During the past three decades, research has focused on understanding the biologi- cal functions and effects of 5a-reductases and their 5a-reduced metabolites in order to obtain compounds able to block androgenic action through the inhibition of 5a- reductases (13). A high degree of selectivity is important to ensure that no other enzyme is affected and that no interactions exist with steroid receptors (10). Devel- opment of specific 5a-reductase inhibitors began in 1983, soon after DHT was reported to be the major androgen acting in the periphery. Great progress has since been made and continuing interests grow in designing and developing more potent and specific 5a-reductase inhibitors. DHT, being 2 to 10 times stronger than testosterone in androgen activity, was supposed to play a more important role than testosterone in many androgen-dependent diseases (14). Approximately 70% to 80% of serum DHT in men is produced by the type 2 isoenzyme, and 20% to 30% by the type 1 isoenzyme (21).
Since sebum production is regulated by DHT, inhibitors of type 1 5a- reductase may, therefore, form a new therapeutic class of agents designed to treat acne. It is expected that 5a-reductase inhibitors will produce fewer side effects than the presently available hormonal therapies. Inhibition of type 1 isoenzyme may represent a means of blocking the local production of DHT within the sebac- eous glands. This inhibition, in turn, may reduce sebum production and improve acne (21).
5a-reductase inhibitors can be classified according to their
1. chemical structures (steroidal/nonsteroidal),
2. type of inhibitor activity (competitive/noncompetitive/uncompetitive), and
3. specificity (type 1/type 2/dual inhibitors).
Notice that the mentioned inhibition constant values (Ki values) and IC50 values vary, depending on the species (e.g., human versus rat), the cells (primary cultures versus cell lines), and the tissue origin (e.g., testis versus prostate) examined.
In the following, a number of 5a-reductase inhibitors are highlighted. Due to the vast quantity of corresponding papers we are far from a position to claim a com- plete listing of all published inhibitors. However, we focus mainly on essential structural requirements for inhibitors, on characterized inhibitors for human type
1 isoenzyme, and on inhibitors that stand in clinical testing phase or that are already commercially available.
Structural Requirements for Inhibitors
The design of most synthetic 5a-reductase inhibitors is based on the transition state inhibitor paradigm, a concept that states that the affinity for the enzyme should be greater for molecules that mimic the transition state of the enzymatic process (Fig. 4B) (16). Molecular modeling studies suggest that there is a require- ment for
1. Groups to mimic the steroid A ring, in particular, the C-3 carbonyl group (Fig. 13) (26). It is concluded that any steroid-derived structure, possessing a C-3 polar group, should exhibit 5a-reductase inhibitory activity, whereas other structural factors determine the extent of the inhibition (27).
2. The space below the steroidal plane area about C-3, C-4, C-5, and C-6. This area should be sterically hindered, since the NADPH moiety requires access to the D4 site (26).
3. The area of the active site of the enzyme about the C-17 position of the steroid substrate, which appears to lack hydrogen bonding groups and is unrestricted (26).
In most synthetic and natural steroids, loss of the C-19 methyl group reduces the potency of the steroid to be recognized by 5a-reductase. Appropriate substi- tution at C-17 is, therefore, of particular importance in inhibitor design (5). That the C-17 hydroxyl group is not essential for inhibition is supported by inhibitors where the D-ring has been altered to a six-membered lactam resulting in potent activity. The volume of space occupied by the C-17 side chain of inhibitors is exten- sive. Therefore, it is postulated that the entry/exit to the active site of 5a-reductase results from this position (26) by lodging the C-17 side chains in specific cavities of the enzyme (28). A further postulation is that the increased inhibitory activity is associated with the lengthening of side chains. It is also proposed that the two iso- enzymes possibly vary in the positioning of the reducing NADPH moiety (26).
The first generation of synthetic inhibitors were steroids with a nitrogen atom located at different positions of the steroidal structure. Because of this characteristic, they are designated as azasteroids (Figs. 14A – C) (10). The presence of the nitrogen atom in the steroid structure increases the nucleophilicity of the carbonyl group at
C-3 and, thus, favors the interaction with an electrophilic residue in the active site. In general, azasteroids interact with the enzyme NADPH complex and are competi- tive with respect to testosterone (16).
4-Azasteroids are mimics of the “product-like” transition state, whereas
6-azasteroids and 10-azasteroids, which have an enone structure in the A ring, are mimics of the “substrate-like” transition state (Fig. 4B). 4- and 6-azasteroids are generally more rigid structures than 10-azasteroids, and thus may be more suit- able for developing models of the 5a-reductase active site (16).
