Synthesis and structure elucidation of potential 6-oxygenatedated metabolites of (22R)-6*9α-difluoro-11β,21-dihydroxy-16α,17α-propylmethylenedioxypregn-4 -ene-3,20-dione, and related glucocorticosteroids

Article abstract of DOI:10.1016/S0039-128X(99)00080-X

(22R)-6α,9α-Difluoro-11β,21-dihydroxy-16α,17α-propylmethylenedioxypregn- 4-ene-3,20-dione (rofleponide) is a synthetic glucocorticosteroid with high affinity for the rat thymus glucocorticoid receptor and a very high biotransformation rate demonstrated through incubation with a human liver S9 subcellular fraction. Because oxidation in the 6-position is an important metabolic pathway of glucocorticosteroids, the potential 6β-hydroxy and 6-oxo metabolites of rofleponide were synthesized to be used as reference compounds. Three alternative routes were used to reach the 6-hydroxy compound: (a) a one-step procedure involving allylic oxidation of rofleponide by selenium dioxide, (b) selenium dioxide oxidation of the corresponding 1,4-diene followed by selective 1,2-hydrogenation using Wilkinson's catalyst, and (c) autoxidation of a 3-methoxypregna-3,5-diene derivative. All three routes proceeded stereospecifically. Routes (a) and (c) gave approximately the same overall yield of the 6β-hydroxy epimer, whereas the overall yield from route (b) was much lower, primarily because of incomplete 1,2-hydrogenation. The 6-oxo compound was prepared through Pfitzner/Moffat oxidation of the 6-hydroxy compound. The stereochemistry of the 6-hydroxy substituent is discussed on the basis of ¹H-NMR spectroscopy and supplementary 2D NOESY experiments. Copyright (C) 2000 Elsevier Science Inc.

Full text of DOI:10.1016/S0039-128X(99)00080-X

Steroids 65 (2000) 16–23
Synthesis and structure elucidation of potential 6-oxygenated
metabolites of (22R)-6␣,9␣-difluoro-11␤,21-dihydroxy-16␣,17␣-
propylmethylenedioxypregn-4-ene-3,20-dione, and related
glucocorticosteroids
Arne Thale´n*, Lars-Inge Wickstro¨m
Department of Medicinal Chemistry, Preclinical R & D, Research Support, Astra Draco AB, S-221 87 Lund, Sweden
Received 31 March 1999; received in revised form 1 June 1999; accepted 3 June 1999
Abstract
(22R)-6␣,9␣-Difluoro-11␤,21-dihydroxy-16␣,17␣-propylmethylenedioxypregn-4-ene-3,20-dione (rofleponide) is a synthetic glu-
cocorticosteroid with high affinity for the rat thymus glucocorticoid receptor and a very high biotransformation rate demonstrated
through incubation with a human liver S9 subcellular fraction. Because oxidation in the 6-position is an important metabolic pathway
of glucocorticosteroids, the potential 6␤-hydroxy and 6-oxo metabolites of rofleponide were synthesized to be used as reference
compounds. Three alternative routes were used to reach the 6-hydroxy compound: (a) a one-step procedure involving allylic oxidation
of rofleponide by selenium dioxide, (b) selenium dioxide oxidation of the corresponding 1,4-diene followed by selective 1,2-
hydrogenation using Wilkinson’s catalyst, and (c) autoxidation of a 3-methoxypregna-3,5-diene derivative. All three routes proceeded
stereospecifically. Routes (a) and (c) gave approximately the same overall yield of the 6␤-hydroxy epimer, whereas the overall yield
from route (b) was much lower, primarily because of incomplete 1,2-hydrogenation. The 6-oxo compound was prepared through
Pfitzner/Moffat oxidation of the 6-hydroxy compound. The stereochemistry of the 6-hydroxy substituent is discussed on the basis of
1H-NMR spectroscopy and supplementary 2D NOESY experiments. © 2000 Elsevier Science Inc. All rights reserved.
Keywords: Synthesis; Corticosteroids; Rofleponide 6-hydroxylation; Rofleponide 6-oxogenation; 1H-NMR; 2D NOESY
1. Introduction
e.g. through enzymatic hepatic oxidation in the 6-posi-
tion similar to that of cortisol [3] and other endogenous
steroids as well as several synthetic glucocorticosteroids
[4–9]. Because, as a rule, 6-oxygenated metabolites have
considerably lower glucocorticoid activities than the par-
ent compounds [5,10,11], this biotransformation pathway
is also expected to result in significant inactivation. Thus,
the 6␤-hydroxy (7R) and 6-oxo (12R) compounds were
needed as reference substances for the proper identifica-
tion and for the evaluation of glucocorticoid activitiy of
the potential metabolites of rofleponide.
The present paper describes several routes for synthesis
of (22R)-9␣-fluoro-6␤,11␤,21-trihydroxy-16␣,17␣-propyl-
methylenedioxypregn-4-ene-3,20-dione (7R), the syntheses
of some structurally related 6-hydroxylated glucocorticoster-
oids, and the synthesis of (22R)-9␣-fluoro-11␤,21-dihydr-
oxy-16␣,17␣-propylmethylenedioxypregn-4-ene-3,6,20-tri-
one (12R).
(22R)-6␣,9␣-Difluoro-11␤,21-dihydroxy-16␣,17␣-pro-
pylmethylenedioxypregn-4-ene-3,20-dione (rofleponide;
Fig. 1) is a synthetic glucocorticosteroid suitable for
local/topical therapy of diseases such as asthma, rhinitis,
and inflammatory bowel disease [1]. The high affinity of
rofleponide for the rat thymus glucocorticoid receptor [1]
has implied its high anti-inflammatory potency, which
has been further confirmed in vivo (unpublished data) in
a model for Sephadex bead-induced inflammation and
edema formation in the rat lung [2]. The high biotrans-
formation rate of rofleponide has been demonstrated
through incubation with human liver S9 subcellular frac-
tion [1]. Rofleponide is expected to undergo metabolism
* Corresponding author. Tel.: ϩ46-4633-6000; fax: ϩ46-4633-6666.
