Industrial synthesis of 4-chloro, 11β-arylestradiol: How to circumvent a poor diastereoselectivity

Article abstract of DOI:10.1021/op0341622

An industrial synthesis of 11β-arylestrone derivatives is described, based on the conjugate opening of a mixture of allylic 5(10)-α and -β epoxides by an aryl cuprate generated catalytically, followed by hydrolysis and subsequent aromatisation of the isomeric mixture of arylation products. An original method for selective 4-chlorination of estrone derivatives is also described.

Full text of DOI:10.1021/op0341622

Organic Process Research & Development 2004, 8, 219−228
Industrial Synthesis of 4-Chloro,11â-arylestradiol: How to Circumvent a Poor
Diastereoselectivity
Denis Prat,* Franc¸oise Benedetti, Gilles Franc Girard, Lahlou Nait Bouda, John Larkin, Christian Wehrey, and
Jacques Lenay
Chemical/Process DeVelopment, AVentis Pharma, 102 Route de Noisy, 93235 RomainVille Cedex, France
Abstract:
First Preparative Synthesis
An industrial synthesis of 11â-arylestrone derivatives is de-
scribed, based on the conjugate opening of a mixture of allylic
5(10)-r and -â epoxides by an aryl cuprate generated catalyti-
cally, followed by hydrolysis and subsequent aromatisation of
the isomeric mixture of arylation products. An original method
for selective 4-chlorination of estrone derivatives is also de-
scribed.
For the supply of the preclinical and phase 1 studies, the
first batches of drug substance were prepared in the pilot
plant from the known chiral norsteroid intermediate 3
(Scheme 1), in a similar synthesis used to prepare the 17-
methyl derivative 1.4 Epoxidation by hydrogen peroxide
catalyzed by hexachloroacetone5 gave a 65/35 mixture of
R-epoxide 4a and â-epoxide 4b. Crystallisation from aceto-
nitrile afforded isomerically pure R-epoxide 4a (33% yield).
The introduction of the aryl group at the 11â-position was
carried out using a Cu(I)-mediated allylic opening of the
R-epoxide by the corresponding Grignard reagent5b,6 (Schemes
2 and 3). The keto group at C-17 was protected as a silyl
enol ether7 prior to this arylation. This protection allowed
us to reduce by half the amount of copper(I) chloride catalyst
and the expensive aryl side chain and also gave an increase
in the overall yield.4
Introduction
The treatment of osteoporosis is a key challenge of this
century. The decrease in bone mineral density after meno-
pause induces a risk of fractures (mainly hip and femur) in
the increasing senior female population.
Ketone 4a was silylated using lithium diisopropylamide8
(LDA; 1.2 equiv) and the inexpensive chlorotrimethylsilane
(TMSCl; 1.4 equiv), at 0 °C (Scheme 1). The silyl enol ether
5a was not isolated but used as a solution in toluene.
The side-chain synthon 6 (4-(2-(diethylamino)ethoxy)-
bromobenzene)9 and the parent 4-(2-piperidino-ethoxy)-
bromobenzene2a,4 were previously prepared in homogeneous
medium (DMF, K2CO3, or NaH) under harsh conditions,
which resulted in modest yields. Therefore, we developed
an improved procedure (Scheme 2), under phase transfer
catalysis conditions,4 under which 4-bromophenol and N-(2-
chloroethyl)diethylamine hydrochloride (1 equiv), in the
presence of 30% sodium hydroxide (2.2 equiv) and triethyl
benzylammonium chloride (TEBAC, 5%), afforded the
The main preventive treatment consisted of the regular
intake of estradiol, which slows down the bone loss, but not
without side effects. The ideal drug would have the benefit
of estradiol on bone without the side effects on the other
tissues. Such a drug is known as a “selective estrogen
receptor modulator” (SERM).1 The first SERM on the
market, Eli Lilly’s raloxifene (Evista) is not a steroid.2
In parallel, Hoechst Marion Roussel (now Aventis Phar-
ma) has been developing a series of 11â-aryl,17â-estradiols,
for the treatment of osteoporosis. Their steroid structure could
be an advantage in terms of tolerance and specificity of
action3 (Figure 1). In 2002, we published in this journal the
synthesis of one candidate, 11â-aryl,17R-methyl-estradiol 1.4
We report here the different synthetic approaches to another
drug candidate, 4-chloro,11â-arylestradiol 2.3b
(5) (a) Costerousse, G.; Teutsch, G. (Roussel Uclaf.). EP 5100, 1979. (b) Scholz,
S.; Hofmeister, H.; Neef, G.; Ottow, E.; Scheidges, C.; Wiechert, R. Liebiegs
Ann. Chem. 1989, 151. (c) Neef, G.; Ast, G.; Michl, G.; Schwede, W.;
Vierhufe, H. Tetrahedron Lett. 1994, 35, 8587. (d) Rao, P. N.; Taraporewala,
I. B. Steroids 1992, 57, 154.
(6) (a) Teutsch, G.; Belanger, A. Tetrahedron Lett. 1979, 2051. (b) Belanger,
A.; Philibert, D.; Teutsch, G. Steroids 1981, 37, 361. (c) Napolitano, E.;
Fiachi, R.; Hanson, R. N. Gazz. Chim. Ital. 1990, 120, 323. (d) Faraj, H.;
Claire, M.; Rondot, A.; Aumelas, A.; Auzou, G. J. Chem. Soc., Perkin Trans.
1 1990, 3045. (e) Nique, F.; Van de Velde, P.; Bre´maud, J.; Hardy, M.;
Philibert, D.; Teutsch, G. J. Steroid Biochem. Mol. Biol. 1994, 50, 21. (f)
Lobaccaro, C.; Pons, J. F.; Duchesne, M. J.; Auzou, G.; Pons, M.; Nique,
F.; Teutsch, G.; Borgna, J. L. J. Med. Chem. 1997, 40, 2217.
(7) (a) Greene, T. W. and Wuts, P. G. ProtectiVe Groups in Organic Synthesis,
3rd ed.; John Wiley & Sons: New York, 1999. (b) For a monograph, see:
Van Look, G.; Simchen, G.; Heberle, J. Silylating Agents; Fluka Chemica,
Fluka Chemie AG: Buchs, Switzerland, 1995.
* To whom correspondence should be addressed. E-mail: denis.prat@
aventis.com. Telephone: (33) 15893 8262. Fax: (33) 15893 8274. Current
address: Aventis Pharma, 13 Quai Jules Guesde, BP 14, 94403 Vitry/ Seine,
France.
(1) (a) Grese, T. A.; Dodge, J. A. Curr. Pharm. Design 1998, 4, 71. (b) Jordan
V. C. J. Med. Chem. 2003, 46, 883. (c) Jordan V. C. J. Med. Chem. 2003,
46, 1081.
(2) (a) Jones, C. D.; Jevnikar, M. G.; Pike, A. J.; Peters, M. K.; Black, L. J.;
Thompson, A. R.; Falcone, J. F.; Clemens, J. A. J. Med. Chem. 1984, 27,
1057. (b) Vincenzi, J. T.; Zhang, T. Y.; Robey, R. L.; Alt, C. A. Org. Process
Res. DeV. 1999, 3, 56. (c) Schmid, C. R.; Sluka, J. P.; Duke, K. M.
Tetrahedron Lett. 1999, 40, 675. (d) Clemett, D.; Spencer, C. M. Drugs
2000, 60, 379.
