Synthesis and biological properties of novel glucocorticoid androstene C-17 furoate esters

Article abstract of DOI:10.1016/j.bmc.2004.06.027

A series of novel corticosteroid derivatives featuring C-17 furoate ester functionality have been synthesised. Profiling in vitro and in vivo has resulted in the identification of a compound with a longer duration of action and a lower oral side effect profile in rodents compared to budesonide.

Full text of DOI:10.1016/j.bmc.2004.06.027

Bioorganic & Medicinal Chemistry 12 (2004) 5213–5224
Synthesis and biological properties of novel glucocorticoid
androstene C-17 furoate esters
a
a
a
a
David A. Sandham,a, Lucy Barker, David Beattie, David Beer, Louise Bidlake,
*
David Bentley,a Keith D. Butler,a Sarah Craig,a David Farr,a Claire Ffoulkes-Jones,a
John R. Fozard,b Sandra Haberthuer,a Colin Howes,a Deborah Hynx,a Sarah Jeffers,a
Thomas H. Keller,a Paul A. Kirkham,a Janet C. Maas,a Lazzaro Mazzoni,b
Andrew Nicholls,a Gaynor E. Pilgrim,a Elisabeth Schaebulin,b Gillian M. Spooner,a
Rowan Stringer,a Pamela Tranter,a Katharine L. Turner,a Morris F. Tweed,a
Christoph Walker,a Simon J. Watsona and Bernard M. Cuenoudb,
*
aNovartis Institutes for Biomedical Research, Horsham Research Centre, Wimblehurst Road, Horsham, West Sussex,
RH12 5AB, United Kingdom
bNovartis Institutes for Biomedical Research, CH-4002 Basel, Switzerland
Received 21 May 2004; accepted 22 June 2004
Available online 4 August 2004
Abstract—A series of novel corticosteroid derivatives featuring C-17 furoate ester functionality have been synthesised. Profiling in
vitro and in vivo has resulted in the identification of a compound with a longer duration of action and a lower oral side effect profile
in rodents compared to budesonide.
ꢀ 2004 Elsevier Ltd. All rights reserved.
1. Introduction
asthma treatment, either as monotherapy or more
recently in combination with a long-acting b-2 agonist
bronchodilator.3
Corticosteroids are clinically proven as the most effec-
tive treatment for all levels of severity of asthma. This
utility resides in their gene transcriptional regulation
effects on multiple biological mediators.1 However,
long-term systemic exposure to corticosteroids results
in well-documented side effects such as childhood
growth retardation, hypertension, osteoperosis, skin
thinning, diabetes, weight gain, cataracts and adreno-
cortical insufficiency.2 This side effect issue has been
addressed by local delivery of corticosteroids with a high
topical to systemic potency ratio, culminating in the
development of low-dose inhaled corticosteroids (ICS).
ICS, particularly those with potential for once daily
dosing, constitute the gold standard in mild to moderate
Most of the ICS currently approved for asthma were
originally developed as topical steroids for dermatolog-
ical indications and subsequently adapted for inhala-
tion.4 Dependent on the nature of the delivery device
and the patientꢀs inhalation technique, up to 80% of
the dose of ICS may be swallowed and hence be availa-
ble for systemic absorption from the gastro-intestinal
(GI) tract, as illustrated in Figure 1. Earlier generation
ICS, such as budesonide 1 (Fig. 2) show significant oral
bioavailability,5 allowing potential for systemic side
effects by this route in addition to those arising from
direct pulmonary absorption. Subsequently, fluticasone
2 has been shown to undergo efficient inactivation via
hepatic metabolism,6 leading to negligible oral bioavail-
ability. More recently, elegant pre-clinical studies by
Biggadike and colleagues demonstrated the potential
for plasma inactivation of compounds such as 3, with
a further reduction in side effect potential.7 Conversely,
the example of mometasone furoate 48 indicates that it
Keywords: Glucocorticoids; Asthma; Duration of action; Side effects.
*
Corresponding authors. Tel.: +44-1403-323827; fax: +44-1403-
323837 (D.A.S.); fax: +41-61-3243991 (B.M.C.); e-mail addresses:
david.sandham@pharma.novartis.com; bernard.cuenoud@pharma.
novartis.com
0968-0896/$ - see front matter ꢀ 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2004.06.027

5214
D. A. Sandham et al. / Bioorg. Med. Chem. 12 (2004) 5213–5224
the in vivo profiles of selected examples as potential no-
vel inhaled corticosteroids for the treatment of asthma.
ICS Dose
Inhaled
Swallowed
2. Compound synthesis
C-21 methyl ester derivatives were prepared according
to Scheme 1. Starting from epoxyparamethasone 7, oxi-
dative cleavage according to literature conditions pro-
vided the acid 8 as a key building block.12 Facile
acylation of the hindered C-17 hydroxyl group was
achieved via the intermediacy of a mixed anhydride,12
either by treatment with an acid chloride in pyridine
(Method A) or by coupling with the corresponding carb-
oxylic acid in the presence of HATU (Method B) to give
9. The synthesis was completed by esterification to fur-
nish 10, followed by acidic epoxide opening to deliver
the target compounds 11a–n, as indicated in Table 1.
The first two steps of this process were amenable to par-
allel synthesis, with purification typically only required
for the final products. Compound 11a was also hydro-
genated in the presence of Wilkinsonꢀs catalyst to pro-
vide 1,2-dihydro analogue 12. The chlorohydrin acid
14 was prepared by treatment of intermediate 8 with
HCl gas, to give chlorohydrin 13, which was then selec-
tively acylated at C-17, as shown in Scheme 2.
Systemic
Systemic
side effects
side effects
Figure 1. Origin of systemic effects in inhaled corticosteroids.
may be possible for an ICS to deliver very low systemic
side effects without recourse to a priori soft drug design.9
The pro-drug ciclesonide 5 is another ICS in develop-
ment, which delivers low systemic side effects, possibly
as a consequence of slow release due to high protein
binding of the active metabolite in vivo.10
We sought to identify novel glucocorticoids with a long
duration of action, suitable for once daily dosing and
having low systemic side effects. A screening paradigm
was devised to identify rapidly an appropriate pre-clini-
cal in vivo profile for inhaled delivery, namely elimina-
tion of side effects upon oral administration,
minimisation of side effects on inhaled administration
and prolonged duration of action in a rodent asthma
model. The starting point for this work was the 6a-
fluoro 9a-chloro androstadiene C-17 propionate ester
6, originally identified some years ago by Ciba-Geigy
as a corticosteroid for potential dermatological indica-
tions with some separation of topical and systemic
effects.11 Following the example of mometasone, we
sought to optimise 6 further by incorporating a range
of substituted furoate esters at the C-17 position, as well
as varying the C-21 methyl ester moiety.
As an analogue of fluticasone propionate, C-21 fluoro-
thioester 19 was prepared according to Scheme 3. Chlo-
rohydrin 13 could be readily converted to thioacid 15
and the C-17 furoate ester was installed analogously to
Scheme 1, to afford 16. Finally, the fluoromethyl moiety
in 19 was installed in a three-step sequence, via chloro-
ester 17 and subsequent conversion to iodoester 18.
3. Biological profiling
In vitro anti-inflammatory glucocorticoid agonist effects
were assessed using an ELISA assay measuring the inhi-
bition of the lipopolysaccaride (LPS) induced release of
the inflammatory cytokine TNFa in a human U937
Herein we disclose the synthesis and in vitro biological
profiles of this novel class of molecules, together with
O
HO
O
O
F
S
O
O
O
S
O
O
HO
HO
O
HO
O
O
F
F
O
O
O
F
2
F
3
1
O
O
O
O
O
O
O
O
O
Cl
O
HO
HO
HO
O
O
O
Cl
Cl
O
O
O
F
6
4
5
Figure 2. Selected corticosteroids.

D. A. Sandham et al. / Bioorg. Med. Chem. 12 (2004) 5213–5224
5215
OH
R
HO
O
HO
O
O
O
OH
O
OH
i)
ii)
O
O
O
95%
O
O
O
F
F
F
7
8
9
R
O
O
R
R
O
O
O
O
O
O
HO
O
O
O
O
HO
iii)
iv)
v)
28%
O
Cl
4%-47%
from 8
Cl
F
O
O
O
F
F
10
11(a-n)
12
Scheme 1. Reagents and conditions: (i) H5IO6, THF–water, rt; (ii) Method A: RCOCl, pyridine 0ꢁC to rt or Method B: RCO2H, HATU, DIPEA,
DMF, rt; (iii) Me2SO4, DBU, EtOAc–DMF, rt; (iv) HCl gas, 1,4-dioxane, rt; (v) (PPh3)3RhCl, H2 (3.5bar), THF, rt.
macrophage-like cell-line.13 This is considered as a func-
tional measure of the transcriptional repression activity
Table 1 (continued)
No.
R1
R2
TNFa potency
relative to 1b
Table 1. Cellular in vitro activity of novel steroidsa
Cl
11i
OCH3
1.5
R1
O
O
O
R2
O
O
HO
CF
3
11j
OCH3
OCH3
4.8
2.4
O
O
Cl
F
O
O
11k
No.
R1
R2
TNFa potency
relative to 1b
O
O
11l
OCH3
OCH3
5.0
1.1
O
O
6
OCH3
OCH3
1.4
4.6
O
11m
11a
O
O
O
O
11n
12
OCH3
OCH3
OH
2.8
0.5
O
O
O
11b
11c
OCH3
OCH3
4.6
5.1
O
O
O
14
<0.001
3.6
O
O
O
O
O
11d
OCH3
5.4
19
SCH2F
O
a All compounds except 12 are D1,2
.
11e
11f
OCH3
OCH3
4.1
2.2
b Ratio derived from the mean of at least two assays; mean IC50 of
budesonide 248 78pM (n=18).
O
O
O
O
O
O
O
O
of the steroid, by which the majority of the beneficial
therapeutic effects are believed to occur.14 The relative
potencies of compounds listed in Table 1 are expressed
as the ratio of the IC50 value of the test compound rel-
ative to that of budesonide as an internal standard.
11g
11h
OCH3
OCH3
0.5
1.0
As an additional in vitro assay, glucocorticoid receptor
binding affinity was determined for selected compounds,

