Novel steroid inhibitors of glucose 6-phosphate dehydrogenase

Article abstract of DOI:10.1021/jm300317k

Novel derivatives of the steroid DHEA 1, a known uncompetitive inhibitor of G6PD, were designed, synthesized, and tested for their ability to inhibit this dehydrogenase enzyme. Several compounds with approximately 10-fold improved potency in an enzyme assay were identified, and this improved activity translated to efficacy in a cellular assay. The SAR for steroid inhibition of G6PD has been substantially developed; the 3β-alcohol can be replaced with 3β-H-bond donors such as sulfamide, sulfonamide, urea, and carbamate. Improved potency was achieved by replacing the androstane nucleus with a pregnane nucleus, provided a ketone at C-20 is present. For pregnan-20-ones incorporation of a 21-hydroxyl group is often beneficial. The novel compounds generally have good physicochemical properties and satisfactory in vitro DMPK parameters. These derivatives may be useful for examining the role of G6PD inhibition in cells and will assist the future design of more potent steroid inhibitors with potential therapeutic utility.

Full text of DOI:10.1021/jm300317k

Article
pubs.acs.org/jmc
Novel Steroid Inhibitors of Glucose 6-Phosphate Dehydrogenase
Niall M. Hamilton,*,† Martin Dawson,‡ Emma E. Fairweather,† Nicola S. Hamilton,† James R. Hitchin,†
Dominic I. James,† Stuart D. Jones,† Allan M. Jordan,† Amanda J. Lyons,† Helen F. Small,†
Graeme J. Thomson,† Ian D. Waddell,† and Donald J. Ogilvie†
‡
†Cancer Research UK Drug Discovery Unit and Department of Clinical and Experimental Pharmacology,
Paterson Institute for Cancer Research, University of Manchester, Wilmslow Road, Manchester, M20 4 BX, U.K.
S
* Supporting Information
ABSTRACT: Novel derivatives of the steroid DHEA 1, a known
uncompetitive inhibitor of G6PD, were designed, synthesized, and tested
for their ability to inhibit this dehydrogenase enzyme. Several
compounds with approximately 10-fold improved potency in an enzyme
assay were identified, and this improved activity translated to efficacy in a
cellular assay. The SAR for steroid inhibition of G6PD has been
substantially developed; the 3β-alcohol can be replaced with 3β-H-bond
donors such as sulfamide, sulfonamide, urea, and carbamate. Improved
potency was achieved by replacing the androstane nucleus with a
pregnane nucleus, provided a ketone at C-20 is present. For pregnan-20-
ones incorporation of a 21-hydroxyl group is often beneficial. The novel compounds generally have good physicochemical
properties and satisfactory in vitro DMPK parameters. These derivatives may be useful for examining the role of G6PD inhibition
in cells and will assist the future design of more potent steroid inhibitors with potential therapeutic utility.
INTRODUCTION
■
In recent years there has been a resurgence of interest in the
role of metabolism in the development and maintenance of
cancer. Over 80 years ago Warburg discovered that tumor cells
prefer to use glycolysis as a source of adenosine triphosphate
(ATP) even when the oxygen levels present would allow
oxidative phosphorylation in the mitochondrial tricarboxylic
acid cycle (TCA) to be used.1 This aerobic glycolysis, although
rapid, generates only two molecules of ATP compared to
32 molecules from the TCA cycle. To redress the energetic
balance, tumor cells can increase the flux of glucose into the
cell. This increased rate of glucose entering tumor cells has
enabled highly glycolytic tumors to be assessed noninvasively in
Figure 1. PPP. Oxidative metabolism of glucose diverges from the
glycolysis pathway to pyruvate at G6P. Generation of NADPH
the clinic using fluorescent glucose analogues.2
It has since become clear that there are many metabolic
reactions that become altered in tumor cells not only to allow
a rapid production of ATP but also to feed the requirement
for new lipids and nucleotides. This enhanced supply of
metabolites allows tumors to both continue their unrestricted
proliferation and adapt to unfamiliar environments.
A key pathway involved in metabolite production is the
pentose phosphate pathway (PPP) (Figure 1). The PPP contains
an initial oxidative arm that diverges from glycolysis at the point
of glucose 6-phosphate (G6P). This substrate is used to generate
a supply of reduced nicotinamide adenine dinucleotide phosphate
(NADPH) and ribulose 5-phosphate (Ru5-P). NADPH is a
cofactor for many enzymes that are used in macromolecular
biosynthesis, while Ru5-P is used for nucleotide synthesis.
The rate-limiting enzyme for the PPP is G6PD, which
catalyzes the conversion of G6P to 6-phosphogluconolactone
maintains glutathione (GSH) in its reduced state. GSH is oxidized to
GSSG with ROS. Production of Ru5-P and subsequent isomerization
to ribose 5-phosphate enable nucleotide synthesis. As glucose 6-
phosphate dehydrogenase (G6PD) is rate-limiting for 6-phosphogluc-
onate (6PG) production, inhibition of G6PD with steroids such as
DHEA and EA will decrease the flux through the PPP and decrease the
accumulation of 6PG. Metabolism of 6PG by 6-phosphogluconate
dehydrogenase (6PGD) is blocked with the inhibitor 6-amino-
nicotinamide (6-AN).
(6PGL) and results in the concomitant production of NADPH.
In addition to its role in synthesis of new molecules, NADPH is
used in the protection of cells against reactive oxygen species
(ROS) which are generated in rapidly proliferating cells and can
Received: March 5, 2012
Published: April 16, 2012
© 2012 American Chemical Society
4431
dx.doi.org/10.1021/jm300317k | J. Med. Chem. 2012, 55, 4431−4445

Journal of Medicinal Chemistry
Article
damage many macromolecules. ROS are dealt with by an
antioxidant defense that uses glutathione and thioredoxin, both
of which rely on NADPH to maintain their activity.
potential maps.22 Although a 16α-bromo substituent increased
the G6PD inhibitory activity of DHEA 1 and EA 2,15,20,21 there
remains considerable scope for further improvement.
DHEA is an abundant steroid in human circulating blood
and exhibits chemopreventive and antiproliferative effects that
have been attributed to its inhibitory G6PD activity,23−25
although other mechanisms associated with DHEA or its meta-
bolites have also been invoked to explain these effects.26−28
These studies highlight the potential for off-target enzyme
activity when using DHEA and the need to identify a more
potent and selective G6PD inhibitor that is less metabolically
labile.
In this study we aimed to identify a novel and potent steroid
inhibitor of G6PD that would be more amenable than DHEA
for examining the role of G6PD modulation in cells. A more
potent and selective compound would help establish whether
inhibition of G6PD does indeed enhance tumor cell sensitivity
to oxidative stress and whether such an inhibitor may also find
therapeutic utility as a treatment for cancer.
The importance of G6PD in cancer has only become
appreciated in the past decade with the development of mouse
models and tumor cell lines that either carry mutations or are
null for G6PD. Cell lines that overexpress G6PD are able to
form tumors in nude mice, and these cells have an altered
morphology.3 In contrast, tumor cells lacking G6PD grow more
slowly and display enhanced apoptosis.4,5 Furthermore,
Chinese hamster ovary cells that are devoid of G6PD display
increased sensitivity to radiation6 and human fibroblasts
deficient in G6PD display increased ROS and radiation
sensitivity.7 Clinically, there is some evidence from studies in
larynx and gastric cancer that tumors increase G6PD activity,8,9
and this correlates to a poor clinical outcome.9 These data
prompted us to initiate a program to develop a potent inhibitor
of G6PD. It is envisaged that such an inhibitor might reduce
the ability of a cell to manage ROS and when combined with
radiotherapy result in increased tumor cell death.10
Currently, there are no drugs that selectively inhibit G6PD,
although a few small molecules with micromolar activity have
been reported.11,12 It has been known since 1960 that steroids
including dehydroepiandrosterone 1 (DHEA) and epiandros-
terone 2 (EA) uncompetitively inhibit enzyme function.13
Following this discovery several publications describing the
SAR of DHEA and other steroids appeared,14 and more
recently there have been some attempts to improve activity
through synthetic modification15−21 and analysis of electrostatic
CHEMISTRY
■
The synthesis of target compounds is summarized in
Schemes 1−8. Scheme 1 illustrates the synthesis of novel
androstane derivatives bearing a 3β-amino substituent. Hydro-
genation of DHEA 1 over palladium/carbon cleanly afforded
the desired 5α-isomer EA 2 in good yield. Following protection
of the 17-ketone as the cyclic ketal 3, the 3β-alcohol was
inverted via the benzoate ester using benzoic acid, diisopropyl
a
Scheme 1
a
Reagents: (a) H2, Pd/C, EtOH (96%); (b) ethylene glycol, PTSA, HC(OMe)3, CH2Cl2 (84%); (c) benzoic acid, DIAD, TTP, THF (81%);
(d) NaOMe/MeOH, DCM (72%); (e) phthalimide, TPP, DIAD, THF (68%); (f) hydrazine hydrate, EtOH, reflux (97%); (g) for 7a, 2 M HCl, THF,
acetone (89%); (h) for 7b, AcCl, Et3N, DCM, then 2 M HCl, THF, acetone (80%); (i) for 7c, (4-nitrophenyl)-N-benzylcarbamate, Et3N, DCM,
then 2 M HCl, acetone and then H2, Pd/C, AcOH (22%); (j) for 7d, (4-nitrophenyl)-N-methylcarbamate, Et3N, DCM, then HCl, acetone (53%);
(k) for 7e, EtNCO, DCM, then 2 M HCl, THF, acetone (77%); (l) for 7f, N-butyl isocyanate, DCM, then 2 M HCl, THF, acetone (58%); (m) for
7g, Me2NCOCl, Et3N, DCM, then 2 M HCl, THF, acetone (75%); (n) for 7h, MeOCOCl, Et3N, DCM, then 2 M HCl, THF, acetone, DCM
(80%); (o) for 7i, MeSO2Cl, Et3N, DCM, then aq MeSO3H, MeOH (46%); (p) for 7j, EtSO2Cl, Et3N, DCM, then 2 M HCl, THF, acetone (75%);
(q) for 7k, PhSO2Cl, Et3N, DCM, then 2 M HCl, THF, acetone (58%); (r) for 7l, sulfamide, dioxane, reflux, then aq MeSO3H, MeOH (62%).
4432
dx.doi.org/10.1021/jm300317k | J. Med. Chem. 2012, 55, 4431−4445

Journal of Medicinal Chemistry
Article
a
Scheme 2
a
Reagents: (a) H2, Pd/C, EtOH (97%); (b) ethylene glycol, PTSA, HC(OMe)3, DCM (100%); (c) Dess−Martin periodinane, DCM (45%); (d) K-
Selectride, THF, −78 °C (91%); (e) phthalimide, TPP, DIAD, THF (70%); (f) hydrazine hydrate, EtOH, reflux (100%); (g) for 14a, aq MeSO3H,
MeOH, DCM (22%); (h) for 14b, AcCl, Et3N, DCM, then aq MeSO3H, MeOH (89%); (i) for 14c, (4-nitrophenyl)-N-benzylcarbamate, Et3N,
DCM, then 2 M HCl, acetone and then H2, Pd/C, AcOH (14%); (j) for 14d, Me2NCOCl, Et3N, DCM, then aq MeSO3H, MeOH (85%); (k) for
14e, MeOCOCl, Et3N, DCM, then 2 M HCl, THF, acetone, DCM (50%); (l) for 14f, MeSO2Cl, Et3N, DCM, then 2 M HCl, THF, acetone, DCM
(60%); (m) for 14g, EtSO2Cl, Et3N, DCM, then 2 M HCl, THF, acetone, DCM (63%); (n) for 14h, PhSO2Cl, Et3N, DCM, then 2 M HCl, THF,
acetone, DCM (86%); (o) for 14i, sulfamide, dioxane, reflux, then aq MeSO3H, MeOH (39%); (p) for 14j, Me2NSO2NH2, DIPEA, DCM, then aq
MeSO3H, MeOH (15%).
a
Scheme 3
a
Reagents: (a) phthalimide, TPP, DIAD, THF (74%); (b) hydrazine hydrate, EtOH, reflux (95%); (c) for 17a, 2 M HCl, THF, acetone (87%);
(d) for 17b, EtNCO, DCM, then 2 M HCl, THF, acetone, DCM (37%); (e) for 17c, MeSO2Cl, Et3N, DCM, then 2 M HCl, THF, acetone, DCM
(73%); (f) for 17d, sulfamide, dioxane, reflux, then 2 M HCl, THF, acetone (50%).
azodicarboxylate (DIAD), and triphenylphosphine (TPP) in
THF, and the ester was hydrolyzed with sodium methoxide in
methanol to give the 3α-ol 4. This alcohol was converted to the
3β-phthalimide 5 using phthalimide, DIAD, and TPP in THF
and then converted in good yield to the 3β-amine 6 through
treatment with hydrazine hydrate in ethanol. The cyclic ketal of
amine 6 was cleaved with acid to yield the desired aminoketone
7a. This method of inversion of the stereochemistry from
the 3α-ol 4 represented a substantial advantage over other litera-
ture methods as the 3β-amine 7a was isolated as a single
4433
dx.doi.org/10.1021/jm300317k | J. Med. Chem. 2012, 55, 4431−4445

