Optimization of the synthesis of a key intermediate for the preparation of glucocorticoids

Article abstract of DOI:10.1016/j.steroids.2018.06.007

A short and efficient synthesis, based on a one-step double elimination, of a key intermediate in the synthesis of various glucocorticosteroids has been developed. This method can be carried out on large scale for further industrial applications. The synthesis allowed us to identify a novel prednisolone derivative 10 and its anti-inflammatory activity was determined in an in vivo model of inflammation. In order to understand the regioselectivity of the double elimination under various conditions, mechanistic studies were undertaken and confirmed the experimental results. We also propose a mechanism for the formation of the new steroid 10 studied by molecular modeling.

Full text of DOI:10.1016/j.steroids.2018.06.007

Accepted Manuscript
Optimization of the synthesis of a key intermediate for the preparation of glu-
cocorticoids
Romain Jouve, Vincent Thery, Sylvie Ducki, Julie Helfenbein, Jean-Christophe
Thiery, Aurélie Job, Elodie Picard, Christophe Mallet, Isabelle Ripoche, Khalil
Bennis
PII:
S0039-128X(18)30116-8
https://doi.org/10.1016/j.steroids.2018.06.007
STE 8277
DOI:
Reference:
To appear in:
Steroids
Received Date:
Revised Date:
Accepted Date:
15 February 2018
15 June 2018
19 June 2018
Please cite this article as: Jouve, R., Thery, V., Ducki, S., Helfenbein, J., Thiery, J-C., Job, A., Picard, E., Mallet,
C., Ripoche, I., Bennis, K., Optimization of the synthesis of a key intermediate for the preparation of glucocorticoids,
Steroids (2018), doi: https://doi.org/10.1016/j.steroids.2018.06.007
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Optimization of the synthesis of a key intermediate for the preparation of
glucocorticoids
a
a
a
b
b
Romain Jouve , Vincent Thery , Sylvie Ducki , Julie Helfenbein , Jean-Christophe Thiery ,
a
c
c
a*
a*
Aurélie Job , Elodie Picard , Christophe Mallet , Isabelle Ripoche , Khalil Bennis .
a
Université Clermont Auvergne, CNRS, SIGMA Clermont, ICCF, 63000 Clermont–Ferrand,
France.
b
Sanofi, Le Bourg, 63480 Vertolaye, France.
c
Université Clermont Auvergne, Inserm, Neuro-Dol, 63000 Clermont-Ferrand, France.
1

Abstract
A short and efficient synthesis, based on a one-step double elimination, of a key intermediate
in the synthesis of various glucocorticosteroids has been developed. This method can be
carried out on large scale for further industrial applications. The synthesis allowed us to
identify a novel prednisolone derivative 10 and its anti-inflammatory activity was determined
in an in vivo model of inflammation. In order to understand the regioselectivity of the double
elimination under various conditions, mechanistic studies were undertaken and confirmed the
experimental results. We also propose a mechanism for the formation of the new steroid 10
studied by molecular modeling.
Keywords
Prednisolone derivatives
Anti-inflammatory activities
Synthesis optimization
Mechanistic studies
Elimination reactions
Wagner-Meerwein rearrangement
Abbreviations
ACN: Acetonitrile
AUC: Area Under Curve
DBU: 1,8-Diazabicyclo[5.4.0]undec-7-ene
DMAP: N,N-dimethyl-4-aminopyridine
Ac: Acetyl
eq: Equivalent
2

1
. Introduction
Tetraene 1 is an important industrial key precursor in the synthesis of numerous
commercial bioactive glucocorticoid drugs, such as dexamethasone 2 and its derivative 3,
fluticasone propionate 4, flumethasone 5 and mometasone furoate 6 (Scheme 1) [1]. These
glucocorticoids are used as therapeutic agents for the treatment of inflammatory skin
disorders and respiratory diseases.
Scheme 1. Examples of glucocorticosteroids synthesized from tetraene 1
The development of an efficient and robust synthesis of this common intermediate 1,
starting from commercially available prednisolone 7 (Scheme 2) is still a challenge for
delivering a resilient and economically viable supply chain for these drugs. Various syntheses
of tetraene 1 have been reported [1-6] but a quicker and cheaper route would offer a
substantially competitive advantage. Indeed, most of these glucocorticosteroids are currently
prepared in multistep syntheses starting from cortisol derivatives or cholesterol analogues [2].
3

Scheme 2. Synthetic route considered
Among the reported syntheses to lead to the tetraene 1 starting from prednisolone 7,
the most efficient synthesis described recently uses the Mattox rearrangement. [1] The
reported strategy consisted in a three-step route using successive protection of the primary
hydroxyl group, elimination of the hydroxyl group at C-11 and rearrangement of the lateral
chain at C-17 affording the tetraene 1 in a moderate yield (54%).
The lack of relevant methods described in the literature, together with the potential
economic issues with the current methods, led us to develop a simultaneous elimination of
both alcohols at C-11 and C-17. [7-10]
The objective of this work is to optimize a short method to synthesize the key
intermediate tetraene 1 starting from commercially available prednisolone 7 and we envisaged
to develop a two-step procedure.
2
. Experimental
.1. Chemistry
Purity was controlled by HPLC (Agilent 1100 Series) and was >95% (unless otherwise
indicated). Melting points (m. p.) were determined using a Büchi® and are given uncorrected.
2
1
13
H and C NMR spectra were recorded on a Brüker Avance spectrometer at 400 or 500 MHz.
Chemical shifts δ are reported in ppm relative to residual solvent. The coupling constants J
are given in Hertz (Hz). The abbreviations used for signal descriptions are as follows: s,
singlet; d, doublet; t, triplet; q, quartet, m, multiplet, dd, doublet of doublets, qd, quartet of
doublets.
4

