Mechanism of base-catalyzed autooxidation of corticosteroids containing 20-keto-21-hydroxyl side chain

Article abstract of DOI:10.1016/j.tetlet.2009.05.074

Corticosteroids and related compounds containing the 20-keto-21-hydroxyl side chain such as betamethasone, betamethasone 9,11-epoxide, dexamethasone, and dexamethasone 9,11-epoxide have been found to undergo facile autooxidation on the 1,3-dihydroxyaceton

Full text of DOI:10.1016/j.tetlet.2009.05.074

Tetrahedron Letters 50 (2009) 4575–4581
Contents lists available at ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
Mechanism of base-catalyzed autooxidation of corticosteroids containing
20-keto-21-hydroxyl side chain
*
Min Li , Bin Chen, Stephanie Monteiro, Abu M. Rustum
Global Quality Services—Analytical Sciences, Schering-Plough Corporation, 1011 Morris Avenue, Union, NJ 07083, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
Corticosteroids and related compounds containing the 20-keto-21-hydroxyl side chain such as betameth-
asone, betamethasone 9,11-epoxide, dexamethasone, and dexamethasone 9,11-epoxide have been found
to undergo facile autooxidation on the 1,3-dihydroxyacetone side chain of their D-rings under strong
alkaline conditions to yield five main degradants (17-formyloxy-17-acid, 17-acid, 21-aldehyde, 20-
hydroxy-21-acid, and 17-ketone). The rate of the autooxidation was correlated with the strength and
concentration of the base used in the reaction. A novel mechanism for the observed autooxidation is pro-
posed, in which the facile oxidation of the presumed enolate (resulting from the carbanion at the 21-posi-
tion) by molecular oxygen is the key step. The proposed autooxidation mechanism, supported by LC–MS
isotope experiments using 18O2 as the oxidant, can satisfactorily explain the oxidative degradation pat-
terns observed for corticosteroids containing the 20-keto-21-hydroxyl side chain.
Received 4 January 2009
Revised 18 May 2009
Accepted 20 May 2009
Available online 25 May 2009
Keywords:
Corticosteroids
Oxidation
Enolate
Betamethasone 21-aldehyde
Betamethasone 17-formyloxy-17-acid
Betamethasone 17-acid
Ó 2009 Elsevier Ltd. All rights reserved.
1. Introduction
gen is the key step. The proposed mechanism can satisfactorily ex-
plain the autooxidative degradation patterns reported for
Forced degradation studies (or stress studies) under various deg-
radative conditions are recommended by the International Confer-
ence on Harmonization (ICH) guidelines during pharmaceutical
development of adrugcandidateinordertoassistintheunderstand-
ing of the stability and/or degradation behavior/pathways of the
drug candidate under long-term stability storage conditions.1 Acid
and base are typically employed to accelerate the evaluation of
hydrolytic stability of a chemical entity; nevertheless, non-hydro-
lytic degradation may also occur and sometimes can become the
dominant degradation process. For example, it has been reported
that rofecoxib undergoes rapid oxidation under alkaline conditions;
a mechanism involving oxidation of the rofecoxib enolate by molec-
ular oxygen was proposed, which appears to be consistent with the
presumption that free radical is not involved during the rather facile
oxidation observed.2 Duringour recent forceddegradation studiesof
betamethasone (1, Fig. 1), an anti-inflammatory corticosteroid, and
its related compounds under strong alkaline conditions, it became
evident that betamethasone and its analogues containing the 20-
keto-21-hydroxyl side chain undergo facile autooxidation. In this
Letter, we present the evidence3 for this relatively less known oxida-
tive degradation process along with a proposed novel degradation
mechanism in which the facile oxidation of the presumed enolate
(resulting from the carbanion at the 21-position) by molecular oxy-
structurally similar corticosteroids containing the 20-keto-21-hy-
droxyl side chain during the past several decades.4–7 A clear under-
standing of the autooxidation mechanism should greatly facilitate
pharmaceutical development and/or improvement of drug products
containing these corticosteroids as they continue to be a very critical
class of therapeutic agents.