5216
D. A. Sandham et al. / Bioorg. Med. Chem. 12 (2004) 5213–5224
HO
O
HO
O
HO
O
O
OH
OH
HO
O
HO
i)
ii)
O
O
51%
98%
Cl
F
Cl
F
O
O
O
F
8
13
14
Scheme 2. Reagents and conditions: (i) HCl gas, dioxane, rt; (ii) 2-furoyl chloride, pyridine, rt.
O
HS
HS
O
HO
O
O
OH
HO
O
HO
OH
HO
i)
ii)
O
85%
Cl
Cl
F
Cl
F
O
O
O
F
13
15
16
Cl
I
F
S
O
O
O
S
S
O
iii)
O
O
O
HO
iv)
O
O
HO
HO
v)
O
87%
from 15
O
O
87%
35%
Cl
Cl
Cl
F
O
O
O
F
F
17
18
19
Scheme 3. Reagents and conditions: (i) NaSH, CDI, DMF, rt; (ii) 2-furoyl chloride, pyridine, rt; (iii) BrCH2Cl, DBU, DMF, rt; (iv) NaI acetone,
reflux; (v) AgF, MeCN, rt.
Table 2. In vitro human glucocorticoid receptor binding affinities
Table 3. In vivo efficacy and side effect profiling
No.
Glucocorticoid Ki/nMa
No.
Thymus weight Eosinophilia reduc-
% controla p.o. tion % controlb i.t.
Thymus weight
% controla i.t.
1
0.3
1.4
11a
11b
11f
19
1
42 4*
57 3*
80 6
63 4*
93 7
77 8
41 4*
29 11*
10 10
93 6**
79 10**
À8 17
5 25
83 16**
10 34
73 35*
21 1*
39 6*
18 1*
30 3*
nd
5.3
9.4
6
11a
11c
11e
11f
11l
15.1
a Mean of at least two assays.
nd
19 2*
nd
24 2*
with a commercial fluorescence polarisation assay based
on human glucocorticoid receptors expressed in insect
SF-9 cells (Panvera Inc). Ki values are shown in Table 2.
11m 84 6
19
66 6*
In each case compounds were dosed at 1mg/kg.
* p<0.05 compared to vehicle.
** p<0.005 compared to vehicle.
nd = not determined.
Compounds were then further profiled in vivo as shown
in Table 3, initially for effects on thymus weight reduc-
tion after chronic oral administration to rats over 4days.
Thymus weight is a very sensitive surrogate marker for
systemic glucocorticoid activity in the rat and thus this
model may be considered as a measure of likely side
effects arising from GI tract absorption. Efficacy in the
sensitised Brown Norway rat, employing ovalbumin
(OA)-induced pulmonary eosinophilia was next
assessed.15 This provides a model for the greatly reduced
levels of eosinophil influx seen in the lungs of asthmatic
patients treated with glucocorticoids. To screen for the
desired long duration of action, compounds were
administered intratracheally (i.t.) as dry powder blends
in lactose at a single dose 24h prior to induction of
eosinophilia by OA challenge. Finally, compounds
a Relative to control animals treated with vehicle.
b Relative to control animals challenged with OA and treated with
vehicle.
showing efficacy in this model were screened for
systemic effects on thymus weight arising from chronic
i.t. administration of lactose dry powder blends over
4days, as a measure of likely glucocorticoid side effects
due to direct pulmonary absorption.
4. Results and discussion
The unsubstituted 2 or 3-furoate esters 11a and 11b
showed a clear increase in in vitro potency over 6. Sub-

D. A. Sandham et al. / Bioorg. Med. Chem. 12 (2004) 5213–5224
5217
stituted analogues of the 2-furoate 11a revealed 3- or 4-
alkyl substituents in 11c–e as being similar or marginally
more potent than the parent compound. In contrast, 5-
substituents in 11f–i with the exception of trifluorometh-
yl derivative 11j were clearly detrimental to potency.
Compared with the 3-furoate parent compound 11b,
monosubstitution at the 2-, 4- or 5-positions appeared
slightly detrimental in 11k, 11m and 11n, whereas the
2,5-disubstituted furoate 11l gave similar potency. In
accord with literature precedent D1,2 dihydro analogue
12 gave reduced potency compared to 11a,16 while the
corresponding C-21 acid 13 was essentially inactive,
presumably as a consequence of the lack of cellular
penetration.9 Finally, compound 19 with the C-21
methyl ester replaced by an a-fluorothioester, showed
a superior potency to both 1 and 6.
***
***
100
60
***
*
20
-20
0.01
0.1
1
10
Dose (mg/kg)
Figure 3. Comparison of the potencies of 1 and 11a in the Brown
Norway rat. Compounds dosed i.t. 24h prior to OA challenge; (d)
11a; (j) 1; *p<0.05, ***p<0.001 compared to OA challenged animals
treated with vehicle.
As expected, all compounds tested showed a high affin-
ity for the human glucocortiocoid receptor (Table 2),
although there appeared to be no correlation between
the rank order of potency in this binding assay and
the functional activity described above.
100
*
Turning to in vivo profiling (Table 3), 1 and 6 both
showed significant glucocorticoid side effects when
administered orally or intratracheally, and 6 lacked any
statistically significant duration of action for inhibition
of eosinophilia in bronchioalveolar lavage (BAL) fluid
at 1mg/kg, while 1 showed only a small effect. In con-
trast, 11a showed no significant oral side effects and com-
plete inhibition of eosinophilia 24h after 1mg/kg dosing.
For this compound, thymus weight reduction via the i.t.
route was similar to 1 and slightly greater than 6. 4-
Methyl 2-furoate derivative 11c, 3-furoate 11l and fluor-
othioester 19, all of which showed similar in vitro
potency to 11a, also gave a 24h duration of action in
the efficacy model, but with increased significant oral side
effects compared to 11a. Compounds 11e, 11f and 11m,
all of which showed lower in vitro potency than 11a
gave no significant oral thymus effects but exhibited little
or no efficacy in the Brown Norway rat eosinophilia
model.
*
50
*
*
*
*
0
*
0.01
0.1
Dose (mg/kg)
1
Figure 4. Comparison of the thymus weight reduction potencies of 1
and 11a after i.t. administration. Compounds dosed i.t. as dry powder
blends for 4days; (d) 11a; (j) 1; *p<0.05 compared to animals treated
with vehicle.
Given the favourable duration of action and oral side
effect in vivo profile of 11a, this compound was further
profiled in the Brown Norway rat efficacy model with
a full dose response (Fig. 3) compared with 1 head to
head, clearly indicating its superior potency for
reduction of eosinophilia in the BAL fluid after OA
challenge.
The minimised thymus weight reduction via the oral
route reflects either a high first pass metabolism and/or
low absorption, and indeed 11a could not be detected
in the plasma of rats when dosed orally either as a sus-
pension or solution at 4.6mg/kg, with a lower limit of
quantification of 0.05lM. By contrast, 1 is reported to
be 15% bioavailable in the rat after a 1.5mg/kg oral
dose.17 The significant thymus weight reduction seen
for all steroids via the i.t. route probably reflects the sta-
bility of the compounds in plasma. This was verified for
compound 11a, which showed no degradation in rat or
human plasma over 30min at 37ꢁC, indicating the C-
21 methyl ester is apparently stable to plasma esterases
under these conditions. One potential explanation for
the duration of action observed in the efficacy model
may be the physical consequence of slow dissolution
of the dry powder formulation. In this regard, the much
lower thermodynamic solubility of 11a (0.0001g/L)
compared to budesonide (0.0045g/L) at pH6.8 in phos-
phate buffer could be a contributory factor.
Further dose–response comparison of the side effect pro-
files of the above compounds underscored the similar
levels of systemic activity after i.t. administration (Fig.
4) and the difference in side effects after oral administra-
tion (Fig. 5). In the latter case, 11a is at least 10-fold less
potent than 1 in eliciting 50% thymus weight reduction.
5. Conclusions
In summary, novel androstadiene C-17 esters have been
prepared and profiled for in vivo anti-inflammatory
activity and side effects. Replacement of propionate by
a furoate ester gave increased potency in vitro, as well
as providing an increased duration of action and reduc-
tion of oral glucocorticoid side effects in vivo. The
in vitro structure–activity relationship indicated a
variety of substitution was allowed on the furan ring

