Air oxidation of 17-hydroxycorticosteroids catalyzed by cupric acetate: Formation of hemiacetal dimers

Article abstract of DOI:10.1016/S0039-128X(03)00033-3

Hydrocortisone, cortexolone, hydrocortisone-17-butyrate, and budesonide were oxidized into α-ketoaldehydes by air exposure in the presence of Cu(OAc)₂. When free hydroxyl functions were present at position 17, hydrocortisone and cortexolone, th

Full text of DOI:10.1016/S0039-128X(03)00033-3

Steroids 68 (2003) 361–365
Air oxidation of 17-hydroxycorticosteroids catalyzed by
cupric acetate: formation of hemiacetal dimers
Cécile Arbez-Gindre, Valérie Berl, Jean-Pierre Lepoittevin∗
Laboratoire de Dermatochimie, UMR 7123, Clinique Dermatologique, CHU, Université Louis Pasteur,
Strasbourg Cedex F-67091, France
Received 6 December 2002; received in revised form 3 February 2003; accepted 13 February 2003
Abstract
Hydrocortisone, cortexolone, hydrocortisone-17-butyrate, and budesonide were oxidized into ␣-ketoaldehydes by air exposure in the
presence of Cu(OAc)2. When free hydroxyl functions were present at position 17, hydrocortisone and cortexolone, the formed oxidation
products, were identified as hemiacetal dimeric structures involving the free hydroxyl functions at position 17 and the newly formed
aldehydes at position 21. Dimeric structures were established by using 1H{13C} correlations (HSQC and HMBC) and 1H–1H correlations
(COSY and ROESY). The hemiacetal function was further confirmed by reaction of the dimer formed from hydrocortisone with two
equivalents of 3-methyl-2-benzotriazolinone hydrazine (MTBH), giving quantitatively two equivalents of the 3-methyl-2-benzotriazolinone
hydrazone of 21-dehydrohydrocortisone. When no free hydroxyl function was present as in the case of hydrocortisone-17-butyrate and
budesonide, the expected ␣-ketoaldehydes were obtained.
© 2003 Elsevier Science Inc. All rights reserved.
Keywords: Corticosteroids; Hydrocortisone; Air oxidation; Cupric acetate; ␣-Ketoaldehyde; Cyclic hemiacetal
1. Introduction
[3]. The processing of these modified proteins by Langer-
hans cells, the main antigen presenting cells of the epider-
mis, results in the selection and activation of T-lymphocytes
with an appropriate T-cell receptor (TcR) able to recog-
nize peptidic sequences modified by the drug metabolite
or hapten [6]. This hypothesis was supported by several
metabolism, biochemical, and clinical studies showing that
these metabolites, even minor, could be formed in the skin
and be involved in the skin sensitization process [7–9].
In order to investigate this hypothesis and try to better
characterize interaction mechanisms of 21-dehydrocortico-
steroids with proteins, we have oxidized different corti-
costeroids, such as hydrocortisone (1), cortexolone (2),
hydrocortisone-17-butyrate (3), and budesonide (4) (Fig. 1),
into their 21-dehydro derivatives. The classical oxidation of
such molecules, as reported in the literature, is based on air
exposure in the presence of cupric acetate, leading smoothly
to ␣-ketoaldehydes in good yields [10]. Nevertheless, these
intermediates have been poorly characterized, and we have
therefore reinvestigated this reaction. We now report our
results, which showed that when a free hydroxyl group was
present at position 17 (hydrocortisone and cortexolone), ox-
idation products could not be isolated as ␣-ketoaldehydes,
but as hemiacetal dimeric structures.
Due to their strong anti-inflammatory and immuno-
suppressive properties, corticosteroids are used for many
therapeutic indications. Despite their beneficial effects, cor-
ticosteroids applied topically have also been associated with
allergic contact dermatitis (ACD) reactions [1,2]. The preva-
lence of corticosteroid related ACD in eczematous patients
has been estimated between 3 and 6% in northern European
countries where the use of topical corticosteroids is high
[1]. This side effect, which can be severe, results from the
formation of reactive intermediates during skin metabolism,
which are able to modify the side chain of nucleophilic
amino acids, and therefore, lead to the formation of modified
proteins recognized as foreign by the skin immune system
[3]. In this respect, it has been suggested that corticosteroid
molecules could be oxidized into 21-dehydro derivatives
and that these ␣-ketoaldehydes, which are potentially strong
electrophiles toward nucleophilic amino acids, such as argi-
nine [4,5], could lead to the formation of modified proteins
∗
Corresponding author. Tel.: +33-388-35-06-64;
fax: +33-388-14-04-47.
