Steroid 5α-reductase: Comparative study of mechanism of inhibition by nonsteroids ONO-3805 and LY191704

Article abstract of DOI:10.1006/bioo.1996.0033

Two nonsteroids, ONO-3805 and LY191704, were evaluated as inhibitors of the human and rat 5α-reductases (5αR). ONO-3805 was prepared in a 12-step convergent synthesis. This compound is a potent inhibitor of the human and rat 5αRs, with more potent inhibition seen against the rat enzymes. The inhibition patterns of this compound were best fit to an uncompetitive model which suggests binding in a ternary complex with enzyme and NADP⁺. Apparent K(i) values of 27, 31, 1, and 0.5 nM versus testosterone were obtained with human type 1, human type 2, rat type 1, and rat type 2 5αR, respectively. Multiple inhibition studies with ONO-3805 and NADP⁺ support synergistic binding of these two inhibitors with all isozymes. LY191704 was also evaluated as an inhibitor of the human and rat 5αRs. This compound is a selective, competitive inhibitor of human type 1 5αR. Poor inhibition was observed with human type 2 and rat types 1 and 2 5αR.

Full text of DOI:10.1006/bioo.1996.0033

BIOORGANIC CHEMISTRY 24, 386–400 (1996)
ARTICLE NO. 0033
Steroid 5Ͱ-Reductase: Comparative Study of Mechanism of
Inhibition by Nonsteroids ONO-3805 and LY191704
GEORGIANNA S. HARRIS, KENNETH
ELLSWORTH, BRUCE E. WITZEL
,
AND
R
ICHARD L. TOLMAN
Departments of Enzymology and Medicinal Chemistry, Merck Research Laboratories,
Rahway, New Jersey 07065
Received July 2, 1996
Two nonsteroids, ONO-3805 and LY191704, were evaluated as inhibitors of the human
and rat 5Ͱ-reductases (5ͰR). ONO-3805 was prepared in a 12-step convergent synthesis. This
compound is a potent inhibitor of the human and rat 5ͰRs, with more potent inhibition
seen against the rat enzymes. The inhibition patterns of this compound were best fit to an
uncompetitive model which suggests binding in a ternary complex with enzyme and NADPϩ.
Apparent Ki values of 27, 31, 1, and 0.5 n
M versus testosterone were obtained with human
type 1, human type 2, rat type 1, and rat type 2 5ͰR, respectively. Multiple inhibition studies
with ONO-3805 and NADPϩ support synergistic binding of these two inhibitors with all
isozymes. LY191704 was also evaluated as an inhibitor of the human and rat 5ͰRs. This
compound is a selective, competitive inhibitor of human type 1 5ͰR. Poor inhibition was
observed with human type 2 and rat types 1 and 2 5ͰR.
1996 Academic Press, Inc.
INTRODUCTION
Dihydrotestosterone (DHT)1 rather than testosterone (T) is considered the pri-
mary androgenic mediator in the development of benign prostatic hyperplasia
(BPH), male pattern baldness, acne, and hirsutism (1,2). As a consequence, consider-
able effort has been devoted toward the design of inhibitors of steroid 5Ͱ-reductase
(5ͰR), the enzyme which catalyzes the conversion of T to DHT. This enzyme is
found in the androgen-responsive tissues such as prostate, seminal vesicle, epididy-
mis, and skin where DHT rather than T is the active androgen (3–7). Two isozymes
of the enzyme are known, which are localized in different tissues (7–10). The ability
to tissue-specifically inhibit the isozymes of 5ͰR has heightened interest in this
enzyme as a therapeutic target.
Various steroidal and nonsteroidal inhibitors of 5ͰR have been reported, and
several are now in human clinical study. Steroidal compounds in the ⌬1-3-oxo-4-
aza series such as finasteride (Fig. 1) are potent, competitive, mechanism-based
inactivators of the enzyme (11, 12). Finasteride, a selective inhibitor of the human
type 2 5ͰR, reduces serum DHT by ȁ70% and produces beneficial effects in men
with BPH (13). The 3-carboxy-androstadienes (Fig. 1) were designed to mimic the
1 Abbreviations used: DHT, dihydrotestosterone; T, testosterone; 5ͰR, 5Ͱ-reductase; BPH, benign
prostatic hyperplasia.
386
0045-2068/96 $18.00
Copyright 1996 by Academic Press, Inc.
All rights of reproduction in any form reserved.

STEROID 5Ͱ-REDUCTASE
387
F
IG. 1. Inhibitors of 5ͰR.
enolate transition state and display uncompetitive kinetics indicating binding to
E:NADPϩ complex (Fig. 2) (14). In human clinical trials, epristeride produces
similar effects on serum DHT as finasteride, albeit at higher doses (15). Long-term
clinical effects of epristeride are not yet known.
Among the many other inhibitors of 5ͰR, there are two major series of nonsteroi-
dal inhibitors: the aryl(oxy)butanoate and benzoquinolinone series. ONO-3805 (Fig.
1) was selected for study as the prototype aryl(oxy)butanoate since it is the earliest
and best described in the literature (16–20). Although chemical synthesis and enzy-
mological studies are only tersely disclosed, a number of in vivo studies have been
published. ONO-3805 has been shown to be effective in reducing prostate size in
FIG. 2. Comparison of competitive and uncompetitive inhibitors of 5ͰR.