High Resolution Electro-Spray Ionisation Mass Spectra (HR-ESI-MS) were obtained from the
UCA-START (Solutions Techniques d’Analyses et Ressources Technologiques).
General procedure for the preparation of compound 1. To a solution of prednisolone acetate 8
in anhydrous DMF (10 volume), at various temperatures, were added anhydrous pyridine (1
eq) then SOCl (see Table 1). The reaction mixture was stirred until consumption of the
2
starting material (TLC), then water and ethyl acetate were added. Layers were separated and
the organic layer was washed with 1N HCl. The organic layer was dried over anhydrous
MgSO , filtered and concentrated in vacuo to give an orange solid. The crude product was
4
purified by flash chromatography (cyclohexane/ ethyl acetate, 6/4, V/V) to furnish 1 as a
white solid, accompanied by compounds 9-11.
2
1-Acetoxypregna-1,4,9(11),16-tetraene-3,20- dione 1 (Prednisolone acetate). m. p.: 171-172
1
°
C. H NMR (400 MHz, CDCl
3
) δ 7.19 (d, J = 10.0 Hz, 1H, H-1), 6.74 (dd, J = 3.2, 1.9 Hz,
1
H, H-16), 6.24 (dd, J = 10.0, 1.8 Hz, 1H, H-2), 6.03 (m, 1H, H-4), 5.53 (m, 1H, H-11), 5.00
(
d, J = 16.1 Hz, 1H, H-21), 4.87 (d, J = 16.0 Hz, 1H, H-21), 2.70 – 2.49 (m, 2H, H-6 + H-12),
.47 – 2.37 (m, 2H, H-15 + H-7), 2.19 – 2.03 (m, 7H, H-23 + H-12 + H-6 + H-15 + H-8), 1.48
2
(
(
m, 1H ,H-14), 1.39 (s, 3H, H-19), 1.26 – 1.12 (m, 1H, H-7), 0.88 (s, 3H, H-18). HR-ESIMS
M+H): m/z calculated: 367.1865; m/z found: 367.1775.
2
1-Acetoxy-17α-hyroxypregna-1,4,9(11)-triene-3,20-dione 9 was obtained as a white
1
powder. m. p.: 220-222 °C. H NMR (400 MHz, CDCl
3
) δ 7.19 (d, J = 10.0 Hz, 1H, H-1),
6
.25 (dd, J = 10.0, 1.9 Hz, 1H, H-2), 6.03 (m, 1H, H-4), 5.55 (m, 1H, H-11), 5.03 (d, J = 17.5
Hz, 1H, H-21), 4.84 (d, J = 17.5 Hz, 1H, H-21), 2.98 – 2.83 (m, 1H, OH), 2.73 (m, 1H, H-16),
2
2
1
1
.69 – 2.59 (m, 2H, H-6 + H-12), 2.44 – 2.35 (m, 1H, H-6), 2.34 – 2.24 (m, 1H, H-8), 2.20 –
.12 (m, 4H, H-23 + H-7), 1.98 – 1.77 (m, 3H, H-15 + H-12 + H14), 1.59 (m, H-16), 1.48 –
.40 (m, 1H, H-15), 1.38 (s, 3H, H-19), 1.19 (qd, J = 12.8, 4.4 Hz, 1H, H-7), 0.64 (s, 3H, H-
1
3
8). C NMR (101 MHz, CDCl
3
) δ 205.5 (C-20), 186.5 (C-3), 170.7 (C-22), 167.5 (C-5),
5