E-mail address: jplepoit@chimie.u-strasbg.fr (J.-P. Lepoittevin).
0039-128X/03/$ – see front matter © 2003 Elsevier Science Inc. All rights reserved.
doi:10.1016/S0039-128X(03)00033-3

362
C. Arbez-Gindre et al. / Steroids 68 (2003) 361–365
the crude product was purified by column chromatography
on silica (dichloromethane 96%, ethanol 3.5%, eau 0.6%)
to give (5) (0.33 g, 0.92 mmol, 97% yield). Yellow solid
m.p. 100–102 ◦C. The 500 MHz proton and the 125 MHz
carbon NMR data are listed in Table 1. IR (CHCl3, cm−1
)
=
=
ν: 3685 (OH); 3605 (OH); 1721 (C O); 1662 (C O); 1613
=
(C C). UV λmax 242 nm (ε 29860 in MeOH). [α]D = +110
(c 2.0, MeOH). Mass spectra (FAB−) m/z 719 (2M+−1),
360 (M+).
MTBH derivative: yellow solid m.p. 160–162 ◦C. 1H
NMR (200 MHz, CDCl3) δ: 0.89 (s, 3H, H18), 1.04–1.25
(m, 4H), 1.43 (s, 3H, H19), 1.59–2.47 (m, 13H), 2.90 (m,
ꢀ
1H, H16␤), 3.60 (s, 3H, H5 ), 4.47 (m, 1H, H11␣), 5.57 (s,
ꢀ
1H, OH17), 5.68 (s, 1H, H4), 7.19 (dd, 1H, J = 7.6 Hz, H2
ꢀ
ꢀ
ꢀ
or H3 ), 7.21 (dd, 1H, J = 7.6 Hz, H2 or H3 ), 7.39 (dd,
ꢀ
ꢀ
ꢀ
1H, J = 7.6 Hz, H1 or H4 ), 7.51 (dd, 1H, J = 7.6 Hz, H1
or H4 ). 13C NMR (50 MHz, CD3OD) δ: 18.2, 20.9, 24.0,
31.6, 31.8, 32.2, 32.9, 33.9, 35.0, 39.3, 41.7, 47.3, 51.8,
55.9, 68.8, 92.9, 110.7, 122.3, 122.6, 123.6, 123.8, 127.2,
140.5, 150.7, 171.9, 172.3, 197.7, 199.5, 206.9. IR (CHCl3,
ꢀ
Fig. 1. Chemical structures of hydrocortisone (1), cortexolone (2), hydro-
cortisone-17-butyrate (3), and budesonide (4).
2. Experimental
cm−1) ν: 3400 (OH); 1662 (C O); 1503 (C N). UV λmax
379 nm (ε 1270 in EtOH). [α]D = +27 (c 0.6, CHCl3).
=
=
Caution: Skin contact with 21-dehydrocortisone deriva-
tives must be avoided. As potential sensitizing substances,
these compounds must be handled with care.
2.3. Dimer of 21-dehydro-cortexolone (6)
The same procedure as for the synthesis of (5) was used,
except starting from cortexolone (2) (0.056 g, 0.016 mmol)
to give (6) (0.05 g, 0.08 mmol, quantitative yield). Solid
m.p. 110–112 ◦C. 1H NMR (200 MHz, CD3OD) δ: 0.65 (bs,
2.1. Chemistry
1H and 13C NMR spectra were recorded on Bruker AC
200 or ARX 500-MHz spectrometers in CD3OD unless oth-
erwise specified. Chemical shifts are reported in ppm (δ)
with respect to TMS, and CHCl3 was used as an internal
standard (δ = 7.26 ppm). Multiplicities are indicated by s
(singlet), d (doublet), t (triplet), and m (multiplet). Infrared
spectra were obtained on a Perkin–Elmer FT-IR 1600 spec-
trometer; peaks are reported in reciprocal centimeters. Ul-
traviolet spectra were determined on a Uvikon 810 spec-
trometer, and mass spectra (FAB−) were obtained using
a ZAB-HF mass spectrometer. Melting points were deter-
mined on a Buchi Tottoli 510 apparatus and are uncorrected.