2. Initial forced degradation studies
For the initial forced degradation study of betamethasone under
an alkalinecondition, approximately5 mgof betamethasone(1) was
dissolved in 3 mL of acetonitrile, followed by addition of 100 lL of
1 N NaOH in water. The resulting solution was monitored at room
temperature by LC–MS.8 Two main oxidative degradants, betameth-
asone 17-formyloxy-17-acid (10, m/z 407) and betamethasone 17-
acid (11, m/z 379),9 were formed with the yields of ꢀ10% and 20%,
respectively, 60 min after the base was added (Fig. 2). The initial
observation that 1 might be susceptible to autooxidation relatively
easily under the alkaline condition led us to design a more system-
atic study in order to elucidate the oxidation mechanism. When
the reaction was performed using a 1 N NaOH solution in methanol
in place of the original 1 N NaOH aqueous solution, the autooxida-
tion was found to proceed much faster: no 1 was detected in
30 min (Reaction 3 in Table 1). In addition to 11 (37% yield), a new
oxidative degradant, betamethasone 21-aldehyde (7, in the form of
its hydrate as detected by LC–MS), was observed along with the
* Corresponding author. Tel.: +1 908 820 3391; fax: +1 908 820 6359.
E-mail address: min.li@spcorp.com (M. Li).
0040-4039/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2009.05.074

4576
M. Li et al. / Tetrahedron Letters 50 (2009) 4575–4581
OH
O
OH
O
OH
OH
HO
HO
F
F
O
O
Betamethasone (1)
Dexamethasone
OH
OH
O
O
OH
OH
HO
O
H
O
O
Dexamethasone 9,11-epoxide
Hydrocortisone
Figure 1. Structures of betamethasone and related compounds, dexamethasone, dexamethasone 9,11-epoxide, and hydrocortisone.
Betamethasone (1)
17-Acid (11)
17-Formyloxy-17-acid (10)
0
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
5.5
Time (min)
m/z 393
10
8
6
m/z 407
m/z 379
4
2
0
0.0
0.5
1.0
1.5
2.0
2.5
3.0
Time (min)
3.5
4.0
4.5
5.0
5.5
Figure 2. Autooxidation of betamethasone (1) under alkaline condition: ꢀ5 mg of betamethasone (1) was dissolved in 3 mL of acetonitrile, followed by addition of ꢀ100
lL of
1 N NaOH in water. LC–MS analysis was performed 60 min after the base was added.6 Top: UV chromatogram at 240 nm. Bottom: the corresponding MS (total ion)
chromatogram showing the molecular ions of the starting material (1, m/z 393) and the two oxidative degradants, betamethasone 17-formyloxy-17-acid (10, m/z 407) and
betamethasone 17-acid (11, m/z 379).
isomer of its hydrate form (7 and 8, in a ratio of ꢀ1:3, were formed in
46% yield collectively) and 17-ketone (5, 17%). Further study sug-
gested that this isomer (8) of betamethasone 21-aldehyde hydrate
is very likely betamethasone 20-hydroxy-21-acid which would be
formed from 21-aldehyde via hydration and a subsequent intramo-
lecular Cannizzaro rearrangement (Scheme 1). Edmonds et al. re-
ported the conversion of dexamethasone 21-aldehyde to
dexamethasone 20-hydroxy-21-acid in strong alkaline aqueous
solution.6 Indeed, when a 1.5 mL solution of 7 in acetonitrile was
treated with 50 lL of 1 N NaOH in methanol at room temperature,
7 was completely converted to 8 within ꢀ10 min.
When subjected to the same stress condition mentioned above,
threesteroidssuch asbetamethasone9,11-epoxide, dexamethasone,
and dexamethasone 9,11-epoxide which are structurally related to
1 were found to undergo the same base-catalyzed autooxidation. In
order to verify that molecular oxygen is the oxidizing agent, we re-
peated the reaction with 1 using the NaOH solution in methanol
under the condition in which oxygen was largely excluded by bub-

M. Li et al. / Tetrahedron Letters 50 (2009) 4575–4581
4577
Table 1
Based-catalyzed autooxidation of betamethasone (1): impact of the strength of the base
Reactiona
Base usedb
Betamethasone% (1) reacted
Distribution of degradation products
21-Aldehyde (7)/20-hydroxy-21-acid (8)c
17-Formyloxy-17-acid (10)
17-Ketone (5)
17-Acid (11)
1
2
3
4
5
Triethylamine
LiOH
NaOH
KOH
NaOMe
0
—
—
—
—
ꢀ20%
>99%
100%
100%
10%
17%
11%
24%
ND
46%
6%
8%
ND
ND
3%
ND
90%
37%
80%
68%
a
All bases were dissolved in solutions of MeOH at a concentration of ꢀ1 M. An aliquot of the base (50
lL) was added to a 1 mL solution of 1 in acetonitrile (1 mg/mL) and
the reaction was monitored at room temperature by LC–MS8 30 min after the base was added. The molar concentration ratio of the base and 1 is ꢀ20:1.