1
4
3
55.4 (C-1), 142.5 (C-9), 127.1 (C-2), 123.9 (C-4), 121.1 (C-11), 89.8 (C-17), 68.2 (C-21),
8.4 (C-14), 46.9 (C-10), 46.1 (C-13), 37.1 (C-8), 35.2 (C-16 or C-7), 34.9 (C-16 or C-7),
2.5 (C-6 or C-12), 32.3 (C-6 or C-12), 26.9 (C-19), 24.8 (C-15), 20.8 (C-23), 14.5 (C-18).
HR-ESIMS (M+H): m/z calculated: 385.1970; m/z found: 385.2001.
2
-((10S)-10,17-Dimethyl-3-oxo-6,7,8,10,12,15,16,17-octahydro-3H-cyclopenta[a]phenanthre
1
n-17-yl)-2-oxoethyl acetate 10 was obtained as white powder. m. p.: 54-56 °C. H NMR (400
MHz, CDCl
.53 (m, 1H, H-11), 4.74 (d, J = 16.5 Hz, 1H, H-21), 4.50 (d, J = 16.5 Hz, 1H, H-21), 3.19
m, 1H, H-8), 2.75 – 2.66 (m, 1H, H-6), 2.66 – 2.54 (m, 2H, H-15 or H-16 + H-12), 2.48 –
.24 (m, 5H, H-6 + H-7 + H-12 + H-15 + H-16), 2.08 (s, 3H, H-23), 1.77 (m, 1H, H-16), 1.42
3
) δ 7.10 (d, J = 10.0 Hz, 1H, H-1), 6.26 (d, J = 10.0, 1H; H-2), 6.00 (s, 1H, H-4),
5
(
2
1
3
(
s, 3H, H-19), 1.32 – 1.23 (m, 1H, H-7), 1.19 (s, 3H, H-18). C NMR (101 MHz, CDCl
3
) δ
2
1
3
2
05.7 (C-20), 186.2 (C-3), 170.3 (C-22), 166.0 (C-5), 154.7 (C-1), 140.7 (C-9), 139.4 (C-13),
33.3 (C-14), 127.6 (C-2), 123.9 (C-4), 118.2, (C-11), 65.2 (C-21), 60.9 (C-10), 46.4 (C-17),
5.5 (C-8), 34.9 (C-7 or C-16), 34.5 (C-7 or C-16), 32.5 (C-6 or C-12), 32.3 (C-6 or C-12),
5.0 (C-15), 24.8 (C-23), 20.6 (C-18 or C-19), 20.4 (C-18 or C-19). HR-ESIMS (M+H): m/z
calculated: 367.1865; m/z found: 367.1801.
1
Dimer 11 was obtained as a brown oil. H NMR (500 MHz, CDCl
3
) δ 7.17 (dd, J = 10.0, 6.4
Hz, 2H, H-1 + H-1’), 6.27 (m, 2H, H-2 + H-2’), 6.09 (d, J = 1.0 Hz, 2H, H-4 + H-4’), 5.56 (d,
J = 5.5 Hz, 2H, H-11 + H-11’), 5.09 (d, J = 17.5 Hz, 1H, H-21 or H-21’), 4.97 (d, J = 17.5
Hz, 1H, H-21 or H-21’), 4.93 (d, J = 17.5 Hz, 1H, H-21 or H-21’), 4.85 (d, J = 17.5 Hz, 1H,
H-21 or H-21’), 2.93 – 2.84 (m, 2H, H-16 + H-16’), 2.72 – 2.61 (m, 3H, H-6 + H-6’ + H-12
or H-12’), 2.49 – 2.41 (m, 3H, H-6 + H-6’ + H-7 or H-7’), 2.39 – 2.21 (m, 5H, H-8 + H-8’ +
H-16 + H-16’ + H-15 or H-15’), 2.19 – 2.06 (m, 9H, H-23 + H-23’ + H-15 or H-15’ + H-12
or H-12’ + H-7 or H-7’), 1.99 – 1.94 (m, 2H, H-12 + H-12’), 1.83 – 1.68 (m, 2H, H-14 + H-
1
4’), 1.57 – 1.47 (m, 2H, H-15 + H-15’), 1.41 (s, 3H, H-19 or H-19’), 1.40 (s, 3H, H-19 or H-
6

1
1
3
9’), 1.26 – 1.16 (m, 2H, H-7 + H-7’), 0.71 (s, 3H, H-18 or H-18’), 0.70 (s, 3H, H-18 or H-
1
3
8’). C NMR (101 MHz, CDCl
3
) δ 199.1 (C-20 or C-20’), 198.9 (C-20 or C-20’), 186.4 (C-
or C-3’), 186.3 (C-3 or C-3’), 170.7 (C-22 or C-22’), 170.4 (C-22 or C-22’), 166.3 (C-5 or
C-5’), 166.1 (C-5 or C-5’), 154.4 (C-1 or C-1’), 154.2 (C-1 or C-1’), 142.5 (C-9 or C-9’),
42.4 (C-9 or C-9’), 127.5 (C-2 or C-2’), 127.4 (C-2 or C-2’), 124.0 (C-4 or C-4’), 123.9 (C-4
1
or C-4’), 119.8 (C-11 or C-11’), 119.7 (C-11 or C-11’), 99.4 (C-17 or C-17’), 99.0 (C-17 or
C-17’), 67.8 (C-21 or C-21’), 67.7 (C-21 or C-21’), 48.7 (C-14 or C-14’), 48.6 (C-14 or C-
1
3
3
4’), 48.4 (C-10 or C-10’), 48.3 (C-10 or C-10’), 45.8 (C-13 or C-13’), 45.7 (C-13 or C-13’),
7.0 (C-8 or C-8’), 34.9 (C-7 or C-7’ or C-16 or C-16’), 34.8 (C-7 or C-7’ or C-16 or C-16’),
4.7 (C-7 or C-7’ or C-16 or C-16’), 34.6 (C-7 or C-7’ or C-16 or C-16’), 32.1 (C-6 or C-6’ or
C-12 or C-12’), 32.0 (C-6 or C-6’ or C-12 or C-12’), 31.9 (C-6 or C-6’ or C-12 or C-12’),
1.8 (C-6 or C-6’ or C-12 or C-12’), 26.7 (C-23 or C-23’), 26.6 (C-23 or C-23’), 24.3 (C-15
or C-15’), 24.2 (C-15 or C-15’), 20.4 (C-19 or C-19’), 20.3 (C-19 or C-19’), 13.7 (C-18 or C-
3
1
8’), 13.5 (C-18 or C-18’). HR-ESIMS (M+H): m/z calculated: 815.3420; m/z found:
8
15.3441.
Scale-up synthesis of compound 1. To a solution of prednisolone acetate 8 (30 g, 74.5 mmol)
in anhydrous DMF (300 ml) at 20 °C were added anhydrous pyridine (59 ml, 731 mmol, 9.8
eq) then SOCl (10.8 ml, 149 mmol, 2 eq). The reaction mixture was stirred for 15 min, then
2
water (300 mL) and ethyl acetate (1000 mL) were added. Layers were separated and the
organic layer was washed with 1N HCl (2 x 200 mL). The organic layer was dried over
MgSO , filtered and concentrated in vacuo to give an orange solid (32 g, 87.4 mmol). The
4
crude was purified by flash chromatography (6/4 cyclohexane/ ethyl acetate) to furnish a
white solid (13.6 g, 37.25 mmol, 50 %).
7