Dried solvents were freshly distilled before use. Methylene
chloride was dried over P2O5 before distillation. All air- or
moisture-sensitive reactions were conducted in flame-dried
glassware under an atmosphere of dry argon. Chromato-
graphic purifications were conducted on silica gel columns
according to the flash chromatography technique.
ꢀ
ꢀ
6H, H18 and H18 ), 0.89–1.13 (m, 4H), 1.19 (s, 6H, H19 and
ꢀ
ꢀ
H19 ), 1.24–2.73 (m, 32H), 5.11 (s, 1H, H21 or H21 ), 5.23
(s, 1H, H21 or H21 ), 5.68 (bs, 2H, H4 and H4 ). 13C NMR
(50 MHz, CD3OD) δ: 15.1, 15.4, 17.7, 21.8, 24.5, 31.7, 33.5,
34.0, 34.4, 34.7, 35.3, 36.9, 37.0, 40.0, 51.6, 51.9, 55.0,
89.9, 93.8, 94.5, 175.1, 202.3, 207.5. IR (CHCl3, cm−1) ν:
ꢀ
ꢀ
=
=
3508 (OH), 1720 (C O), 1663 (C O). [α]D = +102 (c 0.4,
MeOH). Mass spectra (FAB−) m/z 687 (2M+ − 1), 344
(M+).
2.4. 21-dehydrohydrocortisone-17-butyrate (7)
The same procedure as for the synthesis of (6) was used,
except starting from the dehydrocortisone-17-butyrate (3)
(0.91 g, 2.10 mmol) and Cu(OAc)2 (0.43 g, 2.37 mmol, 1.1
equiv.) in ethanol (200 ml) to give (7) (0.90 g, 2.10 mmol,
quantitative yield). Solid m.p. 181–183 ◦C. 1H NMR
ꢀ
(200 MHz, CDCl3) δ: 0.92 (t, 3H, J = 7.5 Hz, H4 ), 0.92 (s,
2.2. Dimer of 21-dehydrohydrocortisone (5)
3H, H18), 0.96–1.27 (m, 2H, NA), 1.44 (s, 3H, H19), 1.60
ꢀ
(qt, 2H, J1 = J2 = 7.5 Hz, H3 ), 1.71–2.21 (m, 14H, NA),
ꢀ
Cu(OAc)2 (0.09 g, 0.5 mmol, 0.5 equiv.) in methanol
(38 ml) was added to a solution of hydrocortisone (1) (0.35 g,
0.95 mmol) in methanol (22 ml). The reaction mixture was
stirred under an air atmosphere for 5 h, hydrolyzed with
a saturated solution of NH4Cl (30 ml), and extracted with
CH2Cl2 (4 × 30 ml). Combined organic layers were dried
over Na2SO4, filtered, concentrated under vacuum, and
2.27 (t, 2H, J = 7.5 Hz, H2 ), 2.38–2.60 (m, 1H, NA), 3.07
(m, 1H, H16␤), 4.50 (m, 1H, H11␣), 5.69 (s, 1H, H4), 9.17 (s,
1H, H21). 13C NMR (50 MHz, CDCl3) δ: 13.6, 17.4, 18.1,
21.0, 23.9, 31.6, 32.0, 32.7, 33.2, 33.8, 35.0, 36.2, 39.2,
41.2, 47.5, 53.6, 55.7, 68.3, 94.0, 122.5, 171.7, 174.8, 187.0,
199.5, 201.6. IR (CHCl3, cm−1) ν: 3460 (OH), 1739 (C O),
=
=
=
1716 (C O), 1662 (C O). [α]D = +66 (c 1.0, CHCl3).