5218
D. A. Sandham et al. / Bioorg. Med. Chem. 12 (2004) 5213–5224
(t, 3H, J=8Hz), 1.27–1.32 (m, 2H), 1.64 (s, 3H), 1.79–
100
50
0
2.01 (m, 3H), 2.23–2.28 (m, 1H), 2.40 (q, 2H, J=8Hz),
2.47–2.62 (m, 2H), 2.70 (dd, 1H, J=4,14Hz), 3.33–
3.36 (m, 1H), 3.73 (s, 3H), 4.57 (br s, 1H), 5.29–5.46
(2m, 1H), 6.37 (dd, 1H, J=2,10Hz), 6.43 (s 1H), 7.12
(d, 1H, J=10Hz); MS(ES+) m/z 483 (M+H)+; Anal.
Calcd for C25H32ClFO6Æ0.2H2O: C, 61.73; H, 6.67.
Found: C, 61.75; H, 6.92.
*
*
*
*
6.2. 9b,11b-Epoxy-6-fluoro-17a-hydroxy-16a-methyl-3-
oxoandrosta-1,4-diene-17b-carboxylic acid (8)
0.1
1
10
100
Dose (mg/kg)
To a suspension of epoxyparamethasone 7 (25.0g,
64.10mmol) in THF (300mL), was added dropwise at
25ꢁC a solution of periodic acid (19.73g, 85.54mmol)
in water (100mL). The reaction mixture was stirred at
25ꢁC for 16h, then concentrated to approximately half
its original volume. Water (200mL) was added slowly
with vigorous stirring and the resulting precipitate was
filtered off and dried under vacuum to give 8 (23.0g,
Figure 5. Comparison of the thymus weight reduction potencies of 1
and 11a after oral administration. Compounds dosed orally as
suspensions for 4days; (d) 11a; (j) 1; *p<0.05 compared to animals
treated with vehicle.
of the ester. In vivo screening enabled rapid identifica-
tion of the unsubstituted 2-furoate ester 11a as a
compound with minimal oral absorption, together with
superior efficacy and duration of action and similar i.t.
side effect profile compared to budesonide in in vivo
rodent models. Based on these results, compound 11a
may have potential to function as a long-acting once-
daily inhaled corticosteroid. Further studies on 11a
and related compounds will be reported in due course.
1
95%) as an off-white solid: H NMR (CDCl3) 0.97 (d,
3H, J=7Hz), 1.08 (s, 3H), 1.42 (s, 3H), 1.64 (q, 2H,
J=12Hz), 1.76–1.87 (m, 2H), 2.14 (d, 2H, J=11Hz),
2.32 (q, 2H, J=8Hz), 2.61–2.70 (m, 2H), 2.94–3.02
(m, 1H), 3.33 (br s, 1H), 5.36–5.52 (2m, 1H), 6.25 (dd,
1H, J=2,8Hz), 6.45 (s, 1H), 6.53 (d, 1H, J=10Hz).
6.3. 9a-Chloro-6a-fluoro-17a-[(2-furanylcarbonyl)oxy]-
11b-hydroxy-16a-methyl-3-oxoandrosta-1,4-diene-17b-
carboxylic acid, methyl ester (11a)
6. Experimental section
Reagents and solvents were purchased from common
commercial suppliers and used as received. Noncom-
mercial acid chlorides were prepared by refluxing the
corresponding furoic acids with thionyl chloride, fol-
lowed by evaporation to dryness in vacuo. All reactions
were carried out under an atmosphere of argon unless
otherwise stated. TLC was performed on Merck Kiesel-
gel 60 F254 plates and column chromatography was con-
ducted on Merck Kieselgel 60 (230–400 mesh).
Analytical HPLC was carried out on a reverse-phase
Hypersil Elite C18 column (10cm·0.3cm) at 40ꢁC,
using a gradient of 30–95% acetonitrile over 8min, flow
rate 0.50mL/min, with detection at 254nm. Mass spectr-
ometry, including accurate mass determination, was per-
General procedure for acylation of C-17 a-hydroxyl
group. Method A.
6.4. 9b,11b-Epoxy-6-fluoro-17a-[(2-furanylcarbonyl)oxy]-
16a-methyl-3-oxoandrosta-1,4-diene-17b-carboxylic acid
(9a)
To a stirred solution of 8 (4.0g, 10.64mmol) in pyridine
(15mL) at 25ꢁC was added dropwise 2-furoyl chloride
(1.53g, 11.70mmol). The reaction mixture was stirred
at room temperature for 2h, then added dropwise to a
vigorously stirred solution of 6M aq HCl (45mL). After
30min of stirring at 25ꢁC, the resulting precipitate was
filtered, dissolved in CH2Cl2, dried (MgSO4), filtered
and concentrated to give 9a (4.9g, 98%) as an off-white
formed using a Micromass Time-of-Flight LCT.
(ES+ve) refers to mass spectra run in positive mode
1
using electrospray techniques. H NMR analyses were
1
solid: H NMR (CDCl3) 1.02 (d, 3H, J=8Hz), 1.08 (s,
performed on Bruker ARX/AV 400MHz instruments
in CDCl3 or DMSO-d6. Chemical shifts (d) are ex-
pressed in ppm relative to residual proton signals in sol-
vent. Melting points were determined on a Gallenkamp
capillary melting point apparatus and are uncorrected.
Elemental analyses were performed on a LECO CHNS
932 instrument.
3H), 1.44 (s, 3H), 1.61–1.77 (m, 2H), 1.84–1.95 (m,
2H), 2.07 (dd, 1H, J=3,15Hz), 2.31–2.41 (m, 2H),
2.66–2.72 (m, 1H), 3.28–3.36 (m, 2H), 4.60 (br s, 1H),
5.39–5.54 (2m, 1H), 6.30 (dd, 1H, J=2,10Hz), 6.49–
6.55 (m, 2H), 6.56 (d, 1H, J=10Hz), 7.18 (d, 1H,
J=3Hz).
6.5. 9b,11b-Epoxy-6a-fluoro-17a-[(2-furanylcarbonyl)-
oxy]-16a-methyl-3-oxoandrosta-1,4-diene-17b-carb-
oxylic acid, methyl ester (10a)
6.1. 9a-Chloro-6a-fluoro-11b-hydroxy-17a-(1-oxoprop-
oxy)-16a-methyl-3-oxoandrosta-1,4-diene-17b-carboxylic
acid, methyl ester (6)
To a stirred solution of 9a (1.10g, 2.34mmol) in EtOAc
(25mL) at 0ꢁC, was added DBU (0.422mL, 2.82mmol),
followed by Me2SO4 (2.67mL, 2.82mmol). The reaction
mixture was stirred at 25ꢁC for 2h, treated with morph-
oline (0.1mL, 1mmol) and stirred for a further 30min,
Prepared as previously described.11
Mp 197–198ꢁC; Anal. HPLC tR 6.82min, 99%; 1H
NMR (CDCl3) 0.98 (d, 3H, J=7Hz), 1.0 (s, 3H), 1.17

