Synthesis and in vitro activity of some epimeric 20α-hydroxy, 20-oxime and aziridine pregnene derivatives as inhibitors of human 17α-hydroxylase/c17,20-lyase and 5α-reductase

Article abstract of DOI:10.1016/S0968-0896(98)00110-2

Some epimeric 20-hydroxy, 20-oxime, 16α, 17α-, 17,20- and 20,21-aziridine derivatives of progesterone were synthesized and evaluated as inhibitors of human 17α-hydroxylase/C_17,20-lyase (P450(17α)) and 5α-reductase (5α-R). The reduction of 16-dehydropregenolone acetate (3a) was reinvestigated. NaBH₄ in the presence of CeCl₃ gave better stereoselectivity for 20β-ol [20α/20β-OH (4α/4β)=1/2.7] than LTBAH or the Meerwein-Pondroff method reported; reduction with Zn in HOAc formed exclusively 20α-ol (4αb). The 20α- and 20β-hydroxy-4,16-pregnadien-3-one (9α) and (9β) were synthesized from the alcohols 4αb and 4βb. Several 20-oxime pregnadienes and 16α,17α-, 17,20- and 20,21-aziridinyl-5-pregnene derivatives were also synthesized. LiAlH₄ reduction of the 16-en-20-oxime (12b) yielded 20 (R)-(13a) and 20(S)-17α,20-aziridine (13b) and 20(R)-17β,20-aziridine (14a). Several compounds inhibited the human P450(17α) with greater potency than ketoconzole. The 5α-R enzyme assay showed that while (9α) did not have any activity, (9β) and (3b) were potent 5α-reductase (IC₅₀=21 and 31nM) inhibitors with activities similar to finasteride. The 20-oximes (17a) and (17b) were potent dual inhibitors for both 5α-R (IC₅₀=63 and 115nM, compared to 33nM for finasteride) and P450(17α) (IC₅₀=43 and 25nM, compared to 78nM for ketoconazole). Copyright (C) 1998 Elsevier Science Ltd.

Full text of DOI:10.1016/S0968-0896(98)00110-2

BIOORGANIC &
MEDICINAL
CHEMISTRY
Bioorganic & Medicinal Chemistry 6 (1998) 1683±1693
Synthesis and In Vitro Activity of some Epimeric 20ꢀ-Hydroxy,
20-Oxime and Aziridine Pregnene Derivatives as Inhibitors of
Human 17ꢀ-Hydroxylase/C17,20-Lyase and 5ꢀ-Reductase
Yang-zhi Ling,y Ji-song Li, Katsuya Kato,{ Yang Liu, Xin Wang,
Gregory T. Klus, Kirk Marat,x Ivo. P. Nnane and Angela M. H. Brodie*
Department of Pharmacology & Experimental Therapeutics, School of Medicine, University of Maryland, Baltimore, MD 21201, USA
Received 29 December 1997; accepted 27 March 1997
AbstractÐSome epimeric 20-hydroxy, 20-oxime, 16a, 17a-, 17,20- and 20,21-aziridine derivatives of progesterone were
synthesized and evaluated as inhibitors of human 17a-hydroxylase/C17,20-lyase (P45017a) and 5a-reductase (5a-R). The
reduction of 16-dehydropregenolone acetate (3a) was reinvestigated. NaBH4 in the presence of CeCl3 gave better stereo-
selectivity for 20b-ol [20a/20b-OH (4a/4b)=1/2.7] than LTBAH or the Meerwein±Pondro€ method reported; reduc-
tion with Zn in HOAc formed exclusively 20a-ol (4ab). The 20a- and 20b-hydroxy-4,16-pregnadien-3-one (9a) and (9b)
were synthesized from the alcohols 4ab and 4bb. Several 20-oxime pregnadienes and 16a,17a-, 17,20- and 20,21-azir-
idinyl-5-pregnene derivatives were also synthesized. LiAlH4 reduction of the 16-en-20-oxime (12b) yielded 20 (R)-(13a)
and 20(S)-17a,20-aziridine (13b) and 20(R)-17b,20-aziridine (14a). Several compounds inhibited the human P45017a
with greater potency than ketoconzole. The 5a-R enzyme assay showed that while (9a) did not have any activity, (9b)
and (3b) were potent 5a-reductase (IC50=21 and 31 nM) inhibitors with activities similar to ®nasteride. The 20-oximes
(17a) and (17b) were potent dual inhibitors for both 5a-R (IC50=63 and 115 nM, compared to 33 nM for ®nasteride)
and P45017a (IC50=43 and 25 nM, compared to 78 nM for ketoconazole). # 1998 Elsevier Science Ltd. All rights
reserved.
Introduction
testes and adrenal gland, and 5a-reductase (5a-R),
which converts testosterone to the more potent andro-
gen, dihydrotestosterone (DHT), in the prostate. Inhi-
bitors of these enzymes have uses in the treatment of
prostatic diseases. Recently, we and others have descri-
Androgens have been implicated in the development
and progression of common disorders of the prostate,
most notably, benign prostatic hypertrophy (BPH) and
prostatic cancer. Two important enzymes in the bio-
synthesis of androgens are 17a-hydroxylase/C17,20-lyase
(P45017a), which regulates an early step in the biosynth-
esis of testosterone (T) and other androgens in both the
bed a number of compounds that inhibit P45017a.
1±11 Of
these compounds, ketoconazole, an imidazole anti-
fungal agent, is currently used to inhibit testosterone
synthesis in patients with advanced prostatic cancer.12,13
However, this compound, which inhibits a number of
other P450 enzymes, is not a very potent inhibitor of
P45017a and is associated with signi®cant side e€ects.
Finasteride, an inhibitor of 5a-R, has recently been
introduced as a new treatment for BPH.14 Finasteride
only reduces DHT levels by ꢀ70% in these patients, but
testosterone levels are often increased.15 Although this is
not a problem for patients with BPH, it could result in
growth stimulation of prostate cancers, since testoster-
one may bind to the androgen receptor, in the absence
of DHT. Several recent studies have demonstrated the
Abbreviations: DHEA: dehydroisoandrosterone; A: androstene-
dione; T: testosterone; P45017a: 17a-hydroxylase/C17,20-lyase;
5a-R: 5a-reductase
*Corresponding author. Fax: 001 410 706 0032.
yCurrent address: School of Pharmceutical Science, Beijing
Medical University, Beijing, PRC.
{Current address: Department of Chemistry, Laboratory of
Bioorganic Chemistry, National Industrial Research Institute
of Nagoya, Hirate-cho, Kita-ku, Nagoya, Japan.
xCurrent address: Department of Chemistry, University of
Manitoba, Winnipeg, Canada R3T 2N2.
0968-0896/98/$Ðsee front matter # 1998 Elsevier Science Ltd. All rights reserved
PII: S0968-0896(98)00110-2

1684
Y. Ling et al./Bioorg. Med. Chem. 6 (1998) 1683±1693
presence of mutations in the androgen receptor, which
may be activated by anti-androgens. Clinical trials using
an LHRH analogue (leuprolide) that inhibits testicular
androgen production have demonstrated increased sur-
vival of patients when used in combination with radia-
tion.16 These studies provide evidence of the value of
hormonal therapy at least in patients with local recur-
rent disease. Therefore, the development of new types of
potent enzyme inhibitors that inhibit P45017a, as well as
the 5a-R, could be e€ective in the treatment of prostatic
cancer by achieving total androgen blockade.
20-hydroxy, 20-oxime pregnadienes and also aziridine
derivatives, as androgen synthesis inhibitors. Some of
these compounds are also very potent dual inhibitors of
P45017a and 5a-R.
Chemistry. We have examined the reduction of
16-DPA(3a) by several reagents and solvent combinations.