b
c
The strengths of the bases are in the following order: NaOMe > KOH > NaOH > LiOH > triethylamine.10
Under the current LC–MS condition, 7 and 8 co-eluted. With a modified LC–MS method, 7 and 8 were found to form in a ratio of ꢀ1:3 in Reaction 3.
O
OH
H
O
HO
O
OH
O
Retro-Aldo
Route d
Betamethasone 17-ketone (5)
3
4
OH
O
H
a,b
O
OH
O
OH
a,b
HO
a
b
O
O
H
O
O
Route a
OH
O
NaOH
O2
OH
OH
OH
F
Acetonitrile
HO-O
O
Betamethasone (1)
Betamethasone 21-aldehyde (7)
2
6
OH
Route b
O
Route c
Baeyer-
Villiger
Intromolecular
Cannizzaro
Rearrangement
OH
OH
H
OH
O
O
O
O
HO
Betamethasone17-acid (11)
OH
O
H
H
OH
OH
O
O
O
Betamethasone
17-diformylanhydride (9)
Betamethasone 20-hydroxy-
21-oic-acid (8)
Retro-Aldo
Betamethasone
17-formyloxy-17-acid (10)
Betamethasone 17-ketone (5)
Scheme 1. Proposed mechanism for the base-catalyzed autooxidation of betamethasone and related degradation reactions. The two acids (8 and 11) should exist as their
carboxylate forms in the basic stress conditions.
bling all the reagent solutions with nitrogen for ꢀ3 min and the reac-
tion was then carried out under a nitrogen atmosphere. Under this
essentially anaerobic condition, oxidation of betamethasone (1)
was mostly inhibited for 30 min at room temperature.
the facile autooxidation observed. Hence, four bases of different
strengths, triethylamine, LiOH, KOH, and NaOMe, were also evalu-
ated in the base-catalyzed autooxidative degradation. The results
as summarized in Table 1 show that the autooxidation is correlated
to the strength of the bases used:10 triethylamine, a weak base, did
not give any reaction; with LiOH, a relatively weak alkali metal
base, ꢀ20% of 1 was consumed; NaOH, KOH, and NaOMe, all very
strong bases, caused all of 1 to react.
3. Strength of the base and solvent effect
Based on the results from the initial base stress studies outlined
above, it appeared that the strength of the base, which should be
impacted by the reaction medium, might play a critical role in
All the reactions described so far were carried out in a medium
consisting of mostly acetonitrile, an aprotic solvent, with a tiny

4578
M. Li et al. / Tetrahedron Letters 50 (2009) 4575–4581
fraction of either of the protic solvents, that is, water and metha-
nol, introduced as part of the base solutions. The solvent effect to-
ward the autooxidative degradation catalyzed by NaOH was
evaluated with two additional aprotic solvents, acetone and tetra-
hydrofuran (THF), versus MeOH, a protic solvent. When the forced
degradation was performed in MeOH instead of acetonitrile while
the concentration of the base remained the same, essentially no
reaction was observed within 2 h. When the same reaction was
carried out in acetone and THF, respectively, similar reaction prod-
uct distribution and reaction rates (vs Reaction 3 in Table 1) were
observed in 30 min. It is known that the strength of an alkali metal
base such as NaOH becomes much stronger in an aprotic environ-
ment than in a protic one, primarily due to the fact that hydroxide
is not solvated in the aprotic medium.10 The non-reactivity in
MeOH can probably be largely attributed to the much reduced ba-
sicity of NaOH in the protic environment. Therefore, the results
from the solvent effect study are consistent with those shown in
Table 1.
