A convenient synthesis of (Z)-4-hydroxy-N-desmethyltamoxifen (endoxifen)

Article abstract of DOI:10.1016/j.bmcl.2010.03.117

A mixture of the (Z)- and (E)-isomers of 4-hydroxy-N-desmethyltamoxifen was conveniently prepared in four steps. These geometrical isomers were then neatly separated by semi-preparative Reverse Phase High Performance Liquid Chromatography (RP-HPLC) using specified conditions. Additionally, the isolated E-isomer could be equilibrated in aqueous strong acid in acetonitrile or trifluoroacetic acid/dichloromethane to give a clean 1:1 mixture of Z/E isomers that was re-subjected to HPLC separation. In this way, most of the undesired (E)-isomer could be readily converted to the desired (Z)-isomer providing quick access to over 200 mg quantities of pure endoxifen (Z-isomer), a potent antiestrogenic metabolite of tamoxifen traditionally used in breast cancer treatment.

Full text of DOI:10.1016/j.bmcl.2010.03.117

Bioorganic & Medicinal Chemistry Letters 20 (2010) 3036–3038
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
A convenient synthesis of (Z)-4-hydroxy-N-desmethyltamoxifen (endoxifen)
*
Abdul H. Fauq , Ghulam M. Maharvi, Dola Sinha
Chemical Synthesis Core Facility, Mayo Clinic Jacksonville, FL 32246, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
A mixture of the (Z)- and (E)-isomers of 4-hydroxy-N-desmethyltamoxifen was conveniently prepared in
four steps. These geometrical isomers were then neatly separated by semi-preparative Reverse Phase
High Performance Liquid Chromatography (RP-HPLC) using specified conditions. Additionally, the iso-
lated E-isomer could be equilibrated in aqueous strong acid in acetonitrile or trifluoroacetic acid/dichlo-
romethane to give a clean 1:1 mixture of Z/E isomers that was re-subjected to HPLC separation. In this
way, most of the undesired (E)-isomer could be readily converted to the desired (Z)-isomer providing
quick access to over 200 mg quantities of pure endoxifen (Z-isomer), a potent antiestrogenic metabolite
of tamoxifen traditionally used in breast cancer treatment.
Received 17 February 2010
Revised 29 March 2010
Accepted 31 March 2010
Available online 3 April 2010
Keywords:
LY411575
Endoxifen
Tamoxifen
Ó 2010 Elsevier Ltd. All rights reserved.
Semipreparative HPLC
1-[4-(2-Dimethylaminoethoxy)-phenyl]-1,2-diphenylbut-1(Z)-
ene (tamoxifen (TAM)) is a non-steroidal antiestrogen drug widely
used for breast cancer treatment.1 The pharmacological profiles of
tamoxifen indicate that it elicits its anti-cancer activity through its
active metabolites 4-hydroxytamoxifen (4-OH-TAM) and its desm-
ethyl analogue endoxifen that are generated by the action of hepatic
CYP 2D6 and 3A4 isozymes on tamoxifen after hydroxylation fol-
lowed by N-demethylation.2,3 It is established that some patients
do not derive therapeutic benefit from the administration of tamox-
ifen or even suffer relapses because of their inherent genotypic con-
straints.3 Co-administration of certain drugs (e.g., paroxetine or
fluoxetine) also have demonstrated interactive inhibitory effect on
CYP 2D6 and other cytochrome P450 enzymes. This of lack of activity
results from reduced availability of therapeutic levels of the active
metabolite (Z)-desmethyl-4-hydroxytamoxifen (endoxifen).4 Re-
cently, Hawse and co-workers have shown that endoxifen is the ac-
tual anti-estrogenic drug that works by degrading the estrogen
receptor and not by its inhibition.3 Therefore, in order to facilitate
human tissue studies as well to explore the possibility that it may
be appropriate in special cases to consider a dose regimen that, in-
stead of tamoxifen, includes the active metabolite endoxifen, rela-
tively large quantities of endoxifen may be required. A search of
the literature showed that while low level milligram quantities of
the endoxifen have been purified by analytical HPLC,5 synthesis
and purification of relatively large quantities of endoxifen have not
been reported. Herein, we report synthesis and purification proto-
cols that have resulted in production of pure endoxifen in excess of
200 mg quantities for animal studies and tissue work in less than
two weeks time. This work involved a short four-step synthesis of
a mixture of endoxifen and the (E)-isomer and the use of semi-pre-
parative reverse phase HPLC columns for their separation. Neverthe-
less, much larger quantities of endoxifen are expected to be
conveniently generated by employing preparative columns.