The 3b,20b-diol (4ba) was prepared by the Meerwein±
Pondro€ reduction of 16-DPA as described by Marker
et al.17 The reduction product actually consisted of 20a-
and 20b-epimers in a ratio of 1:1.14, as determined by
the integration of their 18-methyl signal.21 Pure 20b-ol
(4ba) was obtained after several crystallizations.
Recent studies in our laboratory have identi®ed several
compounds that inhibit both P45017a and 5a-R.89 Such
compounds could block all androgen biosynthesis (T,
DHT and androstenedione) and could be more e€ective
as alternatives or additions to orchidectomy in treating
prostate cancer patients. We previously reported that
epimeric 20-hydroxy pregnane derivative (1a and 1b; 2a
and 2b) demonstrate mild to moderate inhibition of the
Because the 3-acetate group was needed as a protecting
group to synthesize the 4-en-3-one derivative (9)
(Scheme 1), the reduction time with aluminum isoprop-
oxide was reduced. This led only to the formation of the
hydrolysis product 3-ol (3b) (due to the ester exchange
in 2-propanol) together with 20-ol (4aa and 4ba). When
16-DPA was reduced with LTBAH,18 allylic 20-ol (4ab
and 4bb) with 20a/20b ratio of 1.41/1 was formed toge-
ther with 16-saturated products (34%), pregnenolone 3-
acetate (5) and its 20b-ol (6). Reduction with sodium
borohydride in methanol in the presence of cerium
chloride19 gave exclusively the allylic alcohol (4ab
and 4bb) in a ratio of 1/2.7, from which pure 20b-ol-3-
acetate (4bb) was obtained by recrystallization. The
10
P45017a (Scheme 1). 5,16-Pregnadien-3-ol-20-oxime
(12a) is a potent inhibitor (IC50=73 nM) for P45017a
but showed no activity against 5a-R. From our experi-
ence,10 the introduction of a 16,17-double bond into the
steroid nucleus enhances the activity of the compound
as a P45017a inhibitor, whereas the basic steroid 4-en-3-
one structure is better than the 5-en-3b-ol for activity
against 5a-R. We now report the activity of some epimeric
Scheme 1. (i) CeCl3, NaBH4, MeOH; or Zn, HOAc; or LTBAH, THF; or AI(i-PrO)3; (ii) t-BuMe2SiCl, imidazole, DMF; (iii) KOH,
MeOH; (iv) PTS, acetone; (v) AI(i-PrO)3, cyclohexanone, toulene. (5) Pregnenolone 3-acetate; (6) pregnen-3b,20b-diol.

Y. Ling et al./Bioorg. Med. Chem. 6 (1998) 1683±1693
1685
20b-hydroxy group was then protected as the t-butyldi-
methylsilyl ether (7b).20 After selective hydrolysis of the
3-acetate group in base, the 3-ol (8b) was converted to
the 4-en-3-one by Oppenauer oxidation with cyclohexa-
none in toluene. Cleavage of the 20-silyl ether using the
conventional method with ¯uoride anion20 was ine€ec-
tive. This was successfully performed with PTS in ace-
tone to give the pure 20b-ol (9b).
These ®ndings are consistent with a b-orientation for C-
20 and R stereochemistry at C-20. In 20(R)-17b,20-azir-
idine (14a), no NOE could be observed from the C-13
methyl proton to H-20, and only a relatively small
(2.1%) NOE was observed to the C-20 methyl protons.
The NOE observed from the C-20 methyl protons to the
C-13 methyl protons was likewise relatively small. These
data are consistent with an a-orientation of C-20 and R
stereochemistry at C-20. The 20(S)-isomer (14b), which
was previously synthesized by the hydrogenation of the
corresponding azirine,26 was not found. The 13C NMR
of compounds 13a, 13b, and 14a were assigned by
COSY and HSQC experiments on an AMX500 spec-
trometer. The procedure of Drefohl27 was followed to
synthesize the 16a,17a-aziridine (16). 16-DPA (3a)
reacted with free methoxylamine* and gave the Michael
adduct 16a-methoxyamine (15),27a which was then treated
with base to give 16a, 17a-aziridine.27b (*It is important
to prepare free methoxylamine base from its HCl salt and
distill out ®rst, otherwise 20-methoxime (12d) will be
invariably obtained.) Because the 21-methyl signal of (16)
appeared as a low broad singlet overlapped with other
proton signals and was hardly observed, COSY and HSQC
experiments were carried out to con®rm the structure.
The 13C NMR of compound 15 and 16 are assigned.
Synthesis of the 20a-ol counterpart was attempted using
LiAlH4 in ether or THF. The reaction produced a
molecular complex of the epimeric 20a- and 20b-diols
which was dicult to separate, together with consider-
able 16-saturated product.21,22 Ercoli and Ruggieri23
reported that treatment of 16-DPA with zinc and acetic
acid reduced predominantly the 16,17 double bond. The
allylic 20a-ol (4a) was obtained from the nonketonic
residue by acetylation and chromatography. Complete
conversion of the 16-DPA was achieved by doubling the
amount of zinc and extending the reaction time. Sur-
prisingly, the 20a-ol (4ab) was obtained as the sole pro-
duct as evidenced by its 18-methyl signal at 0.9 ppm and
a clean doublet for the C-21 methyl group seen at
1.33 ppm (J=6.4±4 Hz).21 The 16-ene group must play
an important part in this stereoselective reduction, as its
16-saturated counterpart pregnenolone, was not
reduced under the same conditions. Hydrolysis in base
gave 3b,20a-diol (4aa). The 20a-hydroxy-4,16-pregna-
diene-3-one (9a) was prepared from the 3-acetate (4ab),
following the same procedure described above for 4bb
(Scheme 1).
Because 5,16-pregnadien-3-ol-20-oxime (12a) demon-
strated potent inhibition of P45017a, derivatives were
synthesized. Oppenauer oxidation of 12a in toluene with
cyclohexanone and aluminum isopropoxide gave the 4-
en-3-one (17a). Acetylation of 17a yielded the oxime
acetate (17b); similarly, 3-acetate (12b) gave the
diacetate (12c). Although 16-DPA reacted with methoxyl-
amine to give the Michael addition product, 16a-
methoxylamino 20-methoxime (12d) was obtained when
reacted with methoxylamine hydrochloride. Dehydro-
progesterone (18) was treated with hydroxylamine, and
produced 3(E)-(19a) and 3(Z)-oxime (19b) which were
separated by chromatography. The determination of the
con®guration of the Z and E isomers was based on their
4-H chemical shift. In the 3(Z)-oxime (19b), the 4-H is
shift down®eld (6.47 ppm) compares to the 4-H in 3(E)-
oxime (19a) (5.77 ppm). Apparently, in the 3(Z)-oxime
isomer, the 4 vinyl proton is close to the electron-with-
drawing oxygen atom of the oxime group, which makes
it less shielding compared to the 4 vinyl proton in the
3(E)-isomer. This assignment is in agreement with cases
reported in the literature.28
The 20-oxime (10) was reduced with LiAlH4 in re¯uxing
TEF, and yielded mainly the epimeric 20,21e-aziridines
(11), which consisted of about 1:1 epimers.24 As the
mixture (11) showed only moderate activity against
human P45017a (see below), no further e€ort was made
to separate the individual isomers. The reduction of 16-
en-20-oxime (12b) with LiAlH4 yielded 17a,20-aziridine
(13) and 17b,20-aziridine (14). NMR studies showed
that 17a,20-aziridine (13) actually consisted of a mixture
of two isomers (13a and 13b), with an approximate ratio
of 1/2, which were separated by repeated chromato-
graphy. Similar ®ndings with the Á4,6 oxime steroid
have been reported.25 In the major isomers 20(S)-
methyl-(13b), NOEs are observed from the C-13 methyl
protons to both H-20 (0.9%) and the C-20 methyl pro-
ton (4.6%), suggesting that C-20 is on the b-face of the
steroid. The relatives sizes of these NOE suggest that the
C-20 methyl is oriented towards the C-13 methyl, or S
stereochemistry at C-20. The reciprocal experiment
involving irradiation of the C-20 methyl protons was
consistent with the above conclusions. In the minor iso-
mer 20(R)-methyl-(13a), a substantial (5.2%) NOE is
observed from the C-13 methyl protons to H-20, but no
NOE could be observed to the C-20 methyl protons.