after just 5 min. In addition, the product distribution also changed
significantly: the initially formed 17-formyloxy-17-acid (10) had
disappeared while 21-aldehyde (7)/20-hydroxy-21-acid (8) and
17-acid (11) continued to increase. On the other hand, 17-ketone
(5), which was not detected under the lower levels of the base, oc-
curred in significant quantities. Betamethasone 17-ketone (5)
could be generated by strong base under an anaerobic condition
through a retro-aldo pathway that was reported for hydrocorti-
sone.5 Nevertheless, 5 could also be an oxidative degradant of 1
or one of its degradants such as enol aldehyde based on our expe-
rience.11 In the current study, although it is difficult to tell exactly
how 5 was formed as to through which pathway(s), it appears that
the majority of 5 should come from the retro-aldo pathway. This
speculation is consistent with the current observation that the for-
mation of 5 increased when the highest amount of the base was
used (Fig. 3) and/or stronger bases were used (Table 1). In addition,
the retro-aldo pathway leading to the formation of 5 could be oper-
ative in two places in the current study (Scheme 1): (1) starting
from the 21-carbanion of betamethasone (2) through ‘Route d’
and (2) starting from betamethasone 20-hydroxy-21-acid (8).
4. Ratio of the base and. betamethasone: impact on the
reaction rate and degradation product distribution
5. Isotope experiments with 18O2
The ratio of the base used and betamethasone (1) was also stud-
ied and three different ratios of the base (1 N NaOH solution in
methanol) were used. The reaction rate and degradation product
distribution as revealed by LC–MS were compared at different
times after the reactions were initiated by addition of the base
solutions. The results showed that the rate of the autooxidation in-
creased as more base was used, although the difference between
the reactions with the molar ratios of 0.4:1 and 0.8:1 (NaOH and
1) did not appear to be obvious enough at 50 min (Fig. 3). At the
highest amount of the base (4:1) used, only ꢀ2% of 1 remained
Other than 17-ketone (5), other degradants observed in the cur-
rent study are all known oxidative degradants. In order to provide
further evidence that the degradation occurred is indeed a base-
catalyzed autooxidation, that is, oxidation inflicted upon directly
by molecular oxygen,12 isotope experiments with 18O2 were per-
formed using NaOMe in methanol as the base. When compared
with the control experiment in which the regular 16O2 was used
at the molar ratio of NaOMe/1 at 0.8:1, the 18O2 isotope experiment
clearly showed that 18O was incorporated into the two oxidative
17-Acid (11)
17-Ketone (5)
8.3
9.3
10
21-Aldehyde (7)/(8)
7.6
8
6
4
2
Betamethasone (1)
13.2
14.5
0.9
10.1
10
7.3
12.6
2.0
2
2.4
4.9
7.0
3.0
0
0
4
6
8
12
14
Time
7.29
10
8
Betamethasone (1)
6
17-Formyloxy-17-acid (10)
4
11
8.3
7
2
9.5
7.6
10.1
10
13.5 14.1
14
9.0
0
0
2
4
6
8
12
Time
7.36
100
1
10
9.60
11
8.45
7
7.74
10.20
10
0
0
2
4
6
8
12
14
Figure 3. Autooxidation of betamethasone (1) using different amounts of base (1 N NaOH solution in methanol). The molar ratio of NaOH and 1 (1 mg/mL in acetonitrile) is
approximately 0.4:1, 0.8:1, and 4:1, respectively, from the bottom panel to the top panel. LC–MS analyses6 were performed at different times after the base was added to the
respective reaction solutions: 50 min for the reactions at the bottom and the middle panels and 5 min for the reaction at the top panel.