(3) Bouali, Y.; Nique, F.; Teutsch, G.; Van De Velde, P. (Hoechst Marion
Roussel.). FR 2,757,519, 1998.
(4) (a) Larkin, J. P.; Wehrey, C.; Boffelli, P.; Lagraulet, H.; Lemaitre, G.;
Nedelec, A.; Prat, D. Org. Process Res. DeV. 2002, 6, 20. (b) Benedetti, F.;
Mazurie, A.; Nique, F.; Prat, D.; Wehrey, C. (Aventis Pharma.). FR 2,-
826,004, 2002.
(8) LDA was prepared in situ from n-butyllithium and diisopropylamine in THF
at 0 °C.
(9) Pruitt, V. L.; Berlin, K. D.; Hirsch, K. S. Org. Prep. Proced. Int. 1990, 22,
235.
10.1021/op0341622 CCC: $27.50 © 2004 American Chemical Society
Published on Web 01/31/2004
Vol. 8, No. 2, 2004 / Organic Process Research & Development
•
219

Figure 1. SERM drug substances.
Scheme 1. Initial steps
Scheme 2. Preparation of the side-chain synthon
) 8), the dienone 9b was crystallised from diisopropyl ether,
providing a 73-79% yield range in both the laboratory and
the pilot plant. The deconjugated dienone 9c was also formed
(5-10%), but this oily product was easily eliminated in the
mother liquors.10
The dienone 9b was chlorinated using sulfuryl chloride
(1.5 equiv) and pyridine (29 equiv), in dichloromethane at
-40 °C11 (Scheme 4). 4-Chlorodienone hydrochloride 10 was
crystallised from ethyl acetate (yield: 82%; purity: 73%).
This reaction was incomplete and gave several secondary
products, which will not be detailed here. Compound 10 was
aromatised using a mixture of acetyl bromide (3.8 equiv)
and acetic anhydride (3 equiv) in dichloromethane at room
temperature,6c,e,f,12 giving an estrone acetate which was
immediately saponified. The aromatisation reaction afforded
also ca. 10% of elimination product 12. Fortunately, crys-
tallisation of the 4-chloroarylestrone hydrochloride 11 from
dichloromethane removed most of the numerous impurities
generated in these chlorination and aromatisation steps.
However, the yields (40% from 9b in the lab, 44% in the
pilot plant) and the purity (ca. 90%) were low.
bromoaryl derivative 6 in 82% yield. Reaction of 6 with
magnesium turnings in THF at 58 °C gave the Grignard
reagent 7 in an almost quantitative yield. These steps were
very reproducible on scale-up.
The 11â-aryl side chain was introduced by a Cu(I)-
mediated addition of Grignard reagent 7 (1.5 equiv), with
copper(I) chloride (0.1 equiv) (Scheme 3). The arylation was
initially carried out at -5 °C, and then the temperature was
raised to +20 °C. Aqueous workup and concentration
afforded the 11â-aryl alcohol 8b which was not isolated.
This allylic opening of the epoxide (SN2′) is known to
proceed in a regio- and stereospecific manner,6 but the NMR
spectrum of the crude arylation mixture was too complex
(partial hydrolysis of silyl ether) to confirm the absence of
other isomers. Acidic hydrolysis of the alcohol 8b cleaved
the ketal and silyl groups and eliminated the 5-hydroxyl
group, thereby yielding the dienone 9b. Such an acidic
treatment also salified the (2-diethylamino)ethoxybenzene
which was formed by hydrolysis of the excess Grignard
reagent, and this hydrochloride salt was eliminated in the
aqueous phase, whereas the hydrochloride salt of the dienone
was extracted into dichloromethane. After neutralisation (pH
(10) This type of dienone has already been described: see ref 6e.
(11) (a) Nickisch, K.; Bittler, D.; Casals-Stenzel, J.; Laurent, H.; Nickolson, R.;
Nishino, Y.; Petzoldt, K.; Wiechert, R. J. Med. Chem. 1985, 28, 546. (b)
Bourban, C. Y. M.; Hanson, J. R.; Hitchcock, P. B. J. Chem. Res. (S) 1990,
274. (c) Christiansen, R. G.; Bell, M. R.; D’Ambra, T. E.; Mallamo, J. P.;
Herrmann, J. L.; Ackerman, J. H.; Opalka, C. J.; Kullnig, R. K.; Winneker,
R. C.; Snyder, B. W.; Batzold, F. H.; Schane, H. P. J. Med. Chem. 1990,
33, 2094. (d) Hasrat, A.; Van Lier, J. E. J. Chem. Soc., Perkin Trans. 1
1991, 2485.
(12) (a) Joly, R.; Joly, J. (Roussel Uclaf.). FR 81840, 1962. (b) Danishefsky, S.;
Cain, P.; Nagel, A. J. Am. Chem. Soc. 1975, 97, 380; c) Danishefsky, S.;
Cain, P. J. Am. Chem. Soc. 1975, 97, 5282.
220
•
Vol. 8, No. 2, 2004 / Organic Process Research & Development

Scheme 3. Arylation of the r-epoxide
Scheme 4. First pilot-scale synthesis: last steps
The 4-chloroarylestrone hydrochloride 11 was neutralised
with sodium hydroxide and reduced with sodium borohydride
in methanol. This reduction is known to give very selectively
the 17â-alcohol,5d,6e,13 but to check the isomer purity, we
prepared the 17R-alcohol, and this diastereomer could not
be detected in the crude product (<0.3%) by HPLC. The
drug substance 2 was crystallised from methanol, in two
crops (yield: 85%). In the pilot plant, however, only the
first crop was pure enough for clinical use. The yield was
thus limited to 66-71%.
A synthesis in the pilot plant without preparative chro-
matography was achieved, largely as a result of the crystal-
linity of most of the intermediates. However, because of the
poorly stereoselective initial epoxidation and the numerous
by-products generated in the chlorination-aromatisation
sequence, the overall yield was unsatisfactory (7-8%).
Therefore, efforts were directed to designing, optimizing, and
scaling up new syntheses of this drug candidate.
Ethylene deltenone 3, an industrial intermediate for nor-
steroids, particularly trimegestone,14 was a suitable starting
material for the synthesis. Epoxide 4a was crystalline, in
contrast to analogous epoxides bearing other functions at
C-17. The introduction of the preformed aryl side chain made
the route convergent, and the reduction was best performed
at the last step.15 Thus, in the pre-industrial phase, we carried
out more detailed investigations of the epoxidation step and
the chlorination-aromatisation sequence.
1. Epoxidation. The introduction of the aromatic moiety
at the 11â-position required an almost diastereomerically
pure 5,10R-epoxide 4a.6 Optimisation of the initial epoxi-
dation permitted us to increase the yield from 33 to 49%,4
first by using hexafluoroacetone6c,d,16 as the catalyst, which
is slightly more selective than hexachloroacetone5 (R/â )
2.1 and 1.8, respectively), and second by crystallizing the
R-epoxide 4a from ethyl acetate instead of acetonitrile.
2. Aromatisation. In the first preparative synthesis,
chlorination was performed first, and then the chlorodienone
was aromatised. As mentioned above, both steps gave by-
products, such as the elimination product 12 formed during
the aromatisation step. On the other hand, aromatisation of
non-chlorinated dienones such as 9b is a classic reaction in
Second Pilot-Scale Synthesis
Despite its drawbacks, the preparative synthesis described
above was not abandoned, as the strategy was sound.
(13) (a) Shimada, K.; Xie, F.; Nambara, T. Chem. Pharm. Bull. 1986, 34, 179.