Apart from a stereoselective synthesis of tamoxifen involving car-
bometalation of alkynylsilanes,6 earlier reported syntheses of (Z)-4-
hydroxytamoxifen,
a precursor of endoxifen, were somewhat
cumbersome and non-stereoselective.7 A ground-breaking stereose-
lective synthesis of (Z)-4-hydroxytamoxifen by Gauthier and Labrie
involved McMurry reaction as a key step.8 Even though the synthesis
of endoxifenwasnotreported, theauthors managedtoobtaina favor-
able 14:1 ratio of the (E)- and (Z)-isomers of 1-(4-hydroxyphenyl)-1-
[4-(trimethylacetoxy)phenyl]-2-phenylbut-1-ene (3b:4a,Fig. 1) after
reacting the monopivaloyl derivative of 4,40-dihydroxybenzophe-
none (1b) with propiophenone in the McMurry reaction simply by
manipulating proportions of TiCl4 and Zn. The ratio 14:1 was en-
hanced to 100:1 by trituration of the crude with methanol. In our
hands, the crude obtained after the McMurry reaction was directly
chromatographed on silica gel to obtain desired (E)-stereoisomer in
88% yield. In this reaction, the phenolic and the ethyl components
were shown to preferentially align trans to each other as prece-
dented.9 We carried out the reported four-step Gauthier–Labrie syn-
thetic sequence through the intermediate 5 to (Z)-4-OH-TAM 3c.
Unfortunately, attempted demethylation of 5 or the (Z)-4-OH-TAM
3c itself using drastic conditions (vinyl chloroformate in dioxane at
135–170 °C in sealed tube) as reported5c or ethyl chloroformate5a,10a
in refluxing toluene produced a significantly stereoscrambled mix-
ture of endoxifen and its (E)-isomer in low yield. In using some other
milder carbamate-mediated N-demethylation conditions,10b–d it was
noticed that lower yields and significant stereorandomization of the
* Corresponding author. Tel.: +1 904 953 8034; fax: +1 904 953 7117.
E-mail address: fauq.abdul@mayo.edu (A.H. Fauq).
0960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.03.117

A. H. Fauq et al. / Bioorg. Med. Chem. Lett. 20 (2010) 3036–3038
3037
O
RO
OH
RO
OR'
RO
OR'
1a, R = H
4a, R = pivaloyl; R' = H
4b, R = H; R' = CH2CH2N(Me)2
((E)-4-OH-TAM)
3a, R = H; R' = H
1b, R = pivaloyl
3b, R = pivaloyl; R' = H
2, R = CH2CH2N(Me)2
3c, R = H; R' = -CH2CH2N(Me)2
7a
, R =CH2CH2N(CH3)C(O)OCH(Cl)CH3;
R' = pivaloyl
((Z)-4-OH-TAM)
5
,
R = pivaloyl; R' = -CH2CH2N(Me)2
R = pivaloyl;
7b
, R = pivaloyl;
6a,
R' =CH2CH2N(CH3)C(O)OCH2CH3
8b, R = H;
R' =CH2CH2N(CH3)C(O)OCH(Cl)CH3
6b, R = pivaloyl;
R' = CH2CH2NHCH3 (E)-isomer
R' = CH2CH2N(CH3)C(O)OCH2CH3
8a, R = H; R' = CH2CH2NHCH3 (endoxifen)
Figure 1. Endoxifen and reported intermediates for its synthesis.
double bond could not be avoided. Of the various carbamate-medi-
ated demethylation protocols, the best reaction conditions entailed
heating a mixture of 2-chloroethyl chloroformate and (Z)-4-OH-
TAM 3c in dichloroethane which resulted in stereorandomization of
the tetrasubstituted double bond to the extent of 20% furnishing a
4:1 mixture of the desired endoxifen carbamoyl precursor (6a) and
its (Z)-isomer (7a) in 79% yield (Fig. 1). The mechanism of the isomer-
ization during carbamate-mediated N-demethylation is currently
speculative, albeit precedented.5a Since the components of this ste-
reoisomeric mixture could not be separated, it was directly reacted
with methyllithium at ꢀ78 °C in tetrahydrofuran which concomi-
tantly cleaved carbamoyl and pivaloyl ester moieties to generate a
4:1 mixture of endoxifen (8a) and its (E)-isomer (8b). The overall
six-step synthetic sequence furnished a mixture of endoxifen 8a
and its (E)-isomer 8b in 26% combined yield.