Journal of Medicinal Chemistry
Article
diastereoisomer in good yield without any need for chromato-
graphic purification of the intermediates.29 The amine 6 was
also converted to a range of target compounds 7b−l that
included an amide 7b, ureas 7c−g, carbamate 7h, sulfonamides
7i−k, and sulfamide 7l.
Scheme 2 shows the synthesis of novel pregnane derivatives
with a 3β-amino substituent. Similar chemistry as described for
the above androstanes was employed to convert pregnenolone 8
to a range of target compounds 14, although initially the
inversion of the 3β-alcohol 9 was effected via oxidation with
Dess−Martin periodinane, followed by reduction of the 3-ketone
10 with K-Selectride in THF. By use of standard conditions, the
amine 13 was converted to the target compounds 14a−j that
included an amine 14a, amide 14b, ureas 14c,d, carbamate 14e,
sulfonamides 14f−h, and sulfamides 14i,j.
The phthalimide-based Mitsunobu inversion chemistry
described above was used to prepare additional target
androstanes with a 3α-substituted amine (Scheme 3). Thus,
the 3β-alcohol 3 afforded the 3α-amine 16 in two steps, and the
ketal was cleaved to give amine 17a. The 3α-amine 16 was also
functionalized using ethyl isocyanate, methanesulfonyl chloride,
and sulfamide and following ketal deprotection afforded urea
17b, sulfonamide 17c, and sulfamide 17d, respectively.
Two routes were employed to convert pregnan-20-ones to
21-hydroxypregnan-20-ones. Oxidation of the silyl enol ether of
ketone 14 and treatment with acid gave the desired urea 18a,
carbamate 18b, and sulfamide 18d (Scheme 4). Alternatively
and sodium ethoxide to give the known α,β-unsaturated ester
19. Hydrogenation over palladium/carbon gave the 17β-alkyl
ester 20 which was hydrolyzed to the acid 21 using aqueous
lithium hydroxide and then converted to the Weinreb amide
22. Reaction of amide 22 with methylmagnesium bromide in
THF and then acidic deprotection of the TBDMS ether
afforded the desired ketone 23.
The oxetane 25 (Scheme 7) was prepared in low yield by
heating the known epoxide 24 with trimethylsulfoxonium
iodide and potassium tert-butoxide in tert-butanol at reflux,
while the methyl ether 26 (Scheme 8) was obtained by reacting
3-hydroxypregnan-20-one 9 with copper(II) bromide in
methanol at reflux.
BIOLOGY
■
The enzyme G6PD converts its substrate, G6P, to 6PGL with
the concomitant reductive conversion of the cofactor NADP+ to
NADPH. Rather than measurement of the intrinsic fluorescence
of NADPH, it was decided to quantify its production via an
Amplite fluorimetric NADPH assay kit through enzymatic
coupling and reductive conversion of a nonfluorescent NADPH
sensor substrate to a product detectable by excitation/emission
at 540/590 nm. The addition of the coupled Amplite step and
subsequent measurement of the assay signal in the red visible
range significantly reduces the potential for compound
interference effects in comparison to directly measuring the
intrinsic fluorescence of NADPH. This method generated a
robust assay with Z′ > 0.7. This G6PD in vitro enzyme assay was
used to test the inhibitory activity of the target compounds, and
the results are reported in Tables 1 and 2.
a
Scheme 4
Following screening in the enzyme assay, the most active
compounds were then tested in a cell assay. Measurement of
flux through the oxidative PPP in living cells has previously
been achieved by using glucose molecules labeled on their first
carbon atom.30 Metabolism of 1-14C-glucose by the PPP results in
14C-labeled CO2 release by the cells, and this release can be
quantified. However, this method measures the combined activity
of both G6PD and 6PGD and was not suited to the higher
specificity required for our experiments. We therefore decided to
use mass spectrometry to quantify production of the metabolite
6PG (Figure 1). Initial experiments showed that the absolute
levels of 6PG varied little after treatment of cells with DHEA 1.
We speculated that the enzyme 6PGD rapidly metabolized 6PG
and therefore masked any change in the flux accomplished by
inhibiting G6PD. We hypothesized that inhibition of 6PGD
should lead to an accumulation of 6PG, and this was tested
using a commercially available inhibitor of 6PGD, 6-AN, which
has been reported to induce 6PG accumulation in breast cancer
cells.31 The addition of 200 μM 6-AN for 4−6 h led to a
substantial accumulation of 6PG. As G6PD is rate-limiting for
6PG production, it is expected that inhibition of G6PD would
decrease the flux through the PPP and decrease the observed
accumulation of 6PG. This was confirmed by inhibition of G6PD
using either DHEA or siRNA (data not shown) in the presence
of 6-AN, which led to a decrease in the accumulation of 6PG.
The ability of selected test compounds to inhibit G6PD activity in
HEK293T cells is reported in Table 3.
a
Reagents: (a) tert-butyldimethyl trifluoromethanesulfonate, triethyl-
amine, DCM; (b) m-CPBA, −10 °C; (c) 5:1 THF/2 M HCl.
bromination of ketone 14f at the 21-position in low yield
using bromine in acidified methanol and then rapid treatment
with aqueous sodium hydroxide in DMF gave sulfonamide 18c
(Scheme 5). This latter route was less successful, as purification
of the 21-bromo intermediate was protracted.
a
Scheme 5
a
Reagents: (a) Br2, HBr, MeOH (13%); (b) NaOH, aq DMF (59%).
Three additional novel compounds bearing a 3β-alcohol were
also synthesized (Schemes 6−8). For synthesis of the
homologated ketone 23 (Scheme 6), the 3β-alcohol of
epiandrosterone 2 was first protected as its TBDMS ether
and the 17-ketone then reacted with triethyl phosphonoacetate
SAR AND COMPOUND DESIGN
■
The activity of DHEA as an inhibitor of G6PD was first
reported by Marks and Banks in 1960.13 In general the SAR
from this and subsequent studies14−22 indicated that G6PD
4434
dx.doi.org/10.1021/jm300317k | J. Med. Chem. 2012, 55, 4431−4445

Journal of Medicinal Chemistry
Article
a
Scheme 6
a
Reagents: (a) TBDMS-Tf, pyridine (90%); (b) triethyl phosphonoacetate, NaOEt (100%); (c) H2, Pd/C, ethyl acetate (92%); (d) LiOH, aq THF
(88%); (e) 1,1′-carbonyldiimidazole, N,O-dimethylhydroxylamine HCl, Et3N, DMF, DCM (92%); (f) MeMgBr, THF, then (g) 2 N HCl, acetone
(24%).
a
incorporating additional polar groups at synthetically accessible
positions. To better understand the activity of our novel steroids,
several literature compounds were purchased or prepared for
comparison in the enzyme assay, including the aforementioned
16α-bromo derivatives. Although the 16α-bromo compounds are
reported to be more potent G6PD inhibitors than either 1 or 2,
we wished to avoid reactive functional groups in newly designed
compounds to minimize potential off-target activity and toxicity.
Scheme 7
a
Reagents: (a) Me3SOI, KOtBu, BuOH, reflux (6.5%).
t
RESULTS AND DISCUSSION
a
■
Scheme 8
We focused on saturated steroids, as the literature compound
EA 2 was as potent as the Δ5-analogue DHEA 1, and
allopregnanolone 9 was more potent than pregnenolone 8. The
removal of the double bond in the B-ring should also help
prevent formation of potent androgens bearing a Δ4-3-one that
are formed by oxidative metabolism from 3β-hydroxy-Δ5-
steroids such as DHEA 1.
a
The compounds were tested in a G6PD enzyme assay and the
results are shown in Tables 1 and 2. For the steroids shown in
Table 1, androstane derivatives are labeled 7 and 17 while
pregnane derivatives are labeled 14 and 18. Direct replacement
of the 3β-alcohol with a 3β-amine abolished activity in both the
androstane and pregnane series (compounds 7a and 14a,
respectively). The lack of activity for the parent 3β-amines may
be due to the large desolvation penalty that would need to be
overcome, as protonation of the amines will occur at the pH of
the assay. Within the androstane series, improved activity relative
to EA 2 was realized upon conversion of the 3β-amine 7a to the
3β-ureas 7c, 7e, and 7f, 3β-carbamate 7h, 3β-sulfonamides 7i
and 7k, and 3β-sulfamide 7l. This important finding
demonstrated that G6PD inhibition was not wholly dependent
on retention of a 3β-alcohol and that improved activity was
possible with alternative H-bond donors. Within the pregnane
series, improved activity relative to allopregnanolone 9 was also
found with sulfonamide 14f and sulfamide 14i, although the
trend did not extend to either of the ureas 14c or 14d or the
carbamate 14e, suggesting binding modes (or perhaps binding
sites) are dependent on the specific steroid nucleus.
Reagents: (a) CuBr2, MeOH, reflux (24%).
inhibitory activity required an androst-5-ene or pregn-5-ene
nucleus bearing a 3β-alcohol and either a 17-ketone (for
androstenes) or 20-ketone (for pregnenes). Saturation of the
steroid nucleus was well tolerated only when the resulting
stereochemistry was 5α (allo configuration). Oxidation or
inversion of the 3β-alcohol reduced activity. Introduction of
other polar (alcohol or ketone) groups at the 7 or 11 position
was detrimental, while alcohols at the 16α or 21 position were
tolerated. Relative to DHEA 1 or EA 2 the only improvement
was incorporation of a 16α-bromo substituent.15,20,21 These
SARs are summarized in Figure 2.
For new G6PD inhibitors we chose to retain the preferred
steroid nuclei and explore the replacement of the 3β-alcohol
with other H-bond donors and the 17- or 20-ketone with other
H-bond acceptors. We were keen to incorporate these features
that we felt were likely to preclude metabolism to potent
androgens (or progestins), a problem exhibited by both DHEA
1 and EA 2, which makes interpretation of biological activity
more speculative. Both 1 and 2 have relatively poor aqueous
solubility32,33 which we hoped to improve by increasing
polarity with new 3β-, 17-, or 20-substituents and, if possible,
Closer examination of the 3β-ureas, 3β-sulfonamides, and
3β-sulfamides revealed that inhibitory activity remains good in
the androstane series with small substituents and tends to fall as
4435
dx.doi.org/10.1021/jm300317k | J. Med. Chem. 2012, 55, 4431−4445

Journal of Medicinal Chemistry
Article
a
Table 1. Inhibition of Glucose 6-Phosphate Dehydrogenase by Steroids
b
c
b
c
compd
3-substituent
β-OH
IC50 (μM)
n
compd
3-substituent
IC50 (μM)
n
1
11.3 2.2
9.4 1.0
>200
4
14b
14c
14d
14e
14f
14g
14h
14i
β-NHAc
>200
2
2
2
8
4
4
4
2
2
2
2
2
2
2
2
2
2
β-OH
2
5
3
2
2
2
2
2
6
2
6
2
2
2
4
2
β-NHCONH2
β-NHCONMe2
β-NHCO2Me
β-NHSO2Me
β-NHSO2Et
>100
7a
7b
7c
7d
7e
7f
7g
7h
7i
β-NH2
>200
β-NHAc
128 18
5.0 0.4
12.2 0.3
1.0 0.1
1.7 0.8
>200
>200
β-NHCONH2
β-NHCONHMe
β-NHCONHEt
β-NHCONHnBu
β-NHCONMe2
β-NHCO2Me
β-NHSO2Me
β-NHSO2Et
β-NHSO2Ph
β-NHSO2NH2
β-OH
3.4 0.7
11.0 7.6
20.3 6.5
2.0 0.4
β-NHSO2Ph
β-NHSO2NH2
β-NHSO2NMe2
α-NH2
14j
37.8 20.4
>200
d
1.5 1.2
6.6 1.2
17a
17b
17c
17d
α-NHCONHEt
α-NHSO2Me
α-NHSO2NH2
β-NHCONHEt
β-NHCO2Me
β-NHSO2Me
β-NHSO2NH2
>200
7j
70.5 58.6
9.9 3.5
1.2 0.1
81.4 19.5
4.3 1.3
>200
>200
7k
7l
>200
e
18a
18b
18c
18d
34.9 3.7
41.6 10.2
3.8 0.5
1.2 0.2
8
9
β-OH
14a
β-NH2
2
a
b
Inhibition of G6PD was determined by measuring the NADPH generated during the assay using an Amplite fluorimetric NADPH assay kit. IC50
c
SD determined from 10 point concentration/effect experiments. Geometric mean of at least two independent duplicate experimental
determinations. 17a tested as hydrochloride salt. 18a tested as 1:1 mixture of 3α- and 3β-isomers.
d
e
in potency for the larger ethanesulfonamide 14g, benzenesulfo-
namide 14h, and dimethylsulfamide 14j.
Table 2. Inhibition of Glucose 6-Phosphate Dehydrogenase
by Steroids Shown in Chart 1
a
Literature results have previously demonstrated that inversion
of the 3β-alcohol abolishes activity.14d,g To confirm whether
this extended to alternative H-bond donors, the 3α-amine 17a,
3α-urea 17b, 3α-sulfonamide 17c, and 3α-sulfamide 17d were
prepared and tested. All of these 3α-isomers failed to inhibit
G6PD activity in the enzyme assay, confirming previous
observations.
Novel 3β-substituents were combined with a 21-hydroxy-
pregan-20-one scaffold to yield compounds 18a−d. As with
3β-alcohols (discussed below), the incorporation of a 21-
hydroxyl group improved the activity of the 3β-carbamate 18b
and 3β-sulfamide 18d relative to the unsubstituted analogues
14e and 14i, respectively. The activity of the 21-hydroxy-3β-
sulfonamide 18c was similar to the unsubstituted analogue 14f.
The ethylurea 18a was more active than the primary urea 14c
but substantially less active than the androstrane ethylurea 7e.
We then turned our attention to the opposite end of the
steroid to establish whether the ketone could be replaced with
other H-bond acceptors (Chart 1). The strategy was to explore
SAR predominantly with literature compounds retaining a
3β-alcohol and then to combine any improvements with the
novel 3β-H-bond donors to investigate if SARs are additive.
Enzyme assay results for compounds depicted in Chart 1 are
shown in Table 2.
In contrast to the literature reporting improved inhibition of
glucose 6-phosphate dehydrogenase from trypanosomal para-
sites with the incorporation of a 16α-bromo group,20,21
androstanes 27 and 29 did not display improved potency in
the enzyme assay relative to DHEA 1 or EA 2, respectively. The
16α-hydroxyandrostane 28 retained activity but was slightly less
active than DHEA 1.
The lack of activity with compounds 24, 25, and 30−32
demonstrated that replacement of the 17-ketone with an oxime,
epoxides, or oxetanes is detrimental and suggested that for
compounds retaining an androstane nucleus, SAR in this region
is tight. Synthetic ketal intermediates were also inactive in the
enzyme assay (results not shown).
b
c
b
c
compd
IC50 (μM)
n
compd
IC50 (μM)
n
1
11.3 2.2
9.4 1.0
4
34
35
36
37
38
39
40
41
42
43
44
45
46
47
0.9 0.5
>200
8
2
2
2
4
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
4
2
2
2
2
2
2
8
81.4 19.5
4.3 1.3
>200
9.3 1.4
2.5 1.0
>200
9
23
24
25
26
27
28
29
30
31
32
33
>200
>200
>200
>200
>200
66.0 29.3
31.6 6.1
15.8 1.9
6.3 0.4
>200
48.0 2.6
15.8 3.3
20.8 4.4
>200
>200
5.2 1.3
47.0 1.3
>200
5.3 1.3
a
Inhibition of G6PD was determined by measuring the NADPH
generated during the assay using an Amplite fluorimetric NADPH
assay kit. IC50 SD determined from 10 point concentration/effect
experiments. Geometric mean of at least two independent duplicate
b
c
experimental determinations.
substituent size increases. Thus, the small monoalkylated ureas
7c−f are more potent than the dimethylurea 7g, with the
ethylurea 7e most potent. Similarly the methyl carbamate 7h
retains good activity, and the methanesulfonamide 7i is more
potent than the ethanesulfonamide 7j. However, the
benzenesulfonamide 7k does not fit this trend, being more
active than the ethanesulfonamide 7j, although the reason is not
obviously apparent. Comparison of androstane sulfamide 7l
with the pregnane sulfamide 14i and dimethylsulfamide 14j
shows that, like the ureas, dialkylation of the sulfamide nitrogen
further from the steroid is deleterious. For the pregnane series
there appears to be a more stringent steric effect as the
analogous ureas 14c and 14d and methyl carbamate 14e are all
inactive. Good activity is only retained with the methanesulfo-
namide 14f and the sulfamide 14i, with a substantial reduction
4436
dx.doi.org/10.1021/jm300317k | J. Med. Chem. 2012, 55, 4431−4445