2
.2. Molecular modelling studies
All computations were performed with GAUSSIAN 09 rev D03 [11] and the two-layer
ONIOM [12, 13] method with mechanical embedding. For the “LOW” layer a classical
mechanics force field level UFF [14] was applied with QEQ charges. [15] For the “HIGH”
layer, the Density Functional Theory (DFT) [16,17] level with B3LYP [18-21] hybrid
functional and 6-31G basis sets were applied. Energy minimizations were performed with the
Quadmac option [22]. The thermodynamic corrections were finally applied. All energy values
presented below are Gibbs free energy.
2
.3. Anti-inflammatory activity
Animals. All experiments and animal handling were conformed to the guidelines of the
Committee for Research and Ethical Issues of IASP [23]. All procedures and animal handling
were carried out in accordance with the European Commission’s directive 2010/63/EU and
were approved by the local Ethics Committee of Auvergne (CEMEA, France). Great care was
taken to avoid or minimize discomfort of the animals. All experiments involving animals
were performed in accordance with the ARRIVE guidelines [24, 25]. Experiments were
carried out using male C57BL6/J mice (20-24 g) from Janvier Labs (Le Genest St Isle,
France). Mice were housed under controlled environmental conditions (21-22 °C; 55%
humidity) and kept under a 12/12 hours light/dark cycle. Food and water were available ad
libitum.
Model of inflammation. Inflammation was induced by an intraplantar (left hindpaw)
subcutaneous injection of 20 µl of Complete Adjuvant of Freund (CFA, Becton, Dickinson
and company, USA).
Edema measurement. Paw edema induced by CFA injection was measured 7 days after
induction of inflammation using a caliper. Assessment of edema was performed before
(
baseline, T0) and 30 minutes, 1, 2, 3, 4, 6, 8 and 24 hours after drugs administration (5
8

groups of 7 mice). The importance of the edema was assessed by determining the percentage
of augmentation in the size of the paw of the mouse calculated by: (paw size at measured time
–
initial paw size at T0) * 100 / initial paw size at T0.
Statistical analysis. Data are expressed as mean ± SEM and recorded using Prism® software.
Data were compared by two-way ANOVA. The post hoc comparisons were performed by the
Tukey method. Values of P<0.05 were considered statistically significant.
3
. Results and discussion
3
.1. Chemistry
As the lateral chain at C-17 is not stable under acidic or basic conditions and in order
to avoid the Mattox rearrangement [26-29], we first envisaged a two-step synthesis consisting
in acetylating the hydroxyl group at C-21 of prednisolone 7 and then performing simultaneous
9
,11,16,17
dehydration of the 11,17-diol 8 to generate the 
-tetraene 1 (Scheme 3).
The selective acetylation of the alcohol at C-21 has already been described [30]. The
primary hydroxyl group at C-21 was first acetylated using pyridine and acetic anhydride to
provide the desired 21-acetate 8 in 98% yield (Scheme 3). We then focused our attention on
the “double” elimination, which is the key step in the synthesis of compound 1.
Scheme 3. Synthesis of tetraene 1
The only method described to eliminate simultaneously both alcohols on similar
structures, without the formation of a leaving group, consisted in using thionyl chloride
(
SOCl ) in the presence of pyridine [8]. The dehydration was performed, under the reported
2
conditions, using 25 eq of thionyl chloride and pyridine as a solvent. Unfortunately these
9

conditions led to the degradation of the reaction mixture, with only traces of the desired
tetraene 1 observed.
Having previously performed similar studies on other steroids (unpublished results),
we decided to carry out the reaction in DMF using 10 eq of pyridine and to study the
influence of various parameters of the reaction such as the temperature and the quantity of
thionyl chloride on the formation of compound 1 (Table 1). The reaction was first carried out
at 0 °C using 3 eq of thionyl chloride (Table 1, Entry 1); the desired compound 1 was
obtained in a low yield (34%). In addition to the expected tetraene 1, we observed the
9
,11
formation of several by-products, of which three were identified (Figure 1):  -alkene 9,
9
,11,13,14

-olefin 10 and the dimer 11 (all were characterized by NMR and mass spectroscopy).
By-product 10 results from a Wagner-Meerwein rearrangement, which has already been
described on similar steroids [31-38].
Figure 1. Main by-products identified
In order to limit the degradation and formation of by-products, the amount of thionyl
chloride was reduced (Table 1, Entry 2); unfortunately the tetraene 1 was obtained in a similar
yield (37%).
1
0

Table 1. Dehydration of compound 8 to generate tetraene 1 (Scheme 3) using 10 eq of pyridine in DMF.
a
Conditions
SOCl
% AUC of each compound
b
Entry
yield (% of 1)
2
T (°C)
9
5
10
12
13
38
40
26
42
11
5
1
(eq)
1
2
3
4
5
6
3
70
57
54
55
37
46
34
37
50
49
29
50
0
15
2
5
2
6
2
0
2.6
4
0
5
0
2
40
2
9
2
a
Analytical method described in the experimental section ; the % of each compound was determined considering
the AUC of the corresponding peak / total AUC (Area Under Curve).
b
Compound 1 was not purified, the chemical yield was deduced by titration of 1 in the crude and the weight of
crude product.
To promote the elimination at C-16,17, the reaction was performed at 20 °C using 2 eq
of thionyl chloride (Table 1, entry 3). A better conversion of diol 8 into tetraene 1 was
9
,11,13,14
observed (50%) concomitantly to an increase in the % AUC of 
-olefin 10. To study
the influence of the amount of thionyl chloride on the formation of 10, we performed the
dehydration of diol 8 using 2.6 and 4.0 eq of thionyl chloride (Table 1, entries 4 and 5
respectively). The ratio between compounds 1 and 10 remained the same but the use of a
higher amount of thionyl chloride furnished a lower yield of tetraene 1 (49% vs 29%),
probably due to the degradation on the media.
Finally, the influence of the temperature on the formation of each compound was
investigated. The reaction was then performed at 40 °C (Table 1, entry 6) and gave a higher
yield of compound 1 (50%).
Since similar yields were obtained when the reaction was carried out at 20° C and
4
0° C, the optimization was performed at 20 °C, for economic and environmental reasons.
1
1