M. Li et al. / Tetrahedron Letters 50 (2009) 4575–4581
4579
17-Formyloxy-17-acid (10)
RT: 4.99 - 7.51
NL:
2.31E6
17-Acid (11)
6.67
nm=239.5-
240.5
6.18
8
6
4
2
0
Betamethasone (1)
PDA
betamethas
one_24_1h
_24Jul06
21-Aldehyde (7)/(8)
5.78
6.99
5.0
5.1 5.2
5.3
5.4
5.5
5.6
5.7
5.8 5.9
6.0
6.1 6.2
6.3 6.4
6.5
6.6
6.7 6.8
6.9
7.0
7.1 7.2 7.3
7.4
7.5
Time (min)
betamethasone_24_1h_24Jul06 # 729-738 RT: 6.21-6.27 AV:
F: ITMS + c ESI Full ms [ 100.00-1000.00]
5
SB: 86.12-6.18, 6.30-6.37 NL: 5.75E5
381.04
MS Spectrum of 11
500000
400000
300000
200000
361.19
426.07
760.94
736.53 783.06 837.50 850.08
100000
115.93
343.25
147.24
199.16
267.25 295.38
453.73
599.01
551.84
606.77 629.63
636.19
829.09 927.81
977.07
0
100
150
200
250
300
350
400
450
500
550
m/z
600
650
700
750
800
850
900
950
1000
betamethasone_24_1h_24Jul06 # 792-800 RT: 6.71-6.75 AV:
F: ITMS + c ESI Full ms [ 100.00-1000.00]
4
SB: 86.64-6.67, 6.80-6.86 NL: 7.60E5
409.02
MS Spectrum of 10
700000
600000
500000
400000
300000
200000
100000
816.77
873.34 878.76
389.11
316.98
297.08
343.24
449.15 471.10
137.20 153.17
199.14 292.14
519.83 546.56
634.73 636.36
715.82 740.90
719.48
938.07 956.73
0
100
150
200
250
300
350
400
450
500
550
m/z
600
650
700
750
800
850
900
950
1000
Figure 4. LC–MS analysis of the autooxidation isotope experiments using 18O2. A 5 mL solution of betamethasone (1) in acetonitrile (1 mg/mL) and a 0.5 M NaOMe solution in
methanol were gently bubbled with 18O2 for a few min; an aliquot of 10
lL of the NaOMe solution was then injected into the sealed solution of 1, followed by shaking of the
resulting solution for a couple of times. LC–MS analysis was performed 60 min after the base was added.6 Top: UV chromatogram at 240 nm showing the starting material (1),
21-aldehyde (7)/20-hydroxy-21-acid (8), 17-acid (11), and 17-formyloxy-17-acid (10), respectively. Middle: MS spectrum of the 17-acid peak displaying the 18O isotope
molecular ion at m/z 381. Bottom: MS spectrum of the 17-formyloxy-17-acid peak displaying the 18O isotope molecular ion at m/z 409.
degradants, 17-formyloxy-17-acid (10) and 17-acid (11) since they
displayed the corresponding M+2 mass peaks as their molecular
ions (Fig. 4). On the other hand, the 21-aldehyde peak did not show
18O incorporation. In a separate isotope experiment in which high-
er molar ratio of the base was used (20:1), 17-ketone (5) was
formed as expected; however, 18O was not incorporated into the
structure of 5 due to the fact that the original molecular ion of 5
at m/z 333 was still observed (results not shown). The results from
these isotope experiments not only clearly demonstrate the nature
of the autooxidation observed but also provide solid evidence re-
quired for elucidating the underlying mechanism.
the reaction than 7 (Fig. 3), it appears that ‘Route b’ should be
the predominant pathway for the formation of 10 and 11. Under
the current alkaline condition, the anhydride (9) would undergo
a quick rearrangement of the formyl group from the 21-position
to the 17-position to give 17-formyloxy-17-acid (10) and/or a di-
rect cleavage to yield 17-acid (11). The evidence for a 17-diformy-
lanhydride moiety in 9 and the subsequent rearrangement of the
formyl group was reported in a study of a pharmaceutical product
which contains dexamethasone as the active pharmaceutical
ingredient.13 With the reaction progressing and/or conducted at a
higher ratio of the base, 10 should be further degraded to 11
through deformylation. The formation of 17-ketone (5), which be-
comes more significant with stronger and/or more base, is most
likely through the known retro-aldo mechanism via ‘Route d’;5
nevertheless, it is also possible that some of 5 may result from ret-
ro-aldo of 8 as discussed above.