Journal of Medicinal Chemistry
Article
a
Table 3. Steroid Inhibition of Glucose 6-Phosphate Dehydrogenase in HEK293T Cells and Aqueous Solubility of the Steroids
b
c
d
compd
3β-substituent
IC50 (μM)
n
(cell IC50)/(enzyme IC50)
solubility (μM)
1
OH
24.9 13.2
9.5 4.7
7.8 1.6
17.4 8.3
4.8 1.2
9.2 1.7
35.2 1.8
14.1
5
2.2
>100
7e
NHCONHEt
NHCO2Me
NHSO2NH2
NHSO2Me
NHSO2NH2
NHSO2Me
NHSO2NH2
OH
3
3
2
3
3
3
1
3
9.5
30−100
>100
7h
7l
5.2
14.5
1.4
30−100
3−10
10−65
10−30
>100
14f
14i
18c
18d
34
4.6
9.2
11.8
13.7
12.3 2.5
>100
a
b
Inhibition of G6PD was determined by measuring the accumulation of 6PG and blocking its metabolism using 200 μM 6-AN. IC50
SD
c
determined from 10 point concentration/effect experiments. Geometric mean of at least two independent experimental determinations (except for
compound 18d). Turbidimetric aqueous solubility, 1% DMSO in PBS.
d
and EA 2 were identified: compounds 7e, 7l, 18d, and 34.
The 3β-alcohol of DHEA can be replaced with alternative
3β-H-bond donors such as sulfamide, sulfonamide, urea, and
carbamate. Good potency is retained with either an androstane
or pregnane nucleus, provided a carbonyl H-bond acceptor
at C-17 (androstane) or C-20 (pregnane) is present. A
21-hydroxyl group is well tolerated and may improve activity.
The activity of the series plateaus at about 1 μM in the enzyme
assay, although the assay is capable of detecting substantially
more potent compounds. The results also demonstrate that
potency can be improved with a reduction in lipophilicity,
best illustrated by the sulfamides 7l (XlogP = 2.78) and 18d
(XlogP = 2.27) compared with DHEA 1 (XlogP = 3.62) or EA
2 (XlogP = 3.78). In addition the alternative H-bond donors
are expected to have better metabolic stability than the 3β-
alcohol of 1 and 2, with reduced propensity to form androgenic
Figure 2. Established SAR for steroid inhibition of G6PD. Functional
groups shown in green areas are essential for good inhibitory activity.
Areas in amber tolerate some modification, and substituents in red
areas typically reduce or abolish activity.
metabolites or phase II conjugates.
On the basis of these considerations, a number of the
compounds were selected and tested in a cellular assay. The
aqueous solubility of these steroids was also determined, and
the combined results are shown in Table 3.
The improved activity of the novel compounds over DHEA 1
in the enzyme assay, with the exception of the sulfonamide 18c,
was repeated in the cell assay. There was, however, no clear
correlation between the inhibitory activity of the compounds in
the two assays, and the reduction in activity on translation from
enzyme to cell assay varied from 1.4-fold (sulfonamide 14f) to
14.5-fold (sulfamide 7l). Five of the novel compounds in Table 3
have moderate−good aqueous solubility. The reduction in
cellular activity does not correlate with poor aqueous solubility.
This is best exemplified by the water-soluble sulfamides 7l and
18d, which are among the least potent compounds in cells.
To establish whether permeability adversely affects cell
activity, a small number of compounds were tested in vitro in
a Caco-2 permeability assay (Table 4). The stability of some
compounds in the presence of P450 (phase 1) metabolic
enzymes was tested in vitro using human liver microsomes,
and the results are shown in Table 4. With the exception of
sulfamide 14i, the Caco-2 permeability was moderate−good for
the steroids tested, and there was no evidence of efflux. The
good permeability of urea 7e and sulfamide 7l suggests that
permeability is unlikely to be the main reason for the reduction
in activity on translation from the enzyme assay to the cell assay.
No clear rationale for the discrepancy between the enzyme and
cell assays was established; the discrepancy may be due to a
combination of factors including solubility, permeability, and
biological or assay variability.
To investigate whether pregnanes offered greater scope for
further elaboration in this region, compounds 33−38 were
examined. Relative to pregnenolone 8 or pregnanolone 9,
substantial improvements in activity were realized with the
21-hydroxyl derivatives 33 and 34, respectively, while 21-halo
derivatives 36 and 37 retained activity. However, the 21-acetate
35 had no activity. As in the androstane series, conversion of
the ketone to the oxime 38 abolished activity and the novel
homologated ketone 23 also lacked activity.
As anticipated from prior art,14d,g the 3α-alcohols 39, 41, and
42 were either inactive or substantially less active than the
corresponding 3β-alcohols, but in contrast to a literature
report,14d 5α-androstan-17-one 40 (which lacks the 3β-ol) in
our hands is also inactive. Pregnanes 43 and 44, which possess
an 11-ketone, retain moderate activity but are less potent than
the corresponding unsubstituted steroids 9 and 34, respectively.
The 17α-methoxypregnane 26 and 16,17-epoxypregnane 45
both lack activity. Moderate activity is retained with the 17α-
hydroxypregnane 46 which also has the beneficial 21-alcohol,
while the 17α-hydroxypregnane 47, which lacks the 21-alcohol
and also has a deleterious 11-ketone, is almost devoid of activity.
From this assessment it was apparent that most modifications
to the 17- or 20-ketones reduce or abolish activity, but a
21-hydroxyl group markedly improved activity.
To summarize the enzyme assay results, four derivatives with
a potency increase of approximately 10-fold relative to DHEA 1
4437
dx.doi.org/10.1021/jm300317k | J. Med. Chem. 2012, 55, 4431−4445

Journal of Medicinal Chemistry
Article
Chart 1
Table 4. In Vitro DMPK Parameters of Selected Steroids
excellent in vitro metabolic stability against human liver
microsomes, although the result should be treated with caution
owing to the low solubility of this compound. Although none of
the compounds in Table 3 have been tested against a broad
range of receptors, preliminary results demonstrate that none of
them have appreciable activity against two related dehydrogenase
enzymes (≪50% inhibition at 200 μM) and the alcohol 34 is
devoid of CYP450 inhibition against Cyp 1A, 2C19, 2C9, 2D6,
3A4 (>25 μM for all five isoforms tested). The improved
potency of the urea 7e, sulfamide 18d, and alcohol 34 relative to
DHEA 1 and EA 2, coupled with their apparent selectivity and
their good physicochemical and in vitro DMPK parameters,
suggests that these compounds may be useful pharmacological
tools to further elucidate the role of G6PD in cells.
a
Pa→b
a
b
c
compd 3β-substituent
(10−6 cm s−1
)
Pb→a/Pa→b
Clint
56.8
T1/2
7e
NHCONHEt
NHSO2NH2
NHSO2Me
NHSO2NH2
NHSO2Me
NHSO2NH2
31.6 3.1
31.8 5.4
0.54
0.66
24.4
d
d
7l
ND
ND
135
ND
ND
d
d
14f
14i
18c
18d
ND
ND
10.3
d
d
2.1 0.1
0.53
0.30
0.49
ND
ND
d
d
17.4 4.2
13.1 0.01
28.3
48.9
a
Pa→b is the permeability when applied to the apical side, while Pb→a is
the permeability when applied to the basolateral side in a Caco-2
membrane assay (10−6 cm s−1). The ratio of these two values is a
measure of Pgp efflux, while the Pa→b value is a measure of
b
permeability. Metabolic stability in human liver microsomes: intrinsic
c
clearance Clint in (μL/min)/(mg protein). Half-life (T1/2) in min.
d
ND = not determined.
CONCLUSION
■
By use of DHEA 1 and EA 2 as literature leads, the SAR for
steroid inhibition of G6PD has been substantially developed
and a number of derivatives with a potency increase of
In addition to enzyme and cell potency compounds 7e and
18d also showed moderate in vitro metabolic stability against
human liver microsomes. Sulfonamide 14f appeared to show
4438
dx.doi.org/10.1021/jm300317k | J. Med. Chem. 2012, 55, 4431−4445

Journal of Medicinal Chemistry
Article
stan-17-one, cyclic 1,2-ethanediyl acetal 4 (8.17 g, 24.4 mmol),
phthalimide (3.59 g, 24.4 mmol), and TPP (7.08 g, 25.6 mmol) were
dissolved in THF (175 mL), and the solution was cooled to 5 °C using
an ice bath. DIAD (5.3 mL, 26.8 mmol) was added slowly maintaining
the temperature below 5 °C. A white precipitate formed during the
reaction. The mixture was stirred and warmed to ambient temperature
overnight under a nitrogen atmosphere. All volatiles were evaporated.
Then methanol (100 mL) was added and the mixture stirred for
30 min. The precipitate was filtered off and washed with methanol.
The solid was dried under vacuum to give the product 5 as a white
approximately 10-fold were identified. In particular this study
has established that the 3β-alcohol can be replaced with
alternative 3β-H-bond donors such as sulfamide, sulfonamide,
urea, and carbamate. Good potency can also be retained by
replacing the androstane nucleus with a pregnane nucleus,
provided a carbonyl H-bond acceptor at C-17 (androstane) or
C-20 (pregnane) is retained. For pregnan-20-ones a 21-
hydroxyl group may improve activity and in combination
with a 3β-alcohol, 3β-sulfamide, or 3β-sulfonamide leads to
compounds with good potency. The urea 7e, sulfamide 18d,
and alcohol 34 offer substantial benefits over DHEA 1, EA 2,
and their 16α-bromo derivatives for examining the role of
inhibiting G6PD in cells and will assist the future design of
more potent steroid inhibitors with potential therapeutic utility.
1
solid (7.67 g, 68%). H NMR (CDCl3): δ 0.87 (s, 3H), 0.98 (s, 3H),
1.95−2.05 (m, 1H), 2.22−2.49 (m, 2H), 3.85−3.97 (m, 4H), 4.13−
4.24 (m, 1H), 7.68−7.74 (m, 2H), 7.79−7.85 (m, 2H).
(3β,5α)-3-Aminoandrostan-17-one, Cyclic 1,2-Ethanediyl
Acetal 6.29 2-[(3β,5α)-17-Oxoandrostan-3-yl, 17-cyclic 1,2-ethanediyl
acetal]-1H-isoindole-1,3(2H)-dione 5 (7.67 g, 15.6 mmol) was mixed
with ethanol (100 mL) and hydrazine hydrate (11 mL, 226 mmol) and
refluxed overnight under a nitrogen atmosphere. The reaction mixture
was cooled, and the solid was filtered off. The filtrate was evaporated
to dryness. The residue was partitioned between DCM and water, and
the solution passed through a hydrophobic frit. The DCM phase was
EXPERIMENTAL SECTION
■
General Methods. All reagents obtained from commercial sources
were used without further purification. Anhydrous solvents were
obtained from the Sigma-Aldrich Chemical Co. Ltd. or Fisher
Chemicals Ltd. and used without further drying. Solutions containing
products were either passed through a hydrophobic frit or dried over
anhydrous MgSO4 or Na2SO4 and filtered, prior to evaporation of the
solvent. Solvents were evaporated under reduced pressure. Thin layer
chromatography (TLC) was conducted with 5 cm × 10 cm plates
coated with Merck type 60 F254 silica gel to a thickness of 0.25 mm.
As the steroids typically lack a UV chromophore, TLC was often used
to monitor reactions and initially assess compound purity; elution was
usually with isohexane/ethyl acetate mixtures. TLC plates were then
sprayed with a 4% H2SO4/methanol solution and heated on a hot
plate to develop. Chromatography was performed on Biotage SNAP
HP-Sil cartridges using a CombiFlash Companion machine.
1
evaporated to dryness to give the product 6 (5.34 g, 97%). H NMR
(CDCl3): δ 0.79 (s, 3H), 0.84 (s, 3H), 1.92−2.04 (m, 1H), 2.61−2.73
(m, 1H), 3.82−3.98 (m, 4H). 13C NMR (CDCl3): δ 11.2, 13.3, 19.4,
21.5, 27.5, 29.6, 30.3, 31.4, 33.1, 34.5, 34.7, 36.6, 38.2, 44.4, 44.9, 49.3,
50.0, 53.2, 63.4, 64.0, 118.4.
(3β,5α)-3-Aminoandrostan-17-one 7a.29 (3β,5α)-3-Aminoan-
drostan-17-one, cyclic 1,2-ethanediyl acetal 6 (100 mg, 0.30 mmol) was
dissolved in acetone (1 mL), DCM (1 mL) and THF (4 mL). Then
2 M HCl (0.5 mL, 1 mmol) was added and the mixture stirred overnight.
The reaction mixture was basified with 1 M NaOH. DCM was added
and the mixture stirred before passing through a hydrophobic frit. The
organic phase was concentrated and chromatographed on silica (10 g
SNAP cartridge, DCM/MeOH gradient elution) to give the product 7a
as a white solid (10 mg, 12%). 1H NMR (CDCl3): δ 0.84 (s, 3H), 0.88
(s, 3H), 2.40−2.52 (m, 1H), 2.70−2.82 (m, 1H). 13C NMR (CDCl3):
δ 11.2, 12.7, 19.3, 20.7, 27.3, 29.8, 30.4, 31.0, 34.0, 34.6, 34.7, 34.8,
36.4, 37.6, 44.4, 46.7, 49.9, 50.4, 220.3. HRMS (ESI) m/z [M + H]+
calcd for C19H32NO: 290.2479. Found: 290.2480.
Proton (1H) and carbon (13C) NMR spectra were recorded on a
300 MHz Bruker spectrometer at ambient temperature. Solutions were
typically prepared in either deuterochloroform (CDCl3) or deuterated
dimethylsulfoxide (DMSO-d6) with chemical shifts referenced to
deuterated solvent as an internal standard. 1H NMR data are reported
indicating the chemical shift (δ), the multiplicity (s, singlet; d, doublet;
t, triplet; q, quartet; m, multiplet; br, broad; dd, doublet of doublets,
etc.), the coupling constant (J) in Hz, and the integration (e.g., 1H).
Deuterated solvents were obtained from the Sigma-Aldrich Chemical
Co., Goss or Fluorochem.
N-[(3β,5α)-17-Oxoandrostan-3-yl]acetamide 7b. To a stirred
solution of (3β,5α)-3-aminoandrostan-17-one, cyclic 1,2-ethanediyl
acetal 6 (200 mg, 0.600 mmol) and triethylamine (121 mg, 1.20 mmol)
in DCM (5 mL) at 20 °C under nitrogen was added acetyl chloride
(49 mg, 0.63 mmol), and the resulting mixture was stirred overnight.
Water (4 mL) was added, and stirring continued for 15 min. The
reaction mixture was filtered through a hydrophobic filter and the
organic phase evaporated to dryness. The 17-ketal was deprotected in a
similar manner to compound 7a, affording the product 7b as a white
foam (160 mg, 80%). 1H NMR (CDCl3): δ 0.82 (s, 3H), 0.86 (s, 3H),
1.96 (s, 3H), 2.38−2.50 (m, 1H), 3.68−3.84 (m, 1H), 5.43 (br d, J =
7.9 Hz, 1H). 13C NMR (CDCl3): δ 12.2, 13.8, 20.4, 21.8, 23.5, 28.2,
28.7, 30.8, 31.5, 35.1, 35.3, 35.6, 35.9, 37.3, 45.3, 47.8, 48.9, 51.4, 54.3,
169.3, 221.3.
N-[(3β,5α)-17-Oxoandrostan-3-yl]urea 7c. To a stirred solution
of (3β,5α)-3-aminoandrostan-17-one, cyclic 1,2-ethanediyl acetal 6
(100 mg, 0.30 mmol) in DCM (1 mL) were added (4-nitrophenyl)-N-
benzylcarbamate (89.8 mg, 0.33 mmol) and triethylamine (0.05 mL,
0.33 mmol). The mixture was heated at 40 °C for 6 h and then stirred
at ambient temperature overnight. The reaction mixture was flash
chromatographed on silica (25 g SNAP cartridge, ethyl acetate/
isohexane gradient elution) to give the benzylurea intermediate
(148 mg) as a viscous oil. To a stirred solution of the benzylurea
intermediate (148 mg, 0.32 mmol) in acetone (4 mL) was added 2 M
HCl (1 mL) and the mixture stirred at ambient temperature for 2 h.
The mixture was evaporated. Acetone (10 mL) was added and the
solution evaporated. The residue was dried at 40 °C under vacuum for
12 h to give a gum which was purified by flash chromatography
(isohexane/ethyl acetate gradient elution), eluting to give the ketone
intermediate (113 mg, 0.27 mmol, 84%) as a white foam. To a stirred
solution of the ketone benzylurea intermediate (93 mg, 0.22 mmol) in
HRMS results were obtained on a Waters QTof Micro instrument,
using positive and negative polarity electrospray with sodium formate
as standard calibrant.
LC−MS spectra with evaporative light scattering detection (ELSD)
were recorded on a Waters Acquity UPLC. This analytical method was
used to determine compound purity which is ≥95% unless otherwise
stated. Mass spectrometry was performed on a Waters Acquity SQD
quadrupolar spectrometer running in dual ES+ and ES− mode. High
pH runs were conducted at pH 10 and low pH runs were conducted at
pH 3, with a run time of 2 mins. The column temperature was 40 °C,
and the flow rate was 0.6 mL/min. Further details including solvent
gradients are given in the Supporting Information. Details of the
preparative HPLC instrument and the solvent gradient used to purify
two compounds are also given in the Supporting Information
DHEA 1, (3α,5α)-3-hydroxypregnan-20-one 41, (3α,5α)-3,21-
dihydroxypregnan-20-one 42, (5α)-3-hydroxypregnane-11,20-dione
43, (3β,5α)-3,21-dihydroxypregnane-11,20-dione 44, and (3β,5α)-
3,17-dihydroxypregnane-11,20-dione 47 were purchased from Sigma-
Aldrich. EA 2 and (3β)-3-hydroxypregn-5-en-20-one 8 were purchased
from Acros Organics. (5α)-Androstan-17-one 40 was purchased from
Interbioscreen. (3β,5α)-3,17,21-Trihydroxypregnan-20-one 46 was
purchased from Steraloids/Makaira.
Preparative methods and spectroscopic data for compounds 1−4,
8−11, 14a, 14b, 16, 17a, 19, 24, and 27−47 and 1H NMR spectra of
target compounds are included in the Supporting Information.
2-[(3β,5α)-17-Oxoandrostan-3-yl, 17-cyclic 1,2-ethanediyl
acetal]-1H-isoindole-1,3(2H)-dione 5. (3α,5α)-3-Hydroxyandro-
4439
dx.doi.org/10.1021/jm300317k | J. Med. Chem. 2012, 55, 4431−4445