0039-128X/00/$ – see front matter © 2000 Elsevier Science Inc. All rights reserved.
PII: S0039-128X(99)00080-X

A. Thale´n, L.-I. Wickstro¨m / Steroids 65 (2000) 16–23
17
the preparative column chromatography was carried out as
flash chromatography on Merck Silica gel 60 (230–400
mesh) with manual fraction collection. Preparative, centrifu-
gally accelerated, radial TLC was conducted on a Chroma-
totron apparatus from Harrison Research.
The HPLC-analyses were performed on a liquid chro-
matograph from Waters Associates equipped with a
M6000A pump, a U6K injector system, and a M400 UV
detector (240 nm). A column (150 ϫ 4.6 mm I.D.), pre-
packed with 3 ␮m Apex octadecylsilane (Jones Chromato-
graphy), was used as the stationary phase, unless otherwise
stated.
2.1. 9␣-Fluoro-11␤,16␣,17␣,21-tetrahydroxypregn-4-ene-
3,20-dione (2)
Fig. 1. Rofleponide.
2. Experimental
A suspension of 9␣-fluoro-11␤,16␣,17␣,21-tetrahydr-
oxypregna-1,4-diene-3,20-dione (1, 5.0 g) in absolute etha-
nol (500 ml) was added to a solution of tris(triphenylphos-
phine)rhodium chloride (3.0 g) in toluene (1000 ml) and
hydrogenated at room temperature and atmospheric press-
ure for 48 h. The reaction mixture was evaporated to dry-
ness and suspended in methylene chloride (50 ml). The solid
was filtered and washed with small portions of methylene
chloride, yielding 4.4 g (88%) of 2. MS (DCI): m/z (relative
intensity) 397 (MHϩ; 18), 379 (MHϩ Ϫ H2O; 15), 377
(MHϩ Ϫ HF; 22). 1H-NMR (300 MHz): ␦ ppm (DMSO-d6)
0.83 (s, 3H, H-18); 1.49 (s, 3H, H-19); 4.13 (m, 1H, H-11);
4.11 and 4.53 (two q, 2H, H-21); 5.41 (d, 1H, H-16); 5.68
(s, 1H, H-4).
Triamcinolone(9␣-fluoro-11␤,16␣,17␣,21-tetrahydroxy-
pregna-1,4-diene-3,20-dione) was purchased from SICOR
S.p.A. (Milan, Italy).
1H-NMR spectra were recorded with solutions prepared
in CDCl3 at ambient temperature on a 300 MHz Varian
VXR-S spectrometer. The chemical shifts are given in ␦
units (ppm) relative to the internal standard tetramethylsil-
ane; d ϭ doublet, dd ϭ double doublet, m ϭ multiplet, q ϭ
quartet, and s ϭ singlet. Supplementary 2D NOESY exper-
iments were recorded on a 500 MHz Varian Unityϩ spec-
trometer using the standard Varian 2D NOESY sequence. A
mixing time of 400 ms and 256 increments were used,
resulting, after zerofilling in f1, in a 2048 ϫ 1024 2D data
matrix. (The abbreviations used are: NOE ϭ Nuclear Over-
hauser Effect; NOESY ϭ Nuclear Overhauser Spec-
troscopy; 2D ϭ 2 Dimensional; f1 ϭ frequency dimension
1, indirect frequency dimension in 2D NOESY experiment).
Mass spectra were recorded on a Finnigan 4510 spec-
trometer with desorption chemical ionization (DCI) using
methane as the reagent gas (direct inlet; filament current
was increased at a rate of 10 mA/s). Alternatively, the mass
spectra were obtained with liquid chromatography thermo-
spray mass spectrometry (TSP-MS) on the same type of
spectrometer equipped with a Finnigan thermospray inter-
face. Mobile phase: 0.1M ammonium acetate buffer, pH 5,
containing 70% methanol. Temperatures: Ion source 222°C
and vaporizer 103°C. Repeller voltage: 90V.
2.2. (22RS)-9␣-Fluoro-11␤,21-dihydroxy-16␣,17␣-pro-
pylmethylenedioxypregn-4-ene-3,20-dione (3)
Freshly distilled butanal (815 mg) and 2 (2.96 g) were
added to a solution of p-toluenesulfonic acid (870 mg) in
acetonitrile (40 ml). The reaction mixture was stirred at
room temperature under argon protection for 3 h. A satu-
rated aqueous solution of sodium hydrogen carbonate (2 ml)
was added to stop the reaction. Methylene chloride (800 ml)
was added, and the solution was washed with a saturated
aqueous solution of sodium hydrogen carbonate and water
to neutral, dried with sodium sulfate, and evaporated. The
residue was purified on a Sephadex LH-20 column (75 ϫ
6.3 cm i.d.) using chloroform as the mobile phase. The
fraction 2310–3105 ml was collected and evaporated. The
residue was precipitated from methylene chloride Ϫ petro-
leum ether yielding 2.66 g (79%) of compound 3. The purity
determined by HPLC-analysis (ethanol/water, 42/58) was
98%, and the 22R/22S ratio was 60/40. MS (DCI): m/z 451
Preparative column chromatography was performed on a
Quickfit glass column equipped with adjustable Teflon end
pieces. A LKB Uvicord I flow analyzer, working at 254 nm,
served as the detection system. The effluent fractions were
collected on a LKB 7000 Ultro Rac automatic fraction
collector equipped with a LKB 3404 B siphon stand, using
a 15 ml siphon. Sephadex LH-20, particle size 25–100 ␮m
(Pharmacia Fine Chemicals, Uppsala, Sweden), was used as
the stationary phase. All solvents used as the mobile phase
were of puriss grade and glass distilled. The ethanol used in
the mixed solvent system was 99.5% pure. Alternatively,
1
(MHϩ; most abundant ion). H-NMR (300 MHz): ␦ ppm
0.90 and 0.96 (two s, 3H, H-18); 0.92 and 0.94 (two t, 3H,
H-25); 1.53 (s, 1H, H-19); 4.38 (m, 1H, H-11); 4.21, 4.27,
4.54 and 4.65 (four q, 2H, H-21); 4.60 and 5.24 (two t, 1H,
H-22); 4.92 and 5.20 (two d, 1H, H-16); 5.80 (m, 1H, H-4).

