An overview on 5α-reductase inhibitors

Introduction

Benign prostatic hyperplasia (BPH) is the noncancerous growth of the prostate gland resulting due to over-proliferation of the stromal and glandular elements of the prostate [1]. It is caused due to the augmented levels of the androgen dihydrotestosterone (DHT). In BPH, microscopic foci within specific regions of the prostate grows to form macroscopic nodules which eventually displace the normal prostatic tissue and results into the uretheral compression. This compression resulting due to increased cell proliferation and/or impaired apoptosis causes physical enlargement of the prostate gland and is referred to as static component. In addition dynamic component involves sympathetic nerve stimulation causing contraction of prostatic and uretheral smooth muscle which results into outflow obstruction [2]. Despite several hypotheses the molecular trigger for BPH remains unknown [3]. The incidence of BPH is about 70% at 70 years of age and becomes nearly universal with advancing age. Clinically, BPH causes a constellation of symptoms known as lower urinary tract symptoms (LUTS). The hallmarks of the LUTS include frequency, hesitancy, urgency, nocturia, slow urinary stream and incomplete emptying [4]. Earlier the choice of treatment in BPH was watchful waiting, transurethral resection of the prostate (TURP) or open prostatectomy but due to the invasive nature and potential side effects many medical therapies have emerged involving the suppression of androgen stimulation of prostatic growth [5]. These therapies delay or eliminate the requirement of surgery. 5␣-Reductase enzyme has emerged as a target for the pharmaceutical treatment of BPH as abnormally high activity of the enzyme in humans results in excessive DHT levels in peripheral tissues and hence suppression of androgen action by 5␣-reductase inhibitors is a logical treatment for BPH [6]. A large number of molecules have been synthesized as potential 5␣-reductase inhibitors over the years. Several analogues may also act as androgen receptor antagonists by preventing the natural ligands of the androgen receptor such as testosterone (T) (1) and DHT (2) from binding to the receptor. Combination of these two categories of inhibitors may provide effective androgen receptor blockage without undesirable side effects of castrate testosterone levels on muscle and bone mass, energy level and libido which are of particular concern [7]. Some earlier reports have been there covering various aspects of 5␣-reductase enzyme and inhibitors [8][9][10] but a comprehensive review of each category and structural features required for 5␣-reductase inhibitory activity were missing. This review is an attempt to cover all categories of inhibitors of 5␣reductase with an aim to list the most potent compounds of each category along with the special structural requirements that led to 5␣-reductase inhibitory activity and in vitro data obtained from the evaluation of steroidal and non-steroidal compounds that have been tested as inhibitors of 5␣-reductase. In particular IC 50 and K i values for relevant compounds have been compared according to the molecular class. The values given are not comparable across the studies and in each comparison a standard taken in the study is mentioned.

Enzyme 5␣-reductase

A significant correlation between the androgens and prostate is well known. Testicular androgens constitute the most important mitogenic factor in vivo for the prostate [11]. Normal circulating levels of androgens are required for the maintenance of structural function, growth and integrity of the prostate tissue. However, androgens have no direct effect on prostatic epithelial cells in culture [12]. Androgens enhance the production of many growth factors in the prostate tissue in vivo through a complex cell to cell interaction involving both epithelial and stromal prostatic cells [13]. Androgen signalling cascade involves the synthesis of T (1) in testes and adrenal glands which gets peripherally converted to DHT (2). DHT formed gets transferred to the target tissues and binds to the target receptor with consequent modulation of gene expression [14]. Both T and DHT bind to and activate the androgen receptor (AR), but DHT shows a higher affinity leading to different kinetic processes. DHT dissociates from AR protein much more slowly than its precursor. Therefore, at a given time ARs are occupied by DHT much more than by testosterone. T is converted to DHT by a steroid 5␣-reductase enzyme (3-oxo-steroid-4-ene dehydrogenase {E.C. 1.3.99.5}) which is a system of two membrane bound nicotinamide dinucleotide phosphate (NADPH) dependent enzymes at the level of prostatic stromal and basal cells. This has led to the development of steroidal and non-steroidal 5␣-reductase inhibitors as they inhibit the conversion of T (1) to DHT (2) as shown in Fig. 1 [15][16][17]. Thus 5␣-reductase dictates the cellular availability of DHT to prostatic epithelial cells and consequently modulates its growth.

Figure 1

Site of action of 5␣-reductase inhibitors.

The proposed chemical mechanism of T (1) reduction to DHT (2) by 5␣-reductase catalysis involves the formation of a binary complex between the enzyme and NADPH, followed by the formation of a ternary complex with the substrate T. A delocalized carbocation is formed due to the activation of the enone system by a strong interaction with an electrophilic residue (E + ) present in the active site. Enolate of DHT is formed by the direct hydride transfer from NADPH to the ␣ face of the delocalized carbocation leading to a selective reduction at C-5. This enolate which is coordinated with NADP + on the ␣ face, is attacked by a proton on the ␤-face at C-4 giving the ternary complex E-NADP + -DHT. Binary NADP + -enzyme complex is formed after departure of DHT and finally the release of NADP + leaves the free enzyme for further catalytic cycles. Three different types of inhibitors could be conceived according to the kinetic mechanism of testosterone reduction: type A inhibitors which are competitive with the cofactor (NADPH) and the substrate (T) and interact with the free enzyme; type B inhibitors which are competitive with the substrate and fit the enzyme-NADPH complex, and type C inhibitors which fits the enzyme-NADP + complex exhibiting uncompetitive mechanism versus the substrate [8,18].

More potent inhibitors of steroid 5␣-reductase have been found among the transition state analogues as molecules mimicking the transition state of the enzymatic processes exhibit a greater binding to the enzyme and hence produce greater inhibition. The enzyme 5␣-reductase binds the 3-keto-4 steroids in such a way that the carbonyl group is brought into vicinity of a positively charged centre on the enzyme whereby the conjugated ketone becomes activated as shown in Fig. 2.

Figure 2

5␣-Reduction of 3-keto-4 steroids.

A hydride ion can then be transferred from the coenzyme NADPH to the 5␣-position of the steroid. The resulting enolate is protonated at the axial 4␤-position by the solvent and the product is released. This evidence for protonation was based on the model studies with the Penicillium decumbens 5␣-reductase enzyme [19].

Modern methods of molecular biology had assisted in identifying two types of 5␣-reductase enzyme: type I and type II from human and rat prostatic complimentary deoxyribonucleic acid (cDNA) libraries and the structures of both genes were elucidated at the beginning of this decade [20,21]. The type I enzyme is not the major species expressed in the prostate and is present mainly in the hair follicles and peripheral skin whereas type II 5␣-reductase is the major isozyme in genital tissues and a deletion in the gene leads to male pseudohermaphroditism [22,23]. Type I enzyme is constitutively expressed in the brain and in adulthood appears mainly localized in the myelin membranes and has a catabolic rather than an activating role in the brain while type II enzyme is transiently expressed in the prenatal period and in males its expression is controlled by androgens and appears to be confined in the hypothalamus and in the hippocampus after stress hence type II enzyme might participate in the perinatal differentiation of brain towards a male pattern [24].

The two isozymes differ in the constitution of amino acids as well as molecular weight. The type I isozymes is active at pH 6-9 while type II is active at pH 5.5. The two isozymes also differ in the location of the gene structure while type I is located at 5p15 while type II is located at 2p22 although they had same gene structure [8,25]. The comparison of the properties of two isozymes is summarized in the Table 1:

Table 1

Comparison of properties of 5␣-reductase isozymes. a

More recently with the development of genome-wide gene expression profile analyses a third type of 5␣-reductase enzyme (type III) has been identified in hormone-refractory prostate cancer cells (HRPC) [26]. This enzyme also converts T to DHT in HRPC cells in a similar way to type I enzyme and was found to be active at pH 6.9 [27]. Type III isozyme has been recognized as a ubiquitous enzyme in mammals. Northern blot and real time RT-PCR analyses have identified this enzyme in both androgen and non-androgen target human tissues such as pancreas, brain, prostate cancer cell lines, skin and adipose tissues [28].

Based on their structure 5␣-reductase inhibitors are discussed below as steroidal and non-steroidal inhibitors.

Steroidal 5␣-reductase inhibitors

As the only information available about the 5␣-reductase isozymes is their primary sequence estimated from c-DNAs the design of novel inhibitors is affected. Due to the unstable nature of enzyme during purification its crystal structure is not known. The first inhibitors have been therefore designed by modifying the structure of natural substrates, including the substitution of one carbon atom of the rings of the steroids by a heteroatom such as nitrogen thereby forming azasteroids. Singh and coworkers [29,30] as well as other groups have published comprehensive reviews on biological activity of azasteroids [31]. Azasteroidal compounds having nitrogens at various positions have also been covered in this review. However, their 5␣-reductase inhibitory activity has either not been done or they are devoid of activity. Some azasteroids have been found to be 5␣-reductase inhibitors. In the following section azasteroidal inhibitors have been discussed depending upon the position of nitrogen in the steroidal nucleus, i.e. nuclear azasteroids.

2-and 3-Azasteroids

Although some of the 3-azasteroids were synthesized by Doorenbos and Wu [32] and Mazur [33] in the early 1960s but Anderson and Liao in 1968 reported for the first time that steroidal N-oxido-3-aza-1,3,5(10)-triene is a good inhibitor of enzyme 5␣reductase [34]. Haffner in 1994, reported the synthesis of some novel 3-pyridyl-N-oxide steroids (3 and 4) [35] which mimic the enolate or enol like transition state of the enzyme-substrate complex. N-Oxide steroids (3) and (4) were assayed against both type I and type II 5␣-reductase and proved to be potent inhibitors of type II 5␣-reductase with the K i (M) being 0.031 and 0.104, respectively.

In 2003, Robinson et al. reported the synthesis of various 2-and 3-azasteroidal derivatives (5-10) as effective and stable transition state 5␣-reductase inhibitors [36].

All the synthesised 2-and 3-azasteroids (5-10) were evaluated for human 5␣-reductase inhibition. Amines (6 and 10) showed poor inhibitory activity against both type I and type II isozymes whereas lactams (5 and 9) displayed only marginal improvement against type II isozymes. However, nitrones (7 and 8) showed significant enhancement in biological activity.