Journal of Medicinal Chemistry
Article
acetic acid (3 mL) under nitrogen was added 10% Pd/C, and the
mixture was stirred vigorously under hydrogen for 3 h. The catalyst
was removed by filtration through Celite and the solvent evaporated.
The residue was purified by flash chromatography (ethyl acetate
elution) and then by preparative HPLC to give the product 7c (13.6
mg, 22%). 1H NMR (CDCl3): δ 0.81 (s, 3H), 0.85 (s, 3H), 2.10 (t, J =
9.0 Hz, 1H), 2.44 (dd, J = 19.4 and 9.0 Hz, 1H), 3.48 (m, 1H), 4.29 (s,
2H), 4.37 (d, J = 9.2 Hz, 1H). 13C NMR (CDCl3): δ 12.3, 13.8, 20.4,
21.8, 28.3, 29.3, 30.8, 31.5, 35.1, 35.6, 35.86, 35.93, 37.4, 45.4, 47.8,
51.4, 54.4, 157.9, 221.3. LC−MS m/z 333.1 [M + 1]+, 100% purity;
m/z 331.1 [M − 1]−, 100% purity.
hydrophobic frit. The DCM was evaporated, and the resulting 17-ketal
intermediate was deprotected and purified in a similar manner to
compound 7e. Evaporation from ethyl acetate/isohexane afforded the
product 7g as a white crystalline solid (81 mg, 75%). 1H NMR
(DMSO-d6): δ 0.78 (s, 3H), 0.80 (s, 3H), 2.32−2.43 (m, 1H), 2.75
(s, 6H), 3.30−3.47 (m, 1H), 5.88 (br d, J = 8.2 Hz, 1H). 13C NMR
(DMSO-d6): δ 11.9, 13.4, 20.0, 21.3, 28.1, 28.6, 30.4, 31.3, 34.5, 35.3
(×2), 35.4, 35.8 (×2), 37.3, 45.2, 47.1, 49.3, 50.6, 53.7, 157.5, 219.8.
LC−MS m/z 361.5 [M + 1]+, 100% purity.
[(3β,5α)-17-Oxoandrostan-3-yl]carbamic Acid, Methyl Ester
7h. To a stirred solution of methyl chloroformate (95 mg, 1.0 mmol)
and (3β,5α)-3-aminoandrostan-17-one, cyclic 1,2-ethanediyl acetal 6
(350 mg, 1.05 mmol) in DCM (5 mL) at 20 °C under nitrogen was
added triethylamine (121 mg, 1.2 mmol); the mixture became very
warm. Stirring was continued overnight under nitrogen. Then DCM
and water were added and the mixture was passed through a
hydrophobic frit. The DCM was evaporated to give a white solid
(432 mg) that was dissolved in THF (8 mL), DCM (1 mL), and
acetone (1 mL). Then 2 M HCl (0.5 mL) was added and the mixture
stirred for 3 days. DCM and water were added, and the mixture was
passed through a hydrophobic frit. The DCM was evaporated to give a
white solid (312 mg) that was chromatographed on silica (25 g SNAP
cartridge, DCM/ethyl acetate gradient elution). Evaporation of the
solvent afforded the product 7h as a white crystalline solid (293 mg,
N-Methyl-N′-[(3β,5α)-17-oxoandrostan-3-yl]urea 7d. To a
stirred solution of (3β,5α)-3-aminoandrostan-17-one, cyclic 1,2-ethane-
diyl acetal 6 (100 mg, 0.30 mmol) in DCM (1 mL) was added
(4-nitrophenyl)-N-methylcarbamate (59 mg, 0.30 mmol), and the
mixture was stirred at ambient temperature for 2 h. Triethylamine
(0.04 mL, 0.30 mmol) was added and the mixture warmed to 40 °C for
6 h and then stirred at ambient temperature overnight. The reaction
mixture was chromatographed on silica (25 g SNAP cartridge, ethyl
acetate elution). Evaporation of the solvent afforded the 17-ketal
intermediate as a white solid (72 mg). To a stirred solution of this
intermediate (72 mg, 0.18 mmol) in acetone (4 mL) was added 2 M HCl
(1 mL), and the mixture was stirred at ambient temperature for 2 h. The
mixture was evaporated, acetone (10 mL) added, and the solution
evaporated. Repeating this process caused precipitation of the product 7d
1
80%). H NMR (DMSO-d6): δ 0.77 (s, 6H), 2.32−2.43 (m, 1H),
3.17−3.33 (m, 1H), 3.49 (s, 3H), 6.98 (d, J = 8.1 Hz, 1H). 13C NMR
(DMSO-d6): δ 11.9, 13.4, 19.9, 21.3, 27.9, 28.2, 30.4, 31.3, 34.5, 34.9,
35.2, 35.3, 36.9, 44.9, 47.0, 49.8, 50.6, 50.9, 53.6, 155.7, 219.7. LC−MS
m/z 348.5 [M + 1]+, 100% purity. HRMS (ESI) m/z [M + H]+ calcd
for C21H34NO3: 348.2534. Found: 348.2539.
1
as a crystalline white solid (55.5 mg, 53%). H NMR (CDCl3): δ 0.84
(s, 3H), 0.86 (s, 3H), 2.05 (dd, J = 19.0 and 8.7 Hz, 1H), 2.45 (dd, J =
19.0 and 8.7 Hz, 1H), 2.90 (s, 3H), 3.51 (m, 1H). 13C NMR (CDCl3): δ
11.2, 12.7, 19.3, 20.6, 26.5, 27.1, 27.8, 29.7, 30.4, 33.9, 34.3, 34.5, 34.5,
36.1, 44.3, 49.8, 50.3, 53.2, 158.1, 220.1. LC−MS m/z 347.2 [M + 1]+,
100% purity; m/z 345.2 [M − 1]−, 100% purity.
N-[(3β,5α)-17-Oxoandrostan-3-yl]methanesulfonamide 7i.
To a stirred solution of (3β,5α)-3-aminoandrostan-17-one, cyclic
1,2-ethanediyl acetal 6 (200 mg, 0.60 mmol) and triethylamine
(0.17 mL, 1.2 mmol) in DCM (5 mL) at 20 °C under nitrogen was
added methanesulfonyl chloride (0.05 mL, 0.60 mmol). The resulting
mixture was stirred at ambient temperature for 24 h.The reaction
mixture was treated with water (2 mL), MeOH (3 mL), and MeSO3H
(0.4 mL) and stirred vigorously for 15 min. The reaction mixture was
basified with saturated NaHCO3 (10 mL), extracted with DCM, and
passed through a hydrophobic frit. Evaporation of the solvent afforded
a residue that was chromatographed on silica (isohexane/ethyl acetate
gradient elution) and recrystallized from ethyl acetate to give the
product 7i as a white crystalline solid (101 mg, 46%). 1H NMR
(CDCl3): δ 0.84 (s, 3H), 0.88 (s, 3H), 2.40−2.52 (m, 1H), 3.00 (s,
3H), 3.22−3.40 (m, 1H), 4.09 (d, J = 7.5 Hz, 1H). 1H NMR (DMSO-
d6): δ 0.77 (s, 3H), 0.78 (s, 3H), 2.32−2.43 (m, 1H), 2.89 (s, 3H),
3.00−3.16 (m, 1H), 6.95 (d, J = 7.5 Hz, 1H). 13C NMR (DMSO-d6):
δ 11.8, 13.4, 19.9, 21.3, 27.9, 29.4, 30.4, 31.3, 34.4, 35.0, 35.3,
35.9, 36.9, 41.2, 44.9, 47.0, 50.6, 52.3, 53.6, 219.7. LC−MS m/z 366.1
[M − 1]−, 100% purity.
N-Ethyl-N′-[(3β,5α)-17-oxoandrostan-3-yl]urea 7e. To a
stirred solution of (3β,5α)-3-aminoandrostan-17-one, cyclic 1,2-
ethanediyl acetal 6 (350 mg, 1.05 mmol) in DCM (5 mL) at 20 °C
under nitrogen was added ethyl isocyanate (112 mg, 1.57 mmol). The
mixture was stirred overnight, and then DCM and water were added
before the mixture was passed through a hydrophobic frit. The DCM
phase was concentrated to give the intermediate ketal as a white solid
(472 mg). The 17-ketal intermediate (472 mg) was deprotected in a
similar manner to compound 7a and the crude product chromato-
graphed on silica (25 g SNAP, DCM/ethyl acetate gradient elution),
1
affording the product 7e as a white solid (293 mg, 77%). H NMR
(CDCl3): δ 0.82 (s, 3H), 0.86 (s, 3H), 1.14 (t, J = 7.3 Hz, 3H), 2.38−
2.50 (m, 1H), 3.20 (q, J = 7.2 Hz, 2H), 3.46−3.59 (m, 1H), 4.35 (br s,
1
1H). H NMR (DMSO-d6): δ 0.78 (s, 6H), 0.96 (t, J = 7.2 Hz, 3H),
2.32−2.44 (m, 1H), 2.92−3.03 (m, 2H), 3.22−3.37 (m, 1H), 5.58−
5.65 (m, 1H). 13C NMR (CDCl3): δ 12.3, 13.8, 15.5, 20.4, 21.8, 28.3,
29.6, 30.8, 31.5, 35.1, 35.3, 35.6, 35.9, 36.2, 37.5, 45.5, 47.8, 49.7, 51.4,
54.4, 157.6, 221.4. LC−MS m/z 361.2 [M + 1]+, 100% purity; m/z
359.2 [M − 1]−, 100% purity. HRMS (ESI) m/z [M + H]+ m/z calcd
for C22H37N2O2: 361.2850. Found: 361.2851.
N-Butyl-N′-[(3β,5α)-17-oxoandrostan-3-yl]urea 7f. 7f was
prepared and purified in a similar manner to compound 7e, using
(3β,5α)-3-aminoandrostan-17-one, cyclic 1,2-ethanediyl acetal 6
(100 mg, 0.30 mmol) and N-butyl isocyanate (44.5 mg, 0.45 mmol).
Evaporation from ethyl acetate/isohexane afforded the product 7f as a
white crystalline solid (67 mg, 58%). 1H NMR (DMSO-d6): δ 0.77 (s,
6H), 0.86 (t, J = 6.0 Hz, 3H), 2.32−2.44 (m, 1H), 2.90−2.99 (br q, J =
6.2 Hz, 2H), 3.22−3.36 (m, 1H), 5.56−5.67 (m, 2H). 13C NMR
(DMSO-d6): δ 11.3, 12.8, 13.1, 18.9, 19.3, 20.7, 27.4, 28.4, 29.8, 30.7,
31.6, 33.9, 34.60, 34.64, 35.2, 36.5, 38.2, 44.5, 46.4, 47.9, 50.0, 53.2,
156.7, 219.1. LC−MS m/z 389.2 [M + 1]+, 100% purity; m/z 387.2
[M − 1]−, 100% purity.
N-[(3β,5α)-17-Oxoandrostan-3-yl]ethanesulfonamide 7j.
(3β,5α)-3-Aminoandrostan-17-one, cyclic 1,2-ethanediyl acetal 6
(100 mg, 0.30 mmol) was dissolved in DCM (5 mL), and triethylamine
(0.14 mL, 1 mmol) was added. The mixture was cooled in an ice bath,
and ethanesulfonyl chloride (0.04 mL, 0.45 mmol) was added. The
mixture was stirred for 3 days under a nitrogen atmosphere. THF
(5 mL), acetone (2 mL), and 2 M HCl (0.5 mL) were added, and
stirring continued overnight. DCM and water were added, and the
mixture was passed through a hydrophobic frit. Evaporation of the
solvent afforded a residue that was chromatographed on silica (10 g
SNAP cartridge, isohexane/ethyl acetate gradient elution). Evaporation
of the solvent afforded the product 7j as a white crystalline solid
(86 mg, 75%). 1H NMR (CDCl3): δ 0.83 (s, 3H), 0.87 (s, 3H), 1.38 (t,
J = 7.4 Hz, 3H), 2.39−2.51 (m, 1H), 3.05 (q, J = 7.4 Hz, 2H), 3.20−
N,N-Dimethyl-N′-[(3β,5α)-17-oxoandrostan-3-yl]urea 7g.
(3β,5α)-3-Aminoandrostan-17-one, cyclic 1,2-ethanediyl acetal 6
(100 mg, 0.30 mmol) was dissolved in DCM (5 mL), and
triethylamine (0.14 mL, 1 mmol) was added. The mixture was cooled
in an ice bath, and dimethylcarbamyl chloride (0.03 mL, 0.33 mmol)
was added. The mixture was stirred for 3 days. DCM and 2 M HCl
were added, and the mixture was stirred before passing through a
1
3.35 (m, 1H), 4.11 (d, J = 7.9 Hz, 1H). H NMR (DMSO-d6): δ 0.76
(s, 3H), 0.77 (s, 3H), 1.18 (t, J = 7.3 Hz, 3H), 2.32−2.44 (m, 1H), 2.96
(q, J = 7.3 Hz, 2H), 2.90−3.10 (m, 1H), 6.98 (d, J = 7.9 Hz, 1H). 13C
NMR (CDCl3): δ 7.4, 11.1, 12.7, 19.3, 20.6, 27.1, 29.3, 29.6, 30.4, 33.9,
34.3, 34.7, 35.9, 36.3, 44.4, 46.6, 47.2, 50.2, 52.2, 53.2, 220.1. LC−MS
m/z 380.0 [M − 1]−, 100% purity.
4440
dx.doi.org/10.1021/jm300317k | J. Med. Chem. 2012, 55, 4431−4445