18
A. Thale´n, L.-I. Wickstro¨m / Steroids 65 (2000) 16–23
2.3. (22R)- and (22S)-21-Acetoxy-9␣-fluoro-11␤-hydroxy-
16␣,17␣-propylmethylenedioxypregn-4-ene-3,20-dione (4R
and 4S)
was 95.6%. 1H-NMR (300 MHz): ␦ ppm 0.92 (s, 3H, H-18);
0.93 (t, 3H, H-25); 1.28 (s, 3H, H-19); 2.18 (s, 3H,
CH3COO); 3.58 (s, 3H, OCH3); 4.40 (m, 1H, H-11); 4.68 (t,
1H, H-22); 4.80 and 4.91 (two d, 2H, H-21); 4.88 (d, 1H,
H-16); 5.17 (d, 1H, H-4); 5.22 (t, 1H, H-6).
A solution of compound 3 (2.66 g) and acetic anhydride
(25 ml) in freshly distilled and dry pyridine (25 ml) was
stirred at room temperature for 1 h. The reaction mixture
was poured into ice water (150 ml) and extracted with
methylene chloride (3 ϫ 100 ml). The extract was serially
washed with hydrochloric acid (1M), saturated aqueous
sodium hydrogen carbonate, and water. The residue ob-
tained after evaporation of the dried extract was purified by
chromatography on a Sephadex LH-20 column (75 ϫ 6.3
cm i.d.) using chloroform as the mobile phase. The fraction
1350–1950 ml was collected, leaving a 22R/22S-mixture.
The C-22 epimers were separated on a Sephadex LH-20
column (76 ϫ 6.3 cm i.d.) with a mixture of heptane/
chloroform/ethanol (20/20/1) as the mobile phase.
2.5. (22R)-21-Acetoxy-9␣-fluoro-6␤,11␤-dihydroxy-
16␣,17␣-propylmethylenedioxypregn-4-ene-3,20-dione
(6R)
A solution of 5R (600 mg) in ethanol (25 ml) was stirred
and exposed to the sun for 10 h. The reaction mixture was
evaporated, and the residue purified by flash chromatogra-
phy on a silica gel column (28 ϫ 2 cm i.d.) using ethyl
acetate/heptane (50/50) as the mobile phase, leaving 323 mg
(54%) of 6R. The purity determined by HPLC (ethanol/
water, 45/55) was 98.6%. MS: m/z (relative intensity) 509
(MHϩ; 100); 491 (MHϩ Ϫ H2O; 10) (MS-FAB). 1H-NMR
(300 MHz): ␦ ppm 0.94 (t, 3H, H-25); 0.94 (s, 3H, H-18);
1.72 (s, 3H, H-19); 2.18 (s, 3H, CH3COO); 4.39 (broad m,
2H, H-6 and H-11); 4.66 (t, 1H, H-22); 4.74 and 4.94 (two
d, 2H, H-21); 4.87 (d, 1H, H-16); 5.90 (s, 1H, H-4).
After evaporation and precipitation from methylene
chloride Ϫ petroleum ether fraction 1500–1815 ml left 900
mg (31%) of the 22S-epimer 4S. The purity determined by
HPLC (ethanol/water, 50/50) was 97%. MS (DCI): m/z 493
1
(MHϩ; most abundant ion). H-NMR (300 MHz): ␦ ppm
0.94 (t, 3H, H-25); 0.97 (s, 3H, H-18); 1.54 (s, 3H, H-19);
2.18 (s, 3H, CH3COO); 4.38 (m, 1H, H-11); 4.78 and 4.94
(two d, 2H, H-21); 5.16 (d, 1H, H-16); 5.24 (t, 1H, H-22);
5.80 (m, 1H, H-4).
2.6. (22R)-21-Acetoxy-9␣-fluoro-6␤,11␤-dihydroxy-
16␣,17␣-propylmethylenedioxypregn-4-ene-3,20-dione
(6R)
After evaporation and precipitation from methylene
chloride Ϫ petroleum ether fraction 1920–2175 ml left
1180 mg (41%) of the 22R-epimer 4R. The purity deter-
mined by HPLC as above was 98.8%. MS (DCI): m/z 493
A stirred suspension of 4R (25 mg) and selenium dioxide
(15 mg) in dioxane (5 ml) was heated at 100°C for 2 h. The
reaction mixture was chilled to room temperature and fil-
tered through a short column (3 ϫ 2 cm i.d.) of Silica gel 60.
The column was washed with dioxane (25 ml). The eluate
was evaporated, and the residue purified by flash chroma-
tography on a silica gel column (12 ϫ 2 cm i.d.) with ethyl
acetate/heptane (80/20) as the mobile phase. The fractions
25–39 ml (A) and 40–90 ml (B) were collected.
1
(MHϩ; most abundant ion). H-NMR (300 MHz): ␦ ppm
0.93 (s, 3H, H-18); 0.94 (t, 3H, H-25); 1.54 (s, 3H, H-19);
2.18 (s, 3H, CH3COO); 4.38 (m, 1H, H-11); 4.66 (t, 1H,
H-22); 4.76 and 4.92 (two d, 2H, H-21); 4.87 (d, 1H, H-16);
5.80 (m, 1H, H-4).
After evaporation, fraction A left 6 mg of solid material,
2.4. (22R)-21-Acetoxy-9␣-fluoro-11␤-hydroxy-3-methoxy-
16␣,17␣-propylmethylenedioxypregna-3,5-dien-20-one
(5R)
which was shown by 1H-NMR spectroscopy to be the start-
ing material 4R. H-NMR of the solid obtained after evap-
oration of fraction B was identical to compound 6R pre-
pared by autoxidation of 5R. Yield: 10 mg (39%). The
purity determined by HPLC (ethanol/water, 50/50) was
88%.