4-Azasteroids

4-Azasteroids is one of the extensively studied and clinically used classes of azasteroidal 5␣-reductase inhibitors. Voigt et al. in 1970, screened a large number of steroids including 23 steroidal hormones for their ability to inhibit the conversion of T (1) into DHT (2) by a crude cell free enzyme system isolated from rat ventral prostate [37]. In 1973, series of effective 5␣-reductase inhibitors were synthesized and evaluated. From the studies it was established that the key structural requirements for the 5␣-reductase inhibitory activity were presence of 4-en-3-one function and 17␤side chain having one or more oxygen functionalities. Molecules possessing these features act as competitive inhibitors of testosterone 5␣-reductase, therefore, all of them could be regarded as a substrate of the enzyme 4-en-3-one steroids [38]. 4-Androsten-3one-17-carboxylic acid (11) was identified as a potent inhibitor of 5␣-reductase.

It has been reported to be a competitive inhibitor of the enzyme and showed 87.7% inhibition for the microsomal enzyme of human skin. None of the compounds from this series could be shown to interfere with in vivo conversion of the dihydrotestosterone, because of their rapid conversion into the inactive 4,5-dihydro form by the enzyme. In search for a nonreducible inhibitor of 5␣-reductase, Merck and Co. in 1980 reported series of 4-azasteroids where C-4 of 3oxo-5␣-steroids was replaced by nitrogen. The studies showed that there was not only an increase in the 5␣-reductase inhibitory activity but also retention of the in vivo activity [39,40]. Therefore, azasteroids were designed to mimic the putative enzyme-bound enolate intermediate by incorporating sp 2 -hybridized center at C-3 and C-4. Thus a lactam was introduced in the ring A of the steroids to mimic the enol transition state of the enzyme-NADPH-substrate (E.NADPH.S) complex. Substitution at C-17 has been found to enhance potency by binding to a lipophilic pocket on the enzyme. These competitive inhibitors strongly interact with the enzyme at the active site and on other hand unlike the substrate cannot be further reduced to 5␣-metabolites thus have in vivo inhibitory activity [41]. The steroidal pharmacophore provides an anchor between the key A-ring lactam and the C-17 substituent while the former acts as a transition state mimic of intermediate enolate, the latter significantly enhances potency via binding at a pocket largely lipophilic in nature. The key 4-aza-3-oxo-5␣-androstane pharmacophore and the basic structure activity relationship (SAR) is outlined below (Fig. 3) [42,43].

Figure 3

Basic SAR of 4-azasteroids.

Taking into consideration that substitution at C-17␤-could dramatically affect the potency; a large number of modifications were carried out to find potent inhibitors [44][45][46].

4-MA {17␤-N,N-diethylcarbamoyl-4-methyl-4-aza-5␣androstan-3-one} (12) was found to be a potent dual inhibitor of both human 5␣-reductase isozymes having IC 50 value of 1.9 nM against human 5␣-reductase II and 1.7 nM against human 5␣-reductase I, however, it was withdrawn from the clinical developments due to hepatic toxicity and lack of selectivity over 3␤-hydroxy steroid dehydrogenase enzyme [47][48][49]. Out of the series its unsaturated analogue 17␤-(N-tert-butylcarbamoyl)-4-aza-5␣-androst-1-en-3-one, MK-906, Finasteride (13) was found to be the best and extensively studied. Finasteride (13) was a potent inhibitor of 5␣-reductase type II with only weak in vitro activity versus 5␣-reductase type I having IC 50 value of 9.4 and 410 nM, respectively. At clinical dose, 5 mg/day, it caused 65-80% lowering of plasma DHT levels [50]. Finasteride (13) was the first drug to be approved in U.S. for BPH. Long-term studies have demonstrated that there is a sustained improvement in BPH disease and reduction in the prostate specific antigen (PSA) level [51].

It was reported by Merck as well as Glaxo in 1996 that Finasteride (13) and close analogues are mechanism-based inactivators of 5␣-reductase II. Although it is accepted as an alternate substrate and is ultimately reduced to dihydrofinasteride (15), this proceeds through an enzyme-bound NADP-dihydrofinasteride adduct. Initially it was believed that Finasteride (13) act as a transition state mimic whereby confirmation of the A-ring lactam closely mimics the enol form of transition state of 5␣-reduced testosterone but now it is understood that the most likely cause of the slow offset inhibition is rate-limiting hydride transfer from NADPH to the 1 -double bond of Finasteride (13). In the case of the 1,2-ene-containing Finasteride (13), reduction of C1 enables the nucleophilic attack of C2 on the nicotinamide C4. This aberrant reduction results in the formation of lactam enolate which is not positioned for efficient protonation by the enzyme. Instead the enolate is trapped by the electrophilic pyridinium cation of the NADP, yielding a covalent adduct to the cofactor and to the protein (Fig. 4). This dihydrofinasteride-NADP adduct is a remarkably potent bisubstrate analog inhibitor and it binds to the free enzyme with a second-order rate constant equal to k cat /K m for turnover of T (1) and has a dissociation constant K i ≤ 1 × 10 −13 M. Finasteride (13) is also a mechanism-based inhibitor of the human skin (type I) isozyme, but it is processed with a much smaller second-order rate constant, k i /K i = 3 × 10 3 M −1 s −1 , which attenuates its activ- (16)(17)(18)(19). These compounds were tested in vitro for inhibition of testosterone 5␣-reductase and were found to be weak inhibitors with K i s in the 10 −7 range. It was argued that replacement of C-2 in the steroid nucleus by oxygen in the case of 4-aza-3-oxo-steroids would convert it to a urethane (17) from a lactam (16), respectively, thereby enhancing the polarity at C-3 carbonyl, and its affinity for the enzyme active site [55]. Table 3 In vitro screening of compounds 28-33 against human and rat prostatic 5␣reductase.

Figure 4

Mechanism of Finasteride inhibition of 5␣-reductase.

Table 3

Compounds

Human 5␣-reductase, Bakshi et al. reported a series of 4-aza-3-oxo-5␣-androstene-l7␤-N-aryl-carboxamides as dual inhibitors of human type I and type II steroid 5␣-reductases [56]. Some of these compounds were found to be potent inhibitors of both isozymes. Variation of the C-17 amide substituent on the 4-aza-3-androstane skeleton has resulted into a fruitful search of the potent dual azasteroid inhibitors (20)(21)(22)(23)(24)(25)(26). (27) had emerged as the most potent dual inhibitor from this group ( Table 2) [57]. It has been approved by U.S. FDA in 2002, for the symptomatic treatment of BPH [58,59]. Unlike Finasteride (13), Dutasteride (27) is a competitive inhibitor of both 5␣-reductase type I and 5␣-reductase type II isozymes, reduced dihydrotestosterone levels >90% following one year oral administration [60]. It is also a time dependent inhibitor as Finasteride (13) and it forms a stable complex with a slow rate of dissociation constant and does not bind to the androgen receptor [61]. By reducing DHT level, it reduces the size of enlarged prostate, so improving the urinary flow rate. It is about 60 times more potent than Finasteride (13) and has been shown to decrease the risk of acute urinary retention and BPH related surgery [62][63][64]. This greater degree of suppression of serum DHT has been found to correlate with the intraprostatic DHT suppression. Dual inhibition of 5␣-reductase is more beneficial than selective type II inhibition as dual inhibition does not allow the escape of DHT which can formed through type I mediated synthesis thus providing greater efficacy as DHT levels are suppressed to a great extent. Long-term studies have shown that Dutasteride (27) a dual inhibitor is well tolerated during daily use for up to 2 years. It had a tolerability profile comparable to that of placebo with the exception of a modestly elevated incidence of impotence and decreased libido compared with placebo. Also Dutasteride did not clinically significantly impact bone metabolism markers, bone mineral density or lipid levels [65].

Table 2

In vitro screening of compounds 20-27 against type I and type II human steroid 5␣-reductase.

In a programme aimed at searching for novel 5␣-reductase inhibitors a series of C-17-acylurea-substituted 4-azasteroids (28)(29)(30)(31)(32)(33) (Table 3) were synthesized by Di Salle et al. in early 1990s exploiting the tolerance of functionality at this position. Significantly greater potency was found with the derivative containing C-4 methyl group and a saturated A-ring [66]. Table 4 In vitro screening of compounds 38-43 against type I and type II human steroid 5␣-reductase.

Table 4

Human 5␣-reductase inhibition K i (nM) In order to have specific and dual inhibitors of 5␣-reductase Labrie and associates synthesized several steroids having lactam in ring A and substitution at 17␤ position. Several of the compounds were found active (Tables 9 and 10) [75]. Table 10 In vitro screening of compounds 67-69 against human steroid 5␣-reductase.

Table 10

Human 5␣-reductase inhibition K i (nM)

Human 5␣-reductase, IC50 (nM)

Panzeri et al. reported the syntheses of several 17␤substituted 4-aza-5␣-androstan-3-one carboxamides with unsaturation between C-1 and C-2 (70)(71)(72). These were found to be highly potent against human 5␣-reductase enzyme (Table 11) [76]. Table 12) as a potent dual inhibitor of both 5␣-reductase isozymes and it was found to cause 74% reduction in the DHT levels [77].

Table 11

Table 12

In vitro screening of compound 73 against human steroid 5␣-reductase.

CIBA-GEIGY Ltd. reported the synthesis of CGP53153 (N-(2-cyano-2-propyl)-3-oxo-4-aza-5␣-androst-1-ene-17␤carboxamide) (74) (Table 13), a novel inhibitor of 5␣-reductase and structurally related to Finasteride (13) was found to be 10 times more potent than Finasteride in reducing prostate weight of rat [78].

