Mode of action of α1-adrenoreceptor antagonists in the treatment of lower urinary tract symptoms
The obstructive (voiding) symptoms of BPH have been attributed to two factors ; the physical mass of the enlarged gland (the static component) and the tone of the smooth muscle of the prostatic stroma (the dynamic component). Irritative (storage) symptoms have been more closely associated with the bladder dysfunction produced by outlet obstruction. However, the evidence for a direct link between BPH, BOO and symptoms is not convincing , and furthermore, both voiding and storage symptoms have also been shown to occur in women [3,4].
A significant number of patients with BPH will respond to α-adrenoreceptor (AR) antagonists with symptom improvement . As prostatic α-ARs are considered to mediate the dynamic component of obstruction, it is natural that these receptors have been assumed to mediate at least part of the symptoms. Shapiro et al. found a direct relationship between the amount of prostatic smooth muscle and dynamic obstruction (as assessed by the response to α1-AR blockade). Such a relation was further supported by in vitro findings showing that the contractile response to phenylephrine of human BPH tissue was positively correlated with the density of smooth muscle . However, many patients who have undergone prostatectomy, which should relieve obstruction, still experience persistent storage symptoms. Interest has therefore begun to focus on mechanisms outside the prostate in the pathogenesis of (particularly) storage symptoms. LUTS can theoretically be caused by several factors, i.e. changes in the smooth muscle of the urinary tract (hypertrophy, denervation), defects in the central processing of afferent information or in efferent neurotransmission, and changes in both afferent and efferent activity can occur.
Spinal α-ARs may be involved in the control of both the sympathetic, somatic (storage) and parasympathetic (voiding) efferent activity to the lower urinary tract [8–14]. α1-AR antagonists can have beneficial effects on the symptoms of BPH, even in the absence of BOO , which supports the involvement of extraprostatic α1-ARs in the pathogenesis of LUTS. Beside the prostate, there are several possible sites of action of α-AR antagonists, i.e. the prostatic stroma, detrusor, trigone, urethra, ganglia or spinal and/or supraspinal structures, some of which will be discussed.
α-ARs in the prostate
Stromal elements  and the density of stromal α1-ARs [16–18] are increased in BPH. α1-ARs occur in the glandular epithelium  and seem to be of the α1B-subtype . Although there may be regional differences in the binding to and functional properties of the prostatic α1-ARs , there is no evidence of clinically significant regional differences in the contractile responses to α1-AR agonists .
All the three high-affinity α1-AR subtypes identified in molecular cloning and functional studies (α1a/A, α1b/B, α1d/D) have been detected in prostatic stromal tissue [23–26]. The α1a subtype predominates, representing ≈ 70% of the α1-AR population [21,23,26,27]. Among the four recently identified isoforms of the α1A-AR (α1a/A–1, α1a/A–2, α1a/A–3, α1a/A–4), the α1a/A–1-AR mRNA/protein was the most abundant . Nasu et al. suggested that there may be differences in subtype populations between normal and hyperplastic prostate. They found the ratio of the numbers of the α1a, α1b, and α1d subtype mRNAs to be 85:1:14 in samples of BPH, and 63:6:31 in samples that were not BPH. Walden et al. found no significant difference in the overall level of α1-AR mRNA between normal and hyperplastic prostate, but they detected a difference in α1-AR subtype expression, with reduced expression of α1b-mRNA in both glandular and stromal hyperplasia.
Which α1-AR subtype mediates the contractile response to noradrenaline in the human prostate has been much debated. The α1A-AR appears to be important [24,29], but the involvement of other subtypes cannot be excluded . Ford et al.[31,32] suggested the possibility that multiple forms of the α1A-AR exist in the human prostate, or that another pharmacologically distinct α1-AR mediates the noradrenaline-induced contraction of prostatic smooth muscle. One candidate may be the receptor with low affinity for prazosin (the α1L-AR) suggested by Muramatsu et al.[33–35]. This receptor has not been cloned; it may represent a functional phenotype of the α1A-AR .
Postjunctional α2-ARs occur in the human prostate and it has been suggested that they may contribute to the dynamic component of prostatic contraction . The density of α2-ARs was higher in BPH than in normal prostate, and higher in symptomatic than in asymptomatic BPH [38–40]. However, at present there is no clear evidence for the involvement of α2-ARs in the pathogenesis of LUTS associated with BPH.