C. Arbez-Gindre et al. / Steroids 68 (2003) 361–365
363
Table 1
1H and 13C NMR data of the dimeric 21-dehydrohydrocortisone (5)
Atom
δ
13Ca
mb
δ
1Hc
CH2-1
CH2-2
C-3
CH-4
C-5
CH2-6
CH2-7
CH-8
CH-9
C-10
CH-11
CH2-12
CH2-12ꢀ
C-13
CH-14
CH2-15
CH2-16
C-17
CH3-18
CH3-18ꢀ
CH3-19
C-20
CH-21
CH-21ꢀ
35.8
34.3
202.0
122.5
176.6
33.3
34.1
32.9
4
4
0
2
0
4
4
2
2
0
2
2
2
0
2
4
4
0
3
3
6
0
1
1
2.20 m
2.50 ddd (13.7/13.7/16.8)
1.86 ddd (4.3/13.7/13.7)
2.30 ddd (4.3/4.3/16.8)
5.65 bs
2.55 ddd (1.7/5.4/14.2)
2.05 m
2.25 dddd (1.7/4.3/14.2/14.2)
1.10 m
2.05 m
57.5
0.98 dd (2.3/11.1)
40.7/40.9
68.7/68.8
40.7/40.8
40.7/40.8
49.8
53.2/53.2
24.6
34.6
89.9/90.0
17.6/17.9
17.6/17.9
21.5
207.3/207.6
93.8/94.6
93.8/94.6
4.40 bs
1.77 m
1.68 dd (2.6/13.7)
1.96 dd (3.5/13.9)
1.96 dd (3.5/13.9)
1.77 m
1.38 ddd (3.9/5.9/11.6)
1.51 ddd (5.9/8.9/14.8)
1.77 m
2.67 m
0.89 s
0.91 s
1.47 s
5.17 s
5.27 s
a Based on 1H{13C} correlation experiments.
b Multiplicity deduced from DEPT 135 experiments.
c Based on 1H{13C} and 1H–1H COSY correlation experiments.
2.5. 21-dehydrobudesonide (8)
ing a catalytic amount of cupric acetate in methanol. The
structures of the oxidized products were established by a
1
combination of H and 13C NMR experiments, and unex-
The same procedure as for the synthesis of (6) was
used, except starting from the budesonide (4) (0.11 g,
0.25 mmol, mixture 50/50 of diastereomers) to give (8)
(0.98 g, 0.23 mmol, 92% yield) as a mixture of diastere-
pected results were observed in the case of hydrocortisone
1
(1) and cortexolone (2). Thus, the H NMR spectrum of
the hydrocortisone oxidation product (5) showed the ab-
sence of an aldehyde H21 proton expected at about 9.8 ppm
together with the appearance of two non expected signals
at 5.17 and 5.27 ppm, respectively. Therefore, a complete
assignment was performed using 1H{13C} correlations
1
omers. Solid m.p. 81–85 ◦C. H NMR (200 MHz, CDCl3)
ꢀ
δ: 0.87–1.02 (m, 12H, 2 × H18 et 2 × H5 ), 1.07–1.36 (m,
6H, NA); 1.44 (s, 6H, 2 × H19); 1.52–2.16 (m, 28H, NA),
2.57 (m, 4H, NA), 3.33 (dd, 2H, J1 = 9.7 Hz, J2 = 3.3 Hz,
1
(HSQC and HMBC) and H–1H correlations (COSY and
ꢀ
2 × H1 ), 4.46 (s, 2H, 2 × H11␣), 5.46 (m, 2H, 2 × OH11␤),
ROESY) for this derivative (5). Several NMR signals,
like methyl-18, appeared doubled as due to the forma-
tion of a dimer (Table 1). 13C NMR data confirmed the
absence of an aldehydic carbon, the presence of two un-
expected signals at 93.8 and 94.6 ppm, respectively, and
the presence of eight doubled carbon signals. The mass
spectrum (FAB−) of compound (5) showed a molecu-
lar peak at 719 (2M+−1) in accordance with a dimeric
form and a peak at 360, which could correspond to the
monomeric form (M+). In addition, the ultraviolet ab-
sorption spectrum of (5) (λmax = 242 nm; ε = 29860 in
MeOH) was similar to that reported in the literature for
21-dehydrohydrocortisone, when prepared by other routes,
but with a two times increased molar extinction coeffi-
cient [10–13]. This was in accordance with the presence of
two ␣,␤-unsaturated carbonyl systems and thus, with the
formation of a dimer.