(b) Bulman Page, P. C.; Hussain, F.; Maggs, J. L.; Morgan, P.; Park, B. K.
Tetrahedron 1990, 46, 2059. (c) Tada, M.; Chiba, K.; Izumiya, K.; Tamura,
M. Bull. Chem. Soc. Jpn. 1993, 66, 3532. (d) Aliau, S.; Delettre, G.; Mattras,
H.; El Garrouj, D.; Nique, F.; Teutsch, G.; Borgna, J. L. J. Med. Chem.
2000, 43, 613. (e) Schneider, M. F.; Harre, M.; Pieper, C. Tetrahedron Lett.
2002, 43, 8751.
(14) Crocq, V.; Masson, C.; Winter J.; Richard, C.; Lemaitre, G.; Lenay, J.;
Vivat, M.; Buendia, J.; Prat, D. Org. Process Res. DeV. 1997, 1, 2.
(15) The intermediates bearing a 17-hydroxyl group were less crystalline.
(16) Teutsch, G.; Belanger, A.; Philibert, D.; Tournemire, C. Steroids 1982, 39,
607.
Vol. 8, No. 2, 2004 / Organic Process Research & Development
•
221

Scheme 5. Second pilot-scale synthesis (X ) Et2NCH2CH2O)
norsteroid chemistry6c,e,f,12 and proceeds cleanly, with little
elimination. Thus, we decided to aromatise first and then to
try to chlorinate the estrone ring as selectively as possible
(Scheme 5).
Aromatisation of dienone 9b was carried out using a
mixture of acetyl bromide (2.5 equiv) and acetic anhydride
(1 equiv) in dichloromethane at 20-25 °C. The estrone
acetate 13b formed was saponified (KOH, MeOH), affording
11â-arylestrone. The hydrochloride salt 14 was crystallised
from methylethyl ketone in a good yield (ca. 85% from 9b),
in the lab and in the pilot plant.
3. Chlorination. Chlorination of estrogen derivatives has
been studied, but not extensively.17 Many reagents have been
tested, such as sulfuryl chloride, chlorine, N-chlorosuccin-
imide (NCS), trichloroisocyanuric acid, tert-butylhypochlo-
rite, and hexachloro-2,4-cyclohexanedienone. Most of these
gave complex mixtures of 4-chloro and 2-chloro substitution
products and also products from para-addition such as 10â-
chloroestra 1,4-diene-3-one. The formation of the latter
product could be prevented, using 11â-substituted estrogen
derivatives. From these substrates, hexachloro-2,4-cyclohex-
anedienone afforded a relatively clean mixture of 4-chloro-
and 2-chloroestrogens and the 4-Cl/ 2-Cl selectivity de-
creased with the size of the 11â-substituent,17c but no
hypothesis has hitherto been proposed to account for this
regioselectivity.
Many reagents and conditions have been described for
the chlorination of phenols,18 most often with the aim of
influencing the regioselectivity. Although chlorodialkyl-
amines and chlorotrialkylammonium chlorides have been
used in acidic solutions,19 the simple use of N-chlorosuc-
cinimide, sometimes with an acid catalyst, has only very
recently been reinvestigated.20
We rapidly abandoned sulfuryl chloride as chlorinating
agent, which favored bis-chlorination of arylestrone hydro-
chloride 14, even at low temperature (-40 °C). On the other
hand, the use of NCS in the presence of strong acids gave
promising results, notably a good 4-Cl/2-Cl selectivity
(Scheme 5). Nevertheless, optimisation of this step required
45 experiments. For example, the reaction mixture was
heterogeneous: both the starting arylestrone hydrochloride
14 and the 4-chloroarylestrone hydrochloride 11 crystallised
from the mixture. As a result, some starting material
cocrystallised with the product, and attempts to consume it
totally resulted in bis-chlorination. Therefore, the reaction
was better carried out in a mixture of methanol and
dichloromethane, in which all the materials were soluble.
Using 1.00 equiv of NCS at 10 °C and in the presence of
hydrochloric acid (0.5 equiv),21 the final reaction mixture
contained (as HCl salts) ca. 1-2% of starting arylestrone
14, 83-88% of the desired 4-chloroerylestrone 11, 10-11%
of the 2-chloro isomer 15 and 1-2% of the 2,4-dichloro
derivative 16 (Scheme 5). After reductive workup (sodium
(18) (a) Guy, A.; Lemaire, M.; Guette´, J. P. Tetrahedron 1982, 38, 2339. (b)
Watson, W. D. J. Org. Chem., 1985, 50, 2145. (c) Schmitz, E.; Pagenkopf,
I. J. Prakt. Chem. 1985, 327, 998. (d) Onyiriuka, S. O.; Suckling, C. J. J.
Org. Chem. 1986, 51, 1900. (e) Smith, K.; Butters, M.; Nay B. Tetrahedron
Lett. 1988, 29, 1319. (f) Kamigata, N.; Satoh, T.; Yoshida, M.; Matsuyama,
H.; Kameyama, M. Bull. Chem. Soc. Jpn 1988, 61, 2226. (g) Chung, K.
H.; Kim, H. J.; Kim, H. R.; Ryu, E. K. Synth. Commun. 1990, 20, 2991.
(h) Gnaim, J. M.; Sheldon, R. A. Tetrahedron Lett. 1995, 36, 3893. (i)
Deshmukh, A. P.; Padiya, K. J.; Jadhav, V. K.; Salunkhe, M. M. J. Chem.
Res. (S) 1998, 828. (j) Mukhopadhyay, S.; Chandnani, K. H.; Chandalia, S.
B. Org. Process Res. DeV. 1999, 3, 196. (k) Mamaghani, M.; Zoligol, M.
A.; Shojaei, M. Synth. Commun. 2002, 32, 735.
(19) (a) Lindsay Smith, J. R.; McKeer, L. C. Tetrahedron Lett. 1983, 24, 3117.
(b) Lindsay Smith, J. R.; McKeer, L. C.; Taylor, J. M. J. Chem. Soc., Perkin
Trans. 1 1988, 385.
(20) (a) Rama Rao, A. V.; Chakraborty, T. K.; Reddy, K. L.; Rao, A. S.
Tetrahedron Lett. 1994, 35, 5043. (b) Madsen, P. J. Med. Chem. 2002, 45,
5755. (c) Speicher, A.; Koltz, J.; Sambanje, R. P. Synthesis 2002, 2503. (d)
Das, S. K.; Reddy, K. A.; Abbineni, C.; Iqbal, J.; Suresh, J.; Premkumar,
M.; Chakrabarti, R. Bioorg. Med. Chem. Lett. 2003, 13, 399.
(17) (a) Schwenk, E.; Castle, C. G.; Erzsebet, J. J. Org. Chem. 1963, 28, 136.
(b) Mukawa, F. J. Chem. Soc., Perkin Trans. 1 1988, 457. (c) Hasrat, A.;
Van Lier, J. Steroids 1994, 59, 498. (d) Cummins, C. H. Steroids 1994, 59,
590.
(21) The hydrochloride salt of the trisubstituted amine on the side chain was
not acidic enough to catalyze the reaction.