1
A solution of 4R (848 mg) in methylene chloride (13 ml),
dry dioxane (6 ml), trimethyl orthoformate (0.8 ml), and
absolute methanol (0.05 ml) was chilled in an ice-water
bath. Concentrated sulfuric acid (two drops) was added.
After 8 min, the reaction was terminated by the addition of
pyridine (eight drops). The reaction mixture was concen-
trated at reduced pressure to 1/3 of its volume. Water (10
ml) was added, and the mixture was extracted with ethyl
ether (4 ϫ 10 ml). The combined extracts were washed with
water (3 ϫ 15 ml), dried over sodium sulfate, and evapo-
rated. The residue was purified by flash chromatography on
a silica gel column (30 ϫ 2 cm i.d.) with ethyl acetate/
heptane (30/70) as the mobile phase. The fraction 100–180
ml was collected and evaporated yielding 610 mg (70%) of
5R. The purity determined by HPLC (ethanol/water, 60/40)
2.7. (22R)-9␣-Fluoro-6␤,11␤,21-trihydroxy-16␣,17␣-
propylmethylenedioxypregn-4-ene-3,20-dione (7R)
A solution of 6R (278 mg) in methanol (30 ml) was
agitated with a stream of nitrogen for a few minutes. A
solution of 10% aqueous potassium carbonate (2 ml) was
added. After stirring for 45 min at room temperature with
nitrogen protection, the reaction mixture was neutralized
with acetic acid and evaporated. After addition of water (50
ml), the residue was extracted with methylene chloride (4 ϫ
50 ml). The combined extracts were washed with water,
dried over sodium sulfate, and evaporated yielding 174 mg

A. Thale´n, L.-I. Wickstro¨m / Steroids 65 (2000) 16–23
19
of solid material. This crude product was applied to a short
silica gel column (6 ϫ 2 cm i.d.) and eluted with a mixture
of ethyl acetate/heptane, 70/30, (100 ml). After evaporation,
the residue was further purified by flash chromatography on
a silica gel column (28 ϫ 2 cm i.d.) with ethyl acetate/
heptane (70/30) as the mobile phase. The fraction 240–400
ml was collected and evaporated yielding 113 mg (44%) of
7R. The purity determined by HPLC (10 ␮m ␮Bondapak
C18 column, (Waters Associates), 150 ϫ 3.9 mm i.d.; eth-
anol/water, 35/65) was 98.6%. M.p. 125–129°C. MS: m/z
(relative intensity) 467 (MHϩ; 100); 449 (MHϩ Ϫ H2O; 9)
(MSϪFAB). 1H-NMR (300 MHz): ␦ ppm 0.91 (t, 3H,
H-25); 0.94 (s, 3H, H-18); 1.72 (s, 3H, H-19); 4.28 and 4.54
(two q, 2H, H-21); 4.40 (m, 2H, H-6 and H-11); 4.61 (t, 1H,
H-22); 4.93 (d, 1H, H-16); 5.90 (s, 1H, H-4).
H-21); 4.87 (d, 1H, H-16); 6.14 (m, 1H, H-4); 6.33 and 6.37
(two d, 1H, H-2); 7.20 (d, 1H, H-1).
2.11. (22R)-21-Acetoxy-9␣-fluoro-6␤,11␤-dihydroxy-
16␣,17␣-propylmethylenedioxypregna-1,4-diene-3,20-
dione (10R)
A stirred suspension of 9R (286 mg) and selenium diox-
ide (555 mg) in dioxane (25 ml) was heated at 100°C for
72 h. The reaction mixture was chilled to room temperature
and filtered through a short column (5 ϫ 2 cm i.d.) of Silica
gel 60. The column was washed with dioxane. The volume
of the combined eluates was reduced to a few millilitres, and
the residue was suspended in ethyl acetate (50 ml), filtered,
and evaporated leaving 330 mg of crude product. This
product was purified by flash chromatography on a silica gel
column (15 ϫ 2 cm i.d.) with ethyl acetate/heptane (60/40)
as the mobile phase. The fractions 15–20 ml (A) and 120–
345 ml (B) were collected and evaporated yielding 88 mg of
starting material 9R and 130 mg of 10R, respectively. The
latter product was further purified by flash chromatography
on a silica gel column (21 ϫ 2 cm i.d.) with ethyl acetate/
heptane (60/40) as the mobile phase. The fraction 180–450
ml was collected and evaporated, yielding 109 mg (37%) of
10R. The purity determined by HPLC (ethanol/water, 48/
52) was 93.7%. MS: m/z 507 (MHϩ; most abundant ion)
(MSϪFAB). 1H-NMR (300 MHz): ␦ ppm 0.92 (t, 3H,
H-25); 0.97 (s, 3H, H-18); 1.74 (s, 3H, H-19); 2.19 (s, 3H,
CH3COO); 4.46 (m, 1H, H-11); 4.56 (m, 1H, H-6); 4.65 (t,
1H, H-22); 4.71 and 4.95 (two d, 2H, H-21); 4.86 (d, 1H,
H-16); 6.22 (d, 1H, H-4); 6.34 (dd, 1H, H-2); 7.20 (d, 1H,
H-1).
2.8. (22RS)-9␣-Fluoro-11␤,21-dihydroxy-16␣,17␣-pro-
pylmethylenedioxypregna-1,4-diene-3,20-dione (8)
The title compound was prepared from 9␣-fluoro-
11␤,16␣,17␣,21-tetrahydroxypregna-1,4-diene-3,20-dione
(5.0 g) and butanal (1.4 g) and purified as previously de-
scribed [12]. Yield: 5.0 g (88%).