D. A. Sandham et al. / Bioorg. Med. Chem. 12 (2004) 5213–5224
5219
then partitioned between EtOAc and water. The organic
layer was washed with brine, dried (MgSO4), filtered and
concentrated. MeOH (20mL) was added from which the
product crystallised, affording 10a (650mg, 58%) as a
0.92 (s, 3H), 1.12–1.18 (m, 1H), 1.55 (s, 3H), 1.55–
1.72 (m, 2H), 1.86 (q, 1H, J=12Hz), 1.96 (s, 3H),
2.19–2.28 (m, 1H), 2.78 (t, 1H, J=10Hz), 3.23–3.50
(m, 2H), 3.71 (s, 3H), 4.53 (br s, 1H), 5.74–5.91 (2m,
2H), 6.34 (s, 1H), 6.55 (d, 1H, J=11Hz), 7.31 (s,
1H), 7.59 (d, 1H, J=11Hz), 8.08 (s, 1H); MS(ES+)
m/z 535 (M+H)+; Anal. Calcd for C28H32ClFO7Æ
0.3H2O: C, 62.18; H, 6.02. Found: C, 62.27; H,
6.05.
1
white solid: H NMR (CDCl3) 0.96 (d, 3H, J=7Hz),
1.44 (s, 3H), 1.48 (s, 3H), 1.60–1.75 (m, 2H), 1.86 (q,
1H, J=10Hz), 1.98 (dd, 1H, J=3,13Hz), 2.29–2.37
(m, 2H), 2.64–2.73 (m, 1H), 3.27–3.34 (m, 2H), 3.75 (s,
3H), 5.37–5.54 (m, 1H), 6.27 (dd, 1H, J=2,7Hz), 6.48
(s, 1H), 6.52–6.54 (m, 1H), 6.56 (s, 1H), 7.17 (d, 1H,
J=4Hz), 7.62 (s, 1H).
6.9. 9a-Chloro-6a-fluoro-17a-[(2-furanylcarbonyl-3-
methyl)oxy]-11b-hydroxy-16a-methyl-3-oxoandrosta-
1,4-diene-17b-carboxylic acid, methyl ester (11d)
6.6. 9a-Chloro-6a-fluoro-17a-[(2-furanylcarbonyl)oxy]-
11b-hydroxy-16a-methyl-3-oxoandrosta-1,4-diene-17b-
carboxylic acid, methyl ester (11a)
Similarly to the procedure described for 11a, the title
compound was prepared starting from 8 and 3-methyl-
2-furancarboxylic acid. Recrystallisation from i-Pr2O/
MeOH, followed by column chromatography eluting
with CH2Cl2/MeOH (25:1) afforded 11d (11% yield for
three steps from 8) as an off-white solid: mp 194ꢁC;
HCl gas was bubbled through a solution of 10a (550mg,
1.14mmol) in toluene (20mL) for 5min. The reaction
flask was stoppered, then stirred at 25ꢁC for 18h. The
solid obtained after concentration of the reaction mix-
ture was recrystallised from i-PrOH/MeOH to give 11a
(420mg, 71%) as a white solid: mp 196–197ꢁC dec;
1
Anal. HPLC tR 7.21min, 97%; H NMR (CDCl3) 0.90
(d, 3H, J=7Hz), 0.98 (s, 3H), 1.20–1.26 (m, 1H), 1.35
(s, 3H), 1.52–1.60 (m, 2H), 1.72–1.91 (m, 2H), 2.10–
2.18 (m, 1H), 2.21 (s, 3H), 2.46–2.57 (m, 2H), 2.72
(dd, 1H, J=3,15Hz), 3.28–3.36 (m, 1H), 3.63 (s, 3H),
4.46 (br s, 1H), 5.17–5.33 (2m, 1H), 6.21 (s, 1H), 6.23
(dd, 1H, J=2,10Hz), 6.31 (s, 1H), 6.98 (d, 1H,
J=1Hz), 7.29 (s, 1H); MS(ES+) m/z 535 (M+H)+; Anal.
Calcd for C28H32ClFO7Æ0.8H2O: C, 61.16; H, 6.18.
Found: C, 61.20; H, 6.21.
1
Anal. HPLC tR 7.20min, 98%; H NMR (CDCl3) 1.03
(d, 3H, J=7Hz), 1.10 (s, 3H), 1.30–1.35 (m, 1H), 1.64
(s, 3H), 1.72 (dd, 1H, J=3,15Hz), 1.84–2.02 (m, 2H),
2.23–2.29 (m, 1H), 2.58–2.71 (m, 2H), 2.82 (dd, 1H,
J=3,15Hz), 3.36–3.46 (m, 1H), 3.72 (s, 3H), 4.59 (br s,
1H), 5.29–5.46 (2m, 1H), 6.38 (dd, 1H, J=2,9Hz),
6.41 (s, 1H), 6.48 (m, 1H), 7.09 (s, 1H), 7.12 (m, 1H),
7.60 (s, 1H); MS(ES+) m/z 521 (M+H)+; Anal. Calcd
for C27H30ClFO7: C, 62.25; H, 5.80. Found: C, 62.13;
H, 5.78.
6.10. 9a-Chloro-6a-fluoro-17a-[(3-ethyl-2-furanylcarbon-
yl)oxy]-11b-hydroxy-16a-methyl-3-oxoandrosta-1,4-di-
ene-17b-carboxylic acid, methyl ester (11e)
6.7. 9a-Chloro-6a-fluoro-17a-[(3-furanylcarbonyl)oxy]-
11b-hydroxy-16a-methyl-3-oxoandrosta-1,4-diene-17b-
carboxylic acid, methyl ester (11b)
General procedure for acylation of C-17a-hydroxyl
group. Method B.
Similarly to the procedure described for 11a, the title
compound was prepared starting from 8 and 3-furan-
carboxylic acid. Recrystallisation from i-PrOH afforded
11b (24% yield for three steps from 8) as a white solid:
mp 197–201ꢁC; Anal. HPLC tR 7.10min, 97%; 1H
NMR (CDCl3) 0.89 (d, 3H, J=7Hz), 0.97 (s, 3H),
1.16–1.22 (m, 1H), 1.52–1.58 (m, 5H), 1.68–1.91 (m,
2H), 2.11–2.19 (m, 1H), 2.46–2.56 (m, 2H), 2.66 (dd,
1H, J=3,15), 3.24–3.29 (m, 1H), 3.63 (s, 3H), 4.47 (br
s, 1H), 5.16–5.34 (2m, 1H), 6.26 (dd, 1H, J=2,10Hz),
6.31 (s, 1H), 6.57 (s, 1H), 6.98 (d, 1H, J=10Hz), 7.29
(s, 1H), 7.87 (s, 1H); MS(ES+) m/z 521 (M+H)+; Anal.
Calcd for C27H30ClFO7Æ0.1H2O: C, 61.98; H, 5.80.
Found: C, 61.93; H, 5.90.
6.11. 9b,11b-Epoxy-17a-[(3-ethyl-2-furanylcarbonyl)-
oxy]-6a-fluoro-16a-methyl-3-oxoandrosta-1,4-diene-
17b-carboxylic acid (9e)
To
a
solution of 3-ethyl-2-furancarboxylic acid19
(400mg, 7.13mmol) at 0ꢁC in DMF (2.5mL) was added
diisopropylethylamine 1.24mL, 7.13mmol), followed by
portionwise addition of HATU (1.19g, 3.13mmol). The
reaction mixture was stirred for 10min after which time
a solution of 8 (1.07g, 2.85mmol) in DMF (3mL) was
added. The reaction mixture was stirred for 2h, allowing
to warm slowly to room temperature. The reaction mix-
ture was poured into water (30mL), filtered to afford an
off-white precipitate, which was washed with water
(3·5mL) and dried in vacuo. The crude material was
purified by column chromatography on silica gel using
1:1 EtOAc/n-hexane as the eluent, finally eluting with
EtOAc to give 9e (540mg, 38%) as an off-white glassy
6.8. 9a-Chloro-6a-fluoro-17a-[(2-furanylcarbonyl-4-
methyl)oxy]-11b-hydroxy-16a-methyl-3-oxoandrosta-
1,4-diene-17b-carboxylic acid, methyl ester (11c)
1
Similarly to the procedure described for 11a, the title
compound was prepared starting from 8 and 2-furan-
4-methylcarboxylic acid.18 Purification by column chro-
matography using EtOAc/hexane (1:2) as the eluent, af-
forded 11c (47% yield for three steps from 8) as a white
solid: mp 188–190ꢁC dec; Anal. HPLC tR 7.43min,
98%; 1H NMR (DMSO-d6) 0.78 (d, 3H, J=8Hz),
solid: H NMR (CDCl3) 0.99 (d, 3H, J=7Hz), 1.04 (s,
3H), 1.15 (t, 3H, J=8Hz), 1.44 (s, 3H), 1.60–1.77 (m,
2H), 1.91 (q, 2H, J=8Hz), 2.32–2.41 (m, 2H), 2.65–
2.78 (m, 3H), 3.24–3.38 (m, 2H), 5.38–5.55 (m, 1H),
6.30 (dd, 1H, J=2,8Hz), 6.42 (d, 1H, J=2Hz), 6.51
(d, 1H, J=2Hz), 6.57 (d, 1H, J=2Hz), 7.50 (d, 1H,
J=2Hz).