Synthesis of the 21-nor oxime was also explored
(Scheme 3). 16a, 17a-Epoxy (20) was treated with CrCl2
to form the 16-ene (21).29 Reduction with LiAlH4 gave
the diol (22), which was cleaved with sodium periodate
to give the 17-aldehyde (23). Condensation with hydroxyl-
amine gave the 21-nor oxime (24). Oppenauer oxidation
gave the corresponding 4-en-3-one (25).

1686
Y. Ling et al./Bioorg. Med. Chem. 6 (1998) 1683±1693
Scheme 2. (i) LiAIH4, THF; (ii) AI(i-PrO)3, cyclohexanone, toulene; (iii) Ac2, Py; (iv) NH2OCH3, K2CO3, MeOH; (v) NaOMe,
MeOH.
Scheme 3. (i) CrCl2, HOAc; (ii) LiAIH4, THF; (iii) NaIO4, MeOH; (iv) NH2OH, Py; (v) AI(i-PrO)3, cyclohexanone.

Y. Ling et al./Bioorg. Med. Chem. 6 (1998) 1683±1693
1687
Inhibitor studies
potent 5a-R inhibitor, but its rapid metabolism in the
body and lack of oral activity, detracts from its value as
a therapeutic agent.30 However, because 9b and 18 both
have a 16,17-double bond, their 17b-acetyl side chain
should be more dicult to degrade in vivo.
Human testicular microsomes were employed to evalu-
ate the potency of pregnene derivatives as inhibitors of
P45017a as previously described.8±11 Two methods were
used to measure IC50 values and both methods gave
similar results. For some compounds, the radiometric
assay was used in which [213H]-17a-hydroxy-
pregnenolone was the substrate for the human enzyme.
Inhibition was assessed by measuring the release of [3H]-
acetic acid during cleavage of the C-21 side chain in the
conversion to DHEA. Thus, the IC50 values assayed by
this method are for C17,21-lyase. For other compounds,
the IC50 values were determined using product isolation.
Conversion of radiolabeled pregnenolone to 17a-hydroxy-
pregnenolone and DHEA by P45017a was measured in
the presence of several concentrations of candidate
inhibitors. Reverse-phase HPLC was employed to sepa-
rate and measure the amount of substrate and metabo-
lites.9 IC50 assays were performed at least twice and up
to four times for the more potent inhibitors. The mean
values are shown in Table 1. As rodent models are most
frequently used for in vivo studies, we determined inhi-
bition of the conversion of [213H]-17a-hydroxy-
pregnenolone by the compounds in rat testicular
microsomes.
Pregnenolone-20-oxime (10) showed moderate potency
(530/570 nM), while the 16-en-20-oxime (12a) is a good
P45017a inhibitor (73 nM) and its 4-en-3-one derivative,
16-dehydroprogesterone-20-oxime (17a), not only
showed greater activity (43 nM) against human P45017a
,
but also showed potent activity against 5a-R (63 nM). It
was quite unexpected to ®nd that (17a) is similar in
potency to its 5-en-3-ol (12a) counterpart as human
P45017a enzyme prefers pregnenolone (i.e., 5-en-3-ol)
rather than progesterone (i.e., 4-en-3-one) as substrate.
Comparison of the weak activity of 16-dehy-
droprogesterone (18) with its 20-oxime (17a) (43 nM),
and the activity of 16-dehydropregnenolone (3b) (510/
490 nM) with its 20-oxime (12a) (73 nm), shows that the
introduction of a 20-oxime group markedly increased
the inhibitory e€ect for P45017a
.
Acetylation of the 20-oxime (17b) increased the inhibi-
tory activity against human P45017a and maintained
activity against 5a-reductase, compared to 17a. The
compound 12d, the 20-methoxy of the oxime (12a) had
no activity at all. The 21-nor oxime (24) had little inhi-
bition for P45017a, suggesting that 20-methyl group is
also critical for maintaining activity.
Inhibition of the 5a-R in human prostatic tissue was
measured from the conversion of radiolabeled testoster-
one to DHT following puri®cation and isolation by
TLC using an assay similar to that reported earlier.8
The synthesis of 16,17a-(16), 17,20-(13 and 14), and
20,21-aziridindyl (11) derivatives originated from the
idea that the aziridine nitrogen might show strong
coordination to the heme iron of the enzyme.31 Addi-
tionally, the highly reactive aziridine ring might react
covalently at the active site of the enzyme, resulting in
irreversible enzyme inhibition, as was shown for inhi-
bition of aromatase by a 10b-aziridinyl steroid.31 How-
ever, the results were rather disappointing, as the
20,21-aziridine (11) showed only weak inhibition for the
human P45017a. Njar et al. recently reported that 20(S)-
and 20(R)-20,21-aziridine were potent inhibitors of
the rat testicular P45017a enzyme.32 Indeed, we did ®nd
that epimeric 20,21e-aziridine (11) is a more potent
inhibitor of rat enzyme (IC50=184 nM, ketoconazole
2096 mM).