222
•
Vol. 8, No. 2, 2004 / Organic Process Research & Development

Scheme 6. Arylation of the â-epoxide
thiosulfate), the 4-chloroarylestrone hydrochloride 11 was
crystallised from dichloromethane in 75-77% yield in the
pilot plant (purity: 98%). The higher solubility of the
2-isomer facilitated purification. In the initial experiments,
traces of the 4-bromo analogue 17 were also detected in
compound 11 and this impurity could not be totally removed,
even by crystallisation. The bromine came from the previous
steps: acetyl bromide was used in the aromatisation reaction,
and some of it survived the intermediate hydrolysis, and was
cleaved during the saponification of the arylestrone acetate.
The bromide ions thus generated were present in the
crystallised arylestrone 14, as the hydrobromide salt. In the
presence of NCS, this hydrobromic acid may have formed
BrCl, an effective brominating agent.22,23 Extensive aqueous
washing of the organic phases, after aromatisation and
saponification, avoided the formation of the 4-bromo impu-
rity.
still suffered from a modest yield in the epoxidation step
(49%). In an ideal route, both R and â epoxides 4a and 4b
would be used.
It had been observed that silyl enol ether 5b derived from
the epoxyketone 4b reacted slower than in R series, requiring
more copper chloride (0.2 equiv), higher temperatures (20
°C), and more Grignard reagent (2 equiv) (Scheme 6). An
11R-aryl,5â-alcohol 8a was obtained, the hydrolysis of which
gave mainly the deconjugated dienone 9c, with minor
amounts (ca. 10%, NMR estimation) of the 11R-aryl dienone
9a. The exact mass balance could not be established,
compound 9c being neither stable (especially on silica gel)
nor crystalline: after chromatography, only 6% of 9c and
3% of 9a were isolated. These results were at first glance of
little synthetic interest.
The aromatisation reaction also gave unexpected results.
During its optimisation, we observed that starting 11â-
dienone 9b disappeared within 1 h in the reaction mixture
(AcBr, Ac2O, CH2Cl2), giving an intermediate which rear-
ranged within 4 h into the 11â-estrone acetate 13b (Scheme
7). According to LC-MS it had the same molecular mass
as the estrone acetate, and we proposed the structure 13d
for this intermediate. Later, we succeeded in isolating it, and
to our surprise, it was the isomeric trienol acetate 13c.
Compound 13d probably formed as a very unstable inter-
mediate.24
As a conclusion, inversion of the chlorination-aromati-
sation steps resulted in an increase in the overall yield (9b
to 11) from 44 to 65% in the pilot plant and an improvement
in the purity of the product 11 from 90 to 98%.
4. Crystallisation of the Drug Substance. The first two
pilot-plant batches were crystallised from methanol. The first
one, in 66% yield; the second, in 71% yield, as a result of
the higher purity of the product. For the subsequent batches
(dedicated to phase 2 studies), we selected 2-propanol as a
crystallisation solvent. The drug substance was obtained in
one crop, in 83% (lab) to 89% (pilot) yield. As the selected
form was a monohydrate (ca. 3.5% water), the 2-propanol
solvate was suspended into water at 80 °C and then filtered
at 20 °C, dried, and micronised under an atmosphere of
nitrogen saturated with steam.
The structure of the trienol acetate 13c was established
unambiguously by NMR spectroscopy. The axial Me18 group
displayed a singlet at 1.02 ppm, in contrast to the 11â-aryl
derivatives, in which the shielding effect of the axial aromatic
ring gave a chemical shift of 0.3 to 0.5 ppm. Moreover, no
benzylic proton (H11) was detected, and only one olefinic
proton, at 5.52 ppm (H4).25
Industrial Synthesis
In compound 13c, the 11-aryl group was no longer â,
but it transformed into the â-aryl product 13b in a highly
Despite the improvements described above, which in-
creased the overall yield from 7-8 to ca. 20%, the synthesis
(24) At first glance, it may look obvious that 13c would be more stable than
13d, in which the axial aryl group interacts with the axial Me18. On the
other hand, conjugation with the aryl group in 13c imposes important strain
between the aryl group and H1.
(25) F. Nique, and J. L. Borgna et al. isolated a deconjugated dienone such as
9c after alkaline treatment of an aromatisation mixture (ref 13d). This
dienone was probably the saponification product of a trienol acetate similar
to 13c.
(22) (a) Fieser, M.; Fieser, L. Reagents for Organic Synthesis; Wiley Interscience
and John Wiley & Sons: New York, 1967; Vol. 1, p 72; Vol. 2, p 38; Vol.
6, p 102. (b) Focht, G. D.; Klobucar, W. D. (Albermarle Corp.). U.S. Patent
6,271,417, 2001. (c) Muatten, H. A. Synthesis 2002, 169.
(23) As bromination of arylestrone hydrochloride 14 using N-bromosuccinimide
(NBS, 1.5 equiv) and HCl (0.5 equiv) resulted in a very slow reaction,
NBS is probably not the actual brominating agent in the presence of NCS.
Vol. 8, No. 2, 2004 / Organic Process Research & Development
•
223

Scheme 7. Proposed mechanism of A-ring aromatisation
Scheme 8. Industrial synthesis of 4-chloro,11â-estradiol 2
stereoselective way. Thus, starting from pure R-epoxide 4a,
only ca. 0.5% of the 11R-aryl isomer was detected in the
crude mixture after aromatisation and saponification.
These observations opened the way to the synthetic use
of the hitherto “undesired” â-epoxide 4b: the deconjugated
dienone 9c was the main product of silylation-arylation of
the epoxide 4b, and we expected the trienol acetate 13c to
be the first intermediate in the aromatisation of dienone 9c
and that compound 13c would eventually transform into the
11â-aryl product 13b in the same way.
224
•
Vol. 8, No. 2, 2004 / Organic Process Research & Development

It would therefore be tempting to submit the crude mixture
of epoxides (R/â ) 65/35) to the sequence silylation,
arylation, aromatisation, and saponification, as the final
product should be the arylestrone hydrochloride 14 (Scheme
8). Moreover, this would put to good use the small amount
of the deconjugated dienone 9c which was formed from
epoxide 5a (Scheme 3) and which was previously lost during
the purification of dienone 9b.
the overall yield of the synthesis (ca. 40%). Silylation and
arylation of the by-product of epoxidation, the â-epoxide 4b,
afforded an unstable deconjugated aryl dienone 9c. Its
structure suggested that it could converge to the same
unstable intermediate 13c in the aromatisation reaction, as
that obtained from the R-epoxide 4a. This intermediate
eventually gave the final product 2 with the desired config-
uration. Hence, these seemingly undesired and unstable
products were the cornerstone of a new industrial synthesis.
Again, norsteroids still reveal surprises to the chemist,
despite all the knowledge accumulated on this old family of
products.
Ethylene deltenone 3 was epoxidised using hydrogen
peroxide and hexachloroacetone, and a single crystallisation
from diisopropyl ether afforded the epoxide mixture (R/â ≈
2/1), leaving minor impurities in the mother liquors. Sily-
lation gave a ca. 2/1 mixture of silyl enol ethers 5a and 5b,
which was not separated. The arylation was carried out using
2 equiv of Grignard reagent 7 and 0.15 equiv of copper(I)
chloride, at 20 °C. Workup (aqueous ammonium chloride)
and acidic hydrolysis resulted in a mixture of aryl dienones
9b, 9c, and 9a (Scheme 8; the ratio of enones was estimated
by NMR spectroscopy). The mixture was rapidly aromatised,
affording, after 4-5 h at room temperature, 11â-arylestrone
acetate 13b as the main product. Again, it was not isolated
but saponified to give the 11â-arylestrone, which was
crystallised as the hydrochloride 14. The yield from the
mixture of epoxides was ca. 68%, on the laboratory and pilot-
plant scale. Thus, from ethylene deltenone 3, this “recycling
process” permitted us to increase the overall yield in
arylestrone hydrochloride 14 from 28 to 58%.26 The purity
was only slightly lower (95%) than the purity (98%) of the
previous batches prepared from 95% pure R-epoxide. Ca.