Journal of Medicinal Chemistry
Article
N-[(3β,5α)-17-Oxoandrostan-3-yl]benzenesulfonamide 7k.
To a stirred solution of (3β,5α)-3-aminoandrostan-17-one, cyclic
1,2-ethanediyl acetal 6 (200 mg, 0.60 mmol) and triethylamine
(121 mg, 1.20 mmol) in DCM (5 mL) at 20 °C under nitrogen was
added benzenesulfonyl chloride (111 mg, 0.63 mmol). The mixture
was stirred overnight. Water (4 mL) was added. Stirring was continued
for 15 min and the mixture filtered through a hydrophobic frit. The
organic phase was evaporated to dryness. The residue was dissolved in
THF (5 mL) and acetone (1 mL), and 1 N HCl solution (1 mL) was
added. This mixture was stirred overnight. Water (4 mL) and DCM
(10 mL) were added, and the mixture was filtered through a
hydrophobic frit. Evaporation of the solvent afforded the product 7k as
a white foam (150 mg, 58%). 1H NMR (CDCl3): δ 0.76 (s, 3H), 0.84
(s, 3H), 2.37−2.50 (m, 1H), 3.05−3.20 (m, 1H), 4.68 (br s, 1H),
7.48−7.61 (m, 3H), 7.87−7.94 (m, 2H). 13C NMR (CDCl3): δ 11.0,
12.7, 19.2, 20.6, 27.0, 28.6, 29.6, 30.4, 33.8, 34.3, 34.7, 35.2, 36.2, 44.4,
46.6, 50.2, 52.2, 53.2, 125.8 (×2), 127.9 (×2), 131.3, 140.3, 220.1.
LC−MS m/z 428.1 [M − 1]−, 100% purity (combined).
0.30 mmol). Evaporation from ethyl acetate and then purification by
preparative HPLC afforded the product 7f as a white solid (13.8 mg,
14%). H NMR (CDCl3): δ 0.60 (s, 3H), 0.78 (s, 3H), 1.57 (s, 3H),
1
2.52 (t, J = 8.8 Hz, 1H), 3.49 (m, 1H), 4.22 (s, 2H), 4.27 (d, J =
8.2 Hz, 1H). 13C NMR (CDCl3): δ 12.3, 13.5, 21.1, 22.8, 24.4, 28.4,
29.4, 31.6, 31.9, 35.5, 35.9, 37.4, 39.0, 44.2, 45.4, 50.3, 54.2, 56.6, 63.8,
84.8, 157.8, 209.7. LC−MS m/z 361.2 [M + 1]+, 100% purity; m/z
359.0 [M − 1]−, 100% purity.
N,N-Dimethyl-N′-[(3β,5α)-20-oxopregnan-3-yl]urea 14d. 14d
was prepared and purified in a similar manner to compound 14b, using
(3β,5α)-3-aminopregnan-20-one, 20-cyclic 1,2-ethanediyl acetal 13
(100 mg, 0.28 mmol) and dimethylcarbamyl chloride (0.04 mL,
0.40 mmol). Crystallization from diethyl ether gave product 14d as a
white solid (93 mg, 85%). 1H NMR (CDCl3): δ 0.60 (s, 3H), 0.80 (s,
3H), 2.12 (s, 3H), 2.50−2.58 (m, 1H), 2.89 (s, 6H), 3.56−3.69 (m,
1H). 13C NMR (CDCl3): δ 12.3, 13.5, 21.1, 22.8, 24.4, 28.4, 29.7, 31.6,
31.9, 34.8, 35.5 (×2), 36.2, 36.3, 37.6, 39.0, 44.3, 45.4, 50.2, 54.1, 56.6,
63.8, 157.8, 209.7. LC−MS m/z 389.2 [M + 1]+, 100% purity; m/z
387.2 [M − 1]−, 100% purity.
N-[(3β,5α)-17-Oxoandrostan-3-yl]sulfamide 7l. A stirred solu-
tion of sulfamide (481 mg, 5.01 mmol) and (3β,5α)-3-amino-
androstan-17-one, cyclic 1,2-ethanediyl acetal 6 (167 mg, 0.50 mmol)
in 1,4-dioxane (5 mL) was heated at reflux for 48 h. The reaction
mixture was diluted with water and extracted with ethyl acetate. The
combined extracts were washed with saturated NaHCO3 solution and
water. The organic phase was dried and evaporated to leave a residue
that was treated with water (2 mL), MeOH (3 mL), and MeSO3H
(0.4 mL). The mixture was stirred for 15 min and then basified with
saturated NaHCO3 solution (∼10 mL) and passed through a
hydrophobic frit. The aqueous phase was washed with DCM. The
organic phase was evaporated affording a residue (220 mg) that was
chromatographed on silica (isohexane/ethyl acetate gradient elution)
[(3β,5α)-20-Oxopregnan-3-yl]carbamic Acid, Methyl Ester
14e. 14e was prepared and purified in a similar manner to compound
7h, using (3β,5α)-3-aminopregnan-20-one, 20-cyclic 1,2-ethanediyl
acetal 13 (350 mg, 0.97 mmol) and methyl chloroformate (0.11 mL,
1.45 mmol). Evaporation of the solvent afforded the product 14e as a
1
white crystalline solid (182 mg, 50%). H NMR (DMSO-d6): δ 0.51
(s, 3H), 0.74 (s, 3H), 2.05 (s, 3H), 2.56 (m, 1H), 3.17−3.33 (m, 1H),
3.49 (s, 3H), 6.98 (d, J = 8.0 Hz, 1H). 13C NMR (DMSO-d6): δ 13.9,
15.1, 22.6, 24.1, 25.9, 28.4, 30.1, 30.2, 33.1, 33.5, 36.9, 37.0, 38.9, 40.1,
45.5, 46.9, 51.8, 52.9, 55.4, 57.8, 64.6, 157.8, 210.5. LC−MS m/z 376.5
[M + 1]+, 100% purity.
N-[(3β,5α)-20-Oxopregnan-3-yl]methanesulfonamide 14f.
14f was prepared and purified in a similar manner to compound 7h,
using (3β,5α)-3-aminopregnan-20-one, 20-cyclic 1,2-ethanediyl acetal
13 (390 mg, 1.08 mmol) and methanesulfonyl chloride (0.13 mL,
1.62 mmol). Evaporation of the solvent afforded the product 14f as a
white solid (255 mg, 60%). 1H NMR (CDCl3): δ 0.62 (s, 3H), 0.81 (s,
3H), 2.13 (s, 3H), 2.50−2.58 (m, 1H), 3.00 (s, 3H), 3.25−3.38 (m,
1H). 13C NMR (CDCl3): δ 12.2, 13.5, 21.2, 22.8, 24.4, 28.4, 30.3, 31.6,
31.8, 35.3, 35.4, 36.7, 37.4, 39.0, 42.3, 44.2, 45.5, 53.5, 54.1, 56.6, 63.8,
209.7. LC−MS m/z 394.1 [M − 1]−, 100% purity. HRMS (ESI) m/z
[M + Cl]− calcd for C22H37ClNO3S: 430.2188. Found: 430.2201.
N-[(3β,5α)-20-Oxopregnan-3-yl]ethanesulfonamide 14g. 14g
was prepared and purified in a similar manner to compound 7h, using
(3β,5α)-3-aminopregnan-20-one, 20-cyclic 1,2-ethanediyl acetal 13
(100 mg, 0.28 mmol) and ethanesulfonyl chloride (0.04 mL, 0.41
mmol). Evaporation of the solvent afforded the product 14g as a white
1
to give the product 7l as a white solid (115 mg, 62%). H NMR
(CDCl3): δ 0.84 (s, 3H), 0.88 (s, 3H), 2.40−2.52 (m, 1H), 3.26−3.39
1
(m, 1H), 4.48 (br s, 2H). H NMR (DMSO-d6): δ 0.76 (s, 3H), 0.77
(s, 3H), 2.31−2.44 (m, 1H), 2.95−3.10 (m, 1H), 6.40 (br s, 2H), 6.42
(br s, 1H). 13C NMR (DMSO-d6): δ 11.9, 13.4, 20.0, 21.3, 28.0, 28.9,
30.4, 31.3, 34.5, 35.1, 35.3, 35.5, 37.1, 45.2, 47.1, 50.6, 52.2, 53.7,
219.7. LC−MS m/z 367.1 [M − 1]−, 100% purity. HRMS (ESI) m/z
[M − H]− calcd for C19H31N2O3S: 367.2060. Found: 367.2050.
2-[(3β,5α)-20-Oxopregnan-3-yl, 20-cyclic 1,2-ethanediyl ace-
tal]-1H-isoindole-1,3(2H)-dione 12. To a stirred solution of
(3α,5α)-3-hydroxypregnan-20-one, cyclic 1,2-ethanediyl acetal 11
(2.03 g, 5.6 mmol) and TPP (1.63 g, 5.88 mmol) in THF (45 mL)
at 20 °C under nitrogen was added phthalimide (0.87 g, 5.88 mmol).
The resulting suspension was cooled to 5 °C and stirred for 5 min.
A solution of DIAD (1.24 mL, 6.27 mmol) in THF (50 mL) was then
slowly added to the reaction mixture, maintaining the temperature
below 9 °C. The mixture was stirred at ambient temperature for 16 h.
Methanol (100 mL) was added to the reaction mixture, and the
resulting slurry was stirred for 30 min. The solid was filtered off and
washed with methanol to give the product 12 as a white solid (1.93 g,
1
solid (72 mg, 63%). H NMR (CDCl3): δ 0.61 (s, 3H), 0.80 (s, 3H),
1.39 and 1.41 (2 × d, J = 7.4 and 7.2 Hz, 3H), 2.50−2.58 (m, 1H),
3.05 and 3.16 (2 × q, J = 7.4 and 7.2 Hz, 2H), 3.21−3.34 (m, 1H). 13C
NMR (CDCl3): δ (8.5, 8.7), 12.2, 13.5, 21.1, 22.8, 24.4, 28.4, 30.5,
31.6, 31.8, 35.3, 35.4, 37.0, 37.5, 39.0, (44.2, 45.5), 46.1, 48.4, 53.4,
54.1, 56.6, 63.8, 209.7. LC−MS no peak.
N-[(3β,5α)-20-Oxopregnan-3-yl]benzenesulfonamide 14h.
14h was prepared and purified in a similar manner to compound
7h, using (3β,5α)-3-aminopregnan-20-one, 20-cyclic 1,2-ethanediyl
acetal 13 (100 mg, 0.28 mmol) and benzenesulfonyl chloride (51 mg,
0.29 mmol). Evaporation of the solvent afforded the product 14g as a
white solid (110 mg, 86%). 1H NMR (CDCl3): δ 0.51 (s, 3H), 0.66 (s,
3H), 2.03 (s, 3H), 2.38−2.48 (m, 1H), 2.99−3.14 (m, 1H), 4.35 (d,
J = 7.3 Hz, 1H), 7.40−7.54 (m, 3H), 7.79−7.84 (m, 2H). 13C NMR
(CDCl3): δ 12.1, 13.4, 21.1, 22.8, 24.4, 28.3, 29.9, 31.5, 31.8, 35.3,
35.4, 36.4, 37.4, 39.0, 44.2, 45.5, 53.4, 54.0, 56.6, 63.8, 126.9 (×2),
129.0 (×2), 132.4, 141.4, 209.6. LC−MS m/z 456.2 [M − 1]−, 100%
purity.
1
70%). H NMR (CDCl3): δ 0.79 (s, 3H), 0.98 (s, 3H), 1.32 (s, 3H),
2.22−2.49 (m, 2H), 3.87−4.06 (m, 4H), 4.13−4.24 (m, 1H), 7.69−
7.73 (m, 2H), 7.79−7.84 (m, 2H).
(3β,5α)-3-Aminopregnan-20-one, Cyclic 1,2-Ethanediyl Ace-
tal 13. 2-[(3β,5α)-20-Oxopregnan-3-yl, 20-cyclic 1,2-ethanediyl
acetal]-1H-isoindole-1,3(2H)-dione 12 (3.61 g, 7.35 mmol) was
treated with ethanol (60 mL) and hydrazine hydrate (5 mL, 103 mmol).
The mixture was heated at reflux overnight under a nitrogen
atmosphere. The mixture was allowed to cool, and the solid was
filtered off. The solid was washed with chloroform and the combined
filtrates were evaporated to give the product 13 as a white solid (2.79 g,
1
100%). H NMR (CDCl3): δ 0.76 (s, 3H), 0.80 (s, 3H), 1.30 (s, 3H),
1.98−2.07 (m, 1H), 2.61−2.74 (m, 1H), 3.83−4.05 (m, 4H). 13C NMR
(CDCl3): δ 12.4, 13.1, 21.0, 22.9, 23.8, 24.6, 28.8, 32.0, 32.4, 35.0, 35.6,
37.6, 39.1, 39.7, 42.0, 45.6, 51.2, 54.5, 56.4, 58.4, 63.2, 65.2, 112.0.
N-[(3β,5α)-20-Oxopregnan-3-yl]urea 14c. 14c was prepared
and purified in a similar manner to compound 7c, using (3β,5α)-3-
aminopregnan-20-one, 20-cyclic 1,2-ethanediyl acetal 13 (100 mg,
0.28 mmol) and (4-nitrophenyl)-N-benzylcarbamate (82.8 mg,
N-[(3β,5α)-20-Oxopregnan-3-yl]sulfamide 14i. 14i was pre-
pared in a similar manner to compound 7l, using (3β,5α)-3-
aminopregnan-20-one, 20-cyclic 1,2-ethanediyl acetal 13 (100 mg,
0.28 mmol) and sulfamide (242 mg, 2.52 mmol). Trituration of
the crude product (111 mg) with diethyl ether gave the product 14i
1
as a white solid (43 mg, 39%). H NMR (CDCl3): δ 0.62 (s, 3H),
4441
dx.doi.org/10.1021/jm300317k | J. Med. Chem. 2012, 55, 4431−4445