388
HARRIS ET AL.
rats (21, 22). Clinical studies in humans are reportedly underway, although recent
results are not available (23, 24). There are a host of other inhibitors in this structural
class including inhibitors from Fujisawa (FK-143) (25–29), Kyowa Hakko Kogyo
(KF18678) (30), and Pfizer (31, 32).
The second major nonsteroidal series is the benzoquinolinones, of which
LY191704 (Fig. 1) is the prototype (33, 34). Although LY191704 is the first example
of a selective human type 1 5ͰR nonsteroidal inhibitor, the in vivo evaluation of
this compound has not yet been reported.
As a continuation of our studies on the mechanism of various inhibitor classes
of 5ͰR, we present results on the mechanism of the nonsteroids ONO-3805 and
LY191704. Furthermore, we describe here a synthesis of ONO-3805 which reveals
details not available previously (17).
EXPERIMENTAL PROCEDURES
[7-3H]Testosterone was purchased from New England Nuclear. Aquasol 2 and
Flo-Scint II were obtained from New England Nuclear and Radiomatic Instruments,
respectively. All other chemicals were obtained from Sigma Chemical Co. or Aldrich
Chemicals. Column chromatography was performed on E. Merck silica gel 60
(70–230 mesh) or grade 62 (60–200 mesh). Mass spectra were obtained with a
Varian MAT 731 instrument and were consistent with proposed structures for all
chemical intermediates (Scheme 1).
Chemical Synthesis
Synthon A
Ethyl 4-(2-nitrophenoxy)butyrate. To a stirred solution of 2-nitrophenol (1.4 g,
10 mmol) and ethyl 4-bromobutyrate (2.1 g, 1.57 ml, 11 mmol) in dry acetone (35
ml) was added anhydrous, ground potassium carbonate (2 g, 14.5 mmol). The
resultant orange-colored mixture was then heated under a nitrogen atmosphere at
gentle reflux until the color due to the phenol anion had dissipated and a yellow
mixture remained. Concentration of the cooled and filtered mixture gave an oil
which, on flash chromatography (silica gel; ethyl acetate/hexane or methylene chlo-
ride as eluant), yielded 2.4 g (96% yield) of product as an oil.
Ethyl 4-(2-aminophenoxy)butyrate. A solution of the nitro compound (1.27 g, 5.0
mmol) in ethyl acetate (15 ml) containing 5% palladium on carbon (0.20 g) was
reacted in a hydrogen atmosphere (40 psi) at room temperature until hydrogen
uptake ceased. The mixture was then filtered through a layer of anhydrous sodium
sulfate and concentrated in vacuo to yield product (1.0 g) as an oil/low-melting solid.
Synthon B
4-Methoxy-2,3-dimethylbenzoic acid. To a stirred solution of 4-methoxy-2,3-di-
methylbenzaldehyde (4.92 g, 30 mmol) in glacial acetic acid (25 ml) at ca. 15ЊC was
added a chromium trioxide solution (15 ml); prepared from chromium trioxide (9.9

STEROID 5Ͱ-REDUCTASE
389
S
CHEME 1. (i) 2-Nitrophenol, K2CO3 , Me2CO; (ii) H2 , 5% Pd/C, EtOAc; (iii) CrO3 , HOAc; (iv) BBr3 ,
CH2Cl2; (v) MeOH, H2SO4; (vi) AcCl, AlCl3 , C2H4Cl2; (vii) NaBH4 , MeOH; (viii) HBr, PhH; (ix)
PhCH2Br, K2CO3 , Me2CO; (x) NaOH, aq. MeOH; (xi) (COCl)2 , CH2Cl2; (xii) H2 , 10% Pd/C, EtOAc-
HOAc; (xiii) K2CO3 , Me2CO; (xiv) NaOH, aq. MeOH; (xv) K2CO3 , Me2CO; (xvi) NaOH, then HCl;
(xvii) DCC, DMAP, CH2Cl2 .
g, 100 mmol) dissolved in water (6 ml) and then diluted with glacial acetic acid (23
ml) dropwise over 15 min. The resultant mixture was allowed to stir with ice-bath
cooling until the internal temperature dropped to 15ЊC, and the bath removed.
After stirring overnight at ambient temperatures, the mixture was diluted with
water (ca. 150 ml), aged, and filtered, and the solid was washed well with water
and dried; 2.2 g of product was obtained as a gray solid.
4-Hydroxy-2,3-dimethylbenzoic acid. (35). To a stirred suspension of the
4-methoxy acid (1.8 g, 10 mmol) in dried dichloromethane (20 ml) at Ϫ70ЊC was
added boron tribromide (13 ml of 1
N
boron tribromide in dichloromethane) drop-
wise over 8 min. After stirring at this temperature for 3 h the cooling bath was
removed and the resultant suspension was allowed to stir overnight at ambient
temperatures. The reaction mixture was added to a stirred ice–water mixture (100
ml), the layers were separated, and the aqueous layer was exhaustively extracted
with dichloromethane to yield 1.15 g product as a tan solid. More compound
could be obtained by concentration of the aqueous solution followed by repeated
dichloromethane extractions.