Table 13

In vitro screening of compound 74 against rat and human steroid 5␣-reductase. In 1996, Giudici et al. reported the synthesis of FCE 28260

In 1996, Ishibashi et al. reported the synthesis of various 11␣-acetoxy, 11␣-hydroxy, 11␤-hydroxy and 11-oxo substituted 4-aza-5␣-androstane analogues (75-79) with a diphenylmethylcarbamoyl moiety at C-17 and their evaluation. Compounds with an 11␤-hydroxy or 11-oxo showed inhibitory activities comparable to Finasteride (13). The 4-methyl 11␤-hydroxy-4-aza-5␣-androstane derivative (77) was found to be most potent against rat and human enzyme and more active than Finasteride (13) ( Table 14) [79]. Merck in 1997 reported the synthesis of several 4-aza 5␣androstan-3-one 17␤-(N-substituted carboxamides) (80-85) as potent human type II 5␣-reductase inhibitors. From the studies it was indicated that the 17-amide N substituent included aromatic residue potent dual inhibitors of type I and type II 5␣-reductase were obtained (Table 15).

Table 14

In vitro screening of compounds 75-79 against human steroid 5␣-reductase.

Table 15

The addition of N 4 -methyl substituent in A-ring increases human androgen receptor affinity while addition of unsaturation to the A-ring ( 1 ) increased human androgen receptor binding. The unsubstituted carbanilides in the 1 -N 4 -methyl series showed some selectivity for type I 5␣-reductase over type II enzyme. Whereas addition of aryl substitution at the 2-position increased type II 5␣-reductase binding, thus providing dual inhibitors with excellent human androgen receptor binding. Compound (87) was found to be the most potent inhibitor from this series (Table 16) [80].

Table 16

In vitro screening of compounds 86-89 against human steroid 5␣-reductase.

In 1998, Salle et al. reported synthesis of a novel compound PNU 157706 [N-(1,1,1,3,3,3-hexafluorophenylpropyl)-3-oxo-4-aza-5␣androst-1-ene-17␤-carboxamide] (90) as a potent dual type I and type II 5␣-reductase inhibitor. PNU 157706 (90) was found to reduce prostate weight 16-fold than Finasteride (13) while the ED 50 values being 0.12 and 1.9 mg/kg/day, respectively (Table 17) [81]. In 2000, Lerner and coworkers reported the synthesis of haptens 91(a,b) and 92(a,b) which belongs to 4-aza steroids. The resulting 5␣-dihydrotestosterone was shown to be the more potent intracellular hormone [82]. Menzenbach et al. reported the synthesis of several 17methylene-4-azasteroids as inhibitors of 5␣-reductase. Azaestranone II (93) was found to be inhibitor of 5␣-reductase with IC 50 = 34 × 10 −10 (for prostate) and IC 50 = 25 × 10 −10 (for seminal vesicle) [83].

Table 17

In vitro screening of compound 90 against human steroid 5␣-reductase. a Results are the mean ± SE of 3 separate assays. Incubations were performed in the presence of 3 or 1 M [ 3 H] testosterone, for type I or type II isozyme, respectively.

Type I 5␣-reductase, IC50 (nM)

Type II 5␣-reductase, IC50 (nM)

Type I 5␣-reductase, IC50 (nM) Type II 5␣-reductase, IC50 (nM) 262 6500 -263 460 -264 30 -265 60 enriched by the presence of halogens in these steroidal derivatives leading to the increase in the inhibition of 5␣-reductase enzyme as determined by the IC 50 values. Compound 252 was found to be most potent with an IC 50 value of 1.8 nM while compounds 251 and 253 showed values 14 and 10 nM, respectively, in comparison to Finasteride (13) having value 8.5 nM. Compound 254 was not that active when compared to halogenated compounds thus further demonstrating the need of a halogen atom in the ester side chain [171].

Table 5

Recently, Cabeza et al. synthesized several C-6 substituted and unsubstituted pregnane derivatives as potential 5␣-reductase inhibitors. It has been found that steroids that lack a chlorine atom in C-6 (255-257) exhibited a higher capacity for inhibition of the activity of 5␣-reductase (IC 50 in the range of 25-63 nM) than the compounds (258-260) having this atom (IC 50 in the range of 920-990 nM in comparison to Finasteride (13) having an IC 50 8.5 nM). The presence of bromine atom in C-6 of compound 261 however doesn't affect the inhibitory activity of enzyme (IC 50 being 33 nM). Also, the presence of an ester moiety in C-17 ␣ on the steroidal skeleton tends to increase the inhibition of the activity of the enzyme [172].

Type I 5␣-reductase, IC50 consistent with the observation that the introduction of the same group on the corresponding position 7 in 4-azasteroids increased their 5␣-reductase I selectivity [70]. Presence of methyl group at position 5 found to decrease the potency of the compounds while introduction of methyl at position 1 as in compound 317 caused only a slight decrease in activity when compared to unsubstituted compound 312.

In 2001, Guarna et al. studied the effect of C-ring modifications in benzo[c]quinolizin-3-ones. They synthesized several octahydroand decahydrobenzo[c]quinolizin-3-one derivatives containing partially or fully saturated C-ring. These compounds were found to be selective 5␣-reductase I inhibitors. Benzo[c]quinolizin-3-one inhibitor lacking the aromatic C-ring but with a double bond at 6a-10a 318 displayed an inhibitory potency 345-fold higher than that of the corresponding 6a-10a saturated, trans-fused compound 319 (Table 34) [192].

A 3D-QSAR model correlating the potency of the inhibitors with their physicochemical features using density functional theory (DFT) was also developed for a series of benzo[c]quinolizin-3-ones derivatives by adding two "non standard" variables (dipole moment and log P) to the classical electrostatic and steric comparative molecular field analysis (CoMFA) fields [193].

With the aim to discover new dual non-steroidal inhibitors of 5␣-reductase I and II a series of benzo[c]quinolizin-3-ones derivatives (320-325) bearing diverse substituents at position 8 were synthesized in 2005 [194]. They were tested towards 5␣-reductase I and II expressed by Chinese Hamster Ovary cells (CHO 1827 and CHO 1829), respectively. It was found out that most potent dual inhibitors were obtained when F atom was introduced on the phenol moiety of these esters. All compounds displayed inhibition towards 5␣-reductase I in the range of 93-165 nM (Table 35). Com-pound 322 was found to be the most potent dual inhibitor of the series with IC 50 values about 100 nM for both enzymes.

Rat 5␣-reductase, IC50 (nM) indole ring was found to be a selective inhibitor of 5␣-reductase I (Table 37) [215]. FK-143 (372) 4-[3-[3-[bis (4-isobutylphenyl) methyl amino] benzoyl]-lH-indol-l-yl] butyric acid was disclosed by Sawada et al. as a potent dual inhibitor of both human 5␣reductase isozymes. It inhibited in vitro human and rat prostatic 5␣-reductase in a dose-dependent manner with an IC 50 of 1.9 and 4.2 nM, respectively, in a non-competitive fashion while in vivo showed potent inhibitory activity against castrated young rat model. This compound can be a potential drug for the treatment of benign prostatic hyperplasia [216][217][218].

Compounds IC50 (nM) or % inhibition at 100nM (given in parenthesis)

Human 5␣-reductase type I (transfected 293 cells)

Human 5␣-reductase type II (transfected SW-13 cells)

31.5 ± 6.0 (48.7 ± 4.2) 43

11.5 ± 1.9 (39.7 ± 0.4) Finasteride (13) 262 ± 43.2 8.5 ± 0.4 One of the most potent compound of this group, Turosteride (30), a close analogue of 4-MA (12), but unlike 4-MA was found to be devoid of binding at the rat androgen receptor and a weak inhibitor of 3␤-hydroxy steroid dehydrogenase [67].

Other azasteroids which retained 5␣-reductase inhibitory activity are 2-substituted (34), A-homo-(35) and 19-nor- (36) analogues [41].

Observation that selectivity of inhibitors can be increased against type I isozyme by making correct choice of hydrophobic substituent at C-17 position led to development of various 17␤-(Nureylene-N,N -disubstituted)-4-methyl-4-aza-3-one 5␣-reductase derivatives (Table 4) (38-43) as 5␣-reductase inhibitors as they have potent selectivity against 5␣-reductase type I enzyme. Azasteroids with N-cyclopropyl ring exhibit potent inhibitory activity against type I 5␣-reductase. Increase in the chain length from Nethyl to N -butyl the compound showed strong inhibitory activity while branching of alkyl chain decreased potency of compounds and introduction of 1,2-double bond significantly reduced the activity. Replacement of N -alkyl chain with phenyl moiety gave the most active compound (41) of the series [68].

From the studies that 17␤-carboxamides at C-17 position have pronounced activity of 5␣-reductase and possess androgen receptor activities, a number of 17␤-(N-alkyl/aryl formamido) (44)(45)(46)(47)(48)(49) and 17␤-[(N-alkyl/aryl)alkyl)aryl amido]-3-oxo-4-aza-5␣-steroids (50-54) were prepared from 17␤-hydroxy-4-azasteroids and were evaluated as 5␣-reductase inhibitors by Li et al. (Tables 5 and 6). Structure activity relationship indicated that 5␣reductase type I enzyme has preference for N-substituted linear alkyl side chain of 4-5 carbon atoms. N-Amyl substituted 17␤formamide (45) was found to be one of the most promising inhibitor of 5␣-reductase type I while N-heptyl (48) and N-octyl (49) showed dual inhibition of both isozymes of 5␣-reductase (Table 5).

Similarly in series of 17␤-[(N-alkyl/aryl) alkyl) arylamido] derivatives (Table 6) exhibited highly potent inhibitory activity for human 5␣-reductase type I [69].

Table 6

In vitro screening of compounds 50-54 against type I and type II human 5␣reductase.

Various 7␤-substituted derivatives have also been prepared. Preliminary screening of the compounds as inhibitors of 5␣reductase from human scalp and prostate revealed that the presence of 7␤-methyl substitution in ring B, presence of cholesterol type side chain at C-17 and ketone functionalities at C-3 in 4-azasteroids resulted in potent selective inhibitor against 5␣reductase type I [70].

4,7␤-Dimethyl-4-aza-5␣-cholestan-3-one (MK 386) (61) emerged as one of the most potent inhibitor of type I 5␣-reductase ( Table 7) [71].