Structure±activity relationships
The inhibitory activities of the 20-hydroxy epimers are
compared in Table 1. In the 5-pregnene-3b-ol series, the
20b-ol (1b) was found previously to be a more potent
inhibitor of P45017a (IC50=180/190 nM) than the 20a-ol
(1a) epimer (720/510 nM). As expected, both showed no
activity against 5a-R. However, 20a-ol of 4-en-3-one
(2a) is a potent 5a-R inhibitor (13.8 nM) and is stronger
than its 20b-ol epimer (2b) (90 nM). The introduction of
the 16,17-double bond gave the 20a-ol (4aa) and 20b-ol
(4ba), which had about equal activity for P45017a (250/
220 nM). While 20b-hydroxy-4,16-pregnadien-3-ol (9b)
is a potent 5a-R inhibitor (21 nM) comparable to ®na-
steride (33 nM), its epimer 20a-ol (9a) showed no activity
at all. The marked in¯uence on the activity of the iso-
meric con®guration at the C-20 position, together with
the 16,17-ene suggest that this area interacts with the
active site of the enzyme which has strict steric require-
ments. Because the 20b-ol (9b) could be metabolized to
the 20-one in vivo, dehydroprogesterone (18) was also
tested. Compound 18 was found to be a relatively weak
inhibitor of P45017a, but was a moderately good inhi-
bitor of 5a-R (96 nM). Progesterone is known to be a
In summary, we have found that 17a and 17b are potent
5a-R inhibitors, and are also very potent human
P45017a inhibitors. We believe that 16-dehydro-
progesterone-20-oxime (17a) is the most potent dual
inhibitor of these enzymes. Other potent inhibitors of
5a-R were 2a, 3b, and 9b. Recently, Potter et al. repor-
ted that the most potent inhibitor of human P45017a in
their series was 17-(3-pyridyl) androst-5,16-diene-3-ol.7

1688
Y. Ling et al./Bioorg. Med. Chem. 6 (1998) 1683±1693
Table 1. IC50 of steroid inhibitors of human and rat testicular P450 17a-hydroxylase/C17,20-lyase and human 5a-reductase
Human
Rat
17a-Hydroxylase/lyasea C17,20-lyase
5a-Reductase
C17,20-lyase
% Inhibit. at 150 nM
IC50 (nM)
IC50 (nM)
(1500 nM, %)
IC50 (nM)
1a
1b
2a
2b
3b
16.4
13.4
8.4
720/510a,b
180/190a,b
1547/788a
204/143a
510/490a,b
250/220a
255/215a
Weaka
NIa
NIc
NI
21.1
26.6
37.9
73.7
22.3
4.8
13.8
90.2
31.3
NTd
NT
NI
23.3
7.2
300
4aa
4ba
9a
9b
10
2.0
0.4
8.7
15.5
16.8
10.5
19.4
80.3
61.1
73.1
0.4
58.0
80.6
55.0
98.6
63.1
57.3
63.0
NI
21
10
Weaka
73
NT
NT
NT
NI
11
12a
184
1300
2754
12b
12c
67
51
NI
12d
13a
NI
339
NT
NT
NT
NT
NT
NT
63
34.8
37.3
6.5
63.5
72.3
24.5
NI
13b
14b
372
NI
15
16
7.0
NI
162
41.9
76.7
83.8
12.4
46.0
79.7
25.7
19.6
69.5
NI
24.5
91.7
96.3
64.0
67.7
59.2
33.8
41.1
65.3
17a
17b
43
25
48
54
115
95.8
1270
Weak
NT
NT
NI
18
19a
Weak
196
75
1100
1300
19b
24
Weak
Weak
78
25
Ketoconazole
Finasteride
900
NI
33
aValues are for 17a-hydroxylase/C17,20-lyase. Compounds 1a-11 were incubated with [73H] pregnenolone and cofactors for 5 min and
products analyzed by HPLC. Inhibition of testicular 17a-hydroxylase/lyase was determined from the conversion radiolabeled preg-
nenolone to DHEA in the presence of the test compounds. All other assays were performed by incubation with [213H] 17a-hydroxy-
pregnenolone for 60 min at 34 ꢁC with cofactors and test compounds. C17,20-lyase activity was determined by measuring [3H] acetic
acid liberated during cleavage at the C-17. Compounds were screened at 150 nM with human testicular microsomes and at 1500 nM
with rat microsomes. IC50 values were determined over a range of six concentrations. Values are averages of 3±5 repeat experiments.
The variation in repeat values were usually <15% of the mean IC50 values. Inhibition of 5a-reductase was determined from the
conversion of testosterone to DHT in the presence of test compounds. Human prostatic, microsomes were incubated with [73H] tes-
tosterone and cofactors for 10 min at 37 ꢁC. The DHT produced was puri®ed by TLC.
bResults previously reported in ref. 10.
cNI, no inhibition.
dNT, not tested.
Experimental
However, 17a is also a potent inhibitor of 5a-R. As the
IC50 of ketoconazole in their assay was 65/26 nM, its
potency appears to be similar to 17a. However, 17a is
also a potent inhibitor of 5a-R, whereas the former
showed no activity against this enzyme. These dual
inhibitory activities of 17a and 17b suggests that they
have potential for the treatment of prostate cancer.
Further studies are under investigation in biological
systems.
Chemical methods. 1H NMR data (300 MHz) (internal
Me4Si=d0) (QE 300, NMR systems, General Electric
Co.) were recorded in CDCl3 unless otherwise stated.
Reactions were monitored by TLC on silica gel plates
(Merck Type 601-1), and components visualized by dipping
in 4% sulfuric acid in ethanol followed by heating at
ca. 120±150ꢁC. Flash column chromatography was carried

Y. Ling et al./Bioorg. Med. Chem. 6 (1998) 1683±1693
1689
out on silica gel (Merck grade 9385, 230±400 mesh 60 A)
in the solvent systems indicated. LP refers to petroleum
fractions bp 35±60ꢁC. Solutions were dried using anhydrous
Na2SO4, Melting points were measured on a Fischer±
Johns Melting Point apparatus and are uncorrected.
Zn in glacial HOAc. 16-DPA (1 g) and zinc dust (5 g), in
glacial acetic acid (50 mL), was stirred vigorously at
25 ꢁC until no starting material was present on TLC
(16 h). The whole mixture was ®ltered and the Zn pre-
cipitate washed with CH2Cl2 (200 mL). The ®ltrate was
washed with water, 5% aqueous KHCO3 and then
water. Evaporation of the solvent gave a solid residue
(0.98 g), 1H NMR showed that it was solely 20a-ol
(4ab). Recrystallization from acetone-LP gave (4ab),
(607 mg, 60%), mp 125±129 ꢁC. Part of the sample
(50 mg) was recrystallized again, to give an analytical
Reduction of 16-dehydropregnenolone-3-acetate (16-DPA)
(3a)
Aluminum isopropoxide. 16-DPA (0.5 g, 1.4 mmol) was
reduced
with
aluminum
isopropoxide
(2.5 g,
sample (35 mg), mp 130±132 ꢁC (lit.23 mp 98±99 ꢁC). H
1
12.38 mmol) and dry iso-propanol (70 mL) as repor-
ted.17 The crude product (450 mg) obtained consisted of
the 20e-hydroxy epimer, (4aa and 4ba) (20a/20b, 1/1.14)
as determined by integration of their 18-methyl signal at
0.91 and 0.87 ppm, respectively. Recrystallization from
MeOH, and then EtOAc-LP, gave pure 20b-ol (4ba)
(102 mg, 20%) as ®ne needle crystals, mp 168±170 ꢁC
(lit. 169±171 ꢁC,17 or 169.5±171.5 ꢁC).21
NMR d: 5.66 (1H, s, 16-H), 5.39 (1H, d, J=4.9 Hz, 6-
H), 4.61 (1H, m, 3-H), 4.36 (1H, m, 20-H), 2.04 (3H, s,
3-OAC), 1.33 (3H, d, J=6.4 Hz, 20-Me), 1.06 (3H, s, 19-
Me), 0.91 (3H, s, 18-Me). Anal. (C23H34O3)C,H.
5,16-Pregnadiene-3ꢁ,20ꢀ-diol (4ꢀa). The 3-acetate (4ab)
(300 mg) was stirred in 0.5 M KOH/MeOH (10 mL) at
25 ꢁC for 1 h. CH2Cl2 (100 mL) was added and the mix-
ture was washed with water. After evaporation of the
solvent, the residue was immediately recrystallized from
acetone-LP, to give 4aa (105 mg, 35%), mp 185±186 ꢁC
When the reaction was re¯uxed for only 3 h, worked-up
as above, and the residue ¯ash chromatographed on
silica gel, elution with 15% acetone-LP, gave 16-dehydro-
pregnenolone (3b) (210 mg), mp 209±211 ꢁC (acetone/
LP) (lit28 212±214 ꢁC), and 20e-ol (4aa and 4ba)
(230 mg), mp 162±168 ꢁC (from MeOH).
(lit. 21 186.5±188.5 ꢁC).
.