390
HARRIS ET AL.
Methyl 4-hydroxy-2,3-dimethylbenzoate. To a solution of the benzoic acid (1.15
g, 6.9 mmol) in dry methanol (35 ml) was added five drops of concentrated sulfuric
acid (prediluted with a small amount of dry methanol) and the mixture was heated
at reflux until TLC analysis indicated the absence of starting material. The methanol
was removed in vacuo, the residue was stirred with a mixture of water–hexane and
filtered, and the resultant solid was washed with water and dried to give 1.1 g
product as a tan solid.
Synthon C
4-Isobutylacetophenone. To a stirred solution of isobutylbenzene (13.4 g, 15.7 ml,
0.1 mol) and acetyl chloride (7.82 ml, 0.11 mol) in dried 1,2-dichloroethane (200
ml) at ϩ3ЊC was added anhydrous aluminum chloride (16.0 g, 0.12 mol) in portions
over 22 min. The temperature rose to ϩ8ЊC during the addition and an orange
color developed. After an additional hour at ice-bath temperatures, the bath was
removed and the mixture allowed to stir at ambient temperature for 2 days. After
cautious addition of the reaction mixture to a stirred mixture of ice, water, and
methylene chloride, the mixture was allowed to warm to room temperature and
separated, and the organic layer was dried over anhydrous sodium sulfate. Concen-
tration followed by chromatography on a silica gel column using dichloromethane
as eluant gave the product as a fluid liquid in 68% yield.
1-(4-Isobutyl)phenylethanol. To a stirred solution of the acetophenone (1.76 g,
10 mmol) in dry methanol (25 ml) at ϩ5ЊC was added sodium borohydride (0.38
g, 10 mmol) and the resultant mixture was stirred in an ice bath for 10 min. The bath
was then removed, and the reaction mixture was stirred at ambient temperatures for
2 h. After removal of the methanol in vacuo the residue was covered with methylene
chloride (50 ml) and treated with water. When both layers cleared, the organic
layer was separated and dried over anhydrous sodium sulfate. Concentration of
the filtered solution yielded 1.74 g of product as a clear liquid.
1-(4-Isobutylphenyl)bromoethane. Anhydrous hydrogen bromide was gently bub-
bled into an ice-cooled stirred solution of the phenylethanol (0.78 g, 4.38 mmol)
in dried benzene (30 ml) for 15 min. After an additional 10 min at ice-bath tempera-
tures, the resultant two-phase mixture was stirred at ambient temperatures for 2 h,
dry nitrogen was then bubbled through the mixture to displace the excess hydrogen
bromide, the mixture was allowed to settle, and the benzene layer was carefully
decanted from the water droplets. Removal of the benzene in vacuo yielded the
product (1.03 g) as a clear, pale-yellow oil that darkened with time.
Coupling of A ϩ B
Methyl 4-benzyloxy-2,3-dimethylbenzoate. A mixture of B (1.8 g, 10 mmol), benzyl
bromide (1.31 ml, 11 mmol) and anhydrous ground potassium carbonate (5.0 g, 36
mmol) in dry acetone (50 ml) was heated and worked up as for A. Flash chromatog-
raphy on silica gel (using 15% ethyl acetate–hexane as eluant) of the residue thus
obtained gave 2.1 g of product as a thick oil which solidified on standing.
4-Benzyloxy-2,3-dimethylbenzoic acid. A mixture of the benzyloxy ester (1.51 g,
5.6 mmol), water (4.2 ml), 2.5
N
sodium hydroxide solution (7.0 ml, 17.5 mmol),

STEROID 5Ͱ-REDUCTASE
391
and methanol (120 ml) was heated at gentle simmer under a nitrogen atmosphere
until TLC analysis showed the absence of ester. The methanol was removed in
vacuo, the residue was stirred with water (150 ml), and the resulting suspension
was treated dropwise with 2
N
hydrochloric acid (11.2 ml) and aged. Filtration
followed by water washing and drying gave 1.39 g of product as a white solid.
Ethyl 4-(2-(4-benzyloxy-2,3-dimethylbenzoylamino)butanoate. To a suspension
of the acid (0.256 g, 1.0 mmol) in anhydrous dichloromethane (5.0 ml) at room
temperature was added oxalyl chloride (2.0 ml, 23 mmol) dropwise over 2 min.
After 1.5 h the volatiles were removed in vacuo to yield a crust. This crust was
dissolved in dichloromethane (2 ml) and added dropwise to a stirred ice-cold mixture
of A (0.245 g, 1.1 mmol) and dry pyridine (1.0 ml) in dichloromethane (15 ml) over
1 min. After 30 min the ice bath was removed and the mixture was stirred at
ambient temperatures overnight. The mixture was then transferred to a separatory
funnel with dichloromethane and washed with 0.5
N
hydrochloric acid (2ϫ) and
water, dried (sodium sulfate), filtered, and concentrated to 0.47 g of product.
Ethyl 4-(2-(4-hydroxy-2,3-dimethylbenzoylamino)-phenoxy)butanoate. A solu-
tion of the benzylated coupling product (0.23 g, 0.5 mmol) in ethyl acetate (20 ml)
containing one small drop of glacial acetic acid and 10% palladium on carbon (0.2
g) was reduced under a hydrogen atmosphere (40 psi, room temperature) until no
benzyl ether remained by TLC. The filtered and concentrated mixture was then
flash chromatographed on silica gel (15–35% ethyl acetate/hexane eluant) to give
0.163 g of product as a thick oil which solidified on standing.
Coupling of B ϩ C
Methyl 4-(1-(4-isobutylphenyl)ethoxy)-2,3-dimethyl-benzoate. A stirred mixture
of B (0.75 g, 4.14 mmol), C (1.0 g, 4.14 mmol), and anhydrous, finely ground
potassium carbonate (0.75 g, 5.4 mmol) in anhydrous acetone (30 ml) was heated
at reflux for 25 h, cooled, and filtered. Concentration of the filtrate followed by
flash chromatography on silica gel (5% ethyl acetate/hexane as eluant) gave 0.92
g of product as a pale yellow oil.
4-(1-(4-Isobutylphenyl)ethoxy)-2,3-dimethylbenzoic acid. A stirred mixture of the
resultant ester from above (0.41 g, 1.2 mmol) in methanol (25 ml), water (1 ml),
and 2.5
N
sodium hydroxide solution (1.3 ml, 3.25 mmol) was simmered under a
nitrogen atmosphere until TLC analysis showed the absence of ester (ca. 48 h).
The cooled solution was concentrated in vacuo, the resultant crust was stirred with
water (75 ml), and the solution was filtered and acidified with 2
N
hydrochloric
acid. After standing for an hour, the solid was filtered, washed with water, and
dried to give 0.37 g of product as a pale, cream-colored solid.
Coupling of C to [A ϩ B]
Ethyl 4-(2-(4-(1-(4-isobutylphenyl)ethoxy)-2,3-dimethylbenzoylamino)phenoxy)-
butanoate. A mixture of [A ؉ B] (0.536 g, 1.44 mmol), C (0.40 g, 1.66 mmol), and
anhydrous ground potassium carbonate (0.6 g) in acetone (40 ml) was heated and
worked up as described above. Flash chromatography on silica gel (10–20% ethyl