Table 7

In vitro screening of compounds 55-61 against type I and type II human 5␣reductase. In vitro screening of compound 62 against human and rat 5␣-reductase inhibition. a

During 1995, some 17␤-hydroxy-17␣-(-hydroxy/haloalkynl -yl)-4-methyl-4-aza-3-oxo-5␣-androst-1-ene-3-ones were synthesised by Li et al. and their antiandrogenic activity was reported [72]. Salle et al. have reported synthesis and 5␣reductase inhibitory properties of various 4-azasteroids with fluoro substituted 17␤ amidic side chains [73] and further investigated FCE 27837 (N-[1,1,1-trifluoro-2-oxobut-3-yl]-3-oxo-4-aza-5␣-androst-1-ene-17␤-carboxamide) (62), for its endocrinological properties in comparison with those of Finasteride (13) ( Table 8) [74]. Table 9 In vitro screening of compounds 63-66 against human steroid 5␣-reductase.

Table 8

Table 9

Table 11

In vitro screening of compounds 70-72 against human steroid 5␣-reductase.

6-Azasteroids

Glaxo was the first to report design of 6-azasteroidal inhibitors based on the 3-keto-4-en-6-amine functionality to mimic the struc- tural and charge polarization features of the transition state for the enzyme catalyzed transfer of hydride from NADPH to testosterone. The higher reduction potential of ketoenamine compared to that of ␣,␤-unsaturated ketone prevents these compounds from acting as substrates for 5␣-reductase and they show slow offset inhibition instead of irreversible as shown by 4-azasteroids [53]. Structure activity relationship has also been reported by Frye and associates at C-4, N-6 and C-17 carbamoyl (Fig. 5) [43,84]. Initially a set of N-6, C-1, C-2, C-4 substituted derivatives of 6aza-androst-4-en-3-ones (94-101) were prepared to explore the structure activity relationship of A-and B-rings versus type I and type II 5␣-reductase (Table 18).

Figure 5

. Guarna et al. also synthesized 17␤-[N-(phenyl) methyl/phenyl-amido] substituted 10-azasteroids. Unexpectedly, 5␤-H compounds were found more active than their 5␣-H counterparts, with (138) (IC 50 = 279 and 2000 nM toward isoenzymes I and II, respectively) and (139) (IC 50 = 913 and 247 nM toward isoenzymes I and II,respectively) being the most potent compounds of the series[104].

Table 18

In vitro screening of compounds 94-101 against human steroid 5␣-reductase.

Methylation at N-6 (95) and substitution of C-4 with small lipophilic groups such as Cl (96), Br (97) and CH 3 (98) increases type I 5␣-reductase activity selectivity 4-fold while type I 5␣-reductase activity was decreased by unsaturation (99), 1,2 cycloproponation (100) and C-2-methylation (101).

By careful optimization of the C-17 substituent along with combining A-and B-ring substitutions, potent dual inhibitors of both isozymes of 5␣ reductase were obtained (102-114) (Table 19) [84].

Table 19

In vitro screening of compounds 102-114 against human steroid 5␣-reductase.

It was found that optimizing C-17 group resulted in 5␣reductase type I inhibitory activity in 103 with 5-7-fold increase of activity in 105. Swapping a methyl ester (106) for an admantyl ester (108) provides selectivity towards 5␣-reductase I. Preparation of analogues of 103 resulted in compounds (109)(110)(111)(112) which were found to be 16-200-fold selectivity towards type I 5␣-reductase. Compounds (103, 111, 113 and 114) having ketone at C-17 proved extremely potent inhibitors of type I 5␣-reductase. Out of the group, compound (105) demonstrated efficacy equivalent to Finasteride (13) in a castrated rat model of DHT dependent prostate growth. In general, large lipophilic groups at C-17 provides selectivity against 5␣-reductase I.

A variety of C-17 amide-substituted 6-aza-androst-4-en-3-ones were prepared and evaluated against human type I and type II steroid 5␣-reductase in order to optimize potency versus both isozymes of 5␣-reductase. Out of the study two series of potent and selective C-17 amides were discovered, 2,5-disubstituted anilides and (arylcycloalky1) amides (115)(116)(117)(118)(119)(120)(121) (Table 20). Evaluation of some optimal compounds from this series in a chronic castrated rat model of 5␣-reductase inhibitor induced prostate involution, and pharmacokinetic measurements identified compounds (117, 118, 120 and 121) with good in vivo efficacy and half-life in the dog [85].

Table 20

In vitro screening of compounds 115-121 against human steroid 5␣-reductase.

B-Homologated analogue of 17␤-N, N -diethylcarboxy-6-azaandrost-4-en-3-one (122) has also been found to be potent inhibitor of 5␣-reductase inhibitor with IC 50 = 318 nM [86].

Bergmann et al. synthesised 7-substituted ␦-4-6-azasteroid derivatives (123) as 5␣-reductase inhibitors [87]. But activity data was not reported for these compounds. R 1 = H or CH 3 ; R 2 = CH 3 ; R 3 = hydrogen, Alk-R 4 , X-Alk, C 1-6 -X-Alk, XCO-Alk, Co-Ar, CO-NH-Ar, CO-NH-Het, etc., where Alk is C 1-12 straight or branched alkyl, Ar is phenyl, X is O, N or S, Het is piperidinyl, piperizinyl, pirrolidinyl, pyrrolyl, etc.

6-Azacholesten-3-ones (Table 21) were assayed against both type I and type II 5␣-reductase by Haffner [88]. All three compounds were found to be potent dual inhibitors of 5␣-reductase. Unlike the 4-azasteroids the cholesterol side chain imparts very little selectivity between type I and type II 5␣-reductase. It was also found that C-7 methyl group might provide a potent 5␣reductase I selective compound. The ␣-C-7 methyl diastereomer (124) proved to be 7-fold more active 5␣-reductase I inhibitor than ␤-diastereomer (125). Fang and Sharp in 1996 synthesized several 6-azaandrostenones of the general structure (127) as 5␣-reductase inhibitors [89].

Table 21

In vitro screening of compounds 124-126 against human steroid 5␣-reductase.

16␤-Aryloxy, -alkoxy and heteroaryloxy 6-azasteroids of the general formula as given below were synthesised by Aster et al. and the compounds were found to be potent inhibitors of 5␣-reductase [90]. Compound (128) was found to be potent inhibitor of human 5␣-reductase type I with IC 50 in the range of 0.1-1000 nM.

Rahier and Taton have reported synthesis of several novel 6aza-B-homosteroids but they were not tested for 5␣-reductase inhibitory activity [91]. Later in 2000 and 2001, Wenge et al. [92] and Xie et al. described the synthesis of 6-azasteroids as potent phosphatidylinositol phosphalipase C (PI-PLC) inhibitors [93]. Kasal et al. in 2005 reported an efficient synthesis of 6aza-allopregnanolone as neurosteroid analogues but not evaluated them for 5␣-reductase inhibitory activity [94].

7-Azasteroids

Some 7-azasteroids were synthesized in early 1970s [95]. Morzycki and Sicinski reported the synthesis of 6,7diazacholestane derivatives but they were not evaluated for the 5␣-reductase inhibitory activity [96].

8-Azasteroids

Several 8-azasteroids have been synthesized [97] and discussed as antifungal agents [98,99] but none has been reported as 5␣reductase inhibitor.

9-Azasteroids

No work has been published on 9-aza steroids as 5␣-reductase inhibitor although some fungicides have been known from this category [100].

19-Nor-10-azasteroids

On the basis of the molecular model of active site for type II isozyme and to increase the activity and selectivity of compounds towards both 5␣-reductase type I and 5␣-reductase type II. Guarna et al. synthesized a novel class of compounds 19-nor-10azasteroids (Tables 22 and 23) [101].

Table 22

In vitro screening of compounds 129-135 against human steroid 5␣-reductase type II.

Best results were obtained with 9:1 mixture of 9(11) (133) and 8(9) (134) 17 ␤-(N-tert-butyl carbamoyl)-19-nor-10-aza-4-androsten-3-one as it was found to be a good inhibitor of 5␣-reductase type I and 5␣-reductase type II. The enamine structure of ring A of 10-aza-steroid (136) is analogous to that of the substrate like transition state. The presence of N atom at position 10 increases the nucleophilic character of the carbonyl group and stabilizes the carbocation intermediate (137) by delocalization at the positive charge. The inhibitory potency of these compounds depends on the presence of bridgehead N-10 atom conjugated with 4-en-3-one moiety in A-ring, unsaturation in C-ring and substituent at C-17 position. 19-nor-10-aza steroids have low transitional barrier energy values and more flexible as compared to 6-aza or 4azasteroids [102].

Some 10a-azasteroids were also synthesized from fusidic acid but were not evaluated for 5␣-reductase inhibitors [103]

11-, 12a-, 13-Azasteroids

Though many 11-azasteroids [105][106][107], 12a-azasteroids [108] and 13-azasteroids [109] have been prepared but 5␣-reductase inhibitory activities have not been reported.

15-and 16-Azasteroids

Many 15-azasterols have been synthesized as antifungal agents [110][111][112] but none of them have been evaluated for 5␣-reductase inhibitory activity. 16-Azasteroids have been synthesized but none of the compound has evolved as 5␣-reductase inhibitor [113,114].

17-and 17a-Aza-D-homosteroids

Regan and Hayes, in their exemplary work, have synthesized several 17-and 17a-aza-D-homosteroids from several 17-ketosteroid oximes [115]. But 17a-azasteroids attracted more attention when chandonium diiodide was established as a potent neuromuscular blocker [116]. 17 and 17a-Azasteroids have been found to possess numerous biological activities like gamma amino butyric acid (GABA) receptor antagonistic [117][118][119], antifungal [120], antineoplastic, mutagenic [121,122] and anti-inflammatory activity [123]. The most interesting aspect concerning 17-D-homoazasteroids is the possibility of "inverted action" or "backbinding" as proposed by MacDonald et al. [124]. Their proposition was based on the fact that the steroids have the potential to bind in two orientations in the active site of various metabolizing enzymes. Marcus and Talalay [125] first reported that 3(17) ␤-hydroxysteroid dehydrogenase converts both T (1) and dehydroepiandrosterone (141) to androst-4-ene-3,17-dione (140) (Fig. 6a). Other examples of enzyme inhibition by inverted steroids have also appeared. The oxiranes 143 and 144 were found to be active site-directed, irreversible inhibitors of 3-oxo-5 -steroid isomerase (Fig. 6b) [126] and the bromoacetates 147 and 148 act as affinity labels for estradiol 17␤-dehydrogenase (Fig. 6c) [127]. Research on 5␣-reductase inhibitors has shown that steroids without side chains can bind to enzymes with the A-ring of the substance simulating the D-ring of the substrate, while the D-ring emulates the A-ring. This could lead to 17-D-homo-azasteroids exhibiting same mechanism of action as 4-azasteroids. However, up to now, research has been concentrated on 4-azasteroids as 5␣-reductase inhibitors and 17a-azasteroids remained unexplored avenue as far as their 5␣-reductase inhibitory activity is concerned. Some 17a-azasteroids (149-152) evaluated for 5␣-reductase inhibitory activity are summarized in the Table 24 [124].