4-Azasteroids have been studied extensively (6). Hundreds of substituted
4-azasteroids have been published and patented since the early 1980s, especially from Merck, Glaxo, and Pharmacia (29). Among these inhibitors, Finasteride (MK906) (Fig. 15A) was the first and, for the last decade, the only available active ingredient in systemic drugs, (Proscar and Propecia, Merck) approved for clinical use. Finasteride is a specific competitive type 2 5a-reductase inhibi- tor (IC50 type 2 ¼ 9.4 nM, IC50 type 1 ¼ 313 nm) (10) and has been shown to be effective for the treatment of benign prostatic hyperplasia (Proscar) and andro- genetic alopecia (Propecia), respectively (14). In humans Finasteride decreases prostatic DHT levels by 70% to 90% and reduces prostate size, while testoster- one tissue levels remain constant (5). In males with benign prostatic hyperpla- sia treated with Finasteride, there was no reduction of sebum production, which may be related to the possibility that 5a-reductase in the sebaceous gland remains unaffected by the drug (30). The corresponding amino acid sequences proposed to be involved in the Finasteride/human 5a-reductase
interaction are Ala26-Val27-Phe28-Ala29 for type 1 isoenzyme and Gly21-Ala22- Leu23-Ala24 for type 2 isoenzyme (Fig. 3) (19).
On the other hand, Dutasteride (GI198745, GG745) (Fig. 15B), one of the most potent dual inhibitors to date, is the first active ingredient in a drug (Avodart, GlaxoSmithKline) that inhibits both the isoenzymes (16). Avodart received Food and Drug Administration (FDA) approval in 2002 and was launched in 2003 so that after a monopoly for a decade as the only available 5a-reductase inhibitor, Finasteride will be joined by a second (20). This dual 5a-reductase inhibitor shows a greater inhibition of type 1 isoenzyme compared with Finasteride and is also potent against type 2 isoenzyme (IC50 type 1 ¼ 2.4 nM, IC50 type 2 ¼ 0.5 nm) (29). Biochemically, Dutasteride achieves greater and more rapid DHT suppression compared with Finasteride. It is reported to reduce circulating DHT levels by
90%. Clinically, it appears to be at least as good in terms of improving symptoms of benign prostatic hyperplasia. However, until these two drugs are for- mally compared, the true benefits of additional type 1 isoenzyme inhibition are unknown (20).
PNU 157706 (Pharmacia & Upjohn) (Fig. 15C) is highly potent in inhibit- ing human recombinant type 1 and 2 5a-reductase, showing IC50 values of 3.9 and 1.8 nM, respectively. PNU 157706 was shown to have no binding affinity for the rat prostate androgen receptor and to be devoid of any antiandrogen activity (31).
In contrast to 4-azasteroid derivatives, compounds with 4-methyl-4-aza functional- ity are generally more potent to type 1 5a-reductase (6). Experimental data showed that azasteroid derivatives containing groups larger than methyl result in poorer inhibition due to the restricted access of NADPH (32).
After the launch of Proscar in the early nineties, Merck has continued the research efforts and conducted clinical trials for at least three further compounds MK963, MK434, and MK386 for the treatment of acne, alopecia, and hirsutism (16), but will probably not bring any of these to the market. MK386 (Fig. 16A) is a selective inhibitor of human type 1 5a-reductase with an IC50 of 20 nM. A four- teen-day oral application with MK386 resulted in a dose-dependent suppression of serum and sebum DHT, without affecting semen DHT concentrations. Serum DHT concentrations were reduced 22% by a daily dose of 50 mg MK386, with no significant change in serum testosterone. Sebum DHT was reduced by 55% and
35% by doses of 50 and 20 mg, respectively (33). However, due a certain degree of hepatic toxicity in men the development of MK386 was stopped (10). A further development is L-751788 (Fig. 16B), which differs from Merck’s other 5a-reductase inhibitors by having a substituent on C-16 than on C-17. This makes it more selec- tive for type 1 isoenzyme.