20ꢀ-[(tert-Butyldimethysilyl)oxy]-(7ꢀ) and 20ꢁ-[(tert-
butyldimethylsilyl)oxy]-5,16-pregnadiene-3ꢁ-ol-3-acetate
(7ꢁ). To a solution of 20a-ol (4ab) (2.0 g, 5.56 mmol)
in DMF (25 mL) were added tert-butyldimethylsilyl
chloride (1.68 g, 11.2 mmol) and imidazole (1.51 g,
22.2 mmol) and the mixture stirred at 25 ꢁC for 14 h
(TLC). The reaction mixture was diluted with ether and
washed with water. After evaporation, the residue was
recrystallized from Et2O-acetone, to give (7a) (1.20 g,
Lithium tri-t-butoxyaluminium hydride (LTBAH).
solution of 1 M LTBAH in THF (3 mL, 3 mmol) was
added dropwise to stirred solution of 16-DPA
A
a
(356 mg, 1 mmol) in dry THF (10 mL) in an ice-bath and
then stirred for 16 h. CH2Cl2 (100 mL) was added and
the whole mixture was washed with 5% HCl followed
with water. After evaporation of the organic layer, the
residue was ¯ash chromatographed on silica gel, eluted
with 7.5% acetone-LP, to give pregnenolone-3-acetate
(5) (50 mg, 14%), mp 144±147 ꢁC (acetone), (lit.23a 146±
147 ꢁC) and 5-pregnene-3b,20b-diol (6) (70 mg, 20%),
mp 160±164 ꢁC (from ether) (lit.33 mp 160±164 ꢁC), and
20e-ol (4b) (20a/b, 1.41/1) (180 mg, 51%), two crystal-
lizations from acetone-LP, gave pure 20b-ol (4bb)
(15 mg), mp 144±147 ꢁC (lit.23b 145±146 ꢁC).
46%), mp 130±132 ꢁC. H NMR d: 5.58 (1H, s, 16-H),
1
5.39 (1H, d, J=4.9 Hz, 6-H), 4.60 (1H, m, 3a-H), 4.31
(1H, m, 20-H),2.04 (3H, s, 3-OAC), 1.25 (3H, d,
J=6.3 Hz, 20-Me), 1.06 (3H, s, 19-Me), 0.90 (9H, s,
-CMe3), 0.88 (3H, s, 18-Me), 0.05, 0.03 (each 3H, s,
SiMe2). Anal. (C29H48O3Si) C,H. Following the proce-
dure above, the 20b-ol (4bb) (2.0 g, 5.56 mmol), was
converted to the silyl ether (7b) (1.8 g) which was
obtained as an oil and used directly in the next reaction.
Sodium borohydride in the presence of cerium trichloride.
A solution of 16-DPA (356 mg, 1 mmol) in CH2Cl2
(3 mL) was combined with a solution of CeCl3-7H2O
(372 mg, 1 mmol) in methanol (8 mL); NaBH4 (60 mg,
1.5 mmol) was added at once in one portion. The
mixture was stirred at 25 ꢁC for 15 min, then worked-up
as above, to give 20e-ol (4ab and 4bb) (310 mg) (20a/20b
1/2.7). Recrystallization from acetone-LP gave pure 20b-
ol (4bb) (143 mg, 40%), mp 144±147 ꢁC, (lit.23b 145±
146 ꢁ). 1H NMR d: 5.64 (3H, s, 16-H), 5.39 (1H, d,
J=4.9 Hz, 6-H), 4.61 (1H, m, 3-H), 4.38 (IH, m, 20-H),
2.03 (3H, s, 3-0Ac), 1.37 (3H, d, J=6.3 Hz, 20-Me), 1.06
(3H, s, 19-Me), 0.87 (3H, s, 18-Me).
20ꢀ-[(tert-Butyldimethylsilyl)oxy]-(8ꢀ) and 20ꢁ-[(tert-
butyldimethylsilyl)oxy]-5,16-pregnadien-3ꢁ-ol (8ꢁ). The
3-acetate (7b) (oil, 1.8 g) was stirred in 0.5 M KOH-
MeOH (30 mL) for 1 h. Ether (200 mL) was added and
the whole mixture was washed with water. After eva-
poration, the residue was recrystallized from Et2O-LP
to give (8b) (1.2 g, 66%), mp 120±122 ꢁC. H NMR d:
1
5.53 (1H, s, 16-H), 5.35 (1H, d, J=4.2 Hz, 6-H), 4.40
(1H, m, 20-H), 3.49 (1H, m, 3-H), 1.25 (3H, d,
J=6.4 Hz, 20-Me), 1.02 (3H, s, 19-Me), 0.87 (12H, s,
18-Me, -CMe3) 0.03, 0.02 (each 3H, s, SiMe2). Anal.
(C27H45O2Si) C,H.

1690
Y. Ling et al./Bioorg. Med. Chem. 6 (1998) 1683±1693
Following the procedure described above, the 3-acetate
(7a) (1 g) was converted to (8a) (0.65 g, 71%), mp 126±
128 ꢁC (from MeOH). 1H NMR d: 536 (1H, s, 16-H),
5.35 (1H, d, J=4.9 Hz, 6-H), 4.30 (1H, m, 20-H), 3.50
(1H, m, 3-H), 1.23 (3H, d, J=6.3 Hz, 20-Me), 1.02 (3H,
s, 19-Me), 0.87 (9H, s, -CMe3), 0.85 (3H, s, 18-Me),
0.03, 0.01 (each 3H, s, SiMe2). Anal. (C27H45O2Si) C,H.
hydrochloride (40 mg, 0.48 mmol), ethanol (1 mL) and
pyridine (0.3 mL) was re¯uxed for 1 h. Water was added
and the precipitate was collected, recrystallized from
acetone-LP to a€ord (12d) (35 mg, 32%), mp 136±
139 ꢁC. H NMR d: 6.00 (1H, s, 16-H), 5.36 (1H, s, 6-
1
H), 3.88 (3H, s, -OMe), 2.53 (1H, m, 3a-H), 1.93 (3H, s,
20-Me), 1.05 (3H, s, 19-Me), 0.97 (3H, s, 18-Me), anal.
(C22H33O2N) C,H,N.
20ꢀ-Hydroxy-(9ꢀ) and 20ꢁ-hydroxy-4,16-pregnadiene-
3-one (9ꢁ). From a solution of 20a-silyl ether (8a)
(500 mg, 1.16 mmol) in toluene (50 mL) and cyclohexa-
none (3 mL), part of the solvent (10 mL) was distilled o€
to eliminate the moisture. Aluminum isopropoxide
(204 mg, 1.0 mmol) was added and the mixture re¯uxed
for 3 h or until TLC showed that the reaction was com-
plete. Saturated sodium tartrate (30 mL) was added and
the mixture steam distilled until no more toluene and
cyclohexanone could be collected. The remaining resi-
due was extracted with EtOAc. After evaporation of the
organic layer, the residue was taken up with acetone
(20 mL) and treated with p-toluenesulfonic acid
(300 mg) for 14 h at 25 ꢁC. CH2Cl2 (200 mL) was added
and the mixture washed with water. After evaporating
the solvent, the residue was ¯ash chromatographed on
silica gel. Elution with 15% acetone/LP, gave the 20a-ol
(9a) (190 mg, 41%), mp 146±149 ꢁC (Et2O-LP). 1H
NMR d: 5.74 (1H, s, 4-H), 5.32 (1H, m, 16-H), 4.78 (1H,
d, 8.5 Hz, 20-H), 1.74 (3H, d, J=6.8 Hz, 20-Me), 1.21
(3H, s, 19-Me), 1.00 (3H, s, 18-Me), 0.79 (1H, s, -OH).
Anal. (C21H30O2) C,H.
20(R)-17ꢀ,20-Aziridinyl-(13a), 20(S)-17 ꢀ,20-aziridinyl-
(13b), and 20(R)-17ꢁ,20-aziridinyl-5-pregnene-3ꢁ-ol (14a).