Figure 6

(a) Action of 3(17) ␤-hydroxysteroid dehydrogenase, (b) action of 3-oxo-5steroid isomerase, and (c) action of estradiol 17␤-dehydrogenase. 8,13-Diaza steroids were also synthesized by Göndös et al. in 1998 but not evaluated for 5␣-reductase inhibitory activity[130].

Table 24

In vitro screening of compounds 149-152 against human steroid 5␣-reductase.

Diazasteroids

Eberbach and coworkers reported in 1996 a novel access to 4,13-diazasteroid derivatives but they were not evaluated for 5␣reductase activity [128]. In the same year Stuart et al. first reported 4,17-diazasteroids as potential inhibitors of 5␣-reductase. The Finasteride 17-aza-isomer (153) proved to be potent inhibitor of 5␣-reductase II although less active than Finasteride (13) and its congeners. 4-Methylation (154) lowered the inhibition of the 5␣-reductase II enzyme. Removal of 1(2) unsaturation led to the formation of compound (155) that is dual inhibitor of 5␣-reductase type I and type II and 4-methylation of 155 led to the increase in activity in 156. While compound with 5(6) (157) showed only a moderate inhibition of 5␣-reductase II activity (Table 25) [129].

Table 25

11,13,15-Triazasteroids

Hirota et al. reported in 1995 the synthesis of 11,13,15triazasteroid derivatives to investigate antidepressive activity. These analogues were also evaluated for anti-platelet aggregation activity and some derivatives exhibited positive action but no 5␣-reductase activity has been investigated in these categories of steroids [131].

B,D-Dihomo-azasteroids

Several steroidal B,D-dihomolactam have been synthesized and evaluated for antitumour activity but no 5␣-reductase activity has been reported from this group till date [132,133].

Des-AB-azasteroids

Trehan et al. synthesized des-AB-azasteroids but 5␣-reductase activity studies were not done [134].

Steroidal 5␣-reductase inhibitors which are extranuclear, i.e. in which nitrogen is not the part of steroidal nucleus but forms part of the side chain or attached group have also been explored as 5␣reductase inhibitors, therefore are discussed next.

Table 25

In vitro screening of compounds 153-157 against human steroid 5␣-reductase.

Steroidal 3-carboxylic/phosphonic/phosphinic acids

A number of 3-androstene-3-carboxylic acids (158-166) ( Table 26) were designed to mimic the putative enzyme-bound enolate intermediate by incorporating sp 2 -hybridized centers at C-3 and C-4 and, most critically, an anionic carboxylic acid at C-3 as a charged replacement for the enolate oxyanion. Because of presumably favorable electrostatic interaction between the carboxylate and the positively charged oxidized cofactor, the acrylate preferentially binds in a ternary complex with enzyme and NADP + , which leads to the uncompetitive kinetic mechanism [135][136][137]. Activity is enhanced in analogues possessing an additional unsaturation at C-5 (162-167) along with 3 -unsaturation. At C-17, diisopropyl (158) and pivalyl (163) amides were optimal. Epristeride (SK&F 105657) (163) entered the clinical trials for treatment of BPH and is a potent inhibitor of 5␣-reductase II while a weak inhibitor of 5␣-reductase I.

Table 26

In vitro screening of compounds 158-166 against human steroid 5␣-reductase.

A series of estratriene-3-carboxylic acids containing an aromatic A-ring had also been synthesized (167)(168)(169)(170)(171) (Table 27) with structural variations in the C-2 and C-4 substituents, in the degrees of unsaturation in the B-and D-rings, and in the C-17 carboxamide alkyl groups. Despite lacking C-19 methyl group these compounds were found to be potent inhibitors of 5␣-reductase [138].

Table 27

In vitro screening of compounds 167-171 against human steroid 5␣-reductase.

Nitro derivatives also showed interesting structure activity relationship patterns compared to carboxylic acids, compound (172) was found to be potent competitive inhibitor by binding to E-NADPH complex [139]. The sulphonic acid (173) [140], phosphonic acid (174) and phosphinic acid (175) also proved to be the potent inhibitors of human 5␣-reductase but not so of rat 5␣reductase [141]. The affinities of phosphonic acid are relatively less than phosphinic acid derivatives because the increased bulk at the 3-substituent, leading to a steric intolerance for binding to enzyme. Overall 3-phosphosteroids were weaker inhibitors than their corresponding steroidal-3-carboxy steroids. The function of the C-3-moiety is presumably to act as an H-bond acceptor from a residue in the enzyme, which would normally donate a hydrogen bond to stabilize the enolate. Since the negatively charged groups (CO 2 − ) or isosteres of the carboxylate (NO 2 ) best mimic this interaction it indicates that a pK a -matched H-bond with a Lys or Arg donor may be operative. In addition, an interaction between the negatively charged C-3-moiety and the positively charged NADP + cofactor after the enzyme has turned over substrate is possible, especially if the cofactor lies directly under the A-ring of the steroidal skeleton in the transition state and mimics thereof. Compound 176 was not as active due to the presence of alcoholic group as it was not able to provide sufficient negative charge and hence was a weak inhibitor of rat and human 5␣-reductase.

Diazoketone steroids

The primary evidence of a dramatic increase in the affinity of 5␣reductase and an inhibitor with a 5-juncture of A-/B-ring and sp 2 hybridization at the C-3 and C-4 positions was obtained from the inhibition with a mechanism-based inhibitor (5,20R)-4-diazo-21hydroxy-20-methyl-pregn-6-en-3-one (177) (RMI-18,341). Diazoketone (177) had been reported to be a potent time-dependent inhibitor with a K i of 35 nM (time-dependency is considered indicative of irreversibility) [142].

A mechanism of inhibition was proposed that the protonation steps implicated in the normal enzymatic transformation activates the diazoketone functionality to a diazonium ion that could further alkylate some nucleophilic residue at the active site [143].

4-Substituted steroids

The observation that an excellent inhibitor possessed a conjugated system (sp 2 -sp 2 -sp 2 ) at C-3, C-4, and C-5 positions of A-ring of steroids together with a lipophilic group at C-17, a range of 4-substituted-3-oxo-4-androstene-17(-carboxamides (180-183) ( Table 28) were prepared and compared with the Finasteride (13). Out of these 4-cyano compounds were found to be potent inhibitors of 5␣-reductase type II enzyme and substitution with groups like thiol led to decreased activity. This series of compounds were also found to be potent androgen antagonists [144]. Fei et al. carried out the synthesis of some novel 4-trifluoromethylsteroids and proposed them as novel 5␣-reductase inhibitors. Out of the series 4-trifluoromethyl-N-(tbutyl)-4-androsten-17␤-carboxamide (184) emerged as the most potent inhibitor being 4 times more active than Finasteride (13) [145]. 4-Cyanoprogesterone (185) was also found to be a potent inhibitor of both rat and human 5␣-reductase enzymes (IC 50 values = 0.045 and 0.050 M, respectively). The mechanism of action of 4-cyano steroidal inhibitor was assumed to be the transition state inhibitor because on reduction by the enzyme compound would form a stable 5-3-enol that would remain tightly bound to the active site [146,147].

Table 28

In vitro screening of compounds 180-183 against human 5␣-reductase.

Steroidal oximes

A number of pregnenolone (186)(187)(188)(189)(190) and progesterone (191)(192)(193)(194)-based steroids were synthesized bearing a oxime group connected directly or via a spacer to the steroidal D-ring, capable to form a coordinate bond with haeme iron of enzyme 5␣-reductase. In contrast to the pregnenolone derivatives which showed no inhibition of 5␣-reductase isozyme I and II, progesterone derivatives possessed marked inhibition towards type II.

Inhibitory potency of synthesized compounds against target enzyme using whole cell assay revealed that C-20 oxime (191) displayed strong inhibition against both isozymes (IC 50 = 1.63 against human type I and 0.58 M against human type II). Unsaturation in ring D (192 and 193) in conjugation with oxime group further enhances inhibitory activity. Almost complete loss of activity has been found toward type I in the derivative with keto group at C-6 (194).

Transferring the oxime group from positions 20 to 21 caused an increased selectivity toward type II isozyme. Z-21-Hydroxyiminopregn-4-en-3-one (195) was found to be a potential inhibitor of the type II (IC 50 = 1.95 M against human type I and 0.30 M against human type II).

However, none of the compounds showed activity near to that of the reference drug Finasteride (13) (IC 50 values being 45 and 3 nM for type I and type II, respectively, in the corresponding study) [148].

Steroidal tetrahydrooxazin-2-ones

Wölfing et al. synthesized a novel series of steroidal tetrahydrooxazin-2-ones (196)(197)(198)(199)(200)(201) containing heterocycles involving O and N heteroatoms at position 17␤ of androst-4-en-3-one, respectively, as 5␣-reductase inhibitors. The IC 50 values of compounds vary between 270 and 600 nM. The relative inhibitory effect of the unsubsituted N-phenyl compound 196 is 0.20. Concerning the effects of substituents at position 4 of the phenyl ring in 196, the introduction of an ethyl (197) or ethoxy (199) group resulted in a weak enhancement of 5␣-reductase inhibition. Substitution with halogens (200 and 201) or methoxy (198) caused lowering of inhibition ability [149].