Although the beneficial effects of α1-AR antagonists in BPH have been attributed to the relaxant action on prostatic and urethral smooth muscle, additional mechanisms have been suggested. Evidence has indicated that the α1-AR antagonist, doxazosin, induces prostatic apoptosis in patients with BPH . It was unclear whether this potentially beneficial effect was specific for doxazosin, or if it was attributable to the principle of α1-AR blockade. Chon et al. showed that the apoptotic response was not specific to doxazosin, but mediated by an α1-AR-dependent mechanism. In patients with BPH given terazosin (1–10 mg/day; n = 42), or doxazosin (2–8 mg/day; n = 61) for periods varying from 1 week to 3 years, there was a significant induction of apoptosis in both the prostatic epithelial and stromal cells within the first month of terazosin and doxazosin therapy, compared with 31 untreated control patients (P < 0.05). The marked induction of prostatic stromal apoptosis in response to both α1-AR antagonists was paralleled by a significant decrease in smooth muscle α-actin expression. The loss of prostatic smooth muscle cells correlated with morphological stromal regression and BPH symptom improvement. Neither terazosin nor doxazosin therapy resulted in significant changes in prostate cell proliferation. The authors concluded that α1-AR antagonists may regulate prostate growth by inducing apoptosis in both epithelial and stromal cells, with little effect on cell proliferation, and suggested that prostatic stromal regression mediated by apoptosis may be an additional molecular mechanism underlying the therapeutic response to α1-AR blockade in patients with BPH.
Boesch et al. investigated the effect of phenylephrine and doxazosin on the expression of smooth muscle myosin heavy-chain isotypes SM-1 and SM-2 in an in vitro model of prostatic smooth muscle cells. After 6 days of treatment, SM-2 expression increased, the increase being highest in the doxazosin-treated cultures. Compared with unstimulated cells, there was a statistically significant 10-fold increase of the SM-2:SM-1 ratio in doxazosin-treated cultures. The authors suggested that doxazosin may have a long-term effect on the differentiation of prostatic stromal cells, underlining that α1-AR antagonists do not act solely on smooth muscle contractility. This was further supported by the study of Smith et al., showing that doxazosin may inhibit not only stromal contraction of differentiated smooth muscle cells in BPH, but also the phenotypic modulation of stromal smooth muscle cell differentiation induced by noradrenaline. It was suggested that these actions together may render prostatic stroma less contractile, and hence less able to respond to sympathetic stimulation, in patients with BPH.
α-ARs in the urethra
Urodynamic studies in humans have suggested that up to half of the intraurethral pressure is maintained by the stimulation of α1-ARs, as judged from results obtained with α1-AR antagonists [45,46]. In human urethral smooth muscle, both functional studies and receptor-binding studies have suggested that the predominant postjunctional α-AR is α1. Most in vitro investigations of human urethral α-ARs have been carried out in the male. However, there appear to be no differences between the sexes in the distribution of α1- and α2-ARs (as can be found in, e.g. rabbits), or in the distribution of α1-AR subtypes . Fukasawa et al. found the α1L-AR subtype to be more prominent in the human male urethra than in the human prostate. In the human female urethra, the expression and distribution of α1-AR subtypes were determined by in situ hybridization and quantitative autoradiography. mRNA for the α1a subtype was predominant and autoradiography confirmed the predominance of the α1A-AR. If it is assumed that urethral α1-ARs are contributing to LUTS irrespective of sex, an effect of α1-AR antagonists should also be expected in women with these symptoms. This was found to be the case in some studies [3,4], but was not confirmed in a randomized, placebo-controlled pilot study , which showed that terazosin was ineffective for treating prostatism-like symptoms in ageing women. On the other hand, Serels and Stein  presented data on a cross-over study comparing hyoscyamine and doxazosin in the treatment of urge, frequency and urge incontinence in women of various ages. Symptoms were assessed by an expanded AUA symptom score. These authors found that both drugs produced an improvement in ≈ 65% of the patients. Placebo-controlled, adequately powered clinical studies are required to establish whether or not LUTS can be relieved by α1-AR antagonists, irrespective of gender. If the results are positive the importance of extraprostatic α1-ARs will be established.
α-ARs in the bladder
In detrusor muscle from most species, including humans, relaxation-mediating β-ARs normally dominate over contraction-mediating α-ARs, and in isolated, normal human detrusor muscle, drugs selectively stimulating α-ARs, preferentially those acting on α1-ARs, produce a small and variable contractile effect . However, extrapolating in vitro experiments with human bladder strips to functional importance in vivo should be cautious. Some studies have shown that in patients with bladder overactivity associated with, e.g. BOO, or with neurogenic bladders and idiopathic bladder instability, there may be a shift from a β-AR-mediated relaxation to an α-AR-mediated contraction [52–54]. However, Smith and Chapple could not confirm any increase in α-AR function associated with the morphological changes occurring in bladder hypertrophy secondary to BOO [54,55].