2.9. (22R)-9␣-Fluoro-11␤,21-dihydroxy-16␣,17␣-pro-
pylmethylenedioxypregna-1,4-diene-3,20-dione (8R)
Compound 8 (3.0 g) was resolved by chromatography on
a Sephadex LH-20 column (76 ϫ 6.3 cm i.d.) with a mixture
of heptane/chloroform/ethanol (20/20/1) as the mobile
phase [13]. The fraction 10 005–10 815 ml was collected
and evaporated, yielding 1.32 g of 8R. The purity deter-
mined by HPLC (ethanol/water, 42/58) was 98%.
2.12. (22R)-21-Acetoxy-9␣-fluoro-6␤,11␤-dihydroxy-
16␣,17␣-propylmethylenedioxypregn-4-ene-3,20-dione
(6R)
2.10. (22R)-21-Acetoxy-9␣-fluoro-11␤-hydroxy-16␣,17␣-
propylmethylenedioxypregna-1,4-diene-3,20-dione (9R)
A suspension of tris(triphenylphosphine)rhodium chlor-
ide (30 mg) in toluene (15 ml) was treated with hydrogen
for 30 min. A solution of 10R (25 mg) in absolute ethanol
(15 ml) was added, and the hydrogenation was continued at
room temperature and atmospheric pressure for 48 h. The
reaction mixture was evaporated, and the residue was dis-
solved in ethyl acetate/heptane (80/20), applied to a short
silica gel column (4 ϫ 2 cm i.d.), and eluted with the same
solvent mixture. The eluent was evaporated leaving 25 mg
of solid material, which by analytical HPLC, was shown to
contain two products whose retention volumes corre-
sponded to the starting material (10R) and the desired com-
pound 6R in the ratio 70/30. The presence of the 1,4-diene
and the 4-ene in the mixture was confirmed by the following
A solution of acetyl chloride (310 ml) in dioxane (7.5 ml)
was added to a solution of compound 8R (450 mg) in
freshly distilled and dry pyridine (15 ml), and the reaction
mixture stirred at room temperature overnight. The volume
was reduced to 5 ml, methylene chloride (25 ml) was added,
and the solution was serially washed with hydrochloric acid
(1M), aqueous sodium hydrogen carbonate solution (10%),
and saturated aqueous sodium chloride solution. The dried
organic phase was evaporated, and the residue was purified
on a Sephadex LH-20 column (87 ϫ 2.5 cm i.d.) with
chloroform as the mobile phase. The fraction 305–350 ml
was collected and evaporated. The residue was precipitated
from methylene chloride Ϫ petroleum ether yielding 9R
(319 mg; 65%). The purity determined by HPLC (ethanol/
water, 55/45) was 99.1%. MS: m/z 491 (MHϩ) (MS-FAB).
1H-NMR (300 MHz): ␦ ppm 0.92 (t, 3H, H-25); 0.96 (s, 3H,
H-18); 1.55 (s, 3H, H-19); 2.19 (s, 3H, CH3COO); 4.43 (m,
1H, H-11); 4.65 (t, 1H, H-22); 4.72 and 4.92 (two d, 2H,
1
characteristic resonances in H-NMR (300 MHz): ␦ ppm
1.74 (s, 3H, H-19), 4.56 (m, 1H, H-6), 6.22 (d, 1H, H-4),
6.34 (dd, 1H, H-2) and 7.20 (d, 1H, H-1) for 10R and 1.72
(s, 3H, H-19), 4.39 (m, 2H, H-6 and H-11) and 5.90 (s, 1H,
H-4) for 6R.

20
A. Thale´n, L.-I. Wickstro¨m / Steroids 65 (2000) 16–23
2.13. (22R)-21-Acetoxy-9␣-fluoro-11␤-hydroxy-16␣,17␣-
propylmethylenedioxpregn-4-ene-3,6,20-trione (11R)
acetate/heptane (50/50). The fractions 25–160 ml (A) and
185–590 ml (B) were collected and evaporated. Fraction A
(140 mg) was identified as starting material.
Pyridine (60 ␮l), trifluoroacetic acid (28 ␮l), and 1-
cyclohexyl-3-(2-morpholinoethyl)carbodiimide metho-p-
toluenesulphonate (1.03 g) were added to a solution of 6R
(370 mg) in dimethylsulfoxide (1.2 ml) and toluene (2.4
ml). The mixture was stirred for 18 h at room temperature
under nitrogen protection, poured into hydrochloric acid
(10%, 6 ml), and extracted with methylene chloride four
times. The combined extracts were washed with water and
saturated aqueous sodium chloride solution, dried over so-
dium sulfate, and evaporated, leaving 384 mg of solid. This
crude product was purified by radial TLC on silica with
ethyl acetate/heptane (1/1) as the mobile phase, yielding
258 mg (70%) of 11R. The purity determined by HPLC
(ethanol/water, 48/52) was 94%. MS: m/z 506 (Mϩ; most
abundant ion) (field desorption due to low stability with
other techniques). 1H-NMR (300 MHz): ␦ ppm 0.94 (s, 3H,
H-18); 0.94 (s, 3H, H-25); 1.51 (s, 3H, H-19); 2.20 (s, 3H,
CH3COO); 4.44–4.52 (broad m, 1H, H-11); 4.67 (t, 1H,
H-22); 4.72 and 4.96 (two d, 2H, H-21); 4.88 (d, 1H, H-16);
6.39 (s, 1H, H-4).
Fraction B yielded 67 mg (30%) of 14R contaminated
with 10% of the corresponding 6␤-OH derivative (15R), as
determined by HPLC (ethanol/water, 48/52). MS of 14R:
m/z (relative intensity) 489 (MHϩ; 100); 488 (Mϩ; 82)
(field desorption due to low stability with other techniques).
1H-NMR (300 MHz) of 14R: ␦ ppm 0.93 (t, 3H, H-25); 0.95
(s, 3H, H-18); 1.43 (s, 3H, H-19); 2.18 (s, 3H, CH3COO);
4.51 (m, 1H, H-11); 4.57 (broad m, 1H, H-6); 4.62 (t, 1H,
H-22); 4.73 and 4.90 (two d, 2H, H-21); 4.86 (d, 1H, H-16);
6.31 (two d, 1H, H-2); 6.41 (d, 1H, H-4); 7.22 (d, 1H, H-1).