5220
D. A. Sandham et al. / Bioorg. Med. Chem. 12 (2004) 5213–5224
6.12. 9b,11b-Epoxy-17a-[(3-ethyl-2-furanylcarbonyl)-
oxy]-6a-fluoro-16a-methyl-3-oxoandrosta-1,4-diene-
17b-carboxylic acid, methyl ester (10e)
6.15. 9a-Chloro-6a-fluoro-17a-[(2-furanylcarbonyl-5-
methyl)oxy]-11b-hydroxy-16a-methyl-3-oxoandrosta-
1,4-diene-17b-carboxylic acid, methyl ester (11g)
To a stirred solution of 9e (121mg, 1.24mmol) in DMF
(4mL) at 0ꢁC, was added DBU (0.204mL, 1.37mmol),
followed by Me2SO4 (0.129mL, 1.37mmol). The reac-
tion mixture was stirred at 25ꢁC for 2h, then poured
into water (20mL) and extracted with EtOAc
(2·15mL). The combined organic extracts were washed
with water (2·15mL), brine (10mL), dried (MgSO4), fil-
tered and concentrated, affording 10e (528mg, 83%) as
an off-white solid: 1H NMR (CDCl3) 0.99 (d, 3H,
J=7Hz), 1.15 (t, 3H, J=8Hz), 1.46 (s, 3H), 1.59–1.76
(m, 5H), 1.89–2.04 (m, 2H), 2.66–2.72 (m, 2H), 2.74–
2.79 (m, 3H), 3.32 (br s, 2H), 3.73 (s, 3H), 5.38–5.55
(m, 1H), 6.28 (dd, 1H, J=2,10Hz), 6.40 (s, 1H), 6.50
(d, 1H, J=2Hz), 6.58 (dd, 1H, J=2,10Hz), 7.26 (d,
1H, J=2Hz).
Similarly to the procedure described for 11e, the title
compound was prepared starting from 8 and 5-methyl-
2-furancarboxylic acid.21 Trituration with MeOH af-
forded 11g (32% yield for three steps from 8) as a white
solid: mp 193–195ꢁC; Anal. HPLC tR 6.90min, 94%; 1H
NMR (DMSO-d6) 0.80 (d, 3H, J=7Hz), 0.92 (s, 3H),
1.10–1.22 (m, 1H), 1.50–1.71 (m, 5H), 1.87 (q, 1H,
J=12), 2.22 (m, 1H), 2.31 (s, 3H), 2.67–2.82 (m, 2H),
3.70 (s, 3H), 4.52 (br s, 1H), 5.74–5.92 (m, 2H), 6.34
(s, 1H), 6.53 (d, 1H, J=2Hz), 6.56 (d, 1H, J=4Hz),
7.34 (s, 1H), 7.58 (d, 1H, J=10Hz); MS(ES+) m/z 535
(M+H)+; Anal. Calcd for C28H32ClFO7ÆH2O: C, 60.76;
H, 6.01. Found: C, 60.77; H, 6.30.
6.16. 9a-Chloro-17a-[(5-ethyl-2-furanylcarbonyl)oxy]-6a-
fluoro-11b-hydroxy-16a-methyl-3-oxoandrosta-1,4-diene-
17b-carboxylic acid, methyl ester (11h)
6.13. 9a-Chloro-6a-fluoro-17a-[(3-ethyl-2-furanylcarbon-
yl)oxy]-11b-hydroxy-16a-methyl-3-oxoandrosta-1,4-di-
ene-17b-carboxylic acid, methyl ester (11e)
Similarly to the procedure described for 11e, the title
compound was prepared starting from 8 and 5-ethyl-2-
furancarboxylic acid.22 Column chromatography eluting
with EtOAc/n-hexane (1:2) afforded 11h (14% yield for
three steps from 8) as a white solid: mp 182–183ꢁC
HCl gas was bubbled through a solution of 10e (509mg,
0.99mmol) in toluene (75mL) over 5min. The reaction
flask was stoppered, then stirred at 25ꢁC for 18h, prior
to concentration to give a green syrup. Trituration with
MeOH (8mL) afforded a white solid in green superna-
tant liquid. The crude product was filtered, washed with
MeOH (3·1mL) and recrystallised from MeOH to af-
ford 11e (183mg, 34%) as a white solid: mp 179–
1
dec; Anal. HPLC tR 7.85min, 97%; H NMR (DMSO-
d6) 0.79 (d, 3H, J=8Hz), 0.92 (s, 3H), 1.08 (t, 3H,
J=8Hz), 1.13–1.18 (m, 1H), 1.55–1.71 (m, 5H), 1.81–
1.90 (q, 1H, J=13Hz), 2.20–2.28 (m, 1H), 2.54 (s,
1H), 2.67 (q, 2H, J=8Hz), 2.77 (t, 1H, J=11Hz),
3.29–3.33 (m, 1H), 3.71 (s, 3H), 4.52 (br s, 1H), 5.74–
5.92 (2m, 2H), 6.34 (s, 1H), 6.55 (d, 1H, J=11Hz),
6.59 (d, 1H, J=3Hz), 7.39 (d, 1H, J=3Hz), 7.59 (d,
1H, J=11Hz); MS(ES+) m/z 549 (M+H)+; Anal. Calcd
for C29H34ClFO7Æ0.2H2O: C, 62.98; H, 6.19. Found: C,
63.01; H, 6.37.
1
180ꢁC dec; Anal. HPLC tR 7.60min, 90%; H NMR
(DMSO-d6) 0.88 (d, 3H, J=7Hz), 1.01 (s, 3H), 1.08 (t,
3H, J=8Hz), 1.23–1.30 (m, 1H), 1.57–1.76 (m, 5H),
1.92 (q, 1H, J=12Hz), 2.21–2.28 (m, 1H), 2.39–2.47
(m, 2H), 2.66–2.84 (m, 3H), 3.64 (s, 3H), 4.39 (br s,
1H), 5.54–5.68 (m, 2H), 6.11 (s, 1H), 6.31 (d, 1H,
J=10Hz), 6.64 (s, 1H), 7.28 (d, 1H, J=10Hz), 7.83 (s,
1H); MS(ES+) m/z 549 (M+H)+; Anal. Calcd for
C29H34ClFO7Æ0.5H2O: C, 62.37; H, 6.14. Found: C,
62.44; H, 6.30.
6.17. 9a-Chloro-17a-[(5-chloro-2-furanylcarbonyl)oxy]-6-
fluoro-11b-hydroxy-16a-methyl-3-oxoandrosta-1,4-diene-
17b-carboxylic acid, methyl ester (11i)
6.14. 9a-Chloro-17a-[(3,5-dimethyl-2-furanylcarbon-
yl)oxy]-6a-fluoro-11b-hydroxy-16a-methyl-3-oxoandro-
sta-1,4-diene-17b-carboxylic acid, methyl ester (11f)
6.17.1. Preparation of 5-chloro-2-furancarboxylic acid.
To diisopropylamine (1.4mL, 10mmol) cooled to
À50ꢁC, under argon, was added n-BuLi (2.5M in hexa-
nes; 4mL, 10mmol) followed by THF (20mL). 2-Furoic
acid (560mg, 5mmol) was then added dropwise as a
solution in THF (10mL). The reaction mixture was stir-
red at À50ꢁC for 30min, then a dispersion of N-chloro-
succinimide (694mg, 5.2mmol) in THF (5mL) was
added. Once addition was complete, the reaction mix-
ture was kept at À50ꢁC for 30min then warmed up to
25ꢁC slowly over 16h. The resultant suspension was fil-
tered, the resultant solid was dissolved in H2O, washed
with Et2O, then acidified to pH1 with 1M HCl. The
product was extracted with Et2O (2·25mL), washed
with brine, dried (Na2SO4), filtered and concentrated.
The residue was repeatedly recrystallised from H2O
affording 5-chloro-2-furancarboxylic acid (60mg, 8%):
1H NMR (CDCl3) 6.37 (d, 1H, J=4Hz), 7.29 (d, 1H,
J=3Hz).
Similarly to the procedure described for 11e, the title
compound was prepared starting from 8 and 3,5-di-
methyl-2-furancarboxylic acid.20 Column chromatogra-
phy eluting with EtOAc/n-hexane (1:2) afforded 11f
(22% yield for three steps from 8) as a pale green solid:
1
mp 125–135ꢁC dec; Anal. HPLC tR 7.68min, 92%; H
NMR (DMSO-d6) 0.89 (d, 3H, J=8Hz), 1.01 (s, 3H),
1.24–1.29 (m, 1H), 1.61–1.78 (m, 5H), 1.90 (q, 1H,
J=12Hz), 2.20–2.30 (m, 7H), 2.40–2.54 (m, 2H), 2.76
(t, 1H, J=10Hz), 3.64 (s, 3H), 4.39 (br s, 1H), 5.55–
5.71 (m, 2H), 6.10 (s, 1H), 6.21 (s, 1H), 6.30 (dd,
1H, J=2,10Hz), 7.28 (d, 1H, J=10Hz); MS(ES+)
m/z 549 (M+H)+; Anal. Calcd for C29H34ClFO7Æ
0.8H2O: C, 61.77; H, 6.23. Found: C, 61.74; H,
6.41.