1% of 11R-aryl isomer was detected in the crude reaction
mixture after aromatisation and saponification, and only
0.15% in the crystallised product 14. Chlorination and
reduction gave the drug substance 2 with the same purity as
previously obtained and in the same yield (ca. 69%).
Maintaining the same profile of impurities in the drug
substance was very important, since the batches were used
in clinical trials. For the same reasons, the last steps
(chlorination and reduction) were not modified. As is usual
in the development of a drug candidate, these last steps were
optimised first, and then the core of the synthesis was
optimised.
Experimental Section
General. Except for the isolation of intermediates, all the
steps described here were launched in the pilot plant on a
multikilogram scale. However, for the sake of confidentiality,
only the experimental procedures from the laboratory will
be given here. Nevertheless, they were reproducible on scale-
up.
Unless stated, NMR spectra were recorded on a Bruker
300 MHz spectrometer, in CDCl3; IR spectra, on a Nicolet
FTIR SSXB spectrometer, in chloroform; MS spectra, on
Micromass Autospec or Micromass Platform spectrometers.
Water content was determined by Karl Fischer titration in
methanol on a Mettler titrator.
For all the TLC monitoring, Merck Si 60 F254 plates were
used.
The aromatisation reactions, and the subsequent steps,
were monitored by HPLC with the following system:
Hypersil DBS 3 µm CN; 150 mm × 4.6 mm; eluent: water
+ 0.1% TFA: 65, acetonitrile: 30, methanol: 5; flow rate:
1 mL/min; detection: UV 210 nm.
3,3-Ethylenedioxy-5(10)-epoxy-estr-9(11)-ene-17-one (2/1
Mixture of Isomers 4a4 and 4b4). Ethylene deltenone 3 (50
g; MW: 314.4; 0.159 mol), hexachloroacetone (Janssen,
98%; MW: 264.7; 4.35 g; 0.1 equiv), pyridine (0.25 mL),
50% hydrogen peroxide (ca. 18 M; 15 mL; 1.7 equiv) and
dichloromethane (250 mL) were stirred vigorously for 18 h
at 20-25 °C (TLC monitoring: n-heptane 6, ethyl acetate
4). After reductive workup (aqueous sodium thiosulfate),
washing (water), and extraction (dichloromethane), the
organic phase was concentrated to a total volume of 150
mL. Then, dichloromethane was exchanged for diisopropyl
ether by continuous distillation at constant volume, until the
temperature of the liquid reached 68 °C. The mixture was
cooled to 20 °C. The mixture of epoxides crystallised
spontaneously. The suspension was cooled to 0 °C and stirred
for 1 h, then 4a,b was filtered and dried under vacuum for
18 h at 40 °C (44.6 g white solid; yield: 84.9%; purity
Conclusions
The first preparative synthesis of 4-chloro, 11â-aryl
estradiol 2 suffered mostly from a poorly stereoselective
epoxidation and an inappropriate chlorination-aromatisation
sequence, thus resulting in a low overall yield (7-8%).
However, it allowed the delivery of the drug batches in time
for the pre-clinical studies and the phase 1 clinical trials.
Optimisation of the epoxidation and implementation of
an original chlorination of an estrone derivative resulted in
a ca. 20% overall yield, which facilitated the delivery of the
drug substance for the phase 2 clinical studies. This synthesis
would have been acceptable for further scale-up, especially
because of the very low dosage of the drug substance.
Nevertheless, we found a way to transform both epoxides
at the same time, thus providing another 2-fold increase in
1
(LC): 97%): C20H26O4; MW: 330.4; H NMR (CDCl3,
ppm): δ 0.87 (s, 3H, 4b), 0.88 (s, 3H, 4a), 3.94 (m, 4H),
5.86 (m, 1H, 4b), 6.05 (m, 1H, 4a).
3,3-Ethylenedioxy-5(10)-epoxy-17-(trimethylsilyloxy)-
estra-9(11),16-diene (2/1 Mixture of Isomers 5a4 and
5b). This procedure was used to silylate any mixture of
epoxides. n-Butyllithium (Chemetall, 17% solution in cy-
clohexane; 68 g; 1.2 equiv) was added over 30 min to a
stirred solution of diisopropylamine (MW: 101.2; d: 0.714;
(26) Using hexachloroacetone as the epoxidation catalyst, the yield in crystallised
4a was 43%.
Vol. 8, No. 2, 2004 / Organic Process Research & Development
•
225

30 mL; 1.4 equiv) in anhydrous THF (100 mL), at -10 °C.
This solution was added over 30 min at 0 °C to a solution
of epoxides 4a,b (50 g; 151 mmol) in THF (150 mL), and
the mixture was stirred for 15 min at 0 °C. Trimethylchlo-
rosilane (MW: 108.6; d: 0.856; 27 mL; 1.4 equiv) was
added over 15 min at 0 °C, and the mixture was stirred for
1 h at 0 °C (TLC monitoring: toluene 1, ethyl acetate 1).
Methanol (MW: 32.0; d: 0.792; 5 mL; 0.8 equiv) was added
at 0 °C, and the mixture was stirred for 30 min at ca. 10 °C
and then poured into a biphasic mixture of aqueous sodium
dihydrogen phosphate (MW: 156.0; 26 g; 1.1 equiv) and
toluene. After decantation, the organic phase was dried over
sodium sulfate and stored as a solution in toluene (ca. 100
mL). Concentration under vacuum afforded a waxy solid in
quantitative yield. C23H34O4Si; MW: 402.6; 1H NMR
(CDCl3, 250 MHz, ppm): δ 0.19 (s, 9H), 0.80 (s, 3H), 3.85-
4.00 (m, 4H), 4.47 (dd, J ) 1.5 and 1 Hz, 1H, 5a), 4.52 (m,
1H, 5b), 5.86 (m, 1H, 5b), 6.03 (m, 1H, 5a).
4-(2-(Diethylamino)ethoxy)bromobenzene (6).9 To a
stirred solution of 4-bromophenol (Fluka; MW: 173.0; 150
g; 0.867 mol), (2-chloroethyl)diethylamine hydrochloride
(ACROS Organics; MW: 172.1; 150 g; 1.00 equiv), and
TEBAC (Merck; 15 g) in dichloromethane (750 mL) at 25-
30 °C was added a 30% aqueous solution of sodium
hydroxide (MW: 40.0; 271 g; 2.3 equiv). The mixture was
stirred vigorously at 30-35 °C for 18 h (TLC monitoring:
dichloromethane 45, ethyl acetate 45, TEA 10). Sodium
chloride crystallised out. The mixture was then diluted with
water; the lower organic phase was decanted, washed with
water, dried over sodium sulfate, and concentrated in vacuo
to dryness (192.6 g yellow oil; yield: 81.6%; purity (GC):
99%): C12H18BrNO; MW: 272.2; IR (CHCl3, cm-1): ν
1592, 1579, 1490; 1H NMR (CDCl3, ppm): δ 1.06 (t, J ) 7
Hz, 6H), 2.62 (q, J ) 7 Hz, 4H), 2.85 (t, J ) 6.5 Hz, 2H),
4.00 (t, J ) 6.5 Hz, 2H), 6.78 and 7.35 (AA′BB′, 4H); MS
(EI; m/z): 271 (M+), 256, 185, 171, 86.