In order to further optimize the formation of the desired compound, the influence of
the nature of the base was investigated: seven bases were selected according to their pKa
values (5.2-12) and their structure (steric hindrance). The results showed that with bases
having pKa values between 6 and 7.4 (Table 2, entries 2 to 4), the dehydration was efficient at
9
,11
position-9,11 to provide  -alkene 9 (probably because the 11-OH is less hindered);
however the elimination at C-16,17 to afford tetraene 1 was not complete. The use of stronger
bases (pKa > 9) such as triethylamine, DMAP or DBU (entries 5 to 7) did not give the desired
9
,11,16,17

-tetraene 1 but furnished mostly the monodehydrated alkene 9 in 26%, 55% and 49%
yield respectively.
Table 2. Optimization of the synthesis of tetraene 1 from diol 8.
a
Conditions
Base (eq)
% AUC of each compound
b
Entry
Yield
SOCl
2
% 9
2
% 10
38
22
29
15
0
% 11
% 1
54
29
46
13
0
(
eq)
1
2
3
4
5
6
7
Pyridine (10)
4-Methylpyridine (10)
2,6-Lutidine (10)
N-Methylmorpholine (10)
Triethylamine (10)
DBU (10)
2
6
7
50% (1)
34% (1)
40% (1)
14% (1)
26% (11)
55% (11)
49% (11)
2
2
2
2
2
2
41
23
50
28
72
80
2
5
10
5
0
0
DMAP (10)
0
2
0
a
Analytical method described in the experimental section ; the % of each compound was determined considering
the AUC of the corresponding peak / total AUC.
b
Compounds 1 or 9 were not purified, the chemical yield was deduced by titration of 1 or 9 in the crude and the
weight of crude product.
On the basis of these results on the double dehydration reaction, we found that the
treatment of diol 8 in DMF, in the presence of pyridine (10 eq) and thionyl chloride (2 eq) at
9
,11,16,17
2
0 °C gave the best results and allowed the formation of 
-tetraene 1 in 50% yield.
1
2

Since our goal was to develop a short and efficient synthesis of tetraene 1 applicable to
industrial scale, these optimized conditions were applied on a larger scale (30 g). The desired
9
,11,16,17

-tetraene 1 was obtained in 50% yield, demonstrating the reproducibility of the
procedure.
.2. Mechanistic study
To our knowledge, no mechanism has been described in the literature for the double
3
dehydration of corticosteroid diols in the presence of thionyl chloride and base. We have
undertaken two molecular modeling studies to simulate the mechanism for the eliminations of
both hydroxyl groups of 8 (to afford tetraene 1) and the formation of the Wagner-Meerwein
product 10.
The chemical system evolved according to the chemical reaction. The common model
contained 11,17-diol 8 (1 eq), at least one eq of thionyl chloride, pyridine or pyridinium ions
and 70 molecules of solvent (DMF). The HIGH layer contained all atoms of diol 8, thionyl
chloride and pyridine. The LOW layer contained all solvent molecules. The solvent is neither
involved directly nor as a catalyst of the reaction, and is therefore excluded from the HIGH
layer. The solvent is necessary because it provides a constraint in the relative position of
pyridine and steroid.
We studied the two-step pathway reaction; the first step consisting in the addition of
the hydroxyl group on thionyl chloride, the second step being the elimination of the leaving
group formed during the first step.
1
3

3
.2.1.
.2.1.1. Mechanism of the addition step
Considering the chemical species present at the beginning of the reaction, the first
Study of the dehydration of diol 8
3
chemical model studied consisted in using one eq of compound 8, one of SOCl and one of
2
pyridine, without pyridinium salt (IRC Reactant).
Table 3. Addition reaction at 11-OH and 17-OH. Relative Gibbs Free energy profile of the considered species
IRC reactant 8, TS 8a, IRC product 8b) using one eq of compound 8, one of SOCl and one of pyridine. For
details see Supporting Information Table S1 and S2. For clarity, the solvent molecules are not displayed.
(
2
-
1
Relative Gibbs Free energy profile (G) in kcal.mol
Position
IRC Reactant 8
TS 8a
IRC Product 8b
1
1
1-OH
7-OH
0
.00
6.01
1.99
-0.62
3.92
0.00
From this first calculation (Table 3), we can observe that the addition reaction of the
-1
hydroxyl group on C-11 on SOCl
with a low activation energy (TS) of 6.01 kcal.mol . This step is fast and the formed adduct
IRC Product) accumulates. Conversely, the energy of the IRC product, in the case of the 17-
2
(IRC Product 8b) is slightly exothermic (–0.62 kcal.mol )
-1
(
-1
-1
OH, is higher (3.92 kcal.mol ) than the energy of its TS (1.99 kcal.mol ), the addition on the
hydroxyl group on C-17 cannot occur. Considering these results, we conclude that, at the
beginning of the reaction, the addition of the hydroxyl group on the thionyl chloride would
occur at C-11 (fast), but is not favored at C-17 (slow). The difference in reactivity between
1
4