3. Results and discussion
3.1. Chemical synthesis
Autoxidation or peracid oxidation of 3-alkoxy-3,5-dienes
followed by 1,2-dehydrogenation and acetalization have
been used for 6-hydroxylation of triamcinolone 16␣,17␣-
acetonide [14] and budesonide [15] for example. When such
oxidations were applied to the 3-methoxy-3,5-diene deriva-
tive of 16␣-hydroxyhydrocortisone 16␣,21-diacetate (16),
mixtures of the 6␣- and 6␤-hydroxy epimers were obtained.
The ratio of 6␣-/6␤-epimers was changed from 43/57 to
84/16 when autoxidation was replaced by peracid oxidation
[15]. The 6␣- and 6␤-hydroxy derivatives were subjected to
chromatographic separation and used as starting material in
the preparation of 6␣- and 6␤-hydroxybudesonides [15],
respectively. The main objective of the present study was to
develop a simple and particularly stereospecific reaction
pathway to introduce a 6␤-hydroxy substituent into rofle-
ponide.
2.14. (22R)-9␣-Fluoro-11␤,21-dihydroxy-16␣,17␣-pro-
pylmethylenedioxypregn-4-ene-3,6,20-trione (12R)
Aqueous potassium carbonate (10%, 0.5 ml) was added
to a solution of 11R (247 mg) in methanol (25 ml). The
reaction mixture was stirred at room temperature for 20 min
under nitrogen protection, neutralized with acetic acid, and
extracted with methylene chloride (4 ϫ 15 ml). The com-
bined extracts were washed with water, dried over anhy-
drous sodium sulfate, and evaporated. The crude solid (210
mg) was purified by radial TLC on silica with ethyl acetate/
heptane (70/30), yielding 126 mg (56%) of 12R. The purity
determined by HPLC (ethanol/water, 48/52) was 95.4%.
MS (TSP - MS): m/z 465 (MHϩ; most abundant ion).
1H-NMR (300 MHz): ␦ ppm 0.92 (s, 3H, H-18); 0.94 (t, 3H,
H-25); 1.51 (s, 3H, H-19); 4.29 and 4.55 (two q, 2H, H-21);
4.44 Ϫ 4.54 (broad m, 1H, H-11); 4.62 (t, 1H, H-22); 4.94
(d, 1H, H-16); 6.39 (s, 1H, H-4).
It is known that olefins are oxidized at the allylic position
with selenium dioxide [16]. This procedure was recently
successfully applied to hydroxylation of some 1,4-dien-3-
one steroids, whereas selenium dioxide oxidation of the
corresponding 4-en-3-ones resulted in considerably lower
yields [17].
(22R)-21-Acetoxy-9␣-fluoro-11␤-hydroxy-16␣,17␣-
propylmethylenedioxypregn-4-ene-3,20-dione (4R) and the
corresponding 1,4-diene (9R) were both oxidized with se-
lenium dioxide in dioxane (Scheme 1). The hydroxylations
occurred stereospecifically to produce single isomers of the
6␤-hydroxy derivatives 6R and 10R. The pregn-4-ene 4R
was oxidized much more rapidly than the more stable 1,4-
diene 9R and to an unexpectedly high yield [17]. Attempts
to further improve the yield of 6R through prolongation of
the reaction time resulted in complete conversion of 4R but
also in degradation, so that no desired product 6R could be
detected. An alternative pathway to synthesis of compound
6R was based on selective 1,2-hydrogenation of the 1,4-
dien-3-one 10R. However, the overall yield of 6R was much
lower with this route than by direct selenium dioxide oxi-
2.15. (22R)-21-Acetoxy-6␣,11␤-dihydroxy-16␣,17␣-pro-
pylmethylenedioxypregna-1,4-diene-3,20-dione (14R)
A stirred suspension of (22R)-21-acetoxy-11␤-hydroxy-
16␣,17␣-propylmethylenedioxypregna-1,4-diene-3,20-di-
one (13R; 218 mg) and selenium dioxide (150 mg) in
dioxane (25 ml) was heated at 100°C for 72 h. The reaction
mixture was chilled to room temperature and filtered
through a short column (3.5 ϫ 2 cm i.d.) of Silica gel 60.
The column was washed with dioxane (30 ml). The filtrate
was evaporated, and the residue was dissolved in a mixture
of ethyl acetate and heptane (50/50, 50 ml), applied to a dry
column of silica gel (13 ϫ 2 cm i.d.), and eluted with ethyl

A. Thale´n, L.-I. Wickstro¨m / Steroids 65 (2000) 16–23
21
Scheme 1. Synthetic pathways to (22R)-9␣-fluoro-6␤,11␤,21-trihydroxy-16␣,17␣-propylmethylenedioxypregn-4-ene-3,20-dione (7R) and (22R)-9␣-fluoro-
11␤,21-dihydroxy-16␣,17␣-propylmethylenedioxypregn-4-ene-3,6,20-trione (12R). (a) HC(OCH3)3/CH2Cl2/dioxane; (b) autoxidation/EtOH; (c) K2CO3/
MeOH; (d) SeO2/dioxane; (e) [(C6H5)3P]3RhCl/H2; (f) Moffat oxidation.
dation of the pregn-4-en-3-one 4R because of an unexpec-
tedly slow oxidation of 9R and incomplete hydrogenation of
10R with Wilkinson’s catalyst.
Therefore, it was of interest to evaluate if allylic oxi-
dation of budesonide 21-acetate with selenium dioxide, fol-
lowed by ester hydrolysis, could offer a simpler and simul-
taneously stereospecific synthetic route to 6␤-
hydroxybudesonide than that previously used [15].