The critical step of the carbanion/enolate oxidation by molecu-
lar oxygen proposed in this mechanism is similar to what was pro-
posed by Harmon et al. in their study of autooxidation of rofecoxib
enolate by molecular oxygen,2 although the latter study was con-
ducted in aqueous solutions. One of the key observations in the
current study is that the base-catalyzed autooxidation is very fast,
which is a strong supporting evidence for the mechanism of the
oxidation of carbanion/enolate by molecular oxygen that has al-
ready been dissolved in the reaction solvent (acetonitrile). The
probability for a free radical-mediated mechanism can be excluded
since it needs an induction period which should result in a slow
6. Proposed mechanism of the autooxidation by molecular
oxygen
Based on all the results obtained, in particular those from the
critical isotope experiments, the mechanism of the base-catalyzed
degradation is proposed in Scheme 1. In this mechanism, once the
carbanion (2) is generated through deprotonation by the base, oxi-
dation of the presumed enolate (3, resulting from 2)/carbanion (2)
by molecular oxygen should give the key peroxide anion interme-
diate (6). Starting from this point, further degradation via ‘Route a’
would produce 21-aldehyde (7) during which process hydrogen
peroxide anion (HOOÀ) should leave as a by-product. On the other
hand, further degradation via ‘Route b’ would yield the Baeyer–Vil-
liger-type degradant, 17-diformylanhydride (9). Due to the fact
that 10 and 11 were formed more quickly in the early stage of

4580
M. Li et al. / Tetrahedron Letters 50 (2009) 4575–4581
O
O
OH
O
OH
OH
O
O
O2
OH
OH
O
OH
OH
+
O2
OH
3
2
Betamethasone radical
/superoxide radical complex
6
Scheme 2. The caged free radical mechanism, alternate to the direct pathway as shown in Scheme 1, for the formation of the peroxide intermediate (6).
reaction; the same conclusion was also reached by Harmon et al.2
One inevitable question regarding the direct formation of the
hydroperoxide anion (6) from carbanion (2)/enolate (3) is that such
a process14 would violate the spin conservation rule.15–17 The pro-
posed answer to this paradoxical concern is that 2 would transfer a
single electron to molecular oxygen to form a pair of a carbon free
radical and superoxide free radical in a ‘caged’ environment, fol-
lowed by spin inversion and then combination to produce 6
(Scheme 2). However, regardless of which pathway may be opera-
tive, the net result is a very efficient oxidation of 2/3 by molecular
oxygen which appears to display no induction period during the
oxidation. It has been shown in the current study that the rate of
the carbanion/enolate oxidation by molecular oxygen is positively
correlated to the strength and concentration of the base used.
When the base strength is strong (particularly in aprotic solvent)
and concentration is high enough, the carbanion/enolate autooxi-
dation appears to be instantaneous (Fig. 3).
Although autooxidation of carbanion/enolate has been studied
and reviewed in the literature,18–20 it seems that its relevancy
has been overshadowed by the predominant free radical mecha-
nism as the latter mechanism would automatically be applied by
default even though the evidences may be clearly against it.21 Part
of the reason may be that the majority of the work in autooxidation
of carbanion/enolate has been conducted with strong alkoxide
bases such as potassium t-butoxide in organic solvents;18,20 there-
fore, the relevance of the work may not become obviously applica-
ble during pharmaceutical development process in which the
stability of drug substances is frequently assessed in aqueous med-
ia. For example, Hansen and Bundgaard studied degradation pat-
tern of hydrocortisone (Fig. 1) in aqueous solutions at pH
between 0 and 11.5 In basic solution, they found the formation of
three oxidative degradants, hydrocortisone 21-aldehyde, 20-hy-
droxy-21-acid, and 17-acid. More recently, Edmonds et al. reported
the formation of the corresponding 21-aldehyde, 20-hydroxy-21-
acid, and 17-acid degradants from autooxidation of dexametha-
sone (the authors’ term for autooxidation is aerial oxidation) under
various alkaline pH conditions.6 In both cases, no detailed mecha-
nistic explanation was given regarding the oxidative degradation.
Edmonds et al. did compare the oxidation with and without labo-
ratory lighting and found no difference, which eliminated the pos-
sibility for the involvement by singlet oxygen.22 By going through
the results obtained by these authors,4–7 it is apparent that the
mechanism proposed here (Scheme 1) can be readily applied to
the oxidation of corticosteroids containing 20-keto-21-hydroxyl
side chain including hydrocortisone and dexamethasone in various
alkaline conditions, in particular at higher pH region where the ob-
served autooxidation is fairly efficient.