6.03 (s, 2H, 2 × H4), 6.27 (d, 2H, J = 9.95 Hz, 2 × H2),
7.27 (d, 2H, J = 9.95 Hz, 2 × H1), 9.51 (s, 1H, H21), 9.54
(s, 1H, H21). 13C NMR (50 MHz, CDCl3) δ: 14.0 (2C),
17.2 (2C), 21.0 (2C), 30.5, 31.2, 31.9 (2C), 32.9 (2C),
33.3 (2C), 34.1 (2C), 35.0 (2C), 37.1 (2C), 41.3 (2C), 44.1
(2C), 55.3 (2C), 65.0 (2C), 70.1 (2C), 83.1, 84.0, 97.4,
98.2, 104.8 (2C), 108.5 (2C), 122.4 (2C), 127.8 (2C), 156.5
(2C), 170.2 (2C), 186.8 (2C), 190.1 (2C), 199.9 (2C). IR
(CHCl3, cm−1) ν: 3504 (OH), 1723 (C O), 1660 (C O).
[α]D = +82 (c 1.3, CHCl3).
=
=
3. Results
Hydrocortisone (1), cortexolone (2), hydrocortisone-17-
butyrate (3), and budesonide (4) were oxidized in very high
yields into their expected ␣-ketoaldehyde derivatives us-

364
C. Arbez-Gindre et al. / Steroids 68 (2003) 361–365
Scheme 1.
the presence of Cu(OAc)2 was already reported by Orr and
Monder in 1975 [14], who associated these results with a
rapid equilibrium between the aldehydic and the predomi-
nant gem-diol form. Additional analytical data were more in
favor of the formation of a dimer with MS (FAB−) show-
ing a molecular peak at 719 (2M+−1) and the UV spectrum
showing a two times increased molar extinction coefficient
compared with data obtained using other preparation meth-
ods [10–13]. All of our analytical data were, thus, consis-
tent with a hemiacetal, dimeric structure (5) that could be
formed by reaction of the free hydroxyl functions at position
17 with the newly formed aldehydic functions at position
21. The hemiacetal nature of this dimer was further con-
firmed by chemical reaction with two equivalents of MTBH,
leading to the quantitative formation of two equivalents of
MTBH-21-dehydrohydrocortisone.
Formation of a eight member ring between two
hydroxy-aldehydes was not expected to be very favorable,
but could be assisted by the presence of cuprous ions that
could play a role as templates. The ring structure was further
supported by the n Oe observed between the C-18 methyl
groups at 0.89 and 0.91 ppm, respectively, with hemiacetal
protons at 5.17 and 5.27 ppm, respectively. Interestingly,
quenching of the oxidation reaction medium with EDTA,
which could contribute to the chelation of cupric ions,
seemed to favor monomeric structures [10].
Scheme 2.
These data were consistent with a dimeric hemiacetal
structure (Scheme 1) formed between the free hydroxyl
goups at position 17 and the aldehydic functions at position
21. The hemiacetal nature of (5) was further confirmed by re-
action with two equivalents of 3-methyl-2-benzotriazolinone
hydrazine (MTBH) in CH2Cl2 or ethanol, giving quantita-
tively two equivalents of the 3-methyl-2-benzotriazolinone
hydrazone of 21-dehydrohydrocortisone.
Similar results were obtained with cortexolone (2), while
the same reaction carried out on corticosteroids without a
free hydroxyl function at position 17, such as hydrocorti-
sone-17-butyrate (3) or budesonide (4) (Scheme 2), gave
quantitatively the expected ␣-ketoaldehydes as confirmed
by the presence of aldehydic signals at 9.17 and 9.51 ppm,
respectively.
