the present paper that closely relate the hypotensive effects of the alpha-1 antagonists tested with their affinity for the alpha-1D AR subtype are in agreement with the hypothesis suggested recently by Kenny et al. (1996a). Whether the hypotension observed in the anesthetized dog is caused by vascular alpha-1d AR blockade or could be related to the blockade of cardiac (reduced cardiac output) or central (reduced sympathetic tone) alpha-1d ARs can not be explained by the present data. We have previously demonstrated (Leonardi et al., 1997, companion paper) that the affinity of a series of compounds for the alpha-1 ARs in dog and rabbit aortas correlated with the alpha-1b and alpha-1d AR subtypes, respectively, and excluded the involvement of the alpha-1a ARs. On the other hand, Vargas and Gorman (1995), based on accumulated evidence, reported that all the cited alpha-1 AR subtypes exist and mediate contractile responses in vascular tissues, the vascular alpha-1A AR subtype having a primary role in the maintenance of arterial pressure, at least in the rat. Although it is not yet known which alpha-1 AR subtype is primarily involved in the maintenance of arterial pressure in dog, our data do not support this premise. If this was true, selective alpha-1a AR antagonists such as SNAP 5089 and RS 17053 would be potent hypotensive agents, but this is not the case. Moreover, preliminary data shown by Gong et al. (1994) suggest that selective alpha-1a AR antagonists have reduced orthostatic liability in the rat, in comparison with nonselective compounds.

In conclusion, on the basis of the data in this study, selectivity for the prostatic alpha-1L AR relative to the alpha-1D subtype appears to be the best index of prostatic versus vascular selectivity, at least in the dog model used.