The dual inhibitor 4-MA (Fig. 16C) (Merck) has been studied extensively and progressed to clinical trials for the treatment of benign prostatic hyperplasia. It showed a high activity for both isoenzymes (IC50 ¼ 8.5 nM) (16) and a very low affi- nity for the androgen receptor and was thus not expected to produce undesirable antiandrogenic effects. However, 4-MA was subsequently shown to be an inhibitor of 3b-hydroxysteroid dehydrogenase, another steroid metabolizing enzyme, and to cause hepatoxicity, like MK386 (15).
Turosteride (FCE 26073, GlaxoSmithKline) (Fig. 16D), a promising new drug currently in clinical studies, is a selective inhibitor of type 2 isoenzyme (29).
Turosteride caused inhibition of the human enzyme with IC50 values of 53 nM and does not show relevant binding affinity to prostate androgen receptor.
EM-402 strongly inhibits human type 1 5a-reductase (IC50 ¼ 7.3 nM) and shows low potency on the type 2 isoenzyme in transfected cells in vitro. Topically applied EM-402 (Fig. 16E) exerts a potent local antiandrogenic effect without any systemic action in the hamster. EM-402 decreased the size of flank organs of ham- sters and also of the underlying sebaceous glands. Twice daily applications for four weeks at doses of 30, 100, and 300 mg reduce the size of the sebaceous glands by 38%, 42%, and 59%, respectively. Comparable results were observed on the sebaceous glands of the ears. In addition, a concentration-dependent inhi- bition of 5a-reductase activity of 47% to 80% in the flank organs and ears, respect- ively, were observed. EM-402 had no effect on prostatic and seminal vesicle weights (34).
6- and 10-Azasteroids
6- and 10-azasteroids 5a-reductase inhibitors have been synthesized showing different degree of 5a-reductase inhibition and different selectivity, some of them are under investigation for clinical application (10).
The presence of the nitrogen atom at position 10, conjugated with the carbonyl at C-3 (Fig. 14C), is an essential feature for good inhibition since, providing an increase of the negative partial charge on the oxygen, it probably determines a strong interaction with an electrophilic residue in the enzyme active site (35).
Among 6-azasteroids, the aryl substituted derivative (Fig. 17A) is a potent dual inhibitor of both isoenzymes and 17b-N-(t-butyl)carbamoyl substituted 10- azasteroid (Fig. 17B) displayed an inhibitory potency against type 2 5a-reductase with that of Finasteride (IC50 ¼ 10 – 122 nM, depending on the isomer), and was more active against type 1 isoenzyme (IC50 ¼ 127 nM) (16).
4-Oxa- and 4-Thiasteroids
A series of irreversible inhibitors related to 4-azasteroids but with a different heteroa- tom at the C-4 position (sulphur and oxygen) has also been reported by Merck (29).
Flores et al. (5) synthesized novel pregnadiene-dione and pregnatriene-dione deriva- tives (Fig. 18A and B). The trienones showed consistently higher 5a-reductase inhibi- tory activity than the corresponding dienones. Unfortunately, there is no hint on the enzyme specificity. It is believed that the trienones inactivate the enzyme by an irre- versible Michael type addition of the nucleophilic portion of the enzyme to the
conjugated double bond of the steroid. The trienones having a more coplanar struc- ture react faster with the enzyme and thus show a higher inhibitory activity.
Steroidal Carboxylic Acids
A number of 3-androsten-3-carboxylic acids (Fig. 19A) were designed to mimic the putative enzyme bound enolate intermediate. Because of favorable electrostatic interactions between the carboxylate and the positively charged oxidized cofactor, the acrylate preferentially binds in a ternary complex with enzyme and NADPþ, which leads to the observed uncompetitive kinetic (6).
Epristeride (ONO-9302, SK&F 105657, GlaxoSmithKline) (Fig. 19B) is an uncompetitive, potent selective 5a-reductase type 2 inhibitor and is under phase III clinical testing for the treatment of benign prostatic hyperplasia (36). Epristeride has exhibited the ability to lower serum DHT levels by 50% (5).
Steroidal Oxime Inhibitors
Novel steroidal oxime inhibitors (Fig. 20) show an excellent inhibitor activity toward both types of the human 5a-reductases (5).
In the field of nonsteroidal inhibitors, the most interesting molecules are the group of benzoquinolinone and benzo[c]quinolizinone derivatives. Benzoquinolinone derivatives have been designed starting from the related parent 4-azasteroids and the benzo[c]quinolizinone derivatives from 10-azasteroids. They are very potent inhibitors of human 5a-reductases with, in most cases, higher activity against the type 1 isoenzyme. Some compounds are dual inhibitors (35).