Steroid 5␣-reductase inhibitors that do not contain any nitrogen either as part of ring or extranuclear but yet found to inhibit 5␣reductase are discussed next.

16-Substituted steroids

A series of 16-methyl substituted derivatives of androst-4ene and estr-4-ene originally prepared as antiandrogens, were tested for their inhibitory activity on rat and human prostatic 5␣-reductase. The inhibitory activity data indicated that IC 50 increases in sequence in derivatives bearing 16␣-methyl (203), 16␤-methyl (204) and 16,16-dimethyl substituents (205). Acylation of 17-hydroxy group significantly increases the inhibitory potency (IC 50 (207) Overall, in human prostate homogenates IC 50 varies between 0.6 and 120 m while in rat prostate it ranges from 1.6 to 1000 M. This shows enzyme of human prostate is more sensitive than that of rat prostate to methyl substituted compounds. Overall 16-methyl steroids were found to be weak inhibitors both in rat and human enzymes compared to the existing ones [150].

Certain 19-nor analogues have also been synthesized in order to improve the inhibitory activity in 16-methylated derivatives. TSAA-291 (16-ethyl-17␤-hydroxy-4-estren-3-one) (208) was found to be the first anti-androgen known to have dual action of competitive inhibition of 5␣-reductase activity and androgen receptor complex formation. It showed a K i of 1400 nM to the purified nuclei from rat prostatic tissues [151].

6-Methylene steroidal derivatives

2 ,3 ␣-Tetrahydrofuran-2 -spiro-17-(6-methylene-4androsten-3-one) (209; L612,710) is a potent time-dependent inhibitor which causes the highest percentage of inhibition (81%) of rat prostatic 5␣-reductase enzyme. The structure activity relationship showed that 3-oxo-4-ene functionally was essential to the inhibitory activity and that substituents at C-17 influenced the inhibitory potency. The presence of the C-19 methyl group was not essential to the activity. The A-ring appeared to interact with the entire active site of the enzyme. Furthermore, the affinity of an inhibitor to the enzyme was greatly enhanced by the introduction of a methylene group at C-6 whereas activity was completely lost with a large radical such as iodomethylene group at C-7. Thus a series of 6-methylene steroids were prepared and examined as irreversible inhibitors of rat prostatic 5␣-reductase [152]. Another 6-methylene steroid that is potent inhibitor due to priming of its dienone group by electrophilic activation toward nucleophilic attack at the 6-methylene group is having structure (210) [153].

Seco steroids

(4R)-5,10-Seco-estra-4,5-diene-3,10,17-trione (211) and (4R)-5,10-seco-19-nor-pregna-4,5-diene-3,10,20-trione (212) were first found to be non-competitive and possibly irreversible inhibitors of epididymal 5␣-reductase. Radiographic crystallography studies of both compounds showed that the conjugated allenic 3-oxo-5,10-secosteriod (211) has a conformation similar to that of the normal tetracyclic steroid dione. Both compounds were non-competitive inhibitors of 5␣-reductase and have an affinity label for the enzyme with K i of 5470 and 980 nM, respectively [154,155].

Derivatives of natural substrate: pregnane

As a consequence of the important observation that progesterone and deoxycortisone inhibits the synthesis of dihydrotestosterone by competing with 4-en-3-one function of the testosterone for the 5␣-reductase enzyme it led Voigt and coworkers to synthesize number of progesterone derivatives [37,38]. The satisfactory result of 4-cyano-progesterone (185) [146], which possessed marked inhibitory activity for 5␣-reductase enzyme, stimulated great deal of interest to synthesize various 4-and 6-halo-progesterone analogs (213)(214)(215)(216)(217). These compounds were found to be potent antiandrogenic in nature when tested against gonadectomized hamster seminal vesicles and were also found to be inhibitors of 5␣-reductase [156][157][158].

Bratoeff et al. evaluated the antiandrogenic and 5␣-reductase inhibitory activity of various 16-phenyl substituted-D-homo compounds (218-219), 16-methyl substituted steroids (220-222), 4-bromo compound (223) without a methyl group at C-16 and the epoxy compounds (224)(225). Compounds 219, 222 and 223 were found to possess both antiandrogenic and 5␣-reductase inhibitory activity better than the Finasteride (13) [158,159]. The trienones having a more coplanar structure reacts faster with the nucleophilic portion of the enzyme in a Michael type addition reaction to form an irreversible adduct with a concomitant inhibition of the enzyme 5␣-reductase than the dienones.

A range of 4-bromo-17-substituted-4-pregnene-3,20 diones were also synthesized and evaluated as 5␣-reductase inhibitors on gonadectomized hamster seminal vesicle and flank organs. Small diameter of the pigmented flank organ and great reduction in the weight of seminal vesicle has been found with the compounds having p-fluorobenzoyloxy (226) and p-chlorobenzoyloxy (227) indicating that the presence of more electronegative substituent at C-17 position (p-halosubstituted phenyl) and halogen at C-4 enhances the antiandrogenic activity as well as 5␣-reductase inhibitory activity [160,161].

Several new pregnane derivatives were also synthesized and evaluated by the conversion of [ 3 H] T to [ 3 H] DHT in

Penicillium crustosum broths and the conversion of [1,2-14 C] sodium acetate into lipids. Compounds 228 and 229 were found out to be potent 5␣-reductase inhibitors as they inhibit conversion of T to DHT and also decreased the incorporation of radiolabeled sodium acetate into lipids of the flank organs [162].

Cabeza et al. also reported the 5␣-reductase inhibitory activity and the antiandrogenic effect of novel 16-bromo substituted trienedione, 16␤ methyl substituted dienedione and the trienedione (230-232). Compounds 230 and 231 were found to Exhibit 5␣-reductase inhibitory activity higher than the commercially available Finasteride (13) [163].

The in vitro inhibitory activity of some novel progesterone derivatives was also determined and they were evaluated as 5␣-reductase inhibitors as well as antagonists for the androgen receptor. The appropriate homologues extended by a methylene group at C-6 (233-236) showed better activity due to the presence of exocyclic double bond that can react faster with the enzyme in a Michael type addition reaction than the corresponding endocyclic diene. Compounds 233 and 234 showed IC 50 values of 19 and 100 nM, respectively [164,165].

In compound 237 which is an epoxy compound the nucleophilic enzyme attacked the electrophilic center at C-6 and in this process inhibits the enzyme. Several 3-substituted 4-pregnene-6,20-dione derivatives (238-246) have been synthesized and were evaluated for antiandrogenic as well as 5␣-reductase inhibitory activity. The synthesized steroids have an ␣,␤-unsaturated carbonyl moiety in common. It was found that accessible electrophilic ␤-carbon of an ␣,␤-unsaturated carbonyl moiety reacted very readily with a variety of nucleophiles to form Michael adducts. The first step involved the formation of an enzyme-steroid activated complex and in a subsequent step the nucleophilic portion of the enzyme (amino group) attacked the conjugated double bond of the steroid in a Michael type addition reaction to form an irreversible adducts [166].

The IC 50 values of the compounds were found to increase progressively as the substituent on the phenyl group of the ester side chain at C-3 became more electropositive, i.e. compound 241 has an IC 50 value of 3.0 nM as compared to compound 243 which has an IC 50 value of 4.0 nM. On the other hand the compounds (244-246) having saturated 4-pregnene-6,20-dione skeleton for, e.g. 245 exhibited higher IC 50 value (3.7 nM) as compared to the unsaturated ones (240-243) for, e.g. 241 with a value of 3.0 nM. It was also found that the presence of halogen substituents in ester moiety at C-3 as well as double bond at C-16 increased the binding affinity for the androgen receptor [167].

Several new progesterone derivatives were also synthesized by the same group having dienone moiety as reported earlier. Some of them (247-248) were found to have high 5␣-reductase inhibitory activity like 248 showing an IC 50 value of 0.5 nM as compared to Finasteride (13) showing an IC 50 value of 8.5 nM in the same study [168,169].

Bratoeff et al. synthesized two novel steroidal carbamates (249-250) which are the esters of carbamic acid with substituents at the amino and esters ends (NHRCOOR 1 ) and proposed them as novel class of inhibitors for human and hamster steroid 5␣reductase. These compounds have a longer half-life and are hydrolyzed slowly in the liver due to which they have a better pharmacological activity when compared to the conventional esters. Compound 249 has got a similar IC 50 value (10 nM) as that of Finasteride (8.5 nM) (13) while compound 250 has a higher IC 50 value of about 50 nM apparently due to the presence of large bromine atom [170].

Working on similar lines some more novel 4,16-pregnadiene-6,20-dione derivatives were synthesized and evaluated as 5␣reductase inhibitors. In this work, it has been demonstrated that compounds containing chlorine (251), bromine (252), iodine (253) atoms, and (254); without any substituent in the ester moiety) at C-3 produce a significant decrease of the prostate weight in castrated animals treated with testosterone (1). Therefore, it was proposed that the ester moiety at C-3 is functioning as a pharmacophore, Table 29 In vitro screening of compounds 262-265 against human 5␣-reductase.

Table 29

Non-steroidal 5␣-reductase inhibitors

A number of classes of non-steroidal inhibitors of 5␣-reductase have now been identified. It was anticipated that the use of nonsteroidal template can decrease the potential interaction with other enzyme or receptor of the steroidal endocrine system and can limit the complexity of target compound synthesis [173]. They have in fact emerged either from the design of compounds mimic of azasteroidal inhibitors, generally by the formal removing of one or more rings from the azasteroidal structure or by early non-steroidal lead (ONO-3805) (261) which was prepared as leukotriene synthesis inhibitor [174] or by high throughput screening. These compounds are generally thought to act all as competitive inhibitors vs. testosterone with exception of the epristeride analogues which are uncompetitive inhibitors. Non-steroidal inhibitors include benzo[f]quinolinones, pyridones and quinolinones which were mimics of 4-azasteroid inhibitors. Benzo[c]quinolinones were synthesized as mimics of 6-azasteroids while benzo[c]quinolizinones were designed as mimics of 10-azasteroids. The most potent and selective inhibitors of human type I 5␣-reductase are found among these classes of compounds. Almost all the other non-steroidal inhibitors can be grouped as carboxylic acid (generally butanoic acid) derivatives which are thought to act as non-competitive inhibitors versus testosterone in analogy to ONO 3805 (261).