392
HARRIS ET AL.
acetate/hexane as eluant) of the residue thus obtained gave product (0.72 g) identical
to that obtained from the coupling of A to [B ؉ C].
ONO-3805: 4-(2-(4-(1-(4-Isobutylphenyl)ethoxy)-2,3-dimethylbenzoylamino)phe-
noxy)-butanoic acid. To a stirred, ice-cooled solution of the above ester (0.72 g,
1.35 mmol) in methanol (35 ml) was added water (seven drops) and 2.5
N
sodium
hydroxide solution (1.3 ml, 3.3 mmol) dropwise. After 5 min the bath was removed
and the mixture was stirred under a nitrogen atmosphere at ambient temperatures
until TLC analysis showed no ester (ethyl or methyl, from ester exchange) remained.
The methanol was removed in vacuo (Ͻ25ЊC), the residue was dissolved in water
(90 ml) and filtered, and the filtrate was acidified dropwise with 2
N
hydrochloric
acid. The resultant precipitate was filtered, washed well with water, and dried to
give 0.66 g of product as a waxy solid; mass spectral data and TLC mobility were
consistent with published data (17). Anal. Calcd for C31H37NO5 (503.64): C, 73.93;
H, 7.41; N, 2.78. Found: C, 74.09; H, 7.63; N, 2.61.
Coupling of A to [B ϩ C]
Ethyl 4-(2-(4-(1-(4-isobutylphenyl)ethoxy)-2,3-dimethylbenzoylamino)phenoxy)
butanoate. To a stirred mixture of [B ؉ C] (0.082 g, 0.25 mmol) and A (0.11 g, 0.49
mmol) in dried dichloromethane (5 ml) at room temperature was added N,N,-
dimethylaminopyridine (0.032g, 0.26 mmol) followed by N,NЈ-dicyclohexylcarbodii-
mide (0.069 g, 0.33 mmol), and the mixture was covered with a nitrogen atmosphere.
A precipitate of N,NЈ-dicyclohexylurea appeared within 5 min. After stirring over-
night the mixture was filtered, the filtrate was concentrated, the residue was tritu-
rated with small amounts of ether and dichloromethane separately, and the com-
bined ether/dichloromethane extracts were concentrated. Chromatography on a
predeveloped silica gel preparative plate using 22% ethyl acetate/hexane as eluant
yielded 0.11 g of ONO-3805 ethyl ester as a clear oil.
S
TEROID 5ͰR ASSAYS
Enzyme assays were carried out as described previously (7) using baculovirus-
expressed human 5ͰRs (36) and COS cell-expressed rat 5ͰRs (37). Assays of human
and rat type 1 5ͰR were conducted in 40 m
buffer (SID) (38), pH 6.5, while 40 m SID, pH 5.5, was used to assay the type 2
5ͰRs. NADPH was routinely used at a final concentration of 0.5 m
M
succinate/imidazole/diethanolamine
M
M.
I
NHIBITION
STUDIES
Stock solutions of ONO-3805 and LY191704 were prepared in ethanol and diluted
to the appropriate concentration in 10% ethanol. In these studies the final concentra-
tion of ethanol was less than 2.5%. IC50 values were determined at Km concentrations
of T using a five-point titration ranging in concentration from 0.1 to 1000 n
M
for
type 1 5ͰR or from 1 to 10,000 n for type 2 5ͰR.
M
DATA
A
NALYSIS
Kinetic data were fit to Eqs. [1]–[4] using nonlinear regression analysis. S reflects
the concentration of T and I represents inhibitor concentration. Ki and Kii are the
kinetic constants for competitive (Eq. [2]) and uncompetitive inhibitors (Eq. [3]),

STEROID 5Ͱ-REDUCTASE
393
respectively. Double inhibition studies were analyzed using Eq. [4] as described by
Yonetani and Theorell (39). For this analysis, I and J are the concentrations of the
variable inhibitors and vi and vo are the velocities in the presence and absence of
these inhibitors. Ki and Kj are the kinetic constants for the inhibitors and ͱ is the
constant which reflects the degree of interaction between I and J.
v ϭ VmaxS/(Km ϩ S)
[1]
[2]
[3]
[4]
v ϭ VmaxS/[Km (1 ϩ I/Ki) ϩ S]
v ϭ VmaxS/[KM ϩ S(1 ϩ I/Kii)]
vi ϭ vo/[1 ϩ I/Ki ϩ J/Kj ϩ IJ/(ͱKiKj)]
RESULTS
CHEMISTRY
ONO-3805 comprises three substituted aryl rings: ring A, an aminophenoxybuta-
noate; ring B, an alkyl-p-hydroxybenzamide, and ring C, a p-isobutylbenzyl ether.
Synthesis of three synthons was devised to support a convergent synthesis of ONO-
3805; synthon A, ethyl 4-(2-aminophenoxy)butanoate, from 2-nitrophenol and ethyl
Ͳ-bromobutyrate; synthon B, methyl 2,3-dimethyl-4-hydroxybenzoate, from 2,3-
dimethyl-4-methoxybenzaldehyde; and synthon C, 1-(4-isobutylphenyl)bromo-
ethane from isobutylbenzene.
In the convergent connection of the three pieces, synthon A was coupled to B
with powdered potassium carbonate in acetone, after protection of the synthon B
phenol as a benzyl ether and conversion of the acid to the acid chloride. The
connection to synthon C was then undertaken with anhydrous potassium carbonate
in acetone followed by saponification of the ethyl ester to afford ONO-3805. Alter-
natively, synthon B could be coupled to C with anhydrous potassium carbonate,
the ester saponified, and the product then covalently attached to synthon A with
carbodiimide in the presence of dimethylaminopyridine followed by saponification
of the ester to afford the final product identical in all respects to the material
described in the ONO patents (17).
INHIBITION OF 5ͰR BY ONO-3805
The effect of ONO-3805 on the conversion of T to DHT by the human and rat
5ͰRs is presented in Table 1. IC50 values listed in Table 1 represent the concentration
of inhibitor required to achieve 50% inhibition of the conversion of T to DHT at
Km levels of T. These studies were conducted at Km levels of T to minimize effects
of differences in mechanism of inhibition on apparent potency. ONO-3805 is an
inhibitor of human types 1 and 2 5ͰR with IC50 values of 9–10 n . Surprisingly,
M
IC50 values are Ͼ11-fold higher with the human 5ͰRs indicating ONO-3805 is more
potent against the rat isozymes of this enzyme.