D. A. Sandham et al. / Bioorg. Med. Chem. 12 (2004) 5213–5224
5221
Similarly to the procedure described for 11e, the title
compound was prepared starting from 8 and 5-chloro-
2-furancarboxylic acid. Trituration with MeOH
afforded 11i (20% yield for three steps from 8) as a white
NMR (DMSO-d6); 0.90 (d, 3H, J=7Hz), 1.03 (s, 3H),
1.24–1.28 (m, 1H), 1.64–1.78 (m, 5H), 1.90 (q, 1H,
J=12Hz), 2.15–2.28 (m, 4H), 2.43–2.53 (m, 5H), 2.79
(t, 1H, J=10Hz), 3.19–3.34 (m, 1H), 3.66 (s, 3H), 4.42
(s, 1H), 5.57–5.74 (2m, 2H), 6.14 (s, 1H), 6.21 (s,
1H), 6.33 (d, 1H, J=10Hz), 7.31 (d, 1H, J=10Hz);
MS(ES+) m/z 549 (M+H)+; Anal. Calcd for
C29H34ClFO7Æ0.1CH2Cl2: C, 62.69; H, 6.18. Found: C,
62.39; H, 6.12.
1
solid: mp 194ꢁC; Anal. HPLC tR 7.75 min, 99%; H
NMR (CDCl3) 0.90 (d, 3H, J=7Hz), 1.02 (s, 3H),
1.24–1.30 (m, 2H), 1.61–1.77 (m, 5H), 1.90 (q, 1H,
J=12Hz), 2.22–2.30 (m, 1H), 2.77 (t, 1H, J=10Hz),
3.66 (s, 3H), 4.40 (br s, 1H), 5.55–5.71 (2m, 1H), 6.11
(s, 1H), 6.30 (dd, 1H, J=2,10Hz), 6.75 (d, 1H,
J=4Hz), 7.26–7.30 (m, 2H); MS(ES+) m/z 555
(M+H)+; Anal. Calcd for C27H29Cl2FO7: C, 58.39; H,
5.26. Found: C, 58.42; H, 5.10.
6.21. 9a-Chloro-6a-fluoro-17a-[(3-furanylcarbonyl-4-
methyl)oxy]-11b-hydroxy-16a-methyl-3-oxoandrosta-1,4-
diene-17b-carboxylic acid, methyl ester (11m)
6.18. 9a-Chloro-6a-fluoro-11b-hydroxy-16a-methyl-17a-
[(2-furanylcarbonyl-5-trifluoromethyl)oxy]-3-oxoandro-
sta-1,4-diene-17b-carboxylic acid, methyl ester (11j)
Similarly to the procedure described for 11e, the title
compound was prepared starting from 8 and 4-methyl-
3-furancarboxylic acid.22 Recrystallisation from MeOH
afforded 11m (28% yield for three steps from 8) as a
beige solid: mp 181–182ꢁC dec; Anal. HPLC tR
7.72min, 99%; 1H NMR (DMSO-d6) 0.90 (d, 3H,
J=7Hz), 1.01 (s, 3H), 1.22–1.28 (m, 1H), 1.61 (s, 3H),
1.66–1.78 (m, 2H), 1.90 (q, 1H, J=12Hz), 2.10 (s,
3H), 2.21–2.29 (m, 1H), 2.40–2.48 (m, 2H), 2.74 (t,
1H, J=11Hz), 3.23–3.30 (m, 1H), 3.65 (s, 3H), 4.39
(br s, 1H), 5.55–5.71 (m, 2H), 6.11 (s, 1H), 6.30 (d,
1H, J=10Hz), 7.28 (d, 1H, J=11Hz), 7.61 (s, 1H),
8.11 (d, 1H, J=2Hz); MS(ES+) m/z 535 (M+H)+; Anal.
Calcd for C28H32ClFO7Æ0.4H2O: C, 62.08; H, 6.02.
Found: C, 62.07; H, 6.01.
Similarly to the procedure described for 11e, the title
compound was prepared starting from 8 and 5-trifluoro-
methyl-2-furancarboxylic acid. Column chromatogra-
phy eluting with CH2Cl2/MeOH (50:1), followed by
trituration with MeOH, afforded 11j (4% yield for three
steps from 8) as a white solid: mp 196ꢁC; Anal. HPLC
1
tR 8.00min, 97%; H NMR (DMSO-d6) 0.89 (d, 3H,
J=7Hz), 1.0 (s, 3H), 1.18–1.26 (m, 1H), 1.59–1.74 (m,
5H), 1.89 (q, 1H, J=12Hz), 2.18–2.28 (m, 1H), 2.75 (t,
1H, J=8Hz), 3.65 (s, 3H), 4.38 (br s, 1H), 5.52–5.70
(m, 2H), 6.09 (s, 1H), 6.28 (dd, 1H, J=2,10Hz), 7.26
(d, 1H, J=10Hz), 7.33 (d, 1H, J=3Hz), 7.41 (d, 1H,
J=3Hz); MS(ES+) m/z 589 (M+H)+; HR-TOFMS m/z
(M+H)+; HRMS Calcd for C28H29ClF4O7: 589.1616.
Found: 589.1616.
6.22. 9a-Chloro-6a-fluoro-11b-hydroxy-17a-[(3-furanyl-
carbonyl-5-methyl)oxy]-16a-methyl-3-oxoandrosta-1,4-
diene-17b-carboxylic acid, methyl ester (11n)
6.19. 9a-Chloro-6a-fluoro-11b-hydroxy-17a-[(2-methyl-3-
furanylcarbonyl)oxy]-16a-methyl-3-oxoandrosta-1,4-di-
ene-17b-carboxylic acid, methyl ester (11k)
Similarly to the procedure described for 11e, the title
compound was prepared starting from 8 and 5-methyl-
3-furancarboxylic acid.21 Trituration with MeOH af-
forded 11n (22% yield for three steps from 8) as an
off-white solid: mp 177–181ꢁC dec; Anal. HPLC tR
7.29min, 99%; 1H NMR (DMSO-d6) 0.79 (d, 3H,
J=7Hz), 0.91 (s, 3H), 1.10–1.17 (m, 2H), 1.55–1.72
(m, 5H), 1.82 (q, 1H, J=8Hz), 2.24 (s, 3H), 2.49 (t,
1H, J=12Hz), 2.97–3.22 (m, 3H), 3.70 (s, 3H), 4.53
(br s, 1H), 5.74–5.92 (2m, 2H), 6.34 (s, 1H), 6.54 (s,
1H), 6.56 (d, 1H, J=2Hz), 7.58 (d, 1H, J=10Hz),
8.43 (s, 1H); MS(ES+) m/z 535 (M+H)+; Anal. Calcd
for C28H32ClFO7: C, 62.86; H, 6.03. Found: C, 63.12;
H, 6.16.
Similarly to the procedure described for 11a, the title
compound was prepared starting from 8 and 2-methyl-
3-furancarboxylic acid.20 Trituration with CH2Cl2 and
MeOH afforded 11k (61% yield for three steps from 8)
as a white solid: mp 195–196ꢁC; Anal. HPLC tR
7.22min, 96%; 1H NMR (DMSO-d6) 0.91 (d, 3H,
J=7Hz), 1.10 (s, 3H), 1.20–1.35 (m, 1H), 1.60–1.78
(m, 5H), 1.79–1.95 (m, 2H), 2.22–2.34 (m, 1H), 2.49 (s,
3H), 2.80 (t, 1H, J=10Hz), 3.68 (s, 3H), 4.43 (br s,
1H), 5.58–5.71 (m, 2H), 6.14 (s, 1H), 6.32 (d, 1H,
J=10Hz), 6.61 (d, 1H, J=2Hz), 7.31 (d, 1H,
J=10Hz), 7.62 (d, 1H, J=2Hz); MS(ES+) m/z 535
(M+H)+; Anal. Calcd for C28H32ClFO7Æ0.6H2O: C,
61.56; H, 6.02. Found: C, 61.59; H, 6.19.
6.23. 9a-Chloro-6a-fluoro-17a-[(2-furanylcarbonyl)oxy]-
11b-hydroxy-16a-methyl-3-oxoandrost-4-ene-17b-carb-
oxylic acid, methyl ester (12)
6.20. 9a-Chloro-17a-[(2,5-dimethyl-3-furanylcarbon-
yl)oxy]-6a-fluoro-11b-hydroxy-16a-methyl-3-oxoandro-
sta-1,4-diene-17b-carboxylic acid, methyl ester (11l)
A suspension of 11a (670mg, 1.29mmol) and tris(triphen-
ylphosphine)chlororhodium (300mg, 0.32mmol) in
EtOH (40mL) and DMF (10mL) was hydrogenated at
3.5bar at 40ꢁC for 72h. A further 0.2equiv tris(triphenyl-
phosphine)chlororhodium was added after 24h to drive
the reaction to completion. The reaction mixture was
concentrated to dryness and the brown residue obtained
was suspended in 50:1 CH2Cl2/MeOH, then eluted
through a silica gel filter column. The filtrate obtained
was concentrated to dryness, then successive trituration
Similarly to the procedure described for 11a, the title
compound was prepared starting from 8 and 2,5-di-
methyl-3-furancarboxylic acid. Column chromatogra-
phy eluting with CH2Cl2/MeOH (50:1) afforded 11l
(25% yield for three steps from 8) as a pale yellow solid:
1
mp 173–175ꢁC dec; Anal. HPLC tR 8.19min, 89%; H