However, such an oxidation applied to the 22R-epimer of
budesonide 21-acetate (13R) yielded 30% of a mixture
consisting of 90% of the 6␣-hydroxy (14R) and 10% of the
6␤-hydroxy (15R) epimers (Scheme 2). This agrees well
with the orientation of the 6-hydroxy group observed on
allylic oxidation of prednisolone derivatives [17]. The
9␣,11␤-halohydrins all produced the 6␤-alcohols as the
single isomer, whereas the 6␣-isomer was formed in the
case of prednisolone itself, which has no substituent at the
9-position.
The substitution mode at the 9-position was found to also
influence the orientation of the 6-hydroxy group introduced

22
A. Thale´n, L.-I. Wickstro¨m / Steroids 65 (2000) 16–23
Scheme 2. Allylic oxidation of the 22R-epimer of budesonide 21-acetate (13R). (a) SeO2/dioxane; (b) separation of the 6␣- and 6␤-hydroxy epimers in the
oxidation product by analytical HPLC.
through autoxidation of the 3-methoxy-3,5-diene derivative
of 16␣-hydroxycortisol 16␣,21-diacetate (16) [15] as well
as that of compound 5R. In the former case, a mixture of the
6␣- and 6␤-epimers was formed. On the other hand, autoxi-
dation of compound 5R, which carries a 9␣-fluoro substitu-
ent, proceeded stereospecifically to produce the single 6␤-
hydroxy epimer 6R (Scheme 1). This reaction pathway gave
approximately the same overall yield of 6R as selenium
dioxide oxidation of 4R did.
the 3,6,20-trione 21-acetate 11R (Scheme 1). Removal of
the protecting 21-acetate group through hydrolysis with
potassium carbonate then gave the corresponding 21-alco-
hol 12R, which is another potential metabolite of rofle-
ponide.
3.2. C-6 stereochemical analysis
1H-NMR analysis has been used to differentiate be-
tween the 6␣- and 6␤-hydroxy epimers of budesonide
and its 21-acetate. The chemical shifts of the C-4 and
C-19 protons were found to have characteristic features
of the ␣- compared with the ␤-configuration of the 6-hydr-
oxy substituent (Table 1) [15]. A cis 1,3-diaxial interac-
tion with the 6␤-hydroxy group gives a chemical shift of
the C-19 protons signal at ␦ 1.68 ppm, compared with ␦
1.48 in budesonide. The 6␣-hydroxy epimer of budes-
onide is characterized by a strong (0.43 ppm) downfield
shift of the C-4 proton signal caused by a paramagnetic
anisotropy effect from the polar 6␣-substituent. Anal-
ogous to this, the chemical shifts of the C-4 and C-19
Analogous to the autoxidation of steroid ⌬5-enes, e.g.
cholest-5-en-3-one [18,19], the autoxidation of steroid ⌬3,5
-
dien-3-ol ethers to 6-hydroxy ⌬4-3-ketones was suggested
to proceed via 6-hydroxyperoxides [20]. This was con-
firmed by the discovery of minor amounts of ⌬4-ene-3,6-
diones and ⌬4-ketones as by-products. It is reasonable to
assume that 6- hydroxylation of 5 and 16 through autoxi-
dation also proceeds via hydroperoxides. The stereochem-
istry at the 6-position is probably dependent on the side
from which the singlet oxygen attacks the C5 - C6 double
bond. An attack on this bond from the ␣-side of the 9␣-
fluoro substituted compound 5 is less likely than an attack
from the ␤-side attributable to electronic rather than steric
reasons [21]. On the other hand, the attack of singlet oxygen
from the ␣-side is equally probable as from the ␤-side of
compound 16, which lacks a substitution at the 9␣-position.
This may explain the different stereoselectivity noticed on
autoxidation of 5 and 16.
1
protons in the H-NMR spectra of compounds 6R, 7R,
and 10R (Table 1) indicate that they are 6␤-hydroxy
derivatives of the corresponding parent compounds 4R,
3, and 9R, respectively. On the other hand, the compound
14R, resulting from selenium dioxide oxidation of the
22R-epimer of budesonide 21-acetate, seems to have a 6␣-
hydroxy configuration, according to this analysis (Table 1).
A more direct proof of the stereochemistry at C-6 is derived
The 6-hydroxyl group in compound 6R was selectively
oxidized using the Pfitzner/Moffatt procedure [22] to yield

A. Thale´n, L.-I. Wickstro¨m / Steroids 65 (2000) 16–23
23
Table 1
[2] Brattsand R, Ka¨llstro¨m L, Wieslander E, Andersson P, Dahlba¨ck M,
Dahl R. A model for particle-induced inflammation and edema for-
mation in the rat lung and the anti-inflammatory action of the gluco-
corticosteroid (GCS) budesonide in this model. Eur J Respir Dis
1983;64 (Suppl. 126):513.
Partial 1H-NMR spectraa
Compound
1H-Assignments
4-H
19-CH3
[3] Abel SM, Maggs JL, Back DJ, Park BK. Cortisol metabolism by
human liver in vitro - I. Metabolite identification and inter-individual
variability. J. Steroid Biochem 1992;43:713–9.
[4] Edsba¨cker S, Jo¨nsson S, Lindberg C, Ryrfeldt Å, Thale´n A. Metabolic
pathways of the topical glucocorticoid budesonide in man. Drug
Metab Dispos 1983;11:590–6.
[5] Chaplin MD, Rooks II W, Swenson EW, Cooper WC, Nerenberg C,
Chu NI. Flunisolide metabolism and dynamics of a metabolite. Clin
Pharmacol Ther 1980;27:402–13.
[6] To¨kes L, Cho D, Maddox ML, Chaplin MD, Chu NI. Isolation and
identification of an oxidatively defluorinated metabolite of flunisolide
in man. Drug Metab Dispos 1981;9:485–6.