394
HARRIS ET AL.
TABLE 1
Inhibition of Human and Rat 5ͰR’s by ONO-3805 and LY191704
IC50 (nM)
KmT
(Ȑ
Isozyme
M
)
ONO-3805
LY191704
Human
Type 1 5ͰR
Type 2 5ͰR
Rat
0.3
7.5
9.4
8.9
3.8
Ͼ1000
Type 1 5ͰR
Type 2 5ͰR
1.7
0.15
0.8
Ͻ0.1
Ͼ1000
390
M
ECHANISM OF
I
NHIBITION BY ONO-3805
ONO-3805 is a potent nonsteroidal inhibitor of both human types 1 and 2 5ͰR.
As presented in Fig. 3, the kinetic patterns obtained with this compound with
human type 2 5ͰR were best fit to a linear uncompetitive model when both
T (Kii ϭ 31 Ϯ 2 mM; Km ϭ 0.63 Ϯ 0.04 m
M
) and NADPH (Kii ϭ 118 Ϯ 6.8 n ,
M
Km ϭ 1.7 Ϯ 0.14 m
M, data not shown) are used as variable substrates. Uncompetitive
inhibition by ONO-3805 was also obtained using human type 1, rat type 1, and rat
type 2 5ͰR as sources of enzyme. However, the magnitude of the inhibition constants
F
M
IG. 3. Uncompetitive inhibition of 5ͰR by ONO-3805. Human type 2 5ͰR was assayed with 0.1–3
Ȑ
3H-T in the presence of ONO-3805 at concentration indicated. Following a 30-min incubation, the
mixture was quenched with cyclohexane:ethylacetate and T separated from DHT by HPLC as previously
described. The data were fit by nonlinear regression to Eq. [3] to give Km ϭ 0.63 Ϯ 0.04 Ȑ and Ki ϭ
31 Ϯ 2.0 n . (᭹) Experimental data, (—) theoretical line.
M
M

STEROID 5Ͱ-REDUCTASE
395
TABLE 2
Inhibition Analysis of Human and Rat 5ͰRs by ONO-3805
Enzyme
Human
Type 1 5ͰR
Type 2 5ͰR
Rat
Variable substrate
Km (Ȑ
M
)
Ki (n
M
)
Pattern
T
T
17 Ϯ 1.6
0.63 Ϯ 0.04
27 Ϯ 1.8
31 Ϯ 2.0
UCa
UC
Type 1 5ͰR
Type 2 5ͰR
T
T
3.6 Ϯ 0.4
0.1 Ϯ 0.01
0.7 Ϯ 0.05
0.5 Ϯ 0.1
UC
UC
a UC, uncompetitive.
varied Ͼ20-fold with the different isozymes. The values obtained in these experi-
ments are included in Table 2 for comparison.
Uncompetitive inhibition, when interpreted in the context of the ordered binding
of substrates by 5ͰR, indicates that ONO-3805 binds to the E:NADPϩ complex
(Fig. 2). Multiple inhibition studies as originally described by Yonetani and Theorell
were undertaken to support that the inhibitor binds to this form of the enzyme
(39). In these studies ONO-3805 and NADPϩ as second inhibitor were varied at
constant levels of T and NADPH and the data were fit to Eq. [4]. In this analysis,
F
IG. 4. Multiple inhibition analysis of 5ͰR with ONO-3805 and NADPϩ. Human type 2 5ͰR was
assayed at Km levels of T and NADPH in the presence of variable levels of ONO-3805 and NADPϩ.
The data were fit by nonlinear regression to Eq. [4] to give Ki ϭ 20 Ϯ 1.4 n
M and Kj ϭ 1300 Ϯ 70 ȐM
for ONO-3805 and NADPϩ, respectively. In this analysis ͱ ϭ 0.5 Ϯ 1 reflecting synergistic binding of
ONO-3805 and NADPϩ. (᭹) Experimental data, (—) theoretical line.