Grignard Reagent (7). A mixture of magnesium turnings
(MW: 24.3; 7.5 g; 1.13 equiv) and 25 mL of a solution of
4-(2-(diethylamino)ethoxy)bromobenzene (MW: 272.2; 74
g; 0.272 mol) in THF (250 mL) was stirred at 60 °C until
the reaction started (exotherm; gray colour). The rest of the
solution was added cautiously over ca. 60 min at 58 °C, and
the mixture was stirred for an additional hour at 58 °C and
then allowed to cool to room temperature. Assay (potenti-
ometry): ca. 1.0 M.
11-(4-(2-(Diethylamino)ethoxy)phenyl)estra-5(10),9-
(11)-diene-3,17-dione (9c) and 11r-(4-(2-(Diethylamino)-
ethoxy)phenyl)estra-4,9-diene-3,17-dione (9a). Copper(I)
chloride (99%; MW ) 99.0; 600 mg; 0.2 equiv) was added
at 20 °C to a solution of silyl enol ether 5b (â isomer; 12.2
g; 30.3 mmol) in THF (30 mL). A solution of 1 M Grignard
reagent 7 in THF (61 mL; 2 equiv) was then added at ca -3
°C. The mixture was stirred for 1 h at 20 °C (TLC
monitoring: toluene 1, ethyl acetate 1) and then poured into
a biphasic mixture of 25% aqueous solution of ammonium
chloride and dichloromethane. The organic phase was washed
with water and then concentrated under vacuum to dryness.
Dichloromethane (80 mL) and aqueous hydrochloric acid
(4.75 equiv; 42 mL) were added at 0-5 °C (exothermic).
The biphasic mixture was stirred vigorously during 2 h at
0-5 °C and then diluted with water (50 mL). The organic
phase was decanted, washed with water, and then neutralised
(pH ) 8) with aqueous 10% sodium bicarbonate solution
(50 mL), washed with water, dried over sodium sulfate, and
concentrated to dryness. The residue (11 g, 79%) was purified
by chromatography on silica gel (eluent: n-heptane 50, ethyl
acetate 45, triethylamine 5). Concentration of the purest
fractions afforded:
800 mg of 9c (5.7%; oil): C30H39NO3; MW: 461.6; IR
1
(CHCl3, cm-1): ν 1733, 1712, 1607, 1568, 1508; H NMR
(CDCl3, ppm): δ 1.03 (s, 3H); 1.12 (t, J ) 7 Hz, 6H); 2.72
(q, J ) 7 Hz, 4H); 2.94 (t, J ) 6 Hz, 2H); 4.09 (t, J ) 6 Hz,
2H); 6.82 and 7.07 (AA′BB′, 4H); MS (EI, m/z): 461 (M+),
100.
1
444 mg of 9a (3.2%; oil): C30H39NO3; MW: 461.6: H
NMR (CDCl3, ppm): δ 1.02 (s, 3H); 1.12 (t, J ) 7 Hz, 6H);
2.69 (q, J ) 7 Hz, 4H); 2.91 (t, J ) 7 Hz, 2H); 4.05 (m,
3H); 5.72 (s, 1H); 6.79 and 6.95 (AA′BB′, 4H); MS (ESP,
m/z): 426 (MH+).
Mixture of 11â-(4-(2-(Diethylamino)ethoxy)phenyl)-
estra-4,9-diene-3,17-dione (9b), and of Diene-diones 9c
and 9a. Copper(I) chloride (MW ) 99.0; 2.25 g; 0.15 equiv)
was added at 20 °C to a ca. 1 M solution of Grignard reagent
7 in THF (300 mL; 2.0 equiv). A solution of silyl enol ethers
(5b/5a ≈ 2/1) (60.8 g; 151 mmol) in toluene (100 mL) and
THF (100 mL) was added during ca. 30 min at 20 °C
(exothermic). The mixture was then stirred for 1 h at 20 °C
(TLC monitoring: toluene-ethyl acetate 1/1) and then
poured into a biphasic mixture of aqueous ammonium
chloride (15 equiv; 600 mL) and dichloromethane at 10-15
°C. The organic phase was separated, washed with water,
concentrated under vacuum to ca. 100 mL and diluted with
dichloromethane (250 mL). Aqueous hydrochloric acid (6
equiv; 200 mL) was added at 0-5 °C (exothermic). The
biphasic mixture was stirred vigorously for 2 h at 0-5 °C
and then diluted with water (250 mL). The organic phase
was decanted and washed with water. The organic phase was
carefully neutralised (pH ) 8) with aqueous 10% sodium
bicarbonate (250 mL) (evolution of carbon dioxide), washed
with water, and dried over sodium sulfate. Concentration to
dryness afforded a semi-crystalline material (67.5 g; 97%
yield). Compound 9b could be isolated by crystallisation
from diisopropyl ether: C30H39NO3; MW: 461.6; mp
(DSC): 188 °C; IR (CHCl3, cm-1): ν 1735, 1658, 1609,
1581, 1509; 1H NMR (CDCl3, ppm): δ 0.56 (s, 3H), 1.06 (t,
J ) 7 Hz, 6H), 2.63 (q, J ) 7 Hz, 4H), 2.85 (t, J ) 6 Hz,
2H), 4.01 (t, J ) 6 Hz, 2H), 4.38 (ld, J ) 7 Hz, 1H), 5.80
(s, 1H), 6.82 and 7.07 (AA′BB′, 4H); MS (EI, m/z): 461
(M+).
11â-(4-(2-(Diethylamino)ethoxy)phenyl)estra-1,3,5(10)-
trien-3-ol-17-one Hydrochloride (14). To the solution of
enones 9b,c,a described above (67.5 g; 146 mmol) in
dichloromethane (200 mL) were slowly added at 20-25 °C
acetic anhydride (MW: 102.1; d: 1.09; 14 mL; 1.0 equiv)
and then acetyl bromide (MW: 123.0; d: 1.66; 27 mL; 2.5
equiv) (exothermic addition). The brown solution was stirred
226
•
Vol. 8, No. 2, 2004 / Organic Process Research & Development

for 5 h at 20-25 °C (HPLC monitoring) and then carefully
poured into an aqueous suspension of sodium bicarbonate
(500 mL; 10 equiv) (evolution of carbon dioxide). The
mixture was stirred vigorously during 18 h at ca. 20 °C, and
then the organic phase was separated and washed with 1 N
aqueous sodium hydroxide (250 mL), and then water. The
organic phase containing mainly arylestrone acetate 13b was
concentrated to 150 mL. Then, dichloromethane was ex-
changed for methanol by azeotropic distillation at constant
volume (150 mL) at ca. 40 °C (the pressure was progressively
reduced). The mixture was cooled to 0-5 °C, and a solution
of potassium hydroxide (12.3 g; MW: 56.0; 1.5 equiv) in
methanol (100 mL) was added (exothermic). The mixture
was stirred for 1.5 h at 0-5 °C (HPLC monitoring) and then
poured into water (250 mL) and dichloromethane (250 mL).