With the hope of retaining stereochemical integrity during
demethylation, we attempted to synthesize pure carbamoyl (E)-
isomer of 6b by reacting 3b (Scheme 1) either with the ethyl (2-
hydroxyethyl)(methyl)carbamate (10, prepared by reacting the
alcohol with ethylchloroformate/triethylamine in DCM; cf.
Supplementary data) under Mitsunobu conditions, or by SN2
reaction between 3b and 2-bromo or 2-iodo-derivative of ethyl
(2-hydroxyethyl)(methyl)carbamate (prepared by reacting ethyl
(2-hydroxyethyl)(methyl)carbamate 10 with CBr4, triphenylpho-
shine/DCM to give 11a, or with I2/triphenylphoshine/imidazole/
toluene to give 11b under basic (Cs2CO3/DMF); cf. Supplementary
data). While the Mitsunobu reaction completely retained stereo-
chemistry of the substituted product 6b, the basic conditions
(Cs2CO3/DMF) employed during bromo- or iodo-displacement re-
sulted in 30% stereoscrambling to give 7b. This unwelcome reac-
tion outcome suggested that both acidic (e.g., treatment with
DCM/TFA mixture, vide infra) as well as the basic milieu compro-
mised the stereochemical integrity of the electron-rich tetrasubsti-
tuted double bond. The labile nature of the stereochemistry under
basic conditions was further evidenced when an attempt was
made to achieve simultaneous removal of the pivaloyl and ethoxy-
carbonyl moieties from pure 6b (derived from Mitsunobu reaction)
with methyllithium under the usual mild (ꢀ78 °C) conditions. This
reaction also produced ca. 30% undesired (E)-isomer 8b. Similarly,
when methyllithium was mixed with pure 8b in THF at ꢀ78 °C,
30% of endoxifen 8a was obtained. While the acid-catalyzed stere-
orandomization of the tetrasubstituted double bond is expected,
the base-mediated partial isomerization during the ethoxycar-
bonyl removal may be explained by invoking the possibility of res-
onance in phenolate anion playing a part in generating a partial
single bond character in the para-substituted sp2-hybridized ben-
zylic carbon of the phenol moiety.
Since it was important to rapidly produce over 200 mg of
endoxifen for use in several ongoing anti-breast cancer investiga-
tions, the problem of separating the final 4:1 endoxifen:(E)-isomer
mixture presented a challenge. Even though RP-HPLC separation of
endoxifen from its (E)-isomer has been achieved in small quanti-
ties for in vitro studies,5c–e protocols for larger scale separation
were not reported. Our attempts using the reported RP-HPLC con-
ditions for larger scale separation using semi-preparative RP-HPLC
columns and phosphate or triethylamine-containing buffers gave
poor resolution of the endoxifen and its (E)-isomer. Additionally,
silica gel chromatography using various eluents containing basic
additives was unsuccessful. After considerable experimentation,
success was finally realized when the RP-HPLC separation of the
I
Br
OH
OH
CBr4/TPP/THF
EtOC(O)Cl/TEA/
N
O
N
O
or
N
O
or, I2/TPP/imidazole
NH
O
O
DCM
O
11b
11a
10
9
O
N
O
10, TPP, DEAD,
or, 11a,/ 11b, Cs2CO3
OEt
OEt
N
+
PivO
OH
PivO
O
PivO
O
7b
3b
6b
Scheme 1.