Mimics of 4-azasteroids: benzo[f]quinolinones

Benzo[f]quinolinones were the first non-steroidal inhibitors prepared by the Lilly's researchers. They were derived by the removal of the D-ring from 4-azasteroids and replacing the C-ring with an aromatic one [175]. Most of these compounds are type I selective, although dual inhibitors can be obtained if an appropriate substitution is present at the position 8 on the aromatic ring. Two main classes of benzo[f]quinolinones have been described, the hexahydro derivatives (262-265), which have an unsaturation at positions 4a-10a, and the octahydro derivatives (266-271). In general octahydro derivatives are more potent inhibitors than the corresponding 4a-10a unsaturated compounds (Tables 29 and 30) and in both series the potency toward type I 5␣-reductase increases if an halogen atom is present at position 8 (in particular a Cl atom) and a methyl group at position 4; in fact the most potent inhibitor of the series is LY191704 (268) with IC 50 = 8 nM (Table 30). This molecule has progressed into human clinical trials.

Table 30

In vitro screening of compounds 266-274 against human 5␣-reductase.

The quantitative structure activity relationship study of these compounds has focused on the effect of 8 substituent on the aromatic ring, which can be accounted for by its lipophilic character [176]. They found that the optimum activity may reside in the property space around the chlorine substituent. The substitution of the 8-Cl atom by a F (270) or a Br atom (271) decreased very slightly the potency. Finally, several kinds of substituents, including complex aromatic groups, were introduced at the position 8 [177], and some potent type I selective inhibitors such as compounds 272-274 were prepared.

Pyridones, quinolinones and piperidines

Abell et al. synthesized a number of tricyclic thiolactams (275-276), aryl acid (277), bicyclic lactams (278-281) and bicyclic thiolactam (282) and evaluated them in vitro as inhibitors of type I and type II steroid 5␣-reductase (Table 31). Removal of two or more rings from 4-azasteroids resulted in a strong decrease of potency. The tricyclic thiolactams were found to be selective type I 5␣-reductase inhibitors and in general were less active than the corresponding lactams. The aryl acid 277 showed good dual isozyme inhibitory properties with significantly enhanced type II activity. Bicyclic lactams, in general, were found to be less active against type I 5␣-reductase than the tricycles. For example, compounds (278-279), lacking the B and D steroidal rings, were poor type I 5␣-reductase inhibitors, with the highest potency associated to the presence of the Cl atom on the aromatic ring of 279 [178]. A styryl (or azo) substituent dramatically enhances type II activity (and indeed type I activity with the bicycles) (280-281) and provided dual inhibitors of type I and type II. Hartmann and coworkers synthesized pyridones of the general formula 283 and 284 where the B-and C-rings of the steroid system have been replaced by an acyclic linker but these compounds display relatively weak activity versus both the rat and human 5␣reductase isozymes (K i > 20 M) [179][180][181].

Table 31

In vitro screening of compounds 275-282 against human 5␣-reductase.

Their poor potency however illustrates the need for both A-and B-rings to be present with the correct fusion pattern for good recognition at the enzyme active site. A series of 5-phenyl substituted 1-methyl-2-pyridones have also been prepared and tested against human and rat 5␣-reductase type I and type II. Compound 285 bearing bulky carboxamide substituents exhibited excellent 5␣reductase type II inhibitory activity with IC 50 value of 10 M [182].

McCarthy and coworkers have recently synthesized a series of 4 -substituted 5-aryl pyridones along with corresponding 1-arylpyridone derivatives and tested them against 5␣-reductase type I and type II expressed in transfected human embryonic kidney cells to examine structure activity relationship for the 4 position in pyridones.

Weak inhibition was observed against the type I isozyme for 4 -N-substituted acetamide compounds (286-289) while potent inhibition of type I isozyme was observed for compound 290 having 4 -benzoyl substituent and also for compounds (291-292) having long carbon chain tethers attached to the 4 -acetamide (Table 32). Thus further proving that large hydrophobic groups are tolerated in a region of the active site not involved in the enzymatic reaction [183].

Table 32

Few quinolinone derivatives such as 6-substituted 1H-quinolin-2-ones (293-294) and 2-methoxy quinolines (295-296) have also been synthesized. The most active inhibitor for the human type II isozyme was 6-[4-(N,N-diisopropylcarbamoyl) phenyl]-1Hquinolin-2-one 293 having K i 800 ± 85 nM, showing mostly competitive inhibitory patterns. A type I selective inhibitor could be identified with 6-[4-(N,N-diisopropylcarbamoyl) phenyl]-N-methyl-quinolin-2-one 294 (IC 50 = 510 nM) but 2-methoxy quinolines were not found to be active [184].

Hartmann and coworkers synthesized and evaluated a series of 2 -substituted 4-(4 -carboxy-or 4 -carboxymethylbenzylidene)-N-acylpiperidines as active steroid 5␣-reductase type II inhibitors. They synthesized several compounds from N-acyl-4-benzylidenepiperidine-4 -carboxylic acids. In the dicyclohexylacetyl series, fluorination in the 2-position of the benzene nucleus (297), exchange of the carboxy group by a carboxymethyl moiety (298) and combination of both structural modifications (299) led to highly active inhibitors of the human 5␣-reductase type II isozyme with IC 50 values of 297, 298 and 299 being 11, 6 and 7 nM, respectively, in comparison to Finasteride (13) having value of 5 nM [185]. In rat, compounds 300 (IC 50 = 3.44 and 0.37 M for type I and type II, respectively) and 302 (IC 50 = 0.54 and 0.69 M for type I and type II, respectively) displayed the best inhibition toward both isozymes. Compound 301 showed a strong inhibition toward type II human and rat enzyme (IC 50 = 60 and 80 nM) but only a moderate activity versus type I enzyme (IC 50 approximately 10 M for rat and human enzyme). Compound 303 (IC 50 in human type II enzyme being 0.26 M and in rat type II enzyme being 0.29 M) was found to be a moderate dual inhibitor probably due to higher flexibility of the open ring substituent [186].

Methyl esters of N-(dicyclohexyl)acetyl-piperidine-4-(benzylidene-4-carboxylic acid) (304) were designed and monitored for dual inhibition toward type II isozyme in BPH cell free preparation and for type I isozyme in DU 145 cells. Methyl esters, applied as hydrolytically stable precursor drugs to facilitate cell permeation, will yield the corresponding carboxylic acids as type II inhibitors after hydrolysis in the target organ. The esters themselves stable in human plasma and Caco-2 cells act as potent drug toward 5␣-reductase type I. Thus, dual inhibition of 5␣-reductase type I and type II can be achieved by applying a single parent compound [187].

Mimics of 6-azasteroids: benzo[c] quinolinones

On the basis of 6-aza-androst-4-en-3-one derivatives (Fig. 5) in which a vinylogous amide was inserted into a steroid nucleus as a transition state mimic for conversion of T (1) to DHT (2) and Lilly's Benzoquinoline derivatives (267-268), novel phenanthridin-3-one derivatives (305-307) were synthesized having vinylogous amide pharmacophore.

Although compounds were found to be 5␣-reductase type I selective and poor inhibitors of 5␣-reductase type II but overall these compounds did not showed promising inhibitory activity. The potency of compounds was found to increased from 305 (K i 10 M) to 306 (K i = 1.1 M) and 307 (K i = 0.92 M) this was due to the presence of methyl group on the A-ring corresponding to the 4-Me of Eli Lilly inhibitors. This effect is due to the presence of hydrophobic pocket in the active site of enzyme which is able to accommodate a small alkyl group located at the position 4 of the A-ring of these tricyclic inhibitors [188].

Mimics of 10-azasteroids: benzo[c] quinolizinones

Guarna et al. had synthesized two series of benzo[c]quinolizin-3-ones as novel inhibitors of human 5␣-reductase type I: 4aH-series with a double bond between the positions 1 and 2 (308-311) and 1H-series with a double bond between the positions 4 and 4a (312-317). The efficacy and selectivity of these compounds have been demonstrated on recombinant human 5␣-reductase type I expressed in CHO cells but they displayed very poor or no inhibition towards 5␣-reductase type II (Table 33). Increased activity of the compounds of 1H-series than those of corresponding inhibitors of 4aH-series has been attributed to the presence of double bond enabling conjugation between carbonyl and nitrogen atom [189][190][191]. The presence of a methyl group at position 4 (311, 313, 314, and 316) associated with a substituent at position 8, gave potent compounds in comparison with Finasteride (13) and the known 5␣-reductase I selective inhibitor LY191704 (268). All compounds inhibited the enzyme through a reversible competitive mechanism. The structure activity relationship carried out on this class especially methyl group at positions 1, 4, 5, 6 and 8 can be summarized as follows (Fig. 7) [191].

Figure 7

SAR for 4 -benzo[c] quinolizin-3-ones.

It was found that presence of substituent at position 8, either a chlorine or methyl group, generally increased the potency of the inhibitors. Thus compound 309 and 310 were potent were more active than unsubstituted compound 308 in 4aH-series. Similarly in 1H-series compounds 312 and 314 were more potent than unsubstituted compounds due to presence of 8-chlorine group. The introduction of a methyl group at position 4 was found to increase the potency in both series with effect being higher in 1H-series. A very strong increase of potency is observed in 8-chloro-4-methyl derivative 314 when compared to 4-methyl unsubstituted compound 312. The substitution with a methyl group at position 6 also affects potency in both series with effect being more prominent in 1H-series with compound 315 being more potent than unsubstituted 312. This effect of methyl substitution at position 6 seems Table 34 In vitro screening of compounds 318 and 319 against recombinant 5␣-reductase I and II expressed in CHO cells.