The organic phase was decanted and washed with water until
no trace of bromide ions was detected (AgNO3 test). The
organic phase was acidified to pH < 2 by addition of aqueous
hydrochloric acid (2.0 equiv; 275 mL). The organic phase
was decanted, dried over sodium sulfate, and adjusted to 300
mL. Dichloromethane was exchanged for methyl ethyl ketone
(MEK) by distillation at constant volume (300 mL), until
the inner temperature reached 78 °C. The resulting suspen-
sion of arylestrone hydrochloride 14 was cooled and stirred
for 1 h at 20-25 °C. The off-white crystals were washed
with MEK (2 × 50 mL) and dried under vacuum at ca. 70
°C for 18 h: 51.0 g (based on dry product); yield: 70%
(68% from 4a/b); purity (LC): 95%; solvation: 1.4%;
C30H40ClNO3; MW: 498.1; mp(DSC): 189 °C; IR (CHCl3,
MW: 496.0: IR (CHCl3, cm-1): ν 1736, 1677, 1609, 1580,
1549, 1509; H NMR (CDCl3, ppm): δ 0.57 (s, 3H), 1.06
(t, J ) 7 Hz, 6H), 2.62 (q, J ) 7 Hz, 4H), 2.85 (t, J ) 6.5
Hz, 2H), 3.25 (dt, J ) 16.5 and 3 Hz, 1H), 4.00 (t, J ) 6.5
Hz, 2H), 4.38 (d, J ) 7 Hz, 1H), 6.82 and 7.05 (AA′BB′,
4H); MS (EI; m/z): 496 (MH+), 495, 86.
1
4-Chloro-11â-(4-(2-(diethylamino)ethoxy)phenyl)estra-
1,3,5(10)-trien-3-ol-17-one Hydrochloride (11). To a solu-
tion of chlorodienone 10 (40 g; 75.1 mmol) in dichlo-
romethane (160 mL) at room temperature was added acetic
anhydride (21 mL; 3.0 equiv) and then acetyl bromide (21
mL; 3.8 equiv) (exothermic addition). The brown solution
was stirred for 6 h at 20-25 °C (HPLC monitoring) and
then carefully poured into an aqueous suspension of sodium
bicarbonate (84 g; 13.4 equiv) in water (400 mL) (evolution
of carbon dioxide). The mixture was stirred vigorously during
2 h at ca. 20 °C; the organic phase was then collected and
acidified (pH: 1) with aqueous hydrochloric acid (3.3 equiv;
240 mL). The organic phase was dried (sodium sulfate),
filtered, and concentrated to 120 mL. Dichloromethane was
exchanged for methanol by azeotropic distillation at constant
volume (120 mL) at ca. 40 °C (the pressure was progressively
reduced). The mixture was cooled to 0-5 °C, and a solution
of potassium hydroxide (10.5 g; MW: 56.0; 2.5 equiv) in
methanol (80 mL) was added (exothermic). The mixture was
stirred for 1.5 h at 0-5 °C (HPLC monitoring) and then
poured into aqueous hydrochloric acid (3.0 equiv; 240 mL)
and dichloromethane (200 mL). The organic phase was
decanted, dried over sodium sulfate, and concentrated to 120
mL. Residual methanol was exchanged for dichloromethane
by azeotropic distillation under normal pressure. The suspen-
sion was cooled and stirred for 1 h at 15-20 °C. The light-
pink crystals of 11 were washed with dichloromethane and
dried under vacuum at ca. 20 °C for 18 h: 19 g (based on
dry product); yield: 47.5% (40% from 9b); purity (LC):
90%; solvation: 4.1%; C30H39Cl2NO3; MW: 532.6; IR
1
cm-1): ν 3601, 2456, 1733, 1610, 1584, 1511; H NMR
(CDCl3, ppm): δ 0.42 (s, 3H), 1.31 (t, J ) 7 Hz, 6H), 3.16
(q, J ) 7 Hz, 4H), 3.31 (t, J ) 6.5 Hz, 2H), 3.96 (t, J ) 6
Hz, 1H), 4.17 (t, J ) 6.5 Hz, 2H), 6.51 (m, 1H), 6.68 (m,
1H), 6.73 (m, 1H), 6.51 and 6.95 (AA′BB′, 4H), 11.36 (1
exch. H); MS (EI; m/z): 461 (M+), 446, 362, 86, 38, and 36
(HCl).
1
4-Chloro,11â-(4-(2-(diethylamino)ethoxy)phenyl)estra-
4,9-diene-3,17-dione Hydrochloride (10). To a solution of
dienone 9b (20 g; 43.3 mmol) in dichloromethane (100 mL)
at 20 °C was added pyridine (100 mL; MW: 79.1; d: 0.98;
28.6 equiv). The solution was cooled to -40 °C and then
sulfuryl chloride (8.85 g; MW: 135.0; 1.5 equiv) was slowly
added. The mixture was stirred for 2 h at -40 °C (TLC
monitoring: heptane 45, ethyl acetate 45, TEA 10) and then
poured into aqueous hydrochloric acid (27 equiv, 500 mL).
The biphasic mixture (pH: 1) was stirred for 15 min and
then the organic phase was collected, washed with water,
dried (sodium sulfate), and concentrated to 80 mL. The
solvent was exchanged for ethyl acetate, under vacuum at
ca. 40 °C. The yellow solid was filtered at 20 °C, washed
with ethyl acetate, and dried in vacuo at room temperature,
affording 18.9 g of compound 10 (based on dry product).
Yield: 82%; purity (LC): 73%; solvation: 3.2%; C30H39-
Cl2NO3; MW: 532.5; mp: 179 °C; 1H NMR (CDCl3,
ppm): δ 0.57 (s, 3H), 1.46 (t, J ) 7 Hz, 6H), 1.40-1.70
(m, 4H), 3.25 (m, 4H), 3.44 (m, 2H), 4.39 (d, J ) 7 Hz,
1H), 4.52 (m, 2H), 6.83 and 7.09 (AA′BB′, 4H), 12.62 (1
exch. H); base (purified by chromatography): C30H38ClNO3;
(CHCl3, cm-1): ν 1727, 1610, 1582, 1568, 1511, 1493; H
NMR (CDCl3, ppm): δ 0.45 (s, 3H), 1.41 (t, J ) 7 Hz, 6H),
3.21 (q, J ) 7 Hz, 4H), 3.37 (t, J ) 6.5 Hz, 2H), 3.99 (t, J
) 6 Hz, 1H), 4.38 (t, J ) 6.5 Hz, 2H), 5.87 (1 exch. H),
6.68 (d, J ) 8.5 Hz, 1H), 6.78 (d, J ) 8.5 Hz, 1H), 6.60 and
6.94 (AA′BB′, 4H), 12.30 (1 exch. H); MS (ES+; m/z): 498,
496 (MH+).
3-Acetoxy-11â-(4-(2-(diethylamino)ethoxy)phenyl)estra-
1,3,5(10)-trien-17-one Hydrochloride (13b, HCl) and
3-Acetoxy-11-(4-(2-(diethylamino)ethoxy)phenyl)estra-
3,5(10),9(11)-trien-17-one Hydrochloride (13c, HCl). To
a solution of dienone 9b (10 g; 21.7 mmol) in dichlo-
romethane (40 mL) were slowly added at 20-25 °C acetic
anhydride (6.1 mL; 3.0 equiv) and then acetyl bromide (6.1
mL; 3.8 equiv) (exothermic addition). The brown solution
was stirred for only 1 h at 20-25 °C (HPLC monitoring)
and then carefully poured into an aqueous suspension of
sodium bicarbonate (13 equiv; 100 mL) (evolution of carbon
dioxide). The mixture was stirred vigorously during 30 min
at ca. 20 °C; the organic phase was then decanted, washed
with water, and dried over sodium sulfate. Concentration
under vacuum afforded an oily residue which was purified
Vol. 8, No. 2, 2004 / Organic Process Research & Development
•
227

by chromatography on silicagel (eluent: cyclohexane 75,
ethyl acetate 20, triethylamine 5). The purest fractions were
concentrated to dryness, diluted into dichloromethane, acidi-
fied to pH < 2 using 1 N hydrochloric acid, dried, and
concentrated under vacuum.