anion intermediate (such as 6). Subsequent pathways from 6
would satisfactorily explain the product distribution among 17-ke-
tone (5), 21-aldehyde (7), 20-hydroxy-21-acid (8), 17-formyloxy-
17-acid (10), and 17-acid (11) under various conditions. The for-
mation of the degradants, analogous to 5, 7, 8, 10, and 11 from a
large number of structurally related corticosteroids such as hydro-
cortisone and dexamethasone, under alkaline conditions has been
reported during the past several decades.3–5 The current proposed
mechanism is able to provide a reasonable explanation for this
decades-old observation, despite the fact that our studies were
performed in a largely organic solvent environment. Water is a
strong protic solvent and as such, its presence will lower the intrin-
sic basicity of a base leading to slower reaction rate, rendering the
carbanion/enolate-mediated autooxidation less obvious but more
susceptible to confusion with free radical-mediated autooxidation.
It appears that the direct autooxidation of carbanion/enolate by
molecular oxygen may play a much more significant role in the
oxidative degradation of compounds that contain ‘acidic’ CH
protons.
Acknowledgments
We would like to thank Dr. T. M. Chan’s group at Schering-
Plough Research Institute for performing all the NMR experiments.
Supplementary data
1H and 13C NMR results of betamethasone 17-acid, betametha-
sone 21-aldehyde, and betamethasone 17-ketone are available.
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.tetlet.2009.05.074.
References and notes
1. ICH Harmonised Tripartite Guideline, Validation of Analytical Procedures: Text
and Methodology, Q2(R1), Current Step 4 version (November 2005), Parent
Guideline dated 27 October 1994.
2. Harmon, P. A.; Biffar, S.; Pitzenberger, S. M.; Reed, R. A. Pharm. Res. 2005, 22,
1716–1726.
3. In our studies of the base-catalyzed autooxidation, betamethasone and the
related corticosteroids containing the same 20-keto-21-hydroxyl side chain,
such as betamethasone 9,11-epoxide, dexamethasone, and dexamethasone
9,11-epoxide, were found to undergo the same oxidative degradation
pathways. In this Letter, we present the data from the stress study of
betamethasone as an exemplary case study.
4. Guttman, D. E.; Meister, P. O. J. Am. Pharm. Assoc. 1958, 47, 773–778.
5. Hansen, J.; Bundgaard, H. Int. J. Pharmaceut. 1980, 6, 307–319.
6. Edmonds, J. S.; Morita, M.; Turner, P.; Skelton, B. W.; White, A. H. Steroids 2006,
71, 34–41.
7. Hidaka, T.; Huruumi, S.; Tamaki, S.; Shiraishi, M.; Ninato, H. Yakugaku Zasshi
1980, 100, 72–80.
8. All the reactions studied in this Letter were monitored by LC–MS and/or high-
resolution LC–MS. The characterization of the degradants formed during all the
forced degradation studies was done through one of the two ways: (1) For the
majority of the degradants formed, comparative LC–MS analyses were
performed against authentic reference compounds available in-house. (2) In
cases where authentic compounds were not available, high-resolution LC–MS
characterization was further confirmed by 1D and 2D NMR determination. All
In summary, we have proposed a mechanism for the autooxida-
tion of corticosteroids containing 20-keto-21-hydroxyl side chain
under alkaline conditions in which direct oxidation of the pre-
sumed carbanion/enolate at the 21-position by molecular oxygen
is the critical step leading to the formation of the key 21-peroxide

M. Li et al. / Tetrahedron Letters 50 (2009) 4575–4581
4581
the compounds generated in this study were found to be within 3 ppm of their
theoretical formulas by high-resolution LC–MS. The LC–MS analyses were
performed under ESI positive mode either on a Thermo Electron Surveyor HPLC
system coupled with a PDA detector and an MSQ Plus MS detector or on a
Thermo Electron Surveyor HPLC system coupled with a PDA detector and an
LTQ MS detector or a high-resolution Orbitrap MS detector, with the MS
conditions similar to those published by our group previously: Li, M.; Chen, B.;
Lin, M.; T.-M. Chan; Rustum, A. Tetrahedron Lett. 2007, 48, 3901–3905. The LC
conditions of the LC–MS methods used are summarized as follows. (1) For the
conditions used in Figure 2, the chromatographic elution was effected
isocratically on a Supelco Supercosil ABZ-plus column (25 cm  4.6 mm ID,
21-dehydrobetamethasone, (11b,16b)-9-fluoro-11,17-dihydroxy-16-methyl-
pregna-1,4-diene-3,20-dione-21-al; betamethasone 20-hydroxy-21-acid (8),
betamethasone glycolic acid, (11b,16b)-9-fluoro-3-oxo-11,17,20-trihydroxy-
16-methylpregna-1,4-diene-21-oic acid; betamethasone 17-formyloxy-17-
acid
methylandrosta-1,4-diene-17-carboxylic acid; betamethasone 17-acid (11),
(11b,16b,17 )-9-fluoro-11,17-dihydroxy-3-oxo-16-methylandrosta-1,4-diene-
(10),
(11b,16b,17a)-9-fluoro-17-(formyloxy)-11-hydroxy-3-oxo-16-
a
17-carboxylic acid.