the 11-OH and 17-OH could be explained by the stabilizing effect of the leaving chloride ion,
less favorable on C-17 than on C-11 (Supporting information: Tables S1, S2 and S3).
We then added pyridinium ion (1 eq) to the reaction system, which represents the state
at half of the reaction. After minimization, the product is obtained directly without going
through a TS (Table 4). In this case, we observed an exothermic reaction (G <0). In both
cases, the addition reaction is also fast (no TS) but more exothermic at 17-OH than at 11-OH
(
Table 4). These results are compatible with the experimental results (Table 1) which showed
that the elimination at C-11 is carried out first (presence of product 9), while the elimination
at C-17 takes place subsequently.
Table 4. Relative Gibbs Free energy profile of the considered species (IRC reactant, TS, IRC product) using one
eq of compound 8, one of SOCl
2
, 1 eq of pyridine and 1 eq of pyridinium ion. The energy of the Reactant
column is obtained with one system in which the leaving chloride does not interact with the pyridium.
-
1
Relative Gibbs Free energy profile (G) in kcal.mol
Position
Reactant
Minimized Product
1
1
1-OH
7-OH
0.00
0.00
-30.20
-71.69
3
.2.1.2. Mechanism of the elimination step
For the second step, the chemical system is constituted of one eqof intermediate 8b
Cl-S(=O)-O-C(11/17)), two eq of pyridinium ion, one eq of pyridine and one eq of chloride
(
ion. The elimination step at both positions (C-11 or C-17) was carried out in the presence of a
free hydroxyl group on the other position. We first envisaged two possible mechanisms (E1
and E2) during a scan of the coordinates of the reaction. For the E1 mechanism, the distance
between C-17 and its substituent –OS(=)OCl was stretched to force the formation of the
carbocation ; unfortunately no TS for this mechanism was found. For the E2 mechanism, the
distance between the pyridine and H-16 was decreased to force the formation of the alkene.
Under these conditions, a TS for the E2 elimination was found.
1
5

Table 5. Elimination reaction at C-9,11 and C-16,17. Relative Gibbs Free energy profile of the considered
species (IRC reactant, TS, IRC product) using one eq of intermediate 8b, 1 eq of pyridine and 2 eq of pyridinium
ion.
-
1
Relative Gibbs Free energy profile (G) in kcal.mol
Position
IRC Reactant 8b
TS 8c
IRC Product 8d
C9-11
0
.00
16.30
18.16
-4.17
-6.15
C16-17
0.00
The energy profiles (Table 5) of the elimination at both positions are very similar. The
-
1
difference in activation energies between the two positions (TS) is low (1.8 kcal.mol ) but
slightly in favor of the elimination at C-9,11, whereas the elimination (IRC product) at C-
-
1
1
6,17 is slightly more stable (2 kcal.mol ). It can be noted that, in the absence of a second
pyridinium eq, (Supporting information Table S7) the elimination reaction at C-16,17 is
-
1
-1
endothermic (+7.6 kcal.mol ) and the energy of the TS is higher by 9.8 kcal.mol (27.1
-1
kcal.mol ).
The availability of a counterion (pyridinium ion) seems to be determinant in the
reactivity at C-17. We studied the mechanism in details, and demonstrated that the pyridinium
ion is formed simultaneously with the departure of the chloride ion. Considering the geometry
of the molecule, the counter ion formed can’t stabilize the leaving chloride. We also showed
that the stabilization of the leaving chloride by one eq of pyridinium chloride is necessary for
the reaction to proceed. This means that two equivalents of pyridinium chloride are formed
during the elimination reaction at C-16,17.
1
6

3
.2.2.
Study of the formation of the Wagner-Meerwein product 10
The Wagner-Meerwein rearrangement, [1,2]-migration of the 13-methyl, seems to be a
competitive reaction. To study this rearrangement, the model system involved the alcohol 11,
pyridinium ion (2 eq), pyridine (2 eq) and chloride ion (1 eq). Like for the elimination step,
two different mechanisms were considered : via a carbocation (2 steps) or concerted (1 step).
Since no TS (carbocation) was found, the Wagner-Meerwein mechanism (Table 5) proved to
be concerted, favoring the E2 mechanism,
Table 6. Formation of the Wagner-Meerwein product 10. Relative Gibbs Free energy profile of the considered
species (IRC reactant, TS, IRC product) using the alcohol 9, pyridinium ion (2 eq), pyridine (1 eq) and chloride
ion (1 eq)
-
1
Relative Gibbs Free energy profile (G) in kcal.mol
Position
IRC Reactant 9
TS 9a
IRC Product 10
[
1.2]-
Migration
of Me from
C-13 to C-
1
7
0
.00
33.84
-30.03
-1
The results (Table 6) showed that the reaction is exothermic (-30.03 kcal.mol ), and
-1
requires a relatively high activation energy (+33.84 kcal.mol ). By comparing the energy
profiles of the two competing reactions, we can conclude that the elimination product 1 is the
-1
kinetic product (TS at 18.16 vs 33.84 kcal.mol ) and the rearranged product 10 is the
-
1
thermodynamic product (IRC product at -30.03 vs -6.15 kcal.mol ). These results are in
agreement with the experimental results (Table 1, entries 1, 3 and 6).
9
,11,13,14
The proposed mechanism (Scheme 4) for the formation of the 
Wagner- Meerwein rearrangement of compound 9.
-olefin 10 is a
1
7

Scheme 4. Proposed mechanism for the formation of compound 10
3
.3. Anti-inflammatory activity
9
,11,13,14
Since the 
-tetraene 10 has never been described, we investigated its anti-
inflammatory effect in an in vivo model of inflammation. For this purpose, intraperitoneal
administration of vehicle, prednisolone 7 or compound 10 was evaluated on paw edema
induced by CFA-treated mice.
In all groups, a significant increase in the paw size is observed seven days after
intraplantar CFA injection. On day 7, while intraperitoneal injection of vehicle did not modify
paw volume, the administration of prednisolone 7 (15 mg/kg) or compound 10 (15, 30 or 60
mg/kg) induced a significant reduction of edema (Figure 2). This effect is maintained up to 24
hours.
1
8