The less polar fractions gave 13c, HCl (1.3 g yellow solid;
yield: 11%): C32H42ClNO4; MW: 540.1; IR (CHCl3, cm-1):
ν 1736, 1664, 1606, 1570, 1507; 1H NMR (CDCl3, ppm): δ
1.02 (s, 3H), 1.47 (t, J ) 7 Hz, 6H), 2.10 (s, 3H), 3.27 (m,
4H), 3.47 (m, 2H), 4.51 (m, 2H), 5.52 (d, J ) 1.5 Hz, 1H),
6.79 and 7.12 (AA′BB′, 4H), 12.4 (1 exch. H); MS (EI; m/z):
503 (M+) 461, 100, 86, 38, and 36 (HCl).
The more polar fractions gave 13b, HCl (6.09 g white
solid; 52%): C32H42ClNO4; MW: 540.1; IR (CHCl3, cm-1):
ν 1734, 1610, 1582, 1512, 1494; 1H NMR (CDCl3, ppm): δ
0.45 (s, 3H), 1.43 (t, J ) 7 Hz, 6H), 2.25 (s, 3H), 3.22 (m,
4H), 3.39 (m, 2H), 4.40 (m, 2H), 4.04 (d, J ) 4.5 Hz, 1H),
6.63 and 6.99 (AA′BB′, 4H), 6.65 (dd, J ) 8.5 and 1.5 Hz,
1H), 6.86 (d, J ) 1,5 Hz, 1H), 6.94 (d, J ) 8,5 Hz, 1H),
12.3 (1 exch. H); MS (EI; m/z): 503 (M+), 86, 38, and 36
(HCl).
4-Chloroarylestrone Hydrochloride (11), by Chlorina-
tion of 14. To a solution of arylestrone 14 (50 g; 100 mmol)
in dichloromethane (250 mL) and methanol (250 mL) was
added at ca. 10 °C 36% hydrochloric acid (4 mL; 0.5 equiv)
and then portions of N-chlorosuccinimide (13.4 g; MW:
133.5; 1.0 equiv) (exothermic). The solution was stirred for
1-3 h at 8-12 °C (HPLC monitoring). As arylestrone 14
was more difficult to eliminate by crystallisation than the
dichloroarylestrone 16, NCS (0.03 equiv) was added por-
tionwise until compound 14 was totally (>99%) consumed.
The solution was poured into a mixture of water (500 mL),
sodium thiosulfate (25 g), sodium hydroxide (1.5 equiv), and
dichloromethane (250 mL). The mixture was stirred for 15
mn at 20-25 °C. The organic phase was decanted and
acidified to pH < 2 with aqueous hydrochloric acid (1.5
equiv; 140 mL). After decantation, the aqueous phase was
extracted with a 2/1 mixture of dichloromethane and
methanol. The organic phases were combined, dried over
sodium sulfate, and concentrated to 250 mL. Methanol was
exchanged for dichloromethane by azeotropic distillation at
normal pressure (ca. 900 mL of DCM were necessary; final
T: 39.8 °C). Chloroarylestrone hydrochloride 11 crystallised
and was filtered at 20 °C and dried at ca. 40 °C under vacuum
for 18 h: 44.1 g (based on dry product) white solid; yield:
82.5%; purity (HPLC): 98%; solvation: 3%; mp (DSC):
175 °C; C30H39Cl2NO3; MW: 532.6; IR, 1H NMR, MS: see
above.
aqueous sodium hydroxide (4.6 mL; 1.05 equiv). The mixture
was stirred until complete dissolution was obtained and then
cooled to ca. 0 °C. Sodium borohydride (1.4 g; MW: 37.8;
37 mmol) was added portionwise (slightly exothermic;
evolution of hydrogen) during 15 min, keeping the temper-
ature below 5 °C. The mixture was stirred for 1 h at 0-5
°C (HPLC monitoring). Excess hydride was consumed by
addition of acetone (20 mL; 5.8 equiv). The suspension was
stirred for 1 h at 5-10 °C and then poured into water (250
mL) and dichloromethane (250 mL). The organic phase was
washed with 2/1 mixture of water and methanol, the
washings were re-extracted with a 1/1 mixture of dichlo-
romethane and methanol. The organic phase was dried over
sodium sulfate and concentrated to 100 mL. The solvents
were exchanged for 2-propanol by distillation at constant
volume (normal pressure). At ca. 56 °C, seeding crystals of
2 (0.1 g) were added. Distillation was continued (final T: 82
°C); the crystallisation developed. The suspension was cooled
to 0-5 °C during 1 h and stirred for 1 h at 0-5 °C. The
white crystals (2-propanol solvate of 2) were filtered, washed
with 2-propanol (25 mL) and then suspended into water (500
mL). The suspension was stirred for 1 h at 80 °C, the white
crystals (hydrate of 2) were filtered at 80 °C, washed with
warm water, and dried under moderate vacuum for 18 h at
40 °C. 19.4 g (based on dry product); yield: 83%; purity
(HPLC): 99.2%; solvation: 3.7% water (monohydrate).
C30H40ClNO3; MW: 498.1; mp (DSC): 107 °C (dehydration
then melting); [R]D ) -95° (c ) 1% m/v in ethanol; based
on dry product); IR (CHCl3, cm-1): ν 3608, 3536, 1610,
1
1580, 1512; H NMR (CDCl3, 400 MHz, ppm): δ 0.32 (s,
3H), 1.03 (t, J ) 7 Hz, 6H), 1.21 (dd, J ) 13 and 2 Hz,
1H), 1.35 (m, 2H), 1.40 (m, 1H), 1.73 (m, 1H), 1.78 (dd, J
) 13 and 5.5 Hz, 1H), 2.10 (m, 3H), 2.50 (dd, J ) 13 and
2 Hz, 1H), 2.61 (q, J ) 7 Hz, 4H), 2.77 (m, 1H), 2.81 (t, J
) 6.5 Hz, 2H), 2.82 (m, 1H), 3.11 (dd, J ) 17.5 and 4.5
Hz, 1H), 3.66 (dd, J ) 8.5 and 7 Hz, 1H), 3.91 (t, J ) 5.5
Hz, 1H), 3.94 (t, J ) 6.5 Hz, 2H), 6.60 (d, J ) 8.5 Hz, 1H),
6.78 (d, J ) 8.5 Hz, 1H), 6.60 and 6.89 (AA′BB′, 4H); MS
(EI, m/z): 497, 482, 86.
Acknowledgment
This development work (lab and pilot) was carried out in
the Romainville Research Center. Many collegues involved
in this work are thanked for their significant contributions.
Among them, we would particularly like to thank Franc¸ois
Nique and Christian Moratille (Medicinal Chemistry) for
fruitful discussions.
4-Chloro-11â-(4-(2-(diethylamino)ethoxy)phenyl)estra-
1,3,5(10)-trien-3,17-diol (2). To a solution of 11 (25 g; 47
mmol) in methanol (250 mL) was added at ca. 10 °C 30%
Received for review November 14, 200.
OP0341622
228
•
Vol. 8, No. 2, 2004 / Organic Process Research & Development