3038
A. H. Fauq et al. / Bioorg. Med. Chem. Lett. 20 (2010) 3036–3038
O
(i) CH3CH(Cl)OCOCl
H
N
DCE, 0 ºC to reflux
.HCl
2
HO
O
(ii) 6M HCl, MeOH
9
reflux, 4 h
equilibration
+
O
H
N
H
N
.HCl
HO
O
HO
O
Zn, TiCl3, THF
reflux, 7h
8a (endoxifen)
8b
(E)-isomer)
Scheme 2. Modified synthesis of endoxifen.
stereoisomeric mixture was attempted with isocratic elution with
a buffer containing 50% of 20 mM triethylammonium bicarbonate
in acetonitrile at pH 8.8. This protocol separated the two peaks well
apart even under significant column overloads (Vydac column, C-8,
2.2 ꢁ 25 cm, FR 8 mL/min: RT for endoxifen, 53 min; for (E)-isomer
81 min). The identities of the two geometrical isomers were con-
firmed by peak matching with reported NMR data5a as well as by
its expected antiestrogenic activity.3
combined with a highly efficient RP-HPLC protocol for separation
of endoxifen from the (E)-isomer. This methodology was used to
conveniently and rapidly prepare over 200 mg quantities of pure
endoxifen as and when needed for animal and tissue studies.
Acknowledgment
The internal financial support provided for this synthetic pro-
ject by Mayo Foundation is gratefully acknowledged.
In spite of the fact that the six-step protocol for the synthesis of
endoxifen can be achieved from the published synthetic proce-
dures from its precursor 4-OH-TAM, the confounding problem of
significant double bond isomerization during the demethylation
still remained and necessitated HPLC purification. It was, therefore,
deemed pragmatic to shorten the overall synthesis of endoxifen by
doing away with protection/deprotection of the hydroxyl group of
4,40-dihydroxybenzophenone (1) altogether. Our overall four-step
strategy that continues to rely on the McMurry reaction8 is given
in Scheme 2. The N,N-dimethylethyl derivative 4,40-dihydroxyben-
zophenone (2), made from 4,40-dihydroxybenzophenone 1 in 46%
yield was demethylated using 2-chloroethyl chloroformate-medi-
ated demethylation methodology as described above in 83% overall
yield. However, instead of decarbamoylation with methyllithium
furnishing lower yield of the deprotected product, the intermediate
2-chloroethyl carbamate was decomposed with 6 M-HCl in reflux-
ing methanol to give the secondary amine hydrochloride salt (9) in
higher yield (83%, Scheme 2). The hydrochloride salt 9 was sub-
jected to McMurry reaction with propiophenone furnishing a 90%
chromatographed combined yield of a mixture of the endoxifen
8a and the (E)-isomer 8b in 1:3 ratio. Fortunately, this unfavorable
ratio was readily and cleanly enhanced to 1:1 by heating the iso-
meric mixture in 6 N-HCl in aqueous acetonitrile for 6 h or by,
more simply, stirring the mixture with 1:1 DCM/TFA for 1 h. The
(Z)-and (E)-isomers were then separated using the RP-HPLC condi-
tions as outlined above. Additionally, the undesired (E)-isomer
(8b)-containing fractions obtained from the HPLC runs were com-
bined and re-equilibrated cleanly to 1:1 mixture of endoxifen and
8b either by heating with equal volume of 6 N-HCl at 60 °C for 4–
6 h, or, by evaporating to dryness and stirring with 1:1 DCM/TFA at
rt. In both cases, the equilibrated mixture was directly subjected to
HPLC purification resulting in enhanced overall yield of the endox-
ifen. The remarkably large RT difference between the two isomers
achieved under specified buffer conditions was critical to their suc-
cessful larger scale separation because semipreparative RP-HPLC
column could be safely overloaded. This protocol also enabled stor-
age of large quantities of endoxifen as 1:1 Z/E mixture at ꢀ15 °C
under dark for extended periods of time, pending fresh isolation
of endoxifen as and when needed.
Supplementary data
Supplementary data (the spectroscopic and RP-HPLC chromato-
graphic data as well as synthetic procedures for all the reported
intermediates, and those of the final products) associated with this
article can be found, in the online version, at doi:10.1016/
j.bmcl.2010.03.117.
References and notes
1. (a) Fisher, B.; Costantino, J. P.; Wickerham, D. L., et al J. Natl. Cancer. Inst. 1998,
90, 1371; (b) Osborne, C. K. N. Eng. J. Med. 1998, 339, 1609; (c) Cuzick, J.;
Powles, T.; Veronesi, U., et al Lancet 2003, 361, 296; (d) Howell, A.; Howell, S. J.;
Evans, D. G. Cancer Chemother. Pharmacol. 2003, 52, S39.