396
HARRIS ET AL.
TABLE 3
Multiple Inhibition Analysis
Enzyme
I
J
Ki (n
M)
Kj (Ȑ
M)
ͱ
Human
Type 1
Type 2
Rat
Type 1
Type 2
ONO-3805
ONO-3805
NADPϩ
NADPϩ
14 Ϯ 1.0
20 Ϯ 1.4
1200
1300 Ϯ 70
0.5 Ϯ 0.1
0.5 Ϯ 0.1
ONO-3805
ONO-3805
NADPϩ
NADPϩ
1.1 Ϯ 0.1
0.3 Ϯ 0.04
79 Ϯ 6
90 Ϯ 12
0.4 Ϯ 0.1
0.5 Ϯ 0.1
ͱ reflects the degree of cooperativity between the two inhibitors. If ͱ Ͻ 1, there
is a positive interaction between the inhibitors, indicating that they bind synergisti-
cally with respect to each other. As ͱ approaches infinity (ȍ), the inhibitors bind
to the enzyme in a mutually exclusive manner. Results of double-inhibition studies
with human type 2 5ͰR, ONO-3805, and NADPϩ are presented in Fig. 4, which
yielded Ki ϭ 20 Ϯ 1.4 n
M
, Kj ϭ 1300 Ϯ 70 Ȑ , and ͱ ϭ 0.5 (see also Table 3). This
M
result is consistent with the synergistic binding of ONO-3805 and NADPϩ to the
enzyme as expected for an uncompetitive inhibitor of 5ͰR. Similar results were
obtained with human type 1 5ͰR as well as rat types 1 and 2 5ͰR (Table 3).
INHIBITION BY LY191704
LY191704 represents another series of nonsteroidal inhibitors of 5ͰR. As indi-
cated in Table 1, LY191704 is a potent, selective inhibitor of the human type 1
5ͰR. Poor inhibition of human type 2 and rat types 1 and 2 5ͰR was observed.
Given the structural similarity of the A and B ring of the benzophenone to the
azasteroid, it is reasonable to assume that LY191704 can compete with T for the
substrate binding site on the enzyme. If this assumption is correct, then LY191704
should be a competitive inhibitor of the human type 1 5ͰR versus T. The mechanism
of inhibition of this compound was reinvestigated to probe the mode of binding to
5ͰR. Despite previous reports of noncompetitive inhibition of the human type 1
5ͰR (33, 34), the data in Fig. 5 are best fit to competitive inhibition by LY191704
with Ki ϭ 4.6 Ϯ 0.3 n .
M
DISCUSSION
In the present study we have compared the mechanism of inhibition of human
and rat 5ͰRs by two nonsteroidal inhibitors. The results indicate that ONO-3805
inhibits both human types 1 and 2 5ͰR while LY191704 selectively inhibits human
type 1 5ͰR. Large differences are observed in IC50 values for human and rat 5ͰRs
with these two nonsteroids. Although species differences in apparent inhibitor
potency are well established in the steroidal series of inhibitors such as 4-azasteroids
(e.g., finasteride) (40) and 3-carboxyandrostadienes (e.g., epristeride) (41), this is

STEROID 5Ͱ-REDUCTASE
397
F
IG. 5. Competitive inhibition human type 1 5ͰR by LY191704. Human type 1 5ͰR was assayed with
1–15 Ȑ
M
3H-T in the presence of LY191704 at the concentrations indicated. Following a 20-min incuba-
tion, the mixture was quenched with cyclohexane:ethylacetate and T separated from DHT by HPLC
as previously described. The data were fit by nonlinear regression to Eq. [2] to give Km ϭ 4.3 Ϯ 0.7
ȐM and Ki ϭ 4.6 Ϯ 0.3 nM. (᭹) Experimental data, (—) theoretical line.
the first direct comparison of the inhibition of human and rat isozymes of 5ͰR by
nonsteroidal inhibitors. From the results in Table 1, it is clear that differences in
potency and isozyme selectivity between human and rat presents a challenge in the
preclinical evaluation of these compounds.
Although the results in this report strongly support an uncompetitive mode of
inhibition by ONO-3805, it is worth noting that there are conflicting reports concern-
ing the mechanism of inhibition ONO-3805 and another related nonsteroid, FK143.
Nagamoto et al. (42) reported that ONO-3805 is noncompetitive with respect to
inhibition of rat pituitary 5ͰR, which they concluded was the type 1 5ͰR. Addition-
ally, Hirosumi et al. (28) reported that FK143 (4-[3-[3-[bis(4-isobutylphenyl)meth-
ylamino]benzoyl]-1H-indolyl-1-yl]butyric acid) inhibits both the rat and human
prostatic 5ͰRs in a noncompetitive fashion. The difference in the mode of inhibition
compared to that reported in this report may reflect the difficulty in carrying out
complex kinetic experiments with tissue sources of 5ͰR containing relatively low
levels of the enzyme of interest.
Our finding of uncompetitive inhibition by ONO-3805 suggests that this com-
pound binds to the enzyme in a ternary complex with NADPϩ (Fig. 2). Multiple
inhibition with studies with NADPϩ support synergistic binding of ONO-3805 with
NADPϩ, as expected for a compound with this mode of inhibition. In this analysis,
a ͱ Ͻ 1, the factor which reflects the degree of cooperativity between the inhibitors,