Non-steroidal aryl acids

Some novel 9,10-dihydrophenanthrene-2-carboxylic acids (326-328) were prepared by formally removing D-ring from parent androstene carboxylic acid inhibitors and contrary to them, were found to be selective 5␣-reductase I inhibitors. Introduction of a bromine atom at position 7 in compound 328 gave the most potent compound of the series. Substitution by chlorine at position 7 in compound 327 does not result in increase in potency as compared to unsubstituted compound 326. These compounds are supposed to interact with the positively charged enzyme-NADP + complex in an uncompetitive manner versus testosterone [195].

Moreover when double bond was introduced in the B-ring as in compound 329 (formally obtained by removing the D-ring from Epristeride 163) the selectivity toward 5␣-reductase I was found to be lost in favor of an increased potency towards 5␣-reductase II (Table 36) [196].

On removal of two or more rings from the parent steroidal compounds several aryl acids mimics of steroidal carboxylic acids have been synthesized (331-336). Hartmann et al. synthesized several N-substituted 4-(5-indolyl) benzoic acids but potent and selective human 5␣-reductase I inhibitors were not found from the series. Only compound 331 with IC 50 value of 67 nM against human 5␣-reductase I was the most potent inhibitor [197]. Similarly, compound 332 was found to possess IC 50 value of 680 nM [198]. [199]. Screening of aryl carboxylate versus the human 5␣-reductase isozymes by SmithKline-Beecham company led to the discovery of two series of selective and potent 5␣-reductase II non-steroidal inhibitors based on benzophenone and indolecarboxylic acids skeleton. In benzophenone series of inhibitors the linker between the A-and B-rings proved to be crucial whereas linker between B-and C-rings was more tolerant of variation. Both the A-ring carboxylic acid and C-ring are critical for activity. Compounds 337 and 338 were found to be most potent among the series with K i,app being 10 and 5 nM, respectively. In indole series there was a strong preference for substitution at the 5-or 6-position of the indole ring. Compounds 339 and 340 were most active among the series with K i,app being 10 and 40 nM, respectively The structure activity relationships of both the series are summarized as below (Figs. 8 and 9) [43,200].

Takami et al. synthesized various indole derivatives with varied substituents on the ␣,␤-unsaturated double bond and evaluated them for activity to inhibit rat prostatic 5␣-reductase. Among the various derivatives they found that (Z)-4-{2-[[3-[1-(4,4 -difluorobenzhydryl)indol-5-yl]-2pentenoyl]-amino]phenoxy}butyric acid (341,KF20405) was the most potent compound having activity about 20 times greater than Finasteride (13) and an IC 50 value of 0.48 ± 0.086 nM [201].

A novel series of indole and benzimidazole derivatives were also synthesized and evaluated for their inhibitory activity of rat prostatic 5␣-reductase. Among the series, 4-{2-[1-(4,4dipropylbenzhydryl)indole-5-carboxamido]phenoxy}butyric acid (342) and its benzimidazole analogue (343) showed potent inhibitory activities with IC 50 values of 9.6 ± 1.0 and 13 ± 1.5 nM, respectively [202]. In 2002, various N-substituted 4 -biphenyl-4-carboxylic acids were synthesized by increasing the conformational flexibility using an ether linker between the steroidal A-C-ring mimetics and were tested against human and rat 5␣-reductase type I and type II. Two compounds were found to be most potent with compound 345 showing an IC 50 value of 60 nM while 346 showed an IC 50 value improved by a factor of 5 from 1.9 to 0.38 M in comparison with the parent biphenyl compound 347 [204]. Novel substituted benzoyl benzoic acids and phenylacetic acids were synthesized by Salem et al. based on the template structure 338 and were evaluated for the inhibition of rat and human steroid 5␣-reductase isozymes I and II. The phenylacetic acid derivatives were more potent than the analogous benzoic acids. Bromination in the 4-position of the phenoxy moiety led to the strongest inhibitor of the series against human 5␣-reductase II (352; IC 50 = 5 nM), which was equipotent to Finasteride (13) while compounds 350 (IC 50 = 23 nM) and 351 (IC 50 = 27 nM) were also found to potent inhibitors against 5␣-reductase type II [206].

Kato and coworkers synthesized a series of pyrrole butyric acid derivatives and evaluated them for inhibitory activity on human and rat steroid 5␣-reductase. In case of para-aminobenzoyl pyrrole derivatives (353-356), the introduction of ethyl (354) or isopropyl (355) at C-4 increased the activity [IC 50 being 32 and 9.2, respectively], whereas the replacement with carboxyl (356) resulted in compounds with decreased inhibitory activity against human 5␣-reductase enzyme, indicating steric restriction in the binding site of the enzyme. Compound with m-amino benzyl pyrrole (357) moiety was found to be more active than the corresponding para isomer (355) [IC 50 = 3.2 nM].

Compound 358, having 2-hexyloctylamino group, was found to be most potent inhibitor among the compounds with IC 50 being 0.60 nM against human and 5.8 nM against rat 5␣-reductase [207]. A novel series of indole-3-alkanoic acids with varied N-benzyl substituents were also synthesized. Amongst these 4-[1-(6,6dimethyl-6H-dibenzo [b,d] pyran-3-yl) methyl indol-3-yl]-butyric acid (359; FR119680) displayed very high inhibitory activity against rat prostatic 5␣-reductase (IC 50 = 5.0 nM) [208].

Igarashi et al. found a novel series of phenoxybenzoic acids derivatives as potent inhibitors of human prostatic 5␣-reductase. It was found that introduction of a chloro (360), fluoro (361) or methoxy (362) group at 3-position of benzene ring (R 1 ) leads to formation of compounds with high inhibitory activity with IC 50 being 0.87, 0.67 and 0.56 nM, respectively [209].

A series of indoline and aniline derivatives have also been synthesized so as to inhibit both human and rat prostatic 5␣-reductase. Among the indoline series, 3-chloro-4-{[1-(4phenoxybenzyl)indolin-5-yl]oxy}benzoic acid (363; YM-36117) was found to be the most potent inhibitor against human enzyme having an IC 50 value of 5.3 and 46 nM against the rat enzyme while in aniline series, 3-chloro-4-{4-[N-(4-phenoxybenzyl)amino] phe-noxy}benzoic acid (364) turned out to be most potent inhibitor with an IC 50 against human and rat enzyme as 10 and 5.5 nM, respectively [210].

Bisubstrate inhibitors

Ishibashi et al. synthesized a series of novel benzofuran derivatives with both carboxy and 5-or 6-diphenylmethylcarbamoyl groups and their inhibitory activities against rat and human testosterone 5␣-reductase were tested in vitro. The derivatives were more active against human type I enzyme than against type II enzyme. The 6-carbamoyl derivative such as 365 tended to be more potent than the 5-carbamoyl ones such as 366 with 365 being the most potent compound having IC 50 value of 37.9, 50 and 340 nM against rat, human type I and human type II isozymes [211].

Later, they also synthesized a series of 2-phenylbenzofuran derivatives with a carbamoyl, alkylamino, or alkyloxy group at the 5 or 6 position of the benzofuran ring. It was found that carbamoyl derivatives had more potent inhibitory activities than the alkylamino or alkyloxy derivatives against the rat enzyme and the 6-carbamoyl derivatives tended to be more potent than the 5-carbamoyl ones. The 6-carbamoyl and 6-alkylamino derivatives were found to be more potent inhibitors against human type I enzyme than type II but on whole compounds were found not to be selective [212]. The non-steroidal o-hydroxyaniline (261; ONO-3805) was a weak compound in vitro versus human 5␣-reductase II. It is a bi-substrate inhibitor in which the butanoic acid moiety is thought to be localized in the region of the phosphate group of NADPH and the lipophilic part could be orientated in the region of the steroidal C and D-ring, thus occupying the hydrophobic pocket of the enzyme. The fact that this compound acts as noncompetitive inhibitor (versus T) and not as uncompetitive one, supports this hypothesis [173,213,214]. This prompted Pfizer to prepare the derivatives of 261 and subsequently C-3 acylindole (367) was prepared which had improved potency versus both human 5␣-reductase enzymes. The benzodioxolane (368) adopts a similar minimum conformation to the ether (367) and proved to be a potent dual inhibitor of both 5␣-reductase enzymes. In common with the steroidal carboxylic acid inhibitors, these compounds require the carboxylic acid moiety for potency and the 3-acylindole motif was found to be crucial for dual activity presumably by allowing access to both the conformations 369 and 370. The corresponding 2-methyl analog 371 which adopts conformation 370 due to the presence of methyl group on the Table 37 In vitro screening of compounds 367-371 against human 5␣-reductase I and II and rat 5␣-reductase.

Miscellaneous non-steroidal inhibitors

In search of novel non-steroidal mimics of steroidal inhibitors of 5␣-reductase, 4-(2-phenylethyl)cyclohex-1-ene carboxylic acids were synthesized with different substituents in para position of the phenyl ring such as N,N-diisopropylcarbamoyl (373), phenyl, phenoxy, etc. They turned out to be good inhibitors of the human prostatic 5␣-reductase isozyme II with 373 being the most potent one (IC 50 Due to the excellent estrogen receptor binding affinity of a series of 2 ,6 -disubstituted 4-hydroxy-4 -hydroxymethyl biphenyl derivatives Lesuisse et al. designed various biphenyls as surrogates of the steroidal backbone. They hypothesized that by introducing appropriate substituents non-steroidal estrogens could be tailored into 5␣-reductase inhibitors. Two compounds (377 and 378) emerged as potent type II 5␣-reductase inhibitors with IC 50 being 71 and 9.8 nM, respectively [221].

Chen et al. evaluated isoflavonoids as potential non-steroidal inhibitors of rat 5␣-reductase by using the hypothetical pharmacophore of 5␣-reductase. They proposed that although these compounds (379-382) were inhibitors of rat 5␣-reductase in the range of 27-49 M they could be evaluated as human 5␣-reductase inhibitors [222].

Hasoda et al. in 2007 designed and synthesized novel type I 5␣-reductase inhibitors by using 3,3-diphenylpentane skeleton as a substitute for the usual steroid skeleton. 4-(3-(4-(N-Methylacetamido) phenyl) pentan-3-yl) phenyldibenzylcarbamate (383) was found to be a competitive 5␣-reductase type I inhibitor with the IC 50 value of 0.84 M among the series [223].

Table 5

Table 23

Table 32