10. Trofimov et al. studied the basicity of alkali metal hydroxides in a dipolar
aprotic solvent DMSO: Trofimov, B. A.; Valsil’tsov, A. M.; Amosova, S. V. Russ.
Chem. Bull. 1986, 35, 682–686.
11. We have observed that treatment of betamethasone or a number of its related
compounds such as enol aldehyde with MnO2 can easily generate 17-ketone.
12. Kozyrev, A. N.; Dougherty, T. J.; Pandey, R. K. Chem. Commun. 1998, 481–482.
13. Conrow, R. E.; Dillow, G. W.; Bian, L.; Xue, L.; Papadopoulou, O.; Baker, J. K.;
Scott, B. S. J. Org. Chem. 2002, 67, 6835–6836.
14. Gersmann, H. R.; Bickel, A. F. J. Chem. Soc. (B) 1971, 11, 2230–2237.
15. Russell, G. A. J. Am. Chem. Soc. 1954, 76, 1595–1600.
16. Russell, G. A.; Bemis, A. G. J. Am. Chem. Soc. 1966, 88, 5491–5497.
17. Barton, D. H. R.; Jones, D. W. J. Chem. Soc. 1965, 3563–3570.
18. Sucrow, W. Chem. Ber. 1967, 100, 259–264.
5
l
m) with a mobile phase consisting of 55% of A (0.2% acetic acid) and 45% of B
(acetonitrile) at a flow rate of 2.0 mL/min. (2) For the conditions used in Figure
3
and Table 1, the chromatographic elution was effected on Supelco
a
Supercosil ABZ-plus column (25 cm  4.6 mm ID,
5
l
m) with a gradient
generated between mobile phase A (0.1% TFA in water) and B (0.1% TFA in
acetonitrile) at a flow rate of 1.3 mL/min according to the following program:
0–10 min, 30–60%B; 10.1–15 min, 30%B. (3) The conditions used in Figure 4
were similar to those used in Figure 3 and Table 1, except that the flow rate was
2.0 mL/min and the gradient program was slightly different: 0–10 min, 30–
75%B; 10.1–15 min, 30%B.
9. A different nomenclature has been used in the literature for the majority of the
degradants mentioned in this Letter. We prefer to use a more descriptive and
self-explanatory name such as betamethasone 21-aldehyde. All the compound
names we used are listed below along with other conventional names (if
available) and the IUPAC names: betamethasone (1), (11b,16b)-9-fluoro-
11,17,21-trihydroxy-16-methylpregna-1,4-diene-3,20-dione; betamethasone
17-ketone (5): (11b,16b)-9-fluoro-11-hydroxy-16-methylandrosta-1,4-diene-
3,17-dione; betamethasone 21-aldehyde (7), betamethasone glyoxal,
19. Frimer, A. A.; Hameiri-Bush, J.; Ripshtos, S.; Gilinksky-Sharon, P. Tetrahedron
1986, 42, 5693–5706.
20. Sosnovsky, G.; Zaret, E. H. In Organic Peroxides; Swern, D., Ed.; J. Wiley and
Sons: New York, 1971; pp 517–560.
21. Mao, B.; Abrahim, A.; Ge, Z.; Ellison, D. K.; Hartman, R.; Prabhu, S. V.; Reamer, R.
A.; Wyvratt, J. J. Pharm. Biomed. Anal. 2002, 28, 1101–1113.
22. We did not see any difference in our own studies either with or without
laboratory lighting.