The predominant postjunctional α-AR type in the human lower urinary tract seems to be α1, but which α1-AR subtype predominates in the detrusor, trigone and bladder base has been unclear, and it is not known whether there are any differences between sexes. Walden et al. reported a predominance of α1a-AR mRNA in the human bladder dome, trigone and bladder base. This contrasts with the results of Malloy et al., who found that among the high-affinity receptors for prazosin, only α1a and α1d-mRNA were expressed in the human bladder. The total α1-AR expression was low (at 6.3 ± 1.0 fmol/mg) but was reproducible. The expression of the different subtypes was α1d 66% and α1a 34%, with no expression of α1b. This is in contrast to what has been found in the human prostate. That drugs acting selectively on prostatic α1-ARs may have little effect on LUTS can theoretically be explained if they do not block the α1-ARs of the bladder (α1D).
Caine  suggested that α2-ARs may contribute to the effects of, e.g. phenoxybenzamine, in the treatment of BPH. However, postjunctional α2-ARs probably have little importance for bladder function. In rat detrusor, clonidine produced concentration-dependent inhibition of contractions evoked by electrical field stimulation [59,60]. This effect was suggested to be caused by stimulation of α2-ARs located on postganglionic nerve endings, leading to reduced output of excitatory neurotransmitters. In conscious rats, the subcutaneous administration of the selective α2-AR agonist dexmedetomidine reduced peak detrusor pressure, increased the frequency of voiding and produced urinary dribbling . Conversely, the selective α2-AR antagonist atipamezole increased bladder baseline pressure and produced urinary dribbling characteristic of overflow incontinence. Given intrathecally, dexmedetomidine decreased micturition pressure, bladder capacity, micturition volume and residual volume, and caused total incontinence .
The α2-AR agonists and antagonists probably affect micturition at both central (brain, spinal cord) and peripheral (ganglia, nerve terminals, lower urinary tract smooth muscle) sites of action, but to what extent they are involved in the pathogenesis of LUTS associated with BPH is unclear.
α-ARs on nerve terminals, in peripheral ganglia and in the spinal cord
Facilitatory α1-ARs have been detected on cholinergic neurones in the vesical ganglia of the cat [63–65], on dissociated bladder neurones from the rat major pelvic ganglion  and on cholinergic terminals in the rat bladder . An effect of α1-AR antagonists at the ganglionic and/or prejunctional level, leading to a decrease of acetylcholine release, cannot be excluded.
α2-ARs have been detected on adrenergic nerve terminals in both detrusor and urethral smooth muscle . When stimulated by noradrenaline these receptors inhibit the further release of the amine. Clonidine is known to reduce intraurethral pressure in humans . This may be attributed partly to a peripheral effect on adrenergic nerve terminals, leading to a decreased noradrenaline release. However, it is more likely that its effect is on the CNS, with a resulting decrease in peripheral sympathetic nervous activity. Supporting this view, clonidine produced a decrease in plasma noradrenaline concentration . Furthermore, in cats, clonidine depressed firing in the hypogastric nerves innervating the bladder, possibly by a spinal site of action . Denys et al. showed that intrathecal clonidine produced a significant decrease in detrusor hyper-reflexia in patients with spinal cord injuries, supporting the view that spinal α2-ARs are important for the control of the micturition reflex.
Descending spinal pathways concerned with micturition include serotonin- and noradrenaline-containing projections from the raphe nuclei and the locus coeruleus, respectively . From the locus coeruleus, the noradrenergic neurones supply sympathetic and parasympathetic nuclei in the lumbosacral spinal cord. Bladder activation through these bulbospinal noradrenergic pathways may involve excitatory α1-ARs . Thus, in the anaesthetized cat, electrical stimulation of the locus coeruleus induced bladder contractions that were antagonized by the intrathecal administration of prazosin [14,72–74]. Destruction of noradrenergic cells in the locus coeruleus by microinjection of 6-OH-dopamine produced a hypoactive bladder, which could be partly reversed by intrathecal injection of the selective α1-AR agonist, phenylephrine. However, in the conscious cat, Downie et al. found that intrathecal prazosin did not alter the micturition reflex. The reasons for these apparently conflicting results are unclear. Intravenous injection with prazosin or phentolamine in cats depressed the external urethral sphincter activity and reduced reflex firing in pudendal nerve efferent pathways by a presumed central site of action .