5222
D. A. Sandham et al. / Bioorg. Med. Chem. 12 (2004) 5213–5224
with DMF (5mL), MeOH (5mL) and MeCN (5mL) gave
a white precipitate, which was filtered, washed with
MeOH (10mL) and dried in vacuo to afford 12 (185mg,
28%) as a white solid: mp 169ꢁC; Anal. HPLC tR
7.09min, 96%; 1H NMR (CDCl3) 0.90 (d, 3H, J=7Hz),
0.94 (s, 3H), 1.18–1.23 (m, 1H), 1.35 (d, 1H, J=3Hz),
1.51–1.56 (m, 4H), 1.73 (q, 1H, J=12Hz), 1.83–1.99 (m,
2H), 2.07–2.14 (m, 1H), 2.32–2.49 (m, 3H), 2.52–2.60
(m, 1H), 2.65–2.75 (m, 2H), 3.24–3.32 (m, 1H), 3.77 (s,
3H), 4.38 (br s, 1H), 5.06–5.22 (m, 1H), 6.01 (s, 1H),
6.39 (d, 1H, J=3Hz), 7.04 (d, 1H, J=3Hz), 7.46 (s,
1H); MS(ES+) m/z 523 (M+H)+; Anal. Calcd for
C27H32ClFO7Æ0.4H2O: C, 61.11; H, 6.11. Found: C,
61.11; H, 6.08.
poured into a mixture of 2M HCl and ice (50mL).
The resulting precipitate was filtered and dried in vacuo
to give 15 (440mg, 85%) as an off white solid: 1H NMR
(DMSO-d6) 0.83 (d, 3H, J=7Hz), 0.97 (s, 3H), 1.07–
1.13 (m, 1H), 1.58–1.81 (m, 6H), 2.17–2.24 (m,
1H), 2.30–2.49 (m, 2H), 2.64 (t, 1H, J=9Hz), 2.88–
2.97 (m, 1H), 4.36 (br s, 1H), 5.53–5.70 (m, 2H),
6.10 (s, 1H), 6.29 (d, 1H, J=10Hz), 7.28 (d, 1H,
J=10Hz).
6.27. 9a-Chloro-6a-fluoro-17a-[(2-furanylcarbonyl)oxy]-
11b-hydroxy-16a-methyl-3-oxoandrosta-1,4-diene-17b-
carbothioic acid, S-(chloromethyl) ester (17)
To a solution of 15 (2.0g, 4.67mmol) in pyridine (6mL)
was added dropwise at 25ꢁC 2-furoyl chloride (640mg,
4.90mmol). After 10min, a solid formed and the
reaction mixture was then partitioned between
EtOAc (100mL) and 1M aq HCl (100mL). The
organic layer was washed with water, brine, dried
(MgSO4), then filtered and concentrated to yield 16
(1.3g, 53%), which was used crude without further
purification.
6.24. 9a-Chloro-11b,17a-dihydroxy-6a-fluoro-16a-
methyl-3-oxoandrosta-1,4-diene-17b-carboxylic acid (13)
HCl gas was bubbled through a solution of 8 (5.0g,
13.3mmol) in dioxane (250mL) over 5min. The reaction
flask was stoppered, stirred at 25ꢁC for 18h and concen-
trated to give a white residue. Trituration with EtOAc,
followed by recrystallisation from MeOH afforded 13
(2.8g, 51%) as a white solid: 1H NMR (DMSO-d6)
0.88 (d, 3H, J=7Hz), 1.0 (s, 3H), 1.05–1.11 (m, 1H),
1.52 (d, 1H, J=13Hz), 1.60–1.75 (m, 5H), 2.17–2.23
(m, 1H), 2.30–2.37 (m, 2H), 2.65 (t, 1H, J=8Hz),
2.82–2.87 (m, 1H), 3.32 (s, 1H), 4.33 (br s, 1H), 5.47–
5.69 (m, 1H), 6.09 (s, 1H), 6.28 (dd, 1H, J=2,10Hz),
7.27 (d, 1H, J=10Hz).
To a solution of 16 (1.0g, 1.91mmol) in dimethylacet-
amide (10mL) was added NaHCO3 (320mg,
3.83mmol), followed by bromochloromethane (990mg,
7.66mmol). The reaction mixture was stirred at 25ꢁC
for 2h, then diluted with EtOAc (100mL), washed with
satd NaHCO3 solution, followed by water and brine
then dried (MgSO4). Filtration followed by concentra-
tion afforded 17 (0.95g, 87%): 1H NMR (DMSO-d6)
0.96 (d, 3H, J=7Hz), 1.09 (s, 3H), 1.28–1.34 (m, 1H),
1.62 (s, 3H), 1.64–1.84 (m, 2H), 1.90 (q, 1H, J=6Hz),
2.23–2.30 (m, 1H), 2.53–2.61 (m, 2H), 2.80 (t, 1H,
J=9Hz), 3.35–3.39 (m, 1H), 4.68 (br s, 1H), 5.20 (s,
2H), 5.55–5.71 (m, 1H), 5.75 (d, 1H, J=4Hz), 6.12 (s,
1H), 6.31 (d, 1H, J=10Hz), 6.71 (d, 1H, J=3Hz),
7.23 (d, 1H, J=3Hz), 7.28 (d, 1H, J=11Hz), 8.01 (s,
1H).
6.25. 9a-Chloro-6a-fluoro-17a-[(2-furanylcarbonyl)oxy]-
11b-hydroxy-16a-methyl-3-oxoandrosta-1,4-diene-17b-
carboxylic acid (14)
To a stirred solution of 13 (500mg, 1.21mmol) in pyr-
idine (2mL) at 25ꢁC was added dropwise 2-furoyl chlo-
ride (0.166g, 1.27mmol). The reaction mixture was
stirred at 25ꢁC for 1h. The reaction mixture was then
added dropwise to a vigorously stirred solution of 6M
HCl (10mL). After 30min, the resulting precipitate
was filtered, washed with THF (10mL), followed by di-
ethyl ether (10mL), then dried in vacuo to give 14
(613mg, 98%) as an off-white solid: mp 175ꢁC; Anal.
HPLC tR 5.77min, 94%; 1H NMR (DMSO-d6) 0.91
(d, 3H, J=Hz), 1.07 (s, 3H), 1.20–1.26 (m, 1H), 1.62–
1.77 (m, 5H), 1.89 (q, 1H, J=12Hz), 2.23–2.30 (m,
1H), 2.78 (t, 1H, J=10Hz), 3.23–3.28 (m, 2H), 4.41
(br s, 1H), 5.56–5.72 (m, 2H), 6.12 (s, 1H), 6.31 (d,
1H, J=10Hz), 6.68 (d, 1H, J=3Hz), 7.16 (d, 1H,
J=3Hz), 7.29 (d, 1H, J=10Hz), 7.98 (s, 1H); MS(ES+)
m/z 507 (M+H)+; Anal. Calcd for C26H28ClFO7Æ0.7-
H2O: C, 60.16; H, 5.57. Found: C, 60.16; H, 5.66.
6.28. 9a-Chloro-6a-fluoro-17a-[(2-furanylcarbonyl)oxy]-
11b-hydroxy-16a-methyl-3-oxoandrosta-1,4-diene-17b-
carbothioic acid, S-(iodomethyl) ester (18)
To a solution of 17 (700mg, 1.22mmol) in acetone
(20mL) at 25ꢁC was added NaI (2.2g, 14.69mmol)
and the reaction mixture was heated at reflux for 6h.
After this time the reaction mixture was diluted with
EtOAc (100mL), washed successively with water, 10%
aq Na2S2O3 solution, satd aq NaHCO3 solution, water
and brine. The organic layer was dried (MgSO4), filtered
and concentrated. Trituration of the residue with
1
MeOH afforded 17 (704mg, 87%): H NMR (DMSO-
6.26. 9a-Chloro-11b,17a-dihydroxy-6a-fluoro-16a-
methyl-3-oxoandrosta-1,4-diene-17b-carbothioic acid (15)
d6) 0.95 (d, 3H, J=7Hz), 1.08 (s, 3H), 1.28–1.33
(m, 1H), 1.61 (s, 3H), 1.63–1.80 (m, 2H), 1.97 (q, 1H,
J=12Hz), 2.22–2.29 (m, 1H), 2.47–2.59 (m, 2H),
2.79 (t, 1H, J=9Hz), 4.43 (br s, 1H), 4.55–4.58 (m,
2H), 5.55–5.71 (m, 1H), 5.79 (d, 1H, J=5Hz), 6.12 (s,
1H), 6.31 (d, 1H, J=10Hz), 6.71 (d, 1H, J=3Hz),
7.21 (d, 1H, J=3Hz), 7.28 (d, 1H, J=10Hz), 8.01 (s,
1H).
To a solution of 13 (500mg, 1.21mmol) in DMF (2mL)
was added carbonyl diimidazole (393mg, 2.42mmol)
and the reaction mixture was stirred at 25ꢁC for 4h.
NaSH (272mg, 4.85mmol) was then added and stirring
continued at 25ꢁC for 16h. The reaction mixture was