A solution of the 20-oxime (12b)34 (1.5 g, 4.0 mmol) in
THF (50 mL) was added to a stirring solution of LiAlH4
(2.0 g, 40 mmol) in THF (100 mL) containing N-methyl-
N-butylamine (0.5 mL) cooled in ice-water and under
nitrogen. The mixture was then re¯uxed under nitrogen
for 2 days (TLC monitor). The excess LiAlH4 was
decomposed by adding aqueous THF (150 mL). The
precipitated inorganic salt was ®ltered and washed with
5% MeOH-EtOAc (200 mL). After the solvents were
evaporated, the residue was ¯ash chromatographed,
eluted with 25% acetone-LP, and gave a mixture of 13a
and 13b (340 mg, 27%) in a ratio of 1:2 as determined
by their 18-methyl signal, mp 191±195 ꢁC (from acetone-
LP), and then 14a (290 mg, 23%), mp 186±189 ꢁC (from
1
acetone). H NMR (CD3OD) d: 5.36 (1H, s, 6-H), 3.38
(1H, m, 3a-H), 1.33 (3H, d, J=6.3 Hz, 20-Me), 1.02
(3H, s, 19-Me), 0.95 (3H, s, 18-Me). Anal. (C21H33O2N)
C,H,N.
The mixture of 13a and 13b (170 mg) was twice rechro-
matographed, eluted with 3% and 5%MeOH/-CH2C12,
to give pure 20(R)-(13a) (20 mg, 1.6%), mp 197±201 ꢁC
(from ether) (1H NMR (CD3OD) d: 5.35 (1H, d,
J=3.9 Hz, 6-H), 3.30 (1H, m, 3a-H), 1.19 (3H, d,
J=5.4 Hz, 20-Me), 1.03 (3H, s, 19-Me), 0.85 (3H, s, 18-
Me)), and then 20 (S)-(13b) (90 mg, 14%), mp 186±
189 ꢁC (from acetone). 1H NMR d 5.35 (1H, s, 6-H),
3.40 (IH, m, 3a-H), 1.26 (3H, d, J=5.9 Hz, 20-Me), 1.02
(3H, s, 19-Me), 0.94 (3H, s, 18-Me), Anal. (C21H33O2N)
C,H,N.
Following the procedure described above, 20b-silyl
ether (8b) (500 mg) gave 4-en-3-one (9b) (201 mg, 55%),
mp 179±180 ꢁC (acetone-LP). 1H NMR d: 5.74 (IH, s,
4H), 5.64 (1H, m, 16-H), 4.38 (1H, brs, 20-H), 1.37 (1H,
d, J=6.3 Hz, 20-Me), 1.22 (3H, s, 19-Me), 0.89 (3H, s,
18-Me). Anal. (C21H30O2) C,H.
20,21"-Aziridinyl-5-pregnene-3ꢁ-ol (11). Following the
same procedure described below for 13, the 20-oxime
(10)34 (300 mg, 0.9 mmol) was reduced with LiAlH4
(500 mg, 13.2 mmol). The crude product was chromato-
graphed on silica gel and eluted with 5% MeOH-
CH2Cl2, to give 11 (140 mg, 49%), mp 210±216 ꢁC (from
acetone-LP). (lit.24 20b,21-aziridine, mp 199±202 ꢁC;
20a,21-aziridine, mp 195±198 ꢁC.) Anal. (C21H33ON)
C,H,N.
The 13C NMR of compounds 13a, 13b, and 14a were
assigned by COSY and HSQC experiments on an
AMX500 spectrometer.
3ꢁ-Hydroxy-16ꢀ-methoxyaminopregna-5,16-dien-20-one
(15) (mp 136±140 ꢁC, lit.27a mp 140±142 ꢁC) and
3ꢁ-Acetoxy-5,16-pregnadiene-20-oxime acetate (12c).
Following the same procedure described below for 17b,
the 3-acetate (12b)34 (100 mg) was acetylated, to give the
oxime acetate (12c) (80 mg, 78%), mp 194±196 ꢁC.
(From acetone/LP) (lit.35 mp, 195±196 ꢁC.)
16ꢀ,17ꢀ-aziridinyl-3ꢁ-hydroxy-5-pregnen-20-one
(16)
(mp 185±187 ꢁC, lit.27b mp 186±188 ꢁC). These were pre-
pared according to ref. 27 compound 15: 1H NMR
8 5.35 (1H, d, J=2.9 Hz, 6-H), 4.14 (1H, m, 16b-H),
3.53 (1H, m, 3a-H), 3.49 (3H, s, OMe), 2.48 (1H, d,
J=7.4 Hz, 17a-H), 2.17 (3H, s, 21-Me), 1.01 (3H, s,
19-Me), 0.69 (311, s, 18-Me). Compound 16: 1H NMR d
5.33 (1H, d, J=4.9Hz, 6-H), 3.52 (1H, m, 3a-H), 2.82
3ꢁ-Hydroxy-5,16-pregnadien-20-methoxime (12d). The
16-en-20-one (3b) (100 mg, 0.32 mmol), methoxylamine

Y. Ling et al./Bioorg. Med. Chem. 6 (1998) 1683±1693
1691
(1H, s, 16b-H), 2.12 (3H, broad s, 21-Me), 1.05 (3H, s, 19-
Me), 1.03 (3H, s, 18-Me).
25 ꢁC for 30 min. Water (200 mL) was added and the
precipitate was recrystallized from MeOH, to give the
17-al (23) (300 mg, 75%), mp 163±168 ꢁC. 1H NMR d
9.72 (1H, s, CHO), 6.82 (1H, s, 16-H), 5.36 (1H, d,
J=4.8 Hz, 5-H), 3.50 (1H, m, 3a-H), 1.06 (3H, s, 19-
Me), 0.74 (3H, s, 18-Me). Anal. (C20H28O2) C,H.
20-Hydroximino-4,16-pregnadien-3-one (17a). Following
the same procedure described above for 9a, the 20-
oxime 3b-ol (12a) (300 mg, 0.9 mmol) was oxidized with
cyclohexanone (3 mL) and aluminum isopropoxide
(303 mg, 1.5 mmol) in toluene (50 mL) for 6 h. The crude
product obtained was ¯ash chromatographed and
eluted with 1% MeOH-CH2C12, to give 4-en-3-one
Androsta-5,16-dien-3ꢁ-ol-20-oxime (24). Following the
procedure for the preparation of 19, the 17-al (23)
(750 mg, 2.5 mmol) gave oxime (24) (250 mg, 32%), mp
196±202 ꢁC. 1H NMR d: 7.82 (1H, s, CH=N-), 6.75
(1H, s, exchangeable OH in D2O), 6.02 (1H, s, 16-H),
5.37 (1H, d, 4.9 Hz, 6-H), 3.70 (1H, m, 3a-H), 1.05
(3H, s, 19-Me), 0.95 (3H, s, 18-Me). Anal. (C20H29O2N)
C,H,N.
(17a) (165 mg, 55%), mp 254±259 ꢁC (MeOH-H2O). H
1
NMR d: 6.05 (1H, s, 16-H), 5.74 (1H, s, 4-H), 2.00 (3H,
s, 20-Me), 1.21 (3H, s, 19-Me), 0.97 (3H, s, 18-Me).
Anal. (C21H29ON) C,H,N.
4,16-Pregnadiene-3,20-dione-20-oxime acetate (17b).
The oxime (17a) (200 mg, 0.61 mmol) was dissolved in
pyridine (5 mL) and acetic anhydride (0.5 mL) and left
overnight. Water (200 mL) was added and the pre-
cipitate was recrystallized from EtOH, to give the acet-
ate (17b) (140 mg, 62%), mp 199±203 ꢁC. 1H NMR d:
6.3 0 (1H, s, 16-H), 5.74 (1H, s, 4-H), 2.23 (3H, s, 0Ac),
2.10 (3H, s, 20-Me), 1.22 (3H, s, 19-Me), 1.03 (3H, s, 18-
Me). Anal. (C23H31O3N) C,H,N.