In the CNS, facilitatory α1-ARs, tonically active in both the sympathetic and somatic neural control of the lower urinary tract, were detected in the cat [8,13]. In normal rats and rats with bladder hypertrophy secondary to outlet obstruction, doxazosin, given intrathecally, decreased micturition pressure in both . The effect was more pronounced in those animals with hypertrophied/overactive bladders. Doxazosin did not markedly affect the frequency or amplitude of the unstable contractions observed in obstructed rats. It was suggested that doxazosin may have an action at the level of the spinal cord and ganglia, thereby reducing activity in the parasympathetic nerves to the bladder, and that this effect was more pronounced in rats with bladder hypertrophy than in normal rats.
Urodynamic studies revealed that spontaneously hypertensive rats (SHR) have a pronounced bladder overactivity . These animals also have an increased noradrenergic bladder innervation, and an increased voiding frequency . Whereas the control rats (Wistar Kyoto rats, WKY) have a regular contraction frequency during continuous cystometry, the SHR show contractions both during micturition and at other times. To investigate whether or not the peripheral adrenergic system was involved in the pathogenesis of the bladder overactivity, SHR were treated with 6-hydroxydopamine to chemically destroy the noradrenergic nerves. In these rats, bladder overactivity was maintained, as shown by continuous cystometry. Furthermore, α1-AR antagonists, injected intra-arterially near the bladder, did not abolish this bladder overactivity. However, when given intrathecally, the same amount of α1-antagonist restored normal micturition. If such actions are also relevant to the effects of α1-AR antagonists in humans, spinal α1-ARs may be appropriate targets for drugs intended to treat LUTS.
The neuronal location of α1-AR subtypes (α1a, α1b, α1d) in the human spinal cord was investigated recently . In situ hybridization studies revealed that α1-AR mRNA was present in ventral grey matter only (ventral > dorsal; sacral > lumbar = thoracic > cervical). Signalling cell bodies were detected in anterior horn motor neurones at all levels: the dorsal nucleus of Clarke and intermediolateral columns in cervical enlargement, thoracic and lumbar spinal cord regions; and the parasympathetic nucleus in sacral spinal cord. Although all three α1-AR subtypes were present throughout the human spinal cord, α1d-AR mRNA predominated overall. Whether this has any clinical significance in the treatment of LUTS remains to be established. The distribution of α1-AR mRNA subtypes in the rat spinal cord differs from that in humans , which should be considered when the clinical relevance of data obtained in rats is discussed.
α1-ARs and α1-AR antagonists in the treatment of BPH
The efficacy of α1-AR antagonists on LUTS in patients with BPH is widely accepted . Of the α1-AR antagonists currently in clinical use for the treatment of BPH (alfuzosin, doxazosin, prazosin and terazosin) have no selectivity for α1-AR subtypes or for the prostate. Tamsulosin is modestly selective for α1A (7–38 fold) and for α1D over α1B-AR [79,80]. As there is disagreement about which receptors and receptor sites are involved in the actions of the α1-AR antagonists, it may be questioned whether drugs with selective effects on, e.g. the α1A-AR, offer therapeutic advantages. A nonselective drug can be expected to have a maximal effect on outlet resistance and on LUTS, provided that the dose is high enough (as side-effects may be dose-limiting). A subtype-selective drug could therefore be expected to increase efficacy only if the maximum effect has not been achieved with the nonselective drugs. The ‘ceiling effect’ of presently used nonselective α1-AR antagonists may or may not have been established. If they have, efficacy in terms of effects on outlet obstruction and LUTS cannot be increased. It may still be possible to obtain a therapeutically advantageous reduction in side-effects.
The sympathetic nervous system has an important role in the regulation of urogenital function; α1-ARs seem to be involved in the control of micturition at several different levels, all of which may contribute to the overall effect profile of an α1-AR antagonist. The clinical significance of this effect has yet to be established, but the potential importance of the extraprostatic actions of α1-AR antagonists should be considered. It is therefore necessary to focus not only on the α1-ARs in the prostate when treating LUTS with α-AR antagonists, but also on extraprostatic α1-ARs.
α1-AR antagonists may have a central site of action. These sites of action are worthy of further study and may be targets for drugs intended to treat LUTS.
The role of bladder α-ARs in the pathogenesis of LUTS remains controversial.
This work was supported by grants from the Swedish Medical Research Council (grant no. 6837), and the Medical Faculty, University of Lund.
Karl-Erik Andersson, Department of Clinical Pharmacology, Lund University Hospital, S-221 85 Lund, Sweden.