Journal of Medicinal Chemistry
Article
0.81 (s, 3H), 2.13 (s, 3H), 2.50−2.58 (m, 1H), 3.25−3.38 (m, 1H),
4.54 (br s, 2H). 13C NMR (CDCl3): δ 12.2, 13.5, 21.1, 22.8, 24.4, 28.4,
29.7, 31.6, 31.9, 35.3, 35.4, 36.2, 37.5, 39.0, 44.2, 45.5, 53.9, 54.1, 56.6,
63.8, 209.7. LC−MS m/z 95.1 [M − 1]−, 100% purity. HRMS (ESI)
m/z [M − H]− calcd for C21H35N2O3S: 395.2373. Found: 395.2373.
N,N-Dimethyl-N′-[(3β,5α)-20-oxopregnan-3-yl]sulfamide
14j. To a stirred solution of (3β,5α)-3-aminopregnan-20-one,
20-cyclic 1,2-ethanediyl acetal 13 (100 mg, 0.266 mmol) and
dimethylsulfamoyl chloride (42 mg, 0.29 mmol) in DCM (5 mL) at
20 °C under nitrogen was added N,N-diisopropylethylamine (69 mg,
0.53 mmol), and the resulting mixture was stirred overnight. The
reaction mixture was evaporated to dryness and the residue dissolved
in MeOH/water (11 mL, 10/1). A few drops of MeSO3H were added
until the solution was acidic. After 10 min the reaction mixture was
diluted with water (50 mL), basified with aqueous NaHCO3, and
extracted with DCM. Evaporation of the solvent afforded the product
14j as a white solid (18 mg, 15%). 1H NMR (CDCl3): δ 0.53 (s, 3H),
0.71 (s, 3H), 2.04 (s, 3H), 2.41−2.49 (m, 1H), 2.72 (s, 6H), 3.05−
3.18 (m, 1H), 3.77 (br s, 1H). 13C NMR (CDCl3): δ 12.2, 13.5, 21.1,
22.8, 24.4, 28.4, 30.1, 31.5, 31.9, 35.3, 35.4, 36.6, 37.5, 38.1 (×2), 39.0,
44.2, 45.5, 53.8, 54.1, 56.6, 63.8, 209.7. LC−MS m/z 425.1 [M + 1]+,
100% purity; m/z 423.1 [M − 1]−, 100% purity.
21.3, 26.2, 27.8, 30.5, 31.4, 31.9, 33.1, 34.4, 35.2, 35.4, 38.7, 47.1, 48.2,
50.8, 53.6, 219.8. LC−MS m/z 367.1 [M − 1]−, 100% purity.
N-Ethyl-N′-[(5α)-21-hydroxy-20-oxopregnan-3-yl]urea 18a.
N-Ethyl-N′-[(5α)-20-oxopregnan-3-yl]urea was prepared from (5α)-
3-aminopregnan-20-one, 20-cyclic 1,2-ethanediyl acetal 13 in a similar
manner to urea 7e. To a stirred solution of N-ethyl-N′-[(5α)-20-
oxopregnan-3-yl]urea (100 mg, 0.26 mmol) and triethylamine
(1.07 mL, 7.72 mmol) in DCM (1 mL) at 20 °C under nitrogen
was added tert-butyldimethylsilyl trifluoromethanesulfonate (0.59 mL,
2.57 mmol), and the resulting mixture was stirred at ambient
temperature overnight. The mixture was diluted with DCM (10 mL)
and washed with saturated NaHCO3 solution (2 × 10 mL).
Evaporation afforded the silyl enol ether intermediate which was
redissolved in DCM (5 mL). A suspension of m-CPBA (95 mg,
0.41 mmol) in isohexane (10 mL) was cooled to −10 °C, and a
solution of the silyl enol ether in DCM was added slowly over 5 min.
The mixture was stirred at −10 °C for 15 min and then stirred at
ambient temperature for 60 min. The solvent was evaporated and the
residue chromatographed on silica (10 g SNAP cartridge, DCM/
MeOH gradient elution). Evaporation of the solvent from different
fractions afforded the TBDMS ether (80 mg) and the product 18a as a
1
white foam (20 mg, 19%). H NMR (DMSO-d6): δ 0.50 (s, 3H, 3α
2-[(3α,5α)-17-Oxoandrostan-3-yl, 17-cyclic 1,2-ethanediyl
acetal]-1H-isoindole-1,3(2H)-dione 15. (3β,5α)-3-Hydroxyandro-
stan-17-one, cyclic 1,2-ethanediyl acetal (5.83 g, 17.4 mmol),
phthalimide (2.56 g, 17.4 mmol), and TPP (5.06 g, 18.3 mmol) were
dissolved in THF (140 mL) and cooled to 5 °C using an ice bath. DIAD
(3.8 mL, 19.2 mmol) was added slowly, maintaining the temperature
below 7 °C. The yellow color was allowed to disappear between
additions. The mixture was allowed to warm to ambient temperature
overnight under a nitrogen atmosphere. The solvent was evaporated and
methanol was added to the oil, forming a white precipitate. The
suspension was stirred for 30 min. The precipitate was filtered off and
washed with methanol (100 mL), affording the product 15 as a white
and 3β), 0.72 (s, 3H, 3α and 3β), 2.88−3.01 (m, 2H, 3α and 3β),
3.67−3.77 (m, 1H, 3β-H of 3α urea), 4.00 (br d, J = 5.3 Hz, 2H, 3α
and 3β), 4.82−4.90 (m, 1H, 3α and 3β), 5.53−5.64 (m, 2H, 3α and 3β),
5.93−6.03 (m, 1H, N−H of 3α urea). LC−MS m/z 405.1 [M + 1]+,
100% purity; m/z 403.1 [M − 1]−, 82% purity.
[(3β,5α)-21-Hydroxy-20-oxopregnan-3-yl]carbamic Acid,
Methyl Ester 18b. 18b was prepared in a similar manner to
compound 18a, using [(3β,5α)-20-oxopregnan-3-yl]carbamic acid,
methyl ester 14e (270 mg, 0.72 mmol), except the intermediate
TBDMS ether was treated with THF (5 mL) and 2 M HCl (1 mL).
The mixture was stirred overnight. The solvent was evaporated and the
residue partitioned between water and DCM. The mixture was passed
through a hydrophobic frit and the DCM evaporated. The residue was
triturated with diethyl ether to give the product 18b as an off-white
1
solid (5.93 g, 74%). H NMR (CDCl3): δ 0.77 (s, 3H), 0.78 (s, 3H),
3.75−3.86 (m, 4H), 4.41−4.42 (m, 1H), 7.60−7.65 (m, 2H), 7.71−7.77
(m, 2H).
1
solid (33 mg, 12%). H NMR (CDCl3): δ 0.54 (s, 3H), 0.70 (s, 3H),
N-Ethyl-N′-[(3α,5α)-17-oxoandrostan-3-yl]urea 17b. 17b was
prepared and purified in a similar manner to compound 7h, using
(3α,5α)-3-aminoandrostan-17-one, cyclic 1,2-ethanediyl acetal 16
(100 mg, 0.30 mmol) and ethyl isocyanate (0.05 mL, 0.60 mmol).
Evaporation of the solvent afforded the product 17b as a white
2.33−2.41 (m, 1H), 3.30−3.45 (m, 1H), 3.57 (s, 3H), 4.05, 4.13 (ABq,
1
JAB = 18.9 Hz, 2H), 4.36−4.48 (br s, 1H). H NMR (DMSO-d6): δ
0.49 (s, 3H), 0.71 (s, 3H), 2.50−2.58 (m, 1H), 3.13−3.30 (m, 1H),
3.46 (s, 3H), 4.00 (br d, J = 5.9 Hz, 2H), 4.87 (t, J = 5.9 Hz, 2H), 6.95
(br d, J = 7.8 Hz, 1H). 13C NMR (DMSO-d6): δ 11.9, 13.3, 20.6, 22.3,
24.1, 28.1, 28.2, 31.5, 34.9, 35.0, 37.0, 38.0, 43.9, 44.9, 45.5, 50.9, 53.4,
56.0, 57.7, 68.7, 160.3, 210.4. LC−MS m/z 392.3 [M + 1]+, 100%
purity; m/z 390.3 [M − 1]−, 100% purity.
1
crystalline solid (40 mg, 37%). H NMR (DMSO-d6): δ 0.78 (s, 6H),
0.97 (t, J = 7.2 Hz, 3H), 2.32−2.46 (m, 1H), 2.93−3.04 (m, 2H),
3.71−3.79 (m, 1H), 5.62 (br t, J = 5.5 Hz, 1H), 6.01 (d, J = 7.9 Hz,
1H). 13C NMR (DMSO-d6): δ 11.2, 13.4, 15.7, 19.6, 21.3, 26.5, 27.9,
30.6, 31.4, 32.4, 33.4, 33.8, 34.4, 35.2, 35.7, 43.7, 47.1, 50.8, 54.1,
157.2, 219.7. LC−MS m/z 361.1 [M + 1]+, 100% purity; m/z 359.1
[M − 1]−, 100% purity.
N-[(3β,5α)-21-Hydroxy-20-oxopregnan-3-yl]methanesulfon-
amide 18c. To a stirred solution of N-[(3β,5α)-20-oxopregnan-
3-yl]methanesulfonamide 14f (500 mg, 1.26 mmol) in methanol
(30 mL) at ambient temperature was added aqueous HBr (3 drops),
and then a solution of bromine (0.13 mL, 2.53 mmol) in methanol
(10 mL) was added slowly over 10 min. The resulting mixture was
stirred for 2.5 h. (Unlike bromination of the corresponding 3-alcohols
the solution did not decolorize upon the slow addition of bromine.)
The mixture was diluted with water, extracted with ethyl acetate,
and evaporated to leave a residue (640 mg). This residue was
chromatographed on silica (isohexane/acetone gradient elution) and
the crude product crystallized from diethyl ether and then ethyl acetate
to give the 21-bromo intermediate (85 mg) in adequate purity for the
next step. To a stirred solution of the 21-bromo intermediate (80 mg,
0.17 mmol) in DMF (10 mL) and water (5 mL) was added sodium
hydroxide (8 mg, 0.20 mmol), and the mixture was stirred at ambient
temperature for 30 min. The mixture was diluted with water and
extracted with ethyl acetate. The extracts were washed with water, and
the solvent was evaporated. The residue was chromatographed on silica
(isohexane/ethyl acetate gradient elution) and the residue recrystallized
from diethyl ether to give the product 18c as a white crystalline solid
N-[(3α,5α)-17-Oxoandrostan-3-yl]methanesulfonamide 17c.
17c was prepared and purified in a similar manner to compound 7h,
using (3α,5α)-3-aminoandrostan-17-one, cyclic 1,2-ethanediyl acetal
16 (100 mg, 0.30 mmol) and methanesulfonyl chloride (0.03 mL,
0.45 mmol). Evaporation of the solvent afforded the product 17b as a
1
white crystalline solid (81 mg, 73%). H NMR (DMSO-d6): δ 0.77
(s, 3H), 0.78 (s, 3H), 2.32−2.44 (m, 1H), 2.87 (s, 3H), 3.49−3.56
(m, 1H), 6.91 (br d, J = 6.0 Hz, 1H). 13C NMR (DMSO-d6): δ 10.7,
12.8, 19.0, 20.7, 26.1, 27.1, 29.8, 30.7, 31.2, 33.0, 33.8, 34.6, 34.9 (×2),
38.2, 46.4, 48.0, 50.1, 53.1, 219.1. LC−MS m/z 366.1 [M − 1]−, 100%
purity.
N-[(3α,5α)-17-Oxoandrostan-3-yl]sulfamide 17d. 17d was
prepared in a similar manner to compound 7l, using (3α,5α)-3-
aminoandrostan-17-one, cyclic 1,2-ethanediyl acetal 16 (214 mg,
0.64 mmol) and sulfamide (685 mg, 7.13 mmol). The 17-ketal was
deprotected in a similar manner to compound 7a. The residue was
chromatographed on silica (10 g SNAP, DCM/MeOH gradient
elution). Evaporation of the solvent afforded the product 17d as a
white solid (119 mg, 50%). 1H NMR (DMSO-d6): δ 0.77 (s, 3H), 0.78
(s, 3H), 2.32−2.44 (m, 1H), 3.42−3.51 (m, 1H), 6.36 (br d, J = 5.9
Hz, 1H), 6.40 (br s, 2H). 13C NMR (DMSO-d6): δ 11.4, 13.4, 19.6,
1
(41 mg, 8%). H NMR (DMSO-d6): δ 0.52 (s, 3H), 0.74 (s, 3H),
2.52−2.61 (m, 1H), 2.88 (s, 3H), 3.00−3.15 (m, 1H), 3.97−4.07 (m,
2H), 4.89 (t, J = 5.7 Hz, 1H), 6.94 (d, J = 7.6 Hz, 1H). 13C NMR
(DMSO-d6): δ 11.8, 13.3, 20.6, 22.3, 24.1, 28.1, 29.4, 31.5, 34.8, 34.9,
4442
dx.doi.org/10.1021/jm300317k | J. Med. Chem. 2012, 55, 4431−4445