2. (a) Jordan, V. C.; Collins, M. M.; Rowsby, L.; Prestwich, G. A. J. Endocrinol. 1977,
75, 305; (b) Furr, B. J.; Jordan, V. C. Pharmacol. Ther. 1984, 25, 27; (c)
Katzenellenbogen, B. S.; Norman, M. J.; Eckert, R. L.; Peltz, S. W.; Mangel, W. F.
Cancer Res. 1984, 44, 112.
3. Wu, X. L.; Hawse, J. R.; Subramaniam, M.; Goetz, M. P.; Ingle, J. N.; Spelsberg, T.
C. Cancer Res. 2009, 69, 1722. and references cited therein.
4. Goetz, M. P.; Ingle, J. Cancer 2007, 110, 2595.
5. (a) Sun, D.; Sharma, A. K.; Dellinger, R. W.; Blevins-Primeau, A. S.; Balliet, R.;
Chen, G.; Boyiri, T.; Amin, S.; Lazarus, P. Drug Metab. Dispos. 2007, 35, 2006; (b)
Johnson, M. D.; Zuo, H.; Lee, K. H.; Trebley, J.; Rae, J. M.; Weatherman, R. V.;
Desta, Z.; Flockhart, D. A.; Skaar, T. C. Breast Cancer Res. 2004, 85, 151; (c)
Stearns, V.; Johnson, M. D.; Rae, J. M.; Morocho, A.; Novielli, A.; Bhargava, P.;
Hayes, D. F.; Desta, Z.; Flockhart, D. A. J. Natl. Cancer Inst. 2003, 95, 1758; (d)
Johnson, M. D.; Zuo, H.; Lee, K. H.; Trebley, J. P.; Rae, J. M.; Weatherman, R. V.;
Desta, Z.; Flockhart, D. A.; Skaar, T. C. Breast Cancer Res. Treat. 2004, 85, 151; (e)
Kushner, P.; Cyrus, H.; Myles, D.; James, P. Publication No. WO2009/120999
(A2).
6. Al-Hassan, M. I. Synth. Commun. 1989, 19, 1619.
7. (a) Grafe, I.; Schickaneder, H.; Jungblut, P. W.; Ahrens, K. H. EP 287690 A1
881026, Heumann Pharma G.m.b.H. und Co.; (b) McCague, R. J. Chem. Res.,
Synop. 1986, 2, 58; (c) McCague, R. J. Chem. Res., Miniprint 1986, 771; (d) Forster,
A. B.; Jarman, M.; Leung, O. T.; McCague, R.; Leclercq, G.; Devleeschouwer, N. J.
Med. Chem. 1985, 28, 1491; (e) Toivola, R. J.; Karjalainen, A. J.; Kurbela, K. O. A.;
Soderwall, M. L.; Kangas, L. V. M.; Blanco, G. L.; Sundquist, H. K. EP 95875 A2
831207.; (f) Richardson, D. N. DE 2807599 780831.
8. Gauthier, S.; Mailhot, J.; Labrie, F. J. Org. Chem. 1996, 61, 3890.
9. Coe, P. L.; Scriven, C. E. J. Chem. Soc., Perkin Trans. 1 1986, 475.
10. Some other N-demethylation protocols employing various chloroformates, see:
(a) Kazuo, O.; Yoh-ichi, M.; Ichiro, Y.; Manabu, M.; Jiro, J.; Toshiyuki, T.; Tetsuzi,
A. Chem. Pharm. Bull. 1991, 39, 911; (b) Ratz, A. M.; Weigel, L. O. Tetrahedron
Lett. 1999, 2239–2242; (c) Robertson, D. W.; Wong, D. T.; Krushinsky, J. R.;
Joseph, H. U.S. patent 5,023,269, 1991.; (d) Floyd, D. M.; Kimball, S. D.; Krapcho,
J., et al J. Med. Chem. 1992, 35, 756; (e) Peat, A. J.; Buckhwald, S. L. J. Am. Chem.
Soc. 1996, 118, 1028.
In sum, a short four-step methodology for rapid generation of
endoxifen/(E)-isomer mixture, obtained in 34% overall yield, was