398
HARRIS ET AL.
indicates that inhibition by ONO-3805 is enhanced in the presence of NADPϩ. It
is reasonable to assume that ONO-3805 forms a charge transfer complex with
NADPϩ and in effect mimics the enolate. Similar results have been reported with
other carboxylic acid steroidal and nonsteroidal inhibitors of 5ͰR including eprister-
ide, benzophenonecarboxylic acid, and KF18678 ((E)-4,4Ј-͕2-[[3-(indol-5-yl)-1-oxo-
2-butenyl]amino]-difluorobenzhydryl͖ butyric acid), another of the aryl-butanoate
class related to ONO-3805 (14, 30, 43). This strongly suggests that carboxy groups
of the members of the nonsteroidal arylbutanoate series may overlay the carboxy
group of the carboxyandrostadiene series in their charge transfer complex interac-
tions with NADPϩ.
As indicated previously, LY191704 is the most widely studied of the nonsteroidal
inhibitors of 5ͰR. Despite this fact, it has not previously been described as a
competitive inhibitor of the human type 1 5ͰR. The reason for the discrepancy in
the mode of inhibition as reported originally by Hirsch et al. is not clear at present
(34). However, as presented in Fig. 5, this inhibitor clearly displays competitive
inhibition of the human type 1 5ͰR versus T. This finding may suggest that the
type 1 isozyme recognizes the benzoquinolone as an abbreviated steroid rather
than as a nonsteroid, thus allowing it to bind at the substrate binding site. This
result may have implications for other tricyclic inhibitors of 5ͰR. Recently, a tricyclic
aryl acid (7-bromo-9,10-dihydrophenanthrene-2-carboxylic acid) was described as
a potent selective inhibitor of the human type 1 5ͰR (44). It is reasonable to assume
that addition of the carboxy group allows the inhibitor to form a complex with
enzyme and NADPϩ, as is well established for epristeride (45). If so, the mode of
inhibition will change from competitive to uncompetitive as found with 4-azasteroids
and diene acids.
ACKNOWLEDGMENTS
The authors thank Nancy Thornberry and Herb Bull for their helpful discussions regarding the
multiple inhibition studies. Donald W. Graham, David Shih, and Derek Von Langen are gratefully
acknowledged for the synthesis of LY191704.
REFERENCES
1. IMPERATO-
M
CG
INLEY, J., GUERRERO, L., GAUTIER, T., AND
PETERSON, R. E. (1974) Science 27, 1213–
1215.
2. PETERSON, R. E., IMPERATO-
MC
G
INLEY, J., GAUTIER, T., AND TURLA, E. (1977) Am. J. Med.
S
62, 170–191.
3. ANDERSON, K. M., AND
4. BRUCHOVSKY, N., AND
LIAO, S. (1968) Nature 219, 277–279.
WILSON, J. D. (1968) J. Biol. Chem. 243, 2012–2021.
5. TAKAYASU, S., AND
6. WILSON, J. D., AND
ADACHI, K. (1972) J. Clin. Endocrinol. Metab. 34, 1098.
W
ALKER, J. D. (1969) Clin. Endocrinol. Metab. 34, 1098–1101.
7. HARRIS, G., AZZOLINA, B., BAGINSKY, W., CIMIS, G., RASMUSSON, G. H., TOLMAN, R. L., RAETZ
C. R. H., AND LLSWORTH, K. (1992) Proc. Nat. Acad. Sci. USA 89, 10787–10791.
8. ANDERSSON, S., BISHOP, R. W., AND USSELL, D. W. (1989) J. Biol. Chem. 16249–16255.
9. ANDERSSON, S., BERMAN, D. M., JENKINS, E. P., AND USSELL, D. W. (1991) Nature 354, 159–161.
,
E
R
R