Budesonide (BUD)b
6.12 m
6.55 m
6.25 m
6.10 m
6.43 mb
6.17 mb
5.80 m
5.80 m
5.90 m
5.90 m
6.14 m
6.22 m
6.41 m
1.48 s
1.47 s
1.68 s
1.47 s
1.45 sb
1.67 sb
1.53 s
1.54 s
1.72 s
1.72 s
1.55 s
1.74 s
1.43 s
6␣-Hydroxy-BUD, epimer 22Rb
6␤-Hydroxy-BUD, epimer 22Rb
21-Acetoxy-BUDb
21-Acetoxy-6␣-hydroxy-BUD, epimer 22R (14R)
21-Acetoxy-6␤-hydroxy-BUD, epimer 22R (15R)
3
4R
6R
7R
9R
10R
14R
[7] Gordon S, Morrison J. The metabolic fate of triamcinolone acetonide
in laboratory animals. Steroids 1978;32:2272–84.
a Chemical shifts are given in ppm relative to the internal standard
tetramethylsilane in CDCl3.
[8] Gerhards E, Neuweboer B, Schultz G, Gibian H, Hecker W. Stoff-
Wechsel von 6␣-fluor-16␣-methylpregna-1,4-dien-11,21-diol-3,20-
dion (fluocortolon) beim menchen. Acta Endocrinol 1971;68:98–126.
[9] Martinelli M, Ferrari P, Ripamonti A, Tuan G, Perazzi A, Assandri A.
Metabolism of deflazacort in the rat, dog and man. Drug Metab
Dispos 1979;7:335–9.
[10] Dahlberg E, Thale´n A, Brattsand R, Gustafsson J-Å, Johansson U,
Roempke K, Saartok T. Correlation between chemical structure, re-
ceptor binding, and biological activity of some novel, highly active
16␣,17␣-acetal substituted glucocorticoids. Mol Pharmacol 1984;25:
70–8.
b From Ref. [15].
from consideration of the NOE’s between the C-4 and C-19
protons and the remaining 6-proton of the 6-hydroxy com-
pounds. It has been shown for the 22R-epimer of budesonide
that the C-4 proton gives a NOE to the equatorial proton at the
6-position (6␣-H), but not to the axial proton (6␤-H). Simul-
taneously, there is a strong NOE between the C-19 protons and
the axial proton in the 6␤-position, but not with the 6-equa-
torial proton (Gunnar Gro¨nberg, unpublished data).
[11] Draper RW, Berkenkopf J, Carlon E, Fernandez X, Monahan M,
Lutsky BN. Synthesis and topical anti-inflammatory potencies of a
series of 6-keto- and ⌬6-6-acyloxybetamethasone derivatives. Arz-
neim-Forsch/Drug Res 1982;32:317–22.
Analyses of the 2D NOESY spectra of compounds 7R and
10R show a strong NOE between the C-4 proton and the
remaining proton in the 6-position, but no NOE between this
proton and the C-19 protons. Accordingly, the 6-hydroxy sub-
stituent in these compounds has the ␤-orientation. The same
type of analysis performed on the 2D NOESY spectrum of the
main component (14R) after 6-hydroxylation of the 22R-
epimer of budesonide 21-acetate (13R) reveals that there is a
strong NOE between the remaining proton in the 6-position
and the C-19 protons, but not between the C-4 and C-6 pro-
tons. This confirms that selenium dioxide oxidation of 13R
predominantly yields the 6␣-hydroxy derivative (14R).
[12] Thale´n A, Brattsand R, Gruvstad E. Synthesis and pharmacological
properties of some 16␣,17␣-acetals of 16␣-hydroxyhydrocortisone,
16␣-hydroxyprednisolone and fluorinated 16␣-hydroxyprednisolo-
nes. Acta Pharm Suec 1984;21:109–24.
[13] Thale´n A, Nylander B. Epimers of budesonide and related cortico-
steroids. I. Preparative resolution by chromatography on Sephadex
LH-20. Acta Pharm Suec 1982;19:247–66.
[14] Dusza JP, Bernstein S. C-6 Hydroxylated steroids. V. 6␤-Hydroxy-
triamcinolone and related compounds. J Org Chem 1963;28:760–3.
[15] Thale´n A, Wickstro¨m L-I. Synthesis and structure elucidation of 6␣-
and 6␤-hydroxylated budesonide and related corticosteroids. Acta
Pharm Suec 1984;21:145–62.
[16] Rabjohn N. Selenium dioxide oxidation. Org React 1976;24:261–
415.
[17] Terasawa T, Okada T. Selenium dioxide oxidation of steroidal 1,4-
dien-3-ones. A simple and convenient route to 6-hydroxycortico-
steroids. Synth Commun 1991;21:307–17.
Acknowledgments
[18] Fieser LF, Greene TW, Bischoff F, Lopez G, Rupp JJ. A carcinogenic
product of cholesterol. J Am Chem Soc 1955;77:3928–9.
[19] Nickon A, Mendelson WL. Reactivity and geometry in allylic sys-
tems. V. Oxygenation of cholest-5-en-3-one. J Org Chem 1965;30:
2087–9.
[20] Gardi R, Lusignani A. Autoxidation of steroid ⌬3,5-dien-3-ol ethers.
A simple route to 6␤-hydroxy ⌬4-3-ketones. J Org Chem 1967;32:
2647–9.
The authors wish to thank Dr Gunnar Gro¨nberg (Astra
Draco AB) for recording and interpreting the 2D NOESY
spectra.
References
[21] Paquette LA, Hertel LW, Gleiter R, Bo¨hm M. Electronic control of
stereoselectivity. 1. Singlet oxygen and related electrophilic additions
to aryl-substituted 7-isopropylidenebenzonorbornenes. J Am Chem
Soc 1978;100:6510–2.
[22] Pfitzner KE, Moffatt JG. Sulfoxide-carbodiimide reactions. I. A facile
oxidation of alcohols. J Am Chem Soc 1965;87:5661–70.
[1] Thale´n BA, Axelsson BI, Andersson PH, Brattsand RL, Nylander B.
6␣-Fluoro- and 6␣,9␣-difluoro-11␤,21-dihydroxy-16␣,17␣-propyl-
methylenedioxypregn-4-ene-3,20-dione: Synthesis and evaluation of
activity and kinetics of their C-22 epimers. Steroids 1998;63:37–43.