Androsta-4,16-dien-3-one-20-oxime (25). Following the
procedure described above for (9b), the 3b-ol (24)
(200 mg, 0.63 mmol) was oxidized with cyclohexanone
(6 mL), aluminum isopropoxide (400 mg, 1.96 mmol) in
toluene (100 mL) for 4 h. The crude product obtained
was chromatographed, eluted with 1% MeOH-CH2Cl2,
to give 3-one (25), (100 mg, 50%), mp 265±268 ꢁC (from
acetone). 1H NMR d 7.82 (1H, s, CH=N-), 6.87 (1H, s,
exchangeable in D2O, -OH), 6.01 (1H, s, 16-H), 5.74
(1H, s, 4-H), 1.22 (3H, s, 19-Me), 0.97 (3H, s, 18-Me).
Anal. (C20H27O2N) C,H,N.
3(E), 20-Dioximido-(19a) and 3(Z),20-dioximido-4,16-
pregnadiene (19b). 16-Dehydroprogesterone (3b)
(100 mg, 0.29 mmol) and hydroxylamine hydrochloride
(84 mg, 1.2 mmol) were re¯uxed in ethanol (5 mL) for
4 h. Water was added the crude mixture of 17a and 17b
(100 mg) was collected. The mixture was separated by
¯ash chromatography on silica gel and eluted with 25±
30% ethyl acetate-LP. The ®rst fraction collected con-
tained 3(E),20-dioxime (19a) (35 mg, 32%), mp 136±
139 ꢁC (from EtOH). 1H NMR d: 6.05 (1H, s, 16-H),
5.77 (1H, s, 4-H), 1.99 (3H, s, 20-Me), 1.09 (3H, s, 19-
Me), 0.96 (3H, s, 18-Me). Anal. (C21H30O2N2) C,H,N.
The second fraction collected contained 3(Z),20-dioxime
Enzyme assays. [7-3H]-Pregnenolone (25 Ci/mmol),
[7-3H]-testosterone, [4-14C]-DHT, and [4-14C]-DHEA
were purchased from New England Nuclear Corp.
(Boston, MA) and checked for purity and/or puri®ed by
TLC or HPLC prior to use. The [21-3H]17a-hydroxy-
pregnenolone was prepared in our lab as described pre-
viously.11 Ketoconazole was purchased from Sigma
Chemical Co. (St. Louis, MO, USA). Finasteride was a
gift from Merck Co. Scintillation cocktail 3a70B was
purchased from RPI Corp. (Mount Prospect, IL, USA).
(19b) (40 mg, 36.5%), mp 225±227 ꢁC (from EtOH). H
1
NMR d: 6.47 (1H, s, 4-H), 6.05 (1H, t, 16-H), 1.99 (3H,
s, 20-Me), 1.13 (3H, s, 19-Me), 0.96 (3H, s, 18-Me).
.
Anal. (C21H30O2N2 1.5H2O) C,H,N. These two isomers
underwent isomerization during recrystalization in
ethanol.
Testicular microsomes. The previously reported proce-
dure9 was followed to prepare testicular microsomes.
Testes were obtained from untreated prostatic cancer
patients undergoing orchidectomy in the University of
Maryland Hospital and the University of North Caro-
lina Hospital. Testes were also obtained from normal
adult rats (200 g bw Sprague±Dawley, Charles River).
The microsomes were stored at 70 ꢁC until assayed.
Just before use, the thawed microsomes were diluted
with 0.1M phosphate bu€er to appropriate concentra-
tions. The protein concentration of microsomes used in
each assay was determined by the method of Lowry et al.36
Androsta-5,16-dien-3ꢁ-ol-17-carboxaldehyde (23). To a
stirring solution of 21-acetoxy (21) (prepared from 16a,
17a-epoxide (20))29 (0.5 g, 1.34 mmol) in THF (20 mL),
a solution of LiAlH4 (200 mg, 15.8 mmol) in THF
(10 mL) was added dropwise at 25 ꢁC and then stirring
continued for 1 h. EtOAc (150 mL) was added and the
whole mixture was washed with 1 N HCl (3Â20 mL), and
then with water. After evaporation of the solvent, the
residue diol (22) (0.4 g) was dissolved in MeOH (30 mL),
and a solution of sodium periodate (360 mg, 1.68 mmol)
in water (5 mL) was added and the mixture stirred at
Human prostate microsomes. Microsomes from prostatic
tissues from patients with benign prostatic hypertrophy
(BPH) were prepared as above and stored at 70 ꢁC.
Just before use, the microsomes were diluted with 0.1 M

1692
Y. Ling et al./Bioorg. Med. Chem. 6 (1998) 1683±1693
phosphate bu€er to the appropriate concentrations. The
protein concentrations were determined as above.36
results were obtained from duplicate sets of experiments
and were repeated at least once.
Measurement of 17ꢀ-hydroxylase/C17,20-lyase activity.
(1) Radiometric assay of C17,20-lyase activity: in this
assay, [213H] 17a-hydroxypregnenolone (13.61mCi/mmol)
was employed as the substrate. Each tube contained
300,000 dpm, a total concentration of 10 mM 17a-
hydroxypregnenolone, various concentrations of the test
compounds in phosphate bu€er (pH 7.4, total volume
1 mL) and the NADPH generating system (NADP
65 mM; glucose-6-phosphate 0.71 mM; glucose-6-phos-
phate dehydrogenase 0.13 IU in 50 mL phosphate bu€er).
The tubes were preincubated for 10 min at 34 ꢁC. The
reaction was initiated by the addition of the microsomes
(approx. 90 mg protein) and the incubation carried out
for 60 min under oxygen at 34 ꢁC. Chloroform was then
added to extract the steroids, an aliquot (0.75 mL) of
aqueous phase was removed and mixed with an equal
volume of 2.5% charcoal suspension. After vortexing,
the tubes were allowed to stand for at least 30 min and
then centrifuged. An aliquot of the supernatant was
removed and the tritium concentration measured by
liquid scintillation counting. The results are listed in
Table 1. IC50 or 50% inhibition values were calculated
from the linear regression line of the plot of logit of
Measurement of 5ꢀ-reductase activity. [7-3H]Testoster-
one (15-60 nM, 6Â105 dpm) was preincubated with dif-
ferent concentrations of the test compounds or
®nasteride in 0.85 mL, phosphate bu€er (pH 7.4) for
10 min at 37 ꢁC. Then, human prostatic microsomes
(approximately 160 mg protein in 100 mL phosphate
bu€er) and the NADPH generating system (NADP
650 mM; glucose-6-phosphate 7.1 mM; glucose-6-phos-
phate dehydrogenase 1.25 IU in 50 mL phosphate bu€er)
were added. The mixture was incubated under oxygen
for 10 min at 37 ꢁC. Steroids [14C]-labeled (T, A, and
DHT) and authentic markers (T, A, DHT, 5a-andros-
tane-3a,17b-diol, and -3b,17b-diol) were added after
incubation. The steroids were extracted with ether and
separated by TLC (chloroform:ether, 80:20). The DHT
and the 3-diols were located by their markers after
exposure of the plate to iodine vapor. Results were cal-
culated from the percentage conversion of [7-3Hl-tes-
tosterone to [3H]-DHT, and corrected for procedural
losses by the recovery of [414C]DHT (the conversion of
DHT to 3-diols was negligible under these experimental
conditions).
C
17,20-lyase activity versus log of four to ®ve inhibitor
Acknowledgements
concentrations. Ketoconazole was also incubated as an
internal control for each compound and for comparison
of the extent of inhibition. The results were obtained
from duplicate sets of experiments and were repeated at
least once and several times for the more active inhibi-
tors.11
This work was supported by NIH grant CA 27440 to
A.B. We are grateful to Dr. Patrick Callery for use of the
1H NMR, MS, and IR in the Department of Biomedic-
inal Chemistry, School of Pharmacy, UMAB. We thank
Dr. Stephen Jacobs, Department of Surgery, UMAB
and Dr. James Mohler, Department of Surgery, Uni-
versity N. Carolina for supplying the tissues used in
these experiments. Thanks are also due to Dr. John F.