STEROID 5Ͱ-REDUCTASE
399
10. THIGPEN, A. E., SILVER, R. I., GUILEYARDO, J. M., CASEY, M. L., MCCONNELL, J. D., AND RUSSELL,
D. W. (1993) J. Clin. Invest. 92, 903–910.
11. BAKSHI, R. K., RASMUSSON, G. H., PATEL, G. F., MOSLEY, R. T., CHANG, B., ELLSWORTH, K.,
ARRIS, G. S., AND OLMAN, R. L. (1995) J. Med. Chem. 38, 3189–3192.
12. BULL, H. G., GARCIA- ALVO, M., ANDERSSON, S. A., BAGINSKY, W. F., CHAN, H. K., ELLSWORTH
D., MILLER, R., STEARNS, R. A., BAKSHI, R. K., RASMUSSON, G. H., TOLMAN, R. L., MYERS, R. W.,
OZARICH, J. W., AND ARRIS, G. S. (1996) J. Am. Chem. Soc. 118, 2359–2365.
13. GORMLEY, G. J. (1995) Biomed. Pharmacother. 49, 319–324.
H
T
C
,
K
H
14. LEVY, M. A., BRANDT, M., HEYS, R., HOLT, D. A., AND
METCALF, B. W. (1990) Biochemistry
29, 2815–2824.
15. AUDET, P., NURCOMBE, H., LAMB, Y., JORKASKY, D., LLOYD-DAVIES, K. MORRIS, R. (1993) Clin.
Pharm. Therapeut. 53, 231.
16. NAKAI, H., TERASHIMA, H., AND
Activity, Ono Pharmaceutical Co. Chem. Abstr. 110(13), 1144752.
17. NAKAI, H., TERASHIMA, H., AND RAI, Y. (1988) Benzoylaminophenoxybutanoate derivs., Ono
Pharmaceutical Co. Chem. Abstr. 110(23), 212384t.
18. NAKAI, H., TERASHIMA, H., AND RAI, Y. (1988) New N-Subst’d and Ring Subst’d 2- or 3-Methylcin-
ARAI, Y. (1988) Novel Cinnamoylamide Derivs Have 5Ͱ-Reductase
A
A
namoylamide Cmpd Having Inhibitory Activity on 5Ͱ-Reductase, Ono Pharmaceutical Co. Chem.
Abstr. 111(3), 23231a.
19. NAKAI, H., TERASHIMA, H., AND
ARAI, Y. (1988). New 4-(2-Benzoylamino-phenoxy, thio or sulph-
inyl)Butanoic Acids Useful as 5Ͱ-Reductase Inhibitors for Treating e.g. Alopecia, Ono Pharmaceuti-
cal Co. Chem. Abstr. 111(1), 7074w.
20. NAKAI, H., KONNO, M., KOSUGE, S., SAKUYAMA, M. T., TODA, M., ARAI, Y., OBATA, T., KATSUBE
,
N., MIYAMOTO, T., OKEGAWA, T., AND AWASAKI, A. (1988) J. Med. Chem. 31, 84–91.
K
21. IMAI, K., TAKAHASHI, O., WATANABE, K., NAKAZAWA, Y., NAKATA, S., AND
YAMANAKA, H. (1991)
Acta Urol. Jpn. 37, 1669–1676.
22. TAKAHASHI, O., IMAI, K., WATANABE, K., NAKAZAWA, Y., NAKATA, S., KURITA, M., SAITO, Y.,
UBOTA, H., AND AMANAKA, H. (1992) Acta Urol. Jpn. 38, 305–310.
K
Y
23. ANONYMOUS (1990) Epalrestat’s Importance for Ono, Scrip. 1514, May 16, p. 32.
24. ANONYMOUS (1991) Ono’s Full-Year Results, Scrip. 1634, July 17, p. 14.
25. OKADA, S. SAWADA, K., KURODA, A., WATANABE, S., AND
TANAKA, H. (1993) Indole Derivative
as Testosterone 5-Alpha-Reductase Inhibitors, Fujisawa Pharmaceutical Co. Chem. Abstr.
120(11), 134279c.
26. HIROSUMI, J. NAKAYOMA, O., KOJO, H., TAKAHASHI, S., CHIDA, N., AND
N
OTSU, Y. (1994) Jpn. J.
Pharmacol. 64(Suppl. 1), 287.
27. KOJO, H. NAKAYAMA, O., HIROSUMI, J., CHIDA, N., NOTSU, Y., AND
OKUHARA, M. (1995) Mol.
Pharmacol. 48, 401–406.
28. HIROSUMI, J. Nakayama, O., FAGAN, T., SAWADA, K., CHIDA, N., INAMI, M., TAKAHASHI, S., KOJO
H., NOTSU, Y., AND KUHARA, M. (1995) J. Steroid Biochem. Mol. Biol. 52, 357–363.
29. HIROSUMI, J. NAKAYAMA, O., CHIDA, N., INAMI, M., FAGAN, T., SAWADA, K., SHIGEMATSU, S., KOJO
H., NOTSU, Y., AND KUHARA, M. (1995) J. Steroid Biochem. Mol. Bio. 52, 365–373.
,
,
,
O
O
30. KUMAZAWA, T., TAKAMI, H., KISHIBAASHI, N., ISHII, A., NAGAHARA, Y., HIRAYAMA, N., AND
OBASE
H. (1995) J. Med. Chem. 38, 2887–2892.
31. BLAGG, J. FINN, P. W., GREENGROSS, C. W., AND
Reductase Inhibitors, in WO 95/05375. Pfizer, Limited. Chem. Abstr. 120(9), 106982h.
32. MAW, G., BLAGG, J., AND AWSON, D. J. (1994) Indole Derivatives as Steroid 5Ͱ-Reductase Inhibi-
MAW, G. N. (1993) Indole Derivatives as 5-Alpha-
R
tors, in WO 94/27990. Pfizer Limited. Chem. Abstr. 122(13), 160689x.
33. NEUBAUER, B. L., GRAY, H. M., HANKE, W., HIRSCH, K. S., HSIAO, K. C., JONEW, C. D., KUMAR
,
M. V., LAWHORN, D. E., LINDZEY, J., M
UDIA, J. E. (1996) J. Clin. Endocrinol. Metab. 81, 2055–2060.
34. HIRSCH, K. S., JONES, C. D., AUDIA, J. E., ANDERSSON, S., M UAID, L., STAMM, N. B., NEUBAUER
B. I., PENNINGTON, P., AND OOMEY, R. E. (1993) Proc. Natl. Acad. Sci. USA 90, 5277–5281.
35. RAPHALEN, D. (1965) Comp. Rend. 261, 2234–2235.
36. CHAN, H. K., GEISSLER, W. M., AND NDERSSON, S. (1994) in Sex Hormones and Antihormones
CQUAID, L., TINDALL, D. J., TOOMEY, R. E., YAO, R. C.,
AND
A
CQ
,
T
A

400
HARRIS ET AL.
in Endocrine Dependent Pathology: Basic and Clinical Aspects (M. Motta and M. Serio, Eds.), pp.
67–76, Elsevier, Amsterdam.
37. ELLSWORTH, K., AZZOLINA, B. A., BULL, H., AND
38. ELLIS, K. J., AND ORRISON, J. F. (1982) Methods Enzymol. 87, 405–426.
39. YONETANI, T., AND HEORELL, H. (1964) Arch. Biochem. Biophys. 106, 243–251.
40. NORMINGTON, K., AND USSELL, D. W. (1992) J. Biol. Chem. 267, 19548–19554.
41. LEVY, M. A., BRANDT, M., SHEEDY, K. M., HOLT, D. A., HEASLIP, J. I., TRILL, J. J., RYAN, P. J.,
ORRIS, R. A., GARRISON, L. M., AND ERGSMA, D. J. (1995) J. Steroid Biochem. Mol. Biol.
HARRIS, G. S. (1996) FASEB J. 10, 2211.
M
T
R
M
B
54, 307–319.
42. NAGAMOTO, A., NOGUCHI, K., MURAI, T., AND
43. HOLT, D. A., YAMASHITA, D. S., KONIALIAN- ECK, A. L., LUENGO, J. I., ABELL, A. D., BERGSMA
D. J., BRANDT, M., AND EVY, M. A. (1995) J. Med. Chem. 38, 13–15.
KINOSHITA, Y. (1994) J. Androl. 15, 521–527.
B
,
L
44. ABELL, A. D., BRANDT, M., LEXY, M. A., AND
HOLT, D. A. (1996) Bioorg. Med. Chem. Lett.
6, 481–484.
45. LEVY, M. A., BRANDT, J., AND
GREWAY, A. T. (1990) Biochemistry 29, 2808–2815.