Journal of Medicinal Chemistry
Article
35.9, 37.0, 38.0, 41.2, 43.9, 44.9, 52.3, 53.4, 56.0, 57.7, 68.7, 210.4.
LC−MS m/z 410.1 [M − 1]−, 88% purity. HRMS (ESI) m/z
[M − H]− calcd for C22H36NO4S: 410.2370. Found: 410.2387.
N-[(3β,5α)-21-Hydroxy-20-oxopregnan-3-yl]sulfamide 18d.
18d was prepared in a similar manner to compound 18c, using
N-[(3β,5α)-20-oxopregnan-3-yl]sulfamide 14i (120 mg, 0.30 mmol).
The residue was triturated with diethyl ether to give the product 18d
as a white solid (55 mg, 44%). 1H NMR (CDCl3): δ 0.56 (s, 3H), 0.72
(s, 3H), 2.34−2.42 (m, 1H), 3.23 (m, 1H), 4.07, 4.15 (ABq, JAB = 19.0
Hz, 2H), 4.38 (br s, 2H). 13C NMR (DMSO-d6): δ 11.3, 12.7, 20.0,
21.7, 23.5, 27.6, 28.3, 31.0, 34.3 (×2), 34.9, 36.6, 37.4, 43.3, 44.5, 51.6,
52.9, 55.3, 57.1, 68.1, 209.8. LC−MS m/z 411.1 [M − 1]−, 95% purity.
HRMS (ESI) m/z [M − H]− calcd for C21H35N2O4S: 411.2323.
Found: 411.2327.
1-[(3β,5α)-3-Hydroxyandrostan-17-yl]-2-propanone 23. To a
stirred solution of (3β,5α,17E)-3-[[(1,1-dimethylethyl)dimethylsilyl]-
oxy]pregn-17(20)-en-21-oic acid, ethyl ester 19 (1. g, 2.11 mmol) in
ethyl acetate (30 mL) was added Pd/C 10% (85 mg) as a suspension
in ethanol (2 mL) under nitrogen. The mixture was then hydro-
genated at ambient temperature for 16 h. The mixture was filtered
through Celite, washed with ethyl acetate and the combined filtrates
were evaporated to give (3β,5α)-3-[[(1,1-dimethylethyl)dimethylsilyl]-
oxy]pregnan-21-oic acid, ethyl ester, 20 (920 mg, 1.93 mmol, 92%).
1H NMR (CDCl3): δ 0.00 (s, 6H), 0.53 (s, 3H), 0.75 (s, 3H), 0.83
(s, 9H), 1.13 (t, J = 7.1 Hz, 3H), 1.99−2.10 (m, 1H), 2.25−2.34 (m,
1H), 3.43−3.56 (m, 1H), 4.06 (q, J = 7.1 Hz, 2H).
(3β,5α,17α)-17,21-Epoxypregnan-3-ol 25. To a stirred suspen-
sion of trimethylsulfoxonium iodide (439 mg, 1.99 mmol) in tert-
butanol (5 mL) at 50 °C was added a solution of potasium tert-
butoxide (224 mg, 1.99 mmol) in tert-butanol (5 mL). After 30 min
a solution of (3β,5α,17β)-spiro[androstane-17,27′-oxiran]-3-ol 24
(304 mg, 0.99 mmol) in tert-butanol (5 mL) was added and the mix-
ture heated at reflux for 72 h. The reaction mixture was concentrated
to low volume and diluted with DCM. This solution was washed with
water and the solvent evaporated. The residue (324 mg) was flash
chromatographed on silica (isohexane/ethyl acetate gradient elution)
and the residue (133 mg) recrystallized from diethyl ether to give
product 25 as a white crystalline solid (41 mg, 6.5%). 1H NMR
(CDCl3): δ 0.76 (s, 3H), 0.83 (s, 3H), 2.72−2.85 (m 1H), 3.54−3.67
(m 1H), 4.27−4.36 (m 1H), 4.39−4.48 (m 1H). 13C NMR (CDCl3):
δ 12.0, 12.4, 18.6, 20.8, 23.1, 28.6, 30.9, 31.5, 32.2, 35.6, 36.1, 37.1,
37.8, 38.2, 44.6, 44.9, 48.3, 54.2, 64.9, 71.3, 96.7. LC−MS no peak.
(3β,5α)-3-Hydroxy-17-methoxypregnan-20-one 26. To a
stirred solution of (3β,5α)-3-hydroxypregnan-20-one 9 (318 mg,
0.99 mmol) in methanol (30 mL) was added aqueous copper(II)
bromide (669 mg, 3.0 mmol), and the resulting mixture was stirred at
reflux for 24 h.34 The mixture was diluted with water and extracted
with ethyl acetate. The extracts were evaporated and the residue
(450 mg) was chromatographed on silica (isohexane/ethyl acetate
gradient elution). Recrystallization of the residue from diethyl ether
1
gave product 26 as a white crystalline solid (83 mg, 24%). H NMR
(CDCl3): δ 0.58 (s, 3H), 0.82 (s, 3H), 2.15 (s, 3H), 2.37−2.49 (m,
1H), 3.15 (s, 3H), 3.54−3.68 (m, 1H). 13C NMR (CDCl3): δ 12.3,
15.0, 21.0, 23.1, 23.8, 26.6, 28.6, 30.9, 31.5, 32.0, 35.4, 35.6, 37.0, 38.1,
44.8, 47.6, 51.2, 52.3, 53.7, 71.4, 97.4, 211.4. LC−MS no peak.
Biological Experimental Methods. Protein Production.
Recombinant human G6PD (NM_000402.3) was cloned into a
pHis expression vector (University of Manchester, U.K.) to generate a
N-terminal His tagged construct and the protein expressed in E. coli
JM109(DE3) cells. Purification was performed via a single step nickel
affinity column (Qiagen) followed by imidazole elution to afford
protein with >70% purity. The determined kinetic parameters of the
recombinant enzyme (G6P Km = 47 μM, NADP+ Km = 5.5 μM) were
in accordance with previously published values for this enzyme.35,36
Enzyme Assay. The assay was performed in a 25 μL final volume
in 384-well black plates (Corning, 3575) with a reaction buffer
composed of 20 mM Tris-HCl, pH 8.0, 10 mM MgCl2, 0.03% BSA,
0.5 mM glucose 6-phosphate, 1% DMSO-d6 plus or minus inhibitory
compound, and 1 ng of G6PD. The reaction was initiated by the
addition of 50 μM NADP, and the assay was allowed to proceed
at 25 °C for 30 min before stopping by the addition of 0.05% SDS.
The NADPH generated during the assay was then quantified via an
Amplite fluorimetric NADPH assay kit (AAT Bioquest) in accordance
with the manufacturer’s instructions, and the assay plate was read on a
BioTek Synergy 2 multimode microplate reader with excitation and
emission at 540 and 590 nm, respectively.
To a stirred solution of (3β,5α)-3-[[(1,1-dimethylethyl)dimethylsilyl]-
oxy]pregnan-21-oic acid, ethyl ester, 20 (835 mg, 1.75 mmol) in THF
(10 mL) was added a solution of lithium hydroxide (84 mg, 3.5 mmol) in
water (4 mL). The mixture was stirred for 4 h, heated at 50 °C for 3 h,
and then left to cool for 72 h. The solvent was evaporated and the residue
treated with 2 N HCl and then extracted with DCM. The organic liquors
were washed with water, dried, and evaporated to give the crude product
(648 mg). Crystallization from diethyl ether/isohexane gave (3β,5α)-3-
[[(1,1-dimethylethyl)dimethylsilyl]oxy]pregnan-21-oic acid 21 as a white
solid (690 mg, 1.54 mmol, 88%). 1H NMR (CDCl3): δ 0.00 (s, 6H), 0.54
(s, 3H), 0.75 (s, 3H), 0.83 (s, 9H), 2.03−2.14 (m, 1H), 2.31−2.40
(m, 1H), 3.43−3.56 (m, 1H).
To a stirred solution of (3β,5α)-3-[[(1,1-dimethylethyl)dimethylsilyl]-
oxy]pregnan-21-oic acid 21 (367 mg, 0.82 mmol) in a mixture of
DMF (3.9 mL) and DCM (3.9 mL) was added 1,1′-carbonyldiimidazole
(159 mg, 0.98 mmol), and the mixture was stirred for 2 h. N,O-
Dimethylhydroxylamine hydrochloride (239 mg, 2.45 mmol) and
triethylamine (0.34 mL, 2.45 mmol) were added, and the mixture was
stirred overnight. The DCM was evaporated and water (30 mL) added.
The precipitated solid was filtered off and washed with water. The crude
product was purified by flash chromatography (isohexane/ethyl acetate
gradient elution) to give (3β,5α)-3-[[(1,1-dimethylethyl)dimethylsilyl]-
oxy]-N-methoxy-N-methyl-pregnan-21-amide 22 (370 mg, 0.75 mmol,
92%). 1H NMR (CDCl3): δ 0.00 (s, 6H), 0.56 (s, 3H), 0.75 (s, 3H), 0.83
(s, 9H), 2.12−2.23 (m, 1H), 2.37−2.46 (m, 1H), 3.12 (s, 3H), 3.44−3.56
(m, 1H), 3.63 (s, 3H).
Cell (6PG) Assay. HEK293T cells (1 mL) were plated at 16 × 104
mL−1 in each well of a 24-well poly-D-lysine coated plate. After 18 h,
compounds at 100 mM in DMSO-d6 were further diluted in DMSO-d6
(to generate a dose response from 1 to 100 μM final concentration),
and 1 μL of compound solution was added to each well. At 30 min
after compound addition, 200 mM 6-AN was diluted 1:10 in serum-
free medium and 10 μL added to all wells. After 5 h of incubation, cells
were rinsed with cold PBS and 500 μL of −80 °C MeOH/H2O
(80:20) was added directly to the monolayer to lyse the cells. Cell
extracts were transferred to Eppendorf tubes and centrifuged at
20000g for 10 min at 4 °C. The supernatant was removed to a fresh
tube and stored at −80 °C prior to analysis by mass spectrometry.
Relative concentrations of 6-PG in each supernatant sample were
measured using LC−MS/MS. Liquid chromatography was performed
using a Macherey-Nagel, Nucleosil 100-5C18 Nautilus column,
particle size 5 μm, 70 mm × 3.0 mm (catalog no. 721134.30), heated
to 35 °C. An isocratic mobile phase of 100% 2 mM ammonium acetate
in water, flow rate of 700 μL/min with a run time of 2.5 min, was used.
Quantitative detection of 6PG was undertaken using a Sciex API4000
To a stirred solution of (3β,5α)-3-[[(1,1-dimethylethyl)dimethylsilyl]-
oxy]-N-methoxy-N-methylpregnan-21-amide 22 (98 mg, 0.20 mmol) in
THF (2 mL) at 0 °C was added methylmagnesium bromide (0.71 mL,
0.99 mmol). The mixture was stirred for 2 h at ambient temperature.
More methylmagnesium bromide (0.71 mL, 0.99 mmol) was added, and
stirring continued for a further 3 h. Then 2 N HCl (10 mL) was added
cautiously and the mixture was stirred for 10 min and then extracted with
DCM. The organic liquors were washed with saturated sodium
bicarbonate, dried, and evaporated. The residue was dissolved in acetone
(5 mL), 2 N HCl (1 mL) added, and the mixture stirred for 2 h.
The mixture was evaporated and the residue dissolved in DCM. This
solution was passed through a hydrophobic frit and purified by flash
chromatography (isohexane/ethyl acetate gradient elution) to give
1
product 23 (16 mg, 0.048 mmol, 24%). H NMR (CDCl3): δ 0.57 (s,
3H), 0.81 (s, 3H), 2.13 (s, 3H), 2.22 (dd, J = 6.7 and 4.0 Hz, 1H), 2.47
(dd, J = 15.5 and 10.1 Hz, 1H), 3.59 (m, 1H). 13C NMR (CDCl3): δ
12.6, 13.1, 21.2, 24.9, 28.6, 28.9, 30.6, 31.8, 32.4, 35.9, 37.3, 37.8, 38.5,
42.5, 45.1, 45.2, 46.3, 54.9, 55.5, 71.6, 84.9, 209.9. LC−MS no peak.
4443
dx.doi.org/10.1021/jm300317k | J. Med. Chem. 2012, 55, 4431−4445