V e h ic le (n = 6 )
P re d n is o lo n e (1 5 m g /k g ) (n = 6 )
C o m p o u n d 1 0 (1 5 m g /k g ) (n = 7 )
C o m p o u n d 1 0 (3 0 m g /k g ) (n = 8 )
C o m p o u n d 10 (6 0 m g /k g ) (n = 7 )
1
00
7
5
2
5
0
5
0
*
**
*
*
**
*
*
*
*
*
**
**
**
*
*
*
**
**
**
**
***
**
***
**
*
*
*
*
B e fo r e
C FA
0
30' 1h 2h 3h 4h 6h 8h 24h
7
d a y s p o s t C F A in je c tio n
Figure 2. Anti-inflammatory effect on CFA-induced edema. Percentage of edema augmentation after
intraperitoneal drug administration of 10. Tests were performed on C57BL6/J mice 7 days after CFA injection.
Data are shown as mean ± SEM.) *P < 0.05, **P < 0.01, ***P < 0.001, compared to vehicle groups.
The results demonstrate that compound 10 possesses anti-inflammatory properties, but
at higher concentration than prednisolone 7, used as a reference.
4
. Conclusion
We described herein a quick method to prepare the key intermediate tetraene 1 in 50% yield
starting from commercially available prednisolone acetate 8. We confirmed the industrial
feasibility on a larger scale (30g). We identified a novel compound, steroid 10, obtained from
compound 9 via a Wagner-Meerwein rearrangement. This compound demonstrated anti-
inflammatory effect in a rodent model of inflammation. To complete our study, we proposed
mechanisms for the formation of tetraenes 1 and 10, supported by molecular modeling.
1
9

Acknowledgments
The authors would like to acknowledge Sanofi for the financial support; Alain Becker and
Isabelle Finet-Bouillou for support and scientific advice, Regis Egrot, Mounir Traïka for
NMR analyses; Martin Leremboure for mass analyses and Mesocentre Clermont Auvergne
for computing resources.
References
[
1] D. G. Hulcoop, P. D. P. Shapland, Expedient synthesis of 17α,21-dihydroxy-9β,11β-
epoxy-16α-methylpregna-1,4-diene-3,20-dione 21-acetate from prednisolone utilising a novel
Mattox rearrangement, Steroids 78 (2013) 1281–1287.
[
2] R. E.Schaub, G. R. Allen, M. J. Weiss, Studies in the Synthesis of Triamcinolone. The
Ethoxalylation of 4,9(11),16-Fregnatriene-3,20-dione and 11 -Hydroxy-4,16-pregnadiene-
,20-dione, J. Am. Chem. Soc. 81 (18) (1959) 4962–4968.
3
1
6
[
3] H. Hofmeister, H. Laurent, G.-A. Hoyer, R. Wiechert, Synthese von 20-0xo- -
pregnenen durch Eliminierung der 17-Nitroox y gruppe, Chem. Ber 107 (1974) 1235–1244.
[
1
4] H. Hofmeister, H. Laurent, K. Prezewowsky, R. Wiechert, Verfahren Zur Herstellung Von
6,17 ungesaettigten Steroiden. DE2207420 (A1), August 16, 1973.
[
5] H. Mendelsohn, L. L. Smith, V. E. Origoni, Pregnatetraenes and method of preparing the
same. US2864834 (A), December 16, 1958.
[
1
6] H. Hofmeister, H. Laurent, K. Prezewowsky, R. Wiechert, Verfahren Zur Herstellung Von
6,17ungesaettigten Steroiden. DE2207420 (A1), August 16, 1973.
1
6
[
7] W. S. Allen, S. Bernstein, Steroidal Cyclic Ketals. X1I.l The Preparation of  -Steroids,
J. Am. Chem. Soc 77 (1955) 1028–1032.
[
8] S. A. Szpilfogel, V. Gerris, Dehydration of 17-hydroxy-20-ketopregnanes, Recl. Trav.
Chim. Pays-Bas 74 (1955) 1462–1466.
[
1
9] K. Annen, H. Hofmeister, H. Laurent, R. Wiechert, , Verfahren zur Uberfuhrung von
9
(11),16
1,17,21-Trihydroxy-20-oxo-Steroiden in die 21-Chlor-20-oxo-
- Derivative,
Liebigs Ann. Chem. 5 (1982) 966–972.
[
10] K. Annen, H.; Laurent, H. Hofmeister, DELTA 9(11) - And DELTA 16-21-chloro-20-
keto steroids of the pregnane and D-homopregnane series, their preparation and use as
intermediates for the synthesis of highly effective corticoids, US4404141 (A), September 13,
1
983.
[
11] http://www.gaussian.com/g_tech/g_ur/m_citation.htm
2
0