Templeton, Faculty of Pharmacy, University of Mani-
toba, Canada, for reviewing this manuscript.
(2) 17a-Hydroxylase/C17,20-lyase activity: human tes-
ticular microsomes were incubated with [7-3H]-pregneno-
lone (400 nM, 5Â1204 dpm), and the NADPH
generating system above, for 5 min at 34ꢁC under oxygen.
Authentic steroid markers and C14-labeled pregneno-
lone, 17a-hydroxypregnenolone and DHEA were added
to correct for procedural losses. The steroids were
extracted with ether and then separated by HPLC using
a NOVA-PAK, C18 reverse phase column and eluting
with acetonitrile/methanol/water (30/10/60). The radio-
activity was measured in each fraction collected, as pre-
viously described.10 The 17a-hydroxylase/C17,20-lyase
activity was determined from the percentage conversion
of [7-3H]-pregnenolone to the total amount of 17a-
References
1. Ayub, M.; Level, M. J. Ster. Biochem. 1987, 28, 521.
2. (a) Nakajin, S.; Takahashi, K.; Shinoda, M. Chem. Pharm.
Bull. (Tokyo) 1989, 37, 1855. (b) Nakajin, S.; Takahashi, K.;
Shinoda, M. Yakugaku Zasshi 1988, 108, 1188.
3. Angelastro, M. R.; Laughlin, M. E.; Schatzman, G. L.; Bey, R;
Blohm, T. R. Biochem. Biophys. Res. Commun. 1989, 162, 1571.
4. Jarman. M.; Barrie, S. E; Deadman, J. J.; Houghton, L;
McCague, R. J. Med. Chem. 1990, 33, 2452.
hydroxypregnenolone and DHEA. The
C17,20-lyase
activity was determined from the percentage conversion
of 17-hydroxypregnenolone to DHEA (the conversion
of substrate to androstenediol and testosterone under
the experimental conditions was negligible). The results
are listed in Table 1. IC50 values were calculated from
the dose-response curves for inhibition of this enzyme
complex at four to ®ve inhibitor concentrations. The
5. (a) McCague, R.; Rowlands, M. G.; Barrie, S. E.;
Houghton, J. J. Med. Chem. 1990, 33, 3050. (b) Rowlands, M.
G.; Barrie, S. E.; Chan, F.; Houghton, L; Jarman, M.; McCa-
gue, R.; Potter, G. A. J. Med. Chem. 1995, 38, 4191.
6. Barrie, S. E.; Potter, G. A.; Goddard, P. M.; Haynes, B. R;
Dowsett, M.; Jarman, M. J. Ster. Biochem. Mol. Biol. 1994, 50,
267.

Y. Ling et al./Bioorg. Med. Chem. 6 (1998) 1683±1693
1693
7. Potter, G. A.; Barrie, S. E; Jarman, M.; Rowlands, M. G. J.
Med. Chem. 1995, 38, 2463.
23. (a) Ercoli, A.; de Ruggieri, P. Farm. Sci. Tec. (Pavia) 1952,
7, 11; Chem. Abstr., 1952, 46, 10186. (b) ibid, 1952, 7, 129±33,
Chem. Abstr. 1953, 47, 2191. (c) ibid, 1952, 7, 287±290. Chem
Abstr. 1953, 47, 3325.
8. Li, L.; Li, Y.; Son, C.; Banks, P.; Brodie, A. J. Ster. Bio-
chem. Mol. Biol. 1992, 42, 313.
24. Tzikas, A.; Tamm, C.; Boller, A.; Furst, A. Helv. Chim.
Acta 1976, 59, 1850.
9. Li, L.; Li, Y.; Son, C.; Brodie, A. The Prostate 1995, 26, 140.
10. Li, J. S.; Li, Y.; Son, Chong.; Brodie, A. J. Med. Chem.
1996, 39, 4335.
25. Njar, V. C. O.; Roi, F.; Hartmann, W.; Robinson, C. J.
Chem. Soc. Perkin Trans. 1995, 1, 985.
11. Ling, Y.; Li, L; Liu, Y.; Kato, K.; Klus, G. T.; Brodie, A.
J. Med. Chem. 1997, 40, 3297.
26. Morrowd, D. R; Butler, M. E; Huang, E. C. Y. J. Org.
Chem. 1965, 30, 579.
12. Trachtenberg, J. J. Urol. 1984, 132, 61.
27. (a) Drefahl, G.; Ponsold, K.; Schonecker, B.; Rott, U.
Chem. Ber. 1966, 99, 386. (b) Drefahl, G.; Ponsold, K.; Scho-
necker, B. Chem. Ber. 1965, 98, 186.
13. Williams, G.; Kerle, D. L.; Doble, A.; Dunlop, H.; Smith,
C.; Alien, L.; Yeo, T.; Bloom S. R. Br. J. Urol. 1986, 58, 45.
14. Stoner, E. J. Ster. Biochem. Mol. Biol. 1990, 37, 375.
15. Geller, J. J. Clin. Endocrinol. Metab. 1990, 71, 1552.
28. Suginome, H.; Kaji, M.; Yamada, S. J. Chem. Soc. Perkin
Trans. I 1988, 321.
16. Bolla, M.; Gonzalez, D.; Warde, R; Dubois, J. D.; Mir-
imano€, R.-O.; Storme, G.; Bernier, J.; Kuten, A.; Sternberg,
C.; Gil, T.; Collette, L.; and Pierart, M. N. Engl. J. Med. 1997,
337, 295.
29. Mills, J. S.; Bowers, A.; Djerasi, C.; Ringold, H. J. J. Am.
Chem. Soc. 1960, 82, 3399.
30. Petrow, V.; Lork, L. In The Prostatic Cell: Structure and
Junction. Part B; Alan R. Liss: New York, 1981; p 283.
17. Marker, R. E; Turner, D. L.; Wagner, R. B.; Ushafer, P.
R.; Crooks, Jr. H. M.; Wittle, E. L. J. Am. Chem. Soc. 1943, 63,
779.
31. Niar, V. C. O.; Safe, E; Silverton, J. V.; Robinson, C. H. J.
Chem. Soc. Perkin Tracs. I, 1993, 1161.
32. Niar, V. C. O.; Hector, M.; Hartmann, R. W. Bioorg. Med.
Chem. 1996, 4, 1447.
18. Ward, M. G.; Orr, J. C.; Engel, L. L. J. Org. Chem. 1965,
30, 1421.
33. Murray, M. F. US Patent 2 707,189,155 Chem. Abstr. 1955,
50 p 4246c.
19. Luche, J.-L.; Rodrigez-Hahn, L.; Crabbe. P. J. C. S. Chem.
Comm. 1978, 601.
34. Tendick, F.H.; Lawson, E. J. Chem. Abstr. 1946, 40, 2941.9
20. Corey, E. L; Venkateswarlu, A. J. Am. Chem. Soc. 1972,
94, 6190.
35. Romo, L.; Romo De Vivar, A. J. Am. Chem. Soc. 1959, 81,
3446.
21. Benn W. R. J. Org. Chem. 1963, 28, 3557.
22. Shapiro, E. L.; Gould, D.; Hershberg, E. B. J. Amer.
Chem. Soc. 1955, 77, 2912.
36. Lowry, O. H.; Roseborough, N. S.; Farr, A. L.; Randall,
R. S. J. Biol. Chem. 1951, 193, 265.