Journal of Medicinal Chemistry
Article
Q-Trap turbospray operating in negative ion mode. Ions monitored in
multiple reaction monitoring (MRM) were m/z 275−177.
ase plays a crucial role in protection from redox-stress-induced
apoptosis. Cell Death Differ. 2004, 11, 823−831.
(6) Tuttle, S.; Stamato, T.; Perez, M. L.; Biaglow, J. Glucose-6-
phosphate dehydrogenase and the oxidative pentose phosphate cycle
protect cells against apoptosis induced by low doses of ionizing
radiation. Radiat. Res. 2000, 153 (6), 781−787.
ASSOCIATED CONTENT
* Supporting Information
■
S
Summary of purity data; LC−MS methods and solvent
gradients; preparative HPLC instrument and solvent gradients;
preparative methods and spectroscopic data for compounds 1−
(7) Cosentino, C.; Grieco, D.; Costanzo, V. ATM activates the
pentose phosphate pathway promoting anti-oxidant defence and DNA
repair. EMBO J. 2011, 30 (3), 546−555.
1
4, 8−11, 14a, 14b, 16, 17a, 19, 24, and 27−47; H NMR
(8) Dessì, S.; Batetta, B.; Pani, P.; Barra, S.; Miranda, F.; Puxeddu, P.
Glucose-6-phosphate dehydrogenase (G6PD) activity in tumoral
tissues of G6PD-deficient subjects affected by larynx carcinoma.
Cancer Lett. 1990, 53 (2−3), 159−162.
(9) Wang, J.; Yuan, W.; Chen, Z.; Wu, S.; Chen, J.; Ge, J.; Hou, F.;
Chen, Z. Overexpression of G6PD is associated with poor clinical
outcome in gastric cancer. Tumour Biol. 2012, 33 (1), 95−101.
(10) Cabello, C. M.; Bair, W. B.,; Wondrak, G. T. Experimental
therapeutics: targeting the redox Achilles heel of cancer. Curr. Opin.
Invest. Drugs 2007, 8 (12), 1022−1037.
(11) Gupte, S. A. Glucose-6-phosphate dehydrogenase: a novel
therapeutic target in cardiovascular diseases. Curr. Opin. Invest. Drugs
2008, 9 (9), 993−1000.
spectra of target compounds; additional references. This
material is available free of charge via the Internet at http://
pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
*Phone: +44 (0)161-918-7447. E-mail: nhamilton@picr.man.
ac.uk.
Notes
The authors declare no competing financial interest.
(12) (a) Shin, E. S.; Park, J.; Shin, J.-M.; Cho, D.; Cho, S. Y.; Shin, D.
W.; Ham, M.; Kim, J. B.; Lee, T. R. Catechin gallates are NADP+-
competitive inhibitors of glucose-6-phosphate dehydrogenase and
other enzymes that employ NADP+ as a coenzyme. Bioorg. Med. Chem.
2008, 16 (7), 3580−3586. (b) Shin, U. S.; Lee H. H.; Lee, T. R.
Glucose-6-phosphate Dehydrogenase Inhibitors for Inhibiting Differ-
entiation of Adipocytes. Repub. Korean Kongkae Taeho Kongbo KR
2009056611 A, 2009; 11 pp (Amorepacific Corp., S. Korea)
(13) Marks, P. A.; Banks, J. Inhibition of mammalian glucose-6-
phosphate dehydrogenase by steroids. Proc. Natl. Acad. Sci. U.S.A.
1960, 46 (4), 447−452.
ACKNOWLEDGMENTS
■
This work was supported by Cancer Research UK (Grant
C480/A11411).
The G6PD enzyme was cloned and purified by E. McKenzie
at the Protein Expression Facility, Manchester Interdisciplinary
Biocentre, U.K.
ELSD analytical data were kindly generated with assistance
from the Waters MS Technology Centre, Manchester, M22
5PP, U.K.
HRMS spectra were generated by R. Sung, School of
Chemistry, University of Manchester, M13 9PL, U.K.
In vitro DMPK data were provided by Cyprotex Discovery,
Macclesfield, Cheshire SK10 2DR, U.K.
(14) (a) Oertel, G. W.; Rebelein, I. Effects of dehydroepiandroster-
one and its conjugates upon the activity of glucose-6-phosphate
dehydrogenase in human erythrocytes. Biochim. Biophys. Acta, Gen.
Subj. 1969, 184 (2), 459−460. (b) Oertel, G. W.; Rebelein, I. Effects of
dehydroepiandrosterone and its conjugates upon the activity of
glucose-6-phosphate dehydrogenase. I. Inhibition of human red blood
cell glucose-6-phosphate dehydrogenase. J. Steroid Biochem. 1970, 1
(2), 93−97. (c) Benes, P.; Freund, R.; Menzel, P.; Starka, L.; Oertel,
Georg, W. Inhibition of glucose-6-phosphate dehydrogenase by
steroids. I. Effects of 3β-hydroxy-Δ5-steroids of the C19- and C21-
series upon human red blood cell glucose-6-phosphate dehydrogenase.
J. Steroid Biochem. 1970, 1 (4), 287−290. (d) Raineri, R. R.; Levy, H.
R. On the specificity of steroid interaction with mammary glucose 6-
phosphate dehydrogenase. Biochemistry 1970, 9 (11), 2233−2243.
(e) Benes, P.; Oertel, G. W. Inhibition of human testicular glucose-6-
phosphate dehydrogenase by unsaturated C19-steroids. Experientia
1971, 27 (10), 1157. (f) Benes, P.; Oertel, G. W. Inhibition of glucose
6-phosphate dehydrogenase by steroids. IV. Effects of C18-steroids
upon human placental glucose 6-phosphate dehydrogenase. J. Steroid
Biochem. 1972, 31 (5), 799−801. (g) Oertel, G. W.; Benes, P. Effects
of steroids on glucose-6-phosphate dehydrogenase. J. Steroid Biochem.
1972, 3 (3), 493−496. (h) Oertel, G. W.; Benes, P. Inhibition of
glucose 6-phosphate dehydrogenase by steroids. VI. Effects of new
synthetic dehydroepiandrosterone conjugates upon the activity of
human placental glucose 6-phosphate dehydrogenase. J. Steroid
Biochem. 1972, 3 (6), 903−905. (i) Belovsky, O.; Benes, P.; Oertel,
G. W. Inhibition of glucose-6-phosphate dehydrogenase by steroids.
VIII. Effects of synthetic C19- and C20-steroids upon placental
glucose-6-phosphate dehydrogenase. J. Steroid Biochem. 1974, 5 (7),
697−700.
JChem for Excel was used for structure property prediction
and calculation, and general data handling (JChem for Excel,
version 5.4.0.411, 2008−2009, ChemAxon (http://www.
chemaxon.com)).
ABBREVIATIONS USED
■
DHEA, dehydroepiandrosterone; EA, epiandrosterone; G6PD,
glucose 6-phosphate dehydrogenase; PPP, pentose phosphate
pathway; NADP+, nicotinamide-adenine dinucleotide phos-
phate; NADPH, reduced nicotinamide adenine dinucleotide
phosphate; 6PGD, 6-phosphogluconate dehydrogenase; SAR,
structure−activity relationship
REFERENCES
■
(1) Koppenol, W. H.; Bounds, P. L.; Dang, C. V. Otto Warburg’s
contributions to current concepts of cancer metabolism. Nat. Rev.
Cancer. 2011, 11, 325−337;Erratum in Nat. Rev. Cancer. 2011, 11
(8), 618.
(2) Jadvar, H.; Alavi, A.; Gambhir, S. S. 18F-FDG uptake in lung,
breast, and colon cancers: molecular biology correlates and disease
characterization. J. Nucl. Med. 2009, 50 (11), 1820−1827.
(3) Kuo, W.; Lin, J.; Tang, T. K. Human glucose-6-phosphate
dehydrogenase (G6PD) gene transforms NIH 3T3 cells and induces
tumors in nude mice. Int. J. Cancer 2000, 85 (6), 857−864.
(4) Li, D.; Zhu, Y.; Tang, Q.; Lu, H.; Li, H.; Yang, Y.; Li, Z.; Tong, S.
A new G6PD knockdown tumor-cell line with reduced proliferation
and increased susceptibility to oxidative stress. Cancer Biother.
Radiopharm. 2009, 24 (1), 81−90.
(15) Abou-Gharbia, M.; Pashko, L.; Schwartz, A.; Swern, D. Synthesis
of dehydroepiandrosterone sulfatide and 16α-halogenated steroids. J.
Pharm. Sci. 1981, 70 (10), 1154−1157.
(16) Boehm, J. C. The Design, Synthesis and Testing of Novel
Steroidal Inhibitors of the Enzyme Glucose-6-phosphate Dehydrogen-
ase. Dissertation, Temple University, Philadelphia, PA, U.S., 1990; 238
(5) Fico, A.; Paglialunga, F.; Cigliano, L.; Abrescia, P.; Verde, P.;
Martini, G.; Iaccarino, I.; Filosa, S. Glucose-6-phosphate dehydrogen-
4444
dx.doi.org/10.1021/jm300317k | J. Med. Chem. 2012, 55, 4431−4445

Journal of Medicinal Chemistry
Article
pp. Available from University Microfilms International, Order No.
DA9107883. Abstract available at Diss. Abstr. Int., B 1991, 51 (11),
5310.
(17) Schwartz, A. G.; Lewbart, M. L. Preparation of 16-Substituted
Androstanes and 16-Substituted Androstenes as Drugs. US 5001119 A,
1991.
structure−property relationship. J. Pharm. Sci. 2002, 91 (8), 1838−
1852.
(33) Kim, J.; Jung, D. H.; Rhee, H.; Choi, S.-H.; Sung, M. J.; Choi, W.
S. Aqueous solubility of poorly water-soluble drugs: prediction using
similarity and quantitative structure−property relationship models.
Korean J. Chem. Eng. 2008, 25 (4), 865−873.
(34) Numazawa, M.; Nagaoka, M. New preparation and controlled
alkaline hydrolysis of 21-bromo-20-ketopregnenes. A facile synthesis of
deoxycorticoids. J. Org. Chem. 1985, 50 (1), 81−84.
(35) Gordon, G.; Mackow, M. C.; Levy, H. R. On the mechanism of
interaction of steroids with human glucose 6-phosphate dehydrogen-
ase. Arch. Biochem. Biophys. 1995, 318, 25−29.
(36) Wang, X. T.; Au, S. W.; Lam, V. M.; Engel, P. C. Recombinant
human glucose-6-phosphate dehydrogenase. Evidence for a rapid-
equilibrium random-order mechanism. Eur. J. Biochem. 2002, 269,
3417−24.
(18) Frederico, J. J. The Synthesis of 16-Fluorosteroids as G6PD
Inhibitors and the Synthesis of Potential Angiostatic Steroids.
Dissertation, Temple University, Philadelphia, PA, U.S., 1993; 204
pp. Available from University Microfilms International, Order No.
DA9332795. Abstract available at Diss. Abstr. Int., B 1994, 54 (7),
3622.
(19) Williams, J. R.; Boehm, J. C. The synthesis and reactivity of 3β-
(2-alkynylsulfonyl)- and 3β-(2-alkynylsulfonylmethyl)androst-5-en-17-
ones as inhibitors of glucose-6-phosphate dehydrogenase. Steroids
1995, 60, 699−708.
(20) Cordeiro, A. T.; Thiemann, O. H. 16-Bromoepiandrosterone, an
activator of the mammalian immune system, inhibits glucose 6-
phosphate dehydrogenase from Trypanosoma cruzi and is toxic to these
parasites grown in culture. Bioorg. Med. Chem. 2010, 18 (13), 4762−
4768.
(21) (a) Gupta, S.; Igoillo-Esteve, M.; Michels, P. A. M.; Cordeiro, A.
T. Glucose 6-phosphate dehydrogenase of trypanosomatids: character-
ization, target validation, and drug discovery. Mol. Biol. Int. 2011, 2011,
135701. (b) Gupta, S.; Cordeiro, A. T.; Michels, P. A. M. Glucose-6-
phosphate dehydrogenase is the target for the trypanocidal action of
human steroids. Mol. Biochem. Parasitol. 2011, 176 (2), 112−115.
(22) Charlton, M. H.; Thomson, C. Theoretical investigation of
steroidal inhibitors of glucose-6-phosphate dehydrogenase. J. Chem.
Soc., Faraday Trans. 1994, 90 (23), 3533−3537.
(23) Gordon, G. B.; Shantz, L. M.; Talalay, P. Modulation of growth,
differentiation and carcinogenesis by dehydroepiandrosterone. Adv.
Enzyme Regul. 1987, 26, 355−382.
(24) Schwartz, A. G.; Pashko, L. L. Mechanism of cancer preventive
action of DHEA: role of glucose-6-phosphate dehydrogenase. Ann.
N.Y. Acad. Sci. 1995, 774, 180−186.
(25) Feo, F.; Pascale, R. M.; Simile, M. M.; De Miglio, M. R. Role of
Dehydroepiandrosterone in Experimental and Human Carcinogenesis.
In Dehydroepiandrosterone (DHEA); Kalimi, M., Regelson, W., Eds.;
Walter de Gruyter GmbH & Co. KG: Berlin, Germany, 2000; pp 215−
236.
(26) Yoshida, S.; Honda, A.; Matsuzaki, Y.; Fukushima, S.; Tanaka,
N.; Takagiwa, A.; Fujimoto, Y.; Miyazaki, H.; Salen, G. Anti-
proliferative action of endogenous dehydroepiandrosterone metabo-
lites on human cancer cell lines. Steroids 2003, 68, 73−83.
(27) Matsuzaki, Y.; Honda, A. Dehydroepiandrosterone and its
derivatives: potentially novel and anti-proliferative and chemo-
preventive agents. Curr. Pharm. Des. 2006, 12 (26), 3411−3421.
(28) Cheng, M.-L.; Shiao, M.-S.; Chiu, D. T.-Y.; Weng, S.-F.; Tang,
H.-Y.; Ho, H.-Y. Biochemical disorders associated with antiprolifer-
ative effect of dehydroepiandrosterone in hepatoma cells as revealed
by LC-based metabolomics. Biochem. Pharmacol. 2011, 82 (11),
1549−1561.
(29) Hitchin, J. R.; Hamilton, N. M.; Jordan, A. M.; Lyons, A. J.;
Ogilvie, D. A novel scalable and stereospecific synthesis of 3α-and 3β-
amino-5α-androstan-17-ones and 3α- and 3β-amino-5α-pregnan-20-
ones. Tetrahedron Lett. [Online early access]. DOI: http://dx.doi.org/
10.1016/j.tetlet.2012.03.124. Published Online: April 4, 2012.
(30) Tuttle, S. W.; Varnes, M. E.; Mitchell, J. B.; Biaglow, J. E.
Sensitivity to chemical oxidants and radiation in CHO cell lines
deficient in oxidative pentose cycle activity. Int. J. Radiat. Oncol., Biol.,
Phys. 1992, 22 (4), 671−675.
(31) Keniry, M. A.; Hollander, C.; Benz, C. C. The effect of gossypol
and 6-aminonicotinamide on tumor cell metabolism: a 31P-magnetic
resonance spectroscopic study. Biochem. Biophys. Res. Commun. 1989,
164 (2), 947−953.
(32) Chen, X.-Q.; Cho, S. J.; Li, Y.; Venkatesh, S. Prediction of
aqueous solubility of organic compounds using a quantitative
4445
dx.doi.org/10.1021/jm300317k | J. Med. Chem. 2012, 55, 4431−4445