[
12] S. Dapprich I. Komáromi, K. S. Byun, K. Morokuma, M. J. Frisch, A new ONIOM
implementation in Gaussian98. Part I. The calculation of energies, gradients, vibrational
frequencies and electric field derivatives, J. Mol. Struct. (Theochem) 462 (1999) 1-21.
[
13] T. Vreven, K. S. Byun, I. Komáromi, S. Dapprich, Jr J. A. Montgomery, K. Morokuma,
M. J. Frisch, Combining Quantum Mechanics Methods with Molecular Mechanics Methods
in ONIOM, J. Chem. Theory and Comput. 2 (2006) 815-826.
[
14] A. K. Rappe, C. J. Casewit, K. S. Colwell, W. A. Goddard, W. M. Skiff, UFF, A Full
Periodic Table Force Field for Molecular Mechanics and Molecular Dynamics Simulations, J.
Am. Chem. Soc. 114 (1992) 10024–10035.
[
15] A. K. Rappe, W. Goddard, A., Charge Equilibration for Molecular Dynamics
Simulations, J. Phys. Chem. 95 (1991) 3358–3363.
[
16] P. Hohenberg, W. Kohn, Inhomogeneous Electron Gas, Phys. Rev., 1964, 136, B864-
B71.
[
17] W. Kohn, L.J. Sham, Self-Consistent Equations Including Exchange and Correlation
Effects, Phys. Rev. 140 (1965) A1133-A38.
[
18] A.D. Becke, Density-functionnal thermochemistry. III. The role of exact exchange, J.
Chem. Phys. 98 (1993) 5648-5652.
[
19] C. Lee, W. Yang, R.G. Parr, Development of the Colle-Salvettu correlation-energy
formula into a functional of the electron density Phys. Rev. 37 (1988) B785-B789.
[
20] S.H. Vosko, L. Wilk, M. Nusair, Accurate spin-dependent electron liquid correlation
energies for local spin density calculations: a critical analysis, Can. J. Phys. 58 (1980) 1200-
211.
1
[
21] P. J. Stephens, F. J. Devlin, M. J. Frisch, C. F. Chabalowski, Ab Initio Calculation of
Vibrational Absorption and Circular Dichroism Spectra Using Density Functional Force
Fields, J. Phys. Chem. 98 (1994) 11623-11627.
[
(
22] T. Vreven, M. J. Frisch, K. N. Kudin, H. B. Schlegel, K. Morokuma, Mol. Phys., 104
2006) 701-704.
[
23] M. Zimmermann, Ethical guidelines for investigations of experimental pain in conscious
animals Pain 16 (1983) 109-110.
[
24] C. Kilkenny, W. Browne, I.C. Cuthill, M. D.G. Emerson, D.G. Altman, Animal research:
reporting in vivo experiments: the ARRIVE guidelines, J. Gene Med. 12 (2010) 561-563.
[
25] J.C. McGrath, G.B. Drummond, E.M. McLachlan, C. Kilkenny, C.L. Wainwright
Guidelines for reporting experiments involving animals: the ARRIVE guidelines, Br. J.
Pharmacol. 160 (2010) 573-1576.
2
1

[
26] V. R. Mattox, Steroids Derived from Bile Acids. XV. The Formation of a Glyoxal Side
1
6
Chain at C-17 from Steroids with Dihydroxyacetone and  -Ketol Side Chains, J. Am.
Chem. Soc. 74 (1952) 4340–4347.
[
27] M. Li, B. Chen, M. Lin, T.-M. Chan, X. Fu, A. Rustum, A variation of Mattox
rearrangement mechanism under alkaline condition, Tetrahedron Lett. 48 (2007) 3901–3905.
[
28] B. Chen, M. Li, M. Lin, G. Tumambac, A. Rustum, A comparative study of enol
aldehyde formation from betamethasone, dexamethasone, beclomethasone and related
compounds under acidic and alkaline conditions, Steroids 74 (2009) 30–41.
[
1
29] H. Bi, R. Massé, G. Just, Studies on anabolic steroids. 9. Tertiary sulfates of anabolic
7-methyl steroids: synthesis and rearrangement, Steroids 57 (1992) 306–312.
[
30] A. S.Veleiro, M. V. Nevado, M. C. Monteserin, and G. Burton, Syntheses of 21-
hydroxy- 11,19- oxidopregn-4-ene-3,20-dione and 21-hydroxy-6,19-oxidopregn-4-ene- 3,20-
dione, Steroids 60 (1995) 268-271.
[
1
31] I. Vincze, M. Lökös, T. Bakos, A. Dancsi, M. Mák, Investigations on the dehydration of
7-ethynyl-17-hydroxysteroids, Steroids 58 (1993) 220–224.
1
3
[
32] V. I. D. Brabandere, L. M. Thienpont, D. Stöckl, A. P. D. Leenheer, C-NMR and mass
spectral data of steroids with a 17,17-dialkyl-l8-nor-l3(14)-ene substructure, J. Lipid Res. 38
(
[
1997) 780–789.
33] R. W. Draper, M. S. Puar, E. J. Vater, A. T. McPhail, An unusual dienone-phenol
rearrangement product formed during the synthesis of mometasone furoate (Sch 32088),
Steroids 63 (1998) 135–140.
[
34] A. Segaloff, R. Bruce Gabbard, Anti-androgenic activity of 17,17-dimethyl-18-
norandrost-13-enes, Steroids 4 (1964) 433–443.
[
35] M. S. Puar, P. A. Thompson, M. Ruggeri, D. Beiner, A. T. McPhail, An unusual
rearrangement product formed during production of mometasone furoate (Sch 32088),
Steroids 60 (1995) 612–614.
[
36] H. L. Herzog, O. Gnoj, L. Mandell, G. Nathansohn, A. Vigevani, Wagner-Meerwein
Rearrangements. IV. Some reassignments of Structure, J. Org. Chem. 32 (1967) 2906–2910.
[
[
37] J. Fried, J. E. Herz, Synthesis of steroids. US2895954 (A), July 21, 1959.
38] F. H. Lincoln, W. P. Schneider, 9,11-epoxy-6-fluoro-delta4,16-pregnadienes and
products produced therefrom. US3210341 (A), October 5, 1965.
2
2