These results were further confirmed by the oxidation
of cortexolone (2), which also has a hydroxyl function at
C-17, and this also led to the formation of a dimeric struc-
ture (6). The determinant role of the free hydroxyl func-
tion at position 17 was confirmed with results obtained with
hydrocortisone-17-butyrate (3) and budesonide (4) for which
the oxidation with air/cupric acetate led to the formation
of the classical ␣-ketoaldehydes (7) and (8), respectively
(Scheme 2).
4. Discussion
The hydroxyl function at position 21 in corticosteroids
is highly sensitive to oxidation, and it has been reported
in the literature that these molecules could be converted
smoothly into their ␣-ketoaldehyde derivatives by a sim-
ple air exposure of a methanolic solution in the presence
of Cu(OAc)2 [10]. Air oxidation of hydrocortisone (1) us-
ing a catalytic amount of cupric acetate in methanol re-
sulted in the quantitative formation of an oxidation product
(5) with analytical data not in accordance with the expected
structure. The main discrepancy was the lack of aldehydic
Acknowledgments
The authors thank the French Ministère de l’Education
Nationale, de la Recherche et de la Technologie for a fel-
lowship to C.A.G.
References
signals in both H and 13C NMR spectra associated with
the presence of some doubled signals. This lack of alde-
hydic signals following air oxidation of hydrocortisone in
1
[1] Dooms-Goossens A, Andersen KE, Brandao FM, Bruynzeel D,
Burrows D, Camarasa JG, et al. Corticosteroid contact allergy: an
EECDRG multicentre study. Contact Dermat 1996;35:40–4.

C. Arbez-Gindre et al. / Steroids 68 (2003) 361–365
365
[2] Lepoittevin JP, Drieghes J, Goossens A. Studies in patients with
corticosteroid contact allergy. Arch Dermatol 1995;131:31–7.
[3] Roberts DW, Lepoittevin JP. Hapten-protein interactions. In:
Lepoittevin JP, Basketter DA, Goossens A, Karlberg AT, editors.
Allergic contact dermatitis: the molecular basis. Berlin: Springer;
1998. p. 81–111.
[4] Wilkinson SM, Jones MF. Corticosteroid usage and binding
to arginine: determinants of corticosteroid hypersensitivity. Br J
Dermatol 1996;135:225–30.
[8] Monder C, Walker MC. Interactions between corticosteroids and
histones. Biochemistry 1970;9:2489–97.
[9] Bundgaard H. The possible implication of steroid–glyoxal degrada-
tion products in allergic reactions to corticosteroids. Arch Pharm
Chem Sci Ed 1980;8:83–90.
[10] Oh SW, Monder C. Synthesis of corticosteroid derivatives containing
the 20␤-ol-21-al side chain. J Org Chem 1976;41:2477–80.
[11] Leanza WJ, Conbere JP, Rogers EF, Pfister K. Aldehydes derived
from cortisone and hydrocortisone. J Am Chem Soc 1954;76:1691–4.
[12] Sloan KB, Little RJ, Bodor N. Acid-catalyzed rearrangements of the
dihydroxyacetone side chain in steroids during ketal exchange. J Org
Chem 1978;43:3405–8.
[5] Matura M, Lepoittevin JP, Arbez-Gindre C, Goossens A. Testing
with corticosteroid aldehydes in corticosteroid-sensitive patients
(preliminary results). Contact Dermat 1998;38:106–8.
[6] Scheynius A. Immunological aspects. In: Lepoittevin JP, Basketter
DA, Goossens A, Karlberg AT, editors. Allergic contact dermatitis:
the molecular basis. Berlin: Springer; 1998. p. 4–18.
[13] Han CA, Monder C. Asymmetric reduction of 20-steroidal ketone:
synthesis of corticosteroid derivatives containing the 20␣-ol-21-al
side chain. J Org Chem 1982;47:1580–4.
[7] Monder C, Bradlow HL. Cortoic acid: explorations at the frontier of
corticosteroid metabolism. Recent Prog Horm Res 1980;36:345–400.
[14] Orr JC, Monder C. The stereospecificity of the enzymic reduction
of 21-dehydrocortisol by NADH. J Biol Chem 1975;250:7547–53.