D. A. Sandham et al. / Bioorg. Med. Chem. 12 (2004) 5213–5224
5223
6.29. 9a-Chloro-6a-fluoro-17a-[(2-furanylcarbonyl)oxy]-
11b-hydroxy-16a-methyl-3-oxoandrosta-1,4-diene-17b-
carbothioic acid, S-(fluoromethyl) ester (19)
standard recombinant TNFa. Curves were fitted and
IC50 values calculated using Origine software.
8.1.2. Human glucocorticoid receptor (GR) binding assay.
Recombinant human GR expressed in baculovirus–in-
fected insect Sf-9 cells was obtained from Panvera
(Madison, WI, USA), as was GR assay buffer and the
proprietary fluorescent ligand Fluormonee-GS1
(200nM methanol solution).
Finely ground AgF (200mg, 1.88mmol) was added to a
dispersion of 18 (250mg, 0.38mmol) in acetonitrile
(10mL). The reaction mixture was stirred at 25ꢁC for
2h in the dark, then diluted with EtOAc (100mL) and
filtered through Celitee. The filtrate was concentrated
and the residue purified by flash chromatography on sil-
ica gel using EtOAc/n-hexane 1:2 as the eluent. Tritura-
tion with i-Pr2O afforded 19 (75mg, 35%): mp 176–
The assay was conducted in 384 well plates by sequential
addition of a serially diluted DMSO solution of test
compound in water (2lL), Fluormonee-GS1 (2.2nM
in GR assay buffer, 10lL) and GR solution (8.8nM in
GR assay buffer, 10lL). The assay was incubated in
the dark at room temperature for 1h, prior to fluores-
cence polarisation measurement using an analyst multi-
well instrument with 485nm excitation and 530nm
emission filters. The concentration of test compound
resulting in half-maximum shift in polarisation gave
the IC50. Curves were fitted using Origine software
and Ki values were calculated using the Cheng–Prussoff
equation.
1
177ꢁC dec; Anal. HPLC tR 7.35min, 99%; H NMR
(DMSO-d6) 0.95 (d, 3H, J=7Hz), 1.07 (s, 3H), 1.28–
1.33 (m, 1H), 1.61 (m, 3H), 1.63–1.77 (m, 1H), 1.84 (d,
1H, J=14Hz), 1.97 (q, 1H, J=12Hz), 2.23–2.29 (m,
1H), 2.50–2.67 (m, 2H), 2.76–2.81 (m, 1H), 3.31–3.42
(m, 1H), 4.44 (br s, 1H), 5.55–5.72 (2m, 1H), 5.76 (d,
1H, J=5Hz), 5.89 (s, 1H), 6.02 (s, 1H), 6.12 (s, 1H),
6.31 (d, 1H, J=10Hz), 6.72 (dd, 1H, J=3,5Hz), 7.24
(d, 1H, J=3Hz), 7.28 (d, 1H, J=10Hz), 8.02 (s, 1H);
MS(ES+) m/z 555 (M+H)+; Anal. Calcd for
C27H29ClF2O6 S: C, 58.43; H, 5.27; S, 5.78. Found: C,
58.33; H, 5.41; S, 5.57.
8.1.3. Stability in plasma. A 10lM solution of 11a in rat
or human plasma was incubated at 37ꢁC for 1h. The
sample was analysed by HPLC after precipitation by
addition of acetonitrile.
7. Determination of thermodynamic solubilities
Thermodynamic solubility was determined in pH6.8
phosphate buffer using the saturation shake-flask meth-
od as described in the literature.23
8.2. In vivo methods
8.2.1. Preparation of dry powder blends. Approximately
100mg of test compound was ball milled in a 1.5mL
agate grinding jar fitted with one 7mm agate ball using
a MM200 mixer mill (Glen Creston) for 20min at 30Hz.
The milled powder was passed through a 212lm sieve.
8. Biological assay methods
8.1. In vitro methods
8.1.1. Inhibition of LPS induced TNFa release. Human
macrophage cell line U937 was obtained from American
Type Culture Collection (Rockville MD) and cultured in
RPMI 1640 (Gibco UK) supplemented with 10% FCS
(Gibco UK). Cell density was adjusted to 4·105 cells/
mL and the cells were differentiated by adding PMA
(20ng/mL) for 4h. The PMA was removed by washing
and the adherent cells were incubated for a further
48h at 37ꢁC in a humidified incubator with 5% CO2.
Differentiated U937 cells were removed using cell disso-
ciation buffer (Gibco UK) and the cell density was ad-
justed to 1·106 cells/mL. 100 lL of the cell suspension
was placed in 96 well culture plates and 50lL of either
medium or compound at the appropriate concentration
in DMSO were added. After a preincubation of 20min
at 37ꢁC, the cells were stimulated with 10ng/mL LPS
(Sigma) and the supernatants were harvested after 24h
of incubation at 37ꢁC in a humidified incubator with
5% CO2. Concentration of TNFa in the supernatants
was determined by sandwich ELISA using two mono-
clonal antibodies recognising different epitopes of the
cytokine (Pharmingen UK). Binding of the second anti-
body was analysed by stepwise incubation with strept-
avidin alkaline phosphatase conjugate (Sigma UK)
and 4-nitrophenylphosphate disodium salt. Optical den-
sity was measured at 405nm and cytokine concentration
calculated based on results from serial diluations of
This material was then sandwiched between 2 aliquots
of 325# lactose monohydrate (DMV) in a glass vessel.
The mixture was tumble mixed (Turbula T2A) for
20min at 50rpm and subsequently sieved through a
212lm sieve. This procedure was repeated a further
two times and then aliquots (5·1mg) removed and
assayed by HPLC to establish blend uniformity.
8.2.2. Thymus involution studies. Male Sprague-Dawley
CD1 rats (250–265g body weight, n=5 to 8) were dosed
once daily over 4days either orally with compounds in
Klucel suspension or i.t. as lactose blends under haloth-
ane anaesthesia. On day 5, animals were sacrificed by
i.p. injection of pentobarbital and the thymus was re-
moved and weighed. The percentage reduction in mean
weights was calculated from the mean weights of vehicle
treated animals. Significant differences were assessed by
use of the SigmaState package using the appropriate
ANOVA analysis and appropriate subsequent analysis,
which varied according to whether the data passed the
Normality and Equal Variance tests.
8.2.3. Inhibition of OA-induced eosinophilia. Male Brown
Norway rats (approximately 200g body weight, n=5 to
8) were sensitised on day 1 with an i.p. injection of
0.5mL of a mixture of ovalbumin (0.02mg/ml) and

5224
D. A. Sandham et al. / Bioorg. Med. Chem. 12 (2004) 5213–5224
4. Hoegger, P. Curr. Med. Chem.: Anti-Inflammatory Anti-
Allergy Agents 2003, 2, 395–408.
5. Derendorf, H.; Hochhaus, G.; Meibohm, B.; Moellmann,
H.; Barth, J. J. Allergy Clin. Immunol. 1998, 101, S440.
6. Staresinic, A. G.; Sorkness, C. A. Exp. Opin. Pharmac-
other. 2000, 1, 1227.
aluminium hydroxide (20mg/mL), followed by Acellul-
are pertussis adsorbat vaccine (0.2mL of a 1:4 dilution
with 0.9% saline). The procedure was repeated on day
15 and day 21. On day 28, the test compound as a dry
powder lactose blend was administered i.t. under isoflu-
rane anaesthesia. Twenty four hours later, the sensitised
dosed animals were exposed to an aerosol of ovalbumin
(5mg/mL) for 60min and after a further 24h they were
sacrificed by i.p. injection of pentobarbital. The lungs
were removed and after lavage with Hankꢀs solution,
eosinophil numbers in the recovered solution were quan-
tified directly using a Corbas Helios 5Diff appara-
tus (Hoffman-LaRoche). Statistical significance was
assessed with Studentꢀs t-test and the Hommel–Hoch-
berg multiple comparison test.
7. Angell, R. M.; Biggadike, K.; Farrell, R. M.; Flack, S. S.;
Hancock, A. P.; Irving, W. R.; Lynn, S. M.; Procopiou,
P. A. J. Chem. Soc., Perkin Trans. 1 2002, 831.
8. Sharpe, M.; Jarvis, B. Drugs 2001, 61, 1325.
9. Bodor, N.; Buchwald, P. Med. Res. Rev. 2000, 20, 58.
10. Belvisi, M. G.; Hele, D. J. Pulm. Pharmacol. Ther. 2003,
16, 321.
11. Schmidlin, J. European Patent Application (Ciba-Geigy
AG) 135476, 1985. Chem. Abstr. 1985, 103, 105215.
12. Kertesz, D. J.; Marx, M. J. Org. Chem. 1986, 51, 2315.
13. Sajjadi, F. G.; Takabayashi, K.; Foster, A. C.; Domingo,
R. C.; Firestein, G. S. J. Immunol. 1996, 156, 3435.
14. Adcock, I. M. Pulm. Pharmacol. Ther. 2000, 13, 115.
15. Chung, K. F. In Allergy and Allergic Diseases; Kay, A. B.,
Ed.; Blackwell: London, 1997; pp 1068–1078.
16. Burgerꢀs Medicinal Chemistry and Drug Discovery, 6th ed.;
Avery, M. A., Woolfrey, J. R., Abraham, D. J., Eds.;
Wiley: New York, 2002; Vol. 3, pp 747–853.
17. Chanoine, F.; Grenot, C.; Heidmann, P.; Junien, J. L.
Drug Metab. Disp. 1991, 19, 546.
8.2.4. Oral absorption study. Male Wistar rats (approx
230g body weight, n=5) were cannulated at the carotid
artery under halothane anaesthesia. The test compound
was dosed orally at 4.6mg/kg, either a solution in 2% N-
methyl pyrrolidinone in Placebo Neoral or a Klucel sus-
pension. Blood samples were withdrawn at intervals for
analysis up 6h post dose and animals were sacrificed
after the final sampling by i.p. injection of pentobarbital.
18. OꢀHanlon, P. J.; Walker, G. European Patent Application
(SmithKlineBeecham) 399645, 1990. Chem. Abstr. 1991,
114, 228892.
19. Datta, A.; Pooranchand, D.; Ila, H.; Junjappa, H.
Tetrahedron 1989, 45, 7631.
References and notes
20. Knight, D. W.; Nott, A. P. J. Chem. Soc., Perkin Trans. 1
1981, 1125.
21. Knight, D. W. Tetrahedron Lett. 1979, 20, 469.
22. Ichikawa, Y.; Naganawa, A.; Isobe, M. Synlett 1993, 10,
737.
1. Lung Biology in Health and Disease; Schleimer, R. P.,
OꢀByrne, P. M., Szefler, S. J., Brattsand, R., Eds.; Marcel
Dekker: New York, 2002; Vol. 163.
2. Buchman, A. L. J. Clin. Gastroenterol. 2001, 33, 289.
3. Rabe, K. F.; Schmidt, D. T. Eur. Respir. J. 2001, 18(Suppl
34), 34S.
23. Avdeef, A.; Berger, C. M.; Brownell, C. Pharm. Res. 2000,
171, 85.