A multi-gram-scale stereoselective synthesis of Z-endoxifen

Article abstract of DOI:10.1016/j.bmcl.2018.03.008

Z-Endoxifen is widely regarded as the most active metabolite of tamoxifen, and has recently demonstrated a 26.3% clinical benefit in a phase I clinical trial to treat metastatic breast cancer after the failure of standard endocrine therapy. Future pharmac

Full text of DOI:10.1016/j.bmcl.2018.03.008

Bioorganic & Medicinal Chemistry Letters xxx (2018) xxx–xxx
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
A multi-gram-scale stereoselective synthesis of Z-endoxifen
b
a
c
c,d
Lech-Gustav Milroy a, , Bartjan Koning , Daphne S.V. Scheppingen , Nynke G.L. Jager , Jos H. Beijnen
,
⇑
Jan Koek b, Luc Brunsveld a,
⇑
a Laboratory of Chemical Biology and Institute for Complex Molecular Systems (ICMS), Department of Biomedical Engineering, Technische Universiteit Eindhoven, Den Dolech 2,
5612 AZ Eindhoven, The Netherlands
b Syncom B.V., Kadijk 3, 9747 AT Groningen, The Netherlands
c Department of Pharmacy & Pharmacology, The Netherlands Cancer Institute - Antoni van Leeuwenhoek and MC Slotervaart, Louwesweg 6, 1066 EC Amsterdam, The Netherlands
d Faculty of Science, Division of Pharmacoepidemiology and Clinical Pharmacology, Department of Pharmaceutical Sciences, Utrecht University, Universiteitsweg 99, 3584 CG
Utrecht, The Netherlands
a r t i c l e i n f o
a b s t r a c t
Article history:
Z-Endoxifen is widely regarded as the most active metabolite of tamoxifen, and has recently demon-
strated a 26.3% clinical benefit in a phase I clinical trial to treat metastatic breast cancer after the failure
of standard endocrine therapy. Future pharmacological and pre-clinical studies of Z-endoxifen would
benefit from reliable and efficient synthetic access to the drug. Here, we describe a short and efficient,
stereoselective synthesis of Z-endoxifen capable of delivering multi-gram (37 g) quantities of the drug
in >97% purity with a Z/E ratio >99% after trituration.
Received 6 February 2018
Revised 25 February 2018
Accepted 2 March 2018
Available online xxxx
Keywords:
Antiestrogens
Ó 2018 Published by Elsevier Ltd.
Tamoxifen analogs
Endoxifen
Nuclear receptors
Stereoselective synthesis
Z-4-hydroxy-N-desmethyltamoxifen, or Z-endoxifen,1 is an
active metabolite of the pioneering anti-breast cancer prodrug,
tamoxifen.2 During phase I metabolism, tamoxifen is oxidized by
hepatic membrane-bound cytochrome P450 enzymes (CYP) into
at least twenty-two structurally unique metabolites,3,4 among
which, Z-4-hydroxytamoxifen and Z-endoxifen (Fig. 1) display
the strongest antiestrogen effect. Z-4-hydroxytamoxifen and Z-
endoxifen are similar in terms of their biochemical and cellular
profiles.5 For instance, both compounds bind with similarly high
affinity to the ligand binding domain of the estrogen receptor
(ER) – the molecular target of ER-dependent of breast cancer –
thereby inducing a conformational change in the protein that
favors co-repressor recruitment and down regulation of gene
expression.6 However, the pharmacological profiles of both com-
pounds are quite different; in one report, the steady-state plasma
serum concentration of Z-endoxifen in humans was reported to
be fivefold higher than for Z-4-hydroxytamoxifen,7 Such findings
have led multiple research groups to investigate Z-endoxifen as
the metabolite most responsible for the therapeutically beneficial
effects of tamoxifen in patients.8,9 For example, the National
Cancer Institute (NCI) in the US is currently running a randomized
phase II clinical trial of Z-Endoxifen Hydrochloride in parallel study
with tamoxifen citrate for the treatment of advanced or metastatic,
ER+ HER2-
breast
cancer
(ClinicalTrials.gov
Identifier:
NCT02311933). Furthermore, Atossa Genetics (http://www.
atossagenetics.com/) are pursuing clinical studies of oral and topi-
cal formulations of endoxifen for the treatment of tamoxifen
refractory breast cancer. Recently, a phase I trial demonstrated a
26.3% clinical benefit for Z-endoxifen against metastatic breast
cancer after the failure of standard endocrine therapy.10
The emergence of Z-endoxifen as a candidate anti-breast cancer
drug is timely given the ongoing global challenge of tamoxifen
resistance.11 A well-documented cause of tamoxifen resistance is
the existence of >75 alleles of CYP2D6 with varying metabolizing
activity. Some ethnic cohorts carry ‘null” alleles that do not express
CYP2D6 or express inactive forms of CYP2D6,12 which has been
linked to reduced blood serum concentrations of Z-endoxifen and
potentially reduced therapeutic benefit in these patients. To com-
plicate matters, selective serotonin uptake inhibitors (SSRIs) such
as fluoxetine and paroxetine, co-prescribed with tamoxifen to alle-
viate hot flashes – a side-effect of tamoxifen therapy – are known
to potently inhibit CYP2D6 thereby lowering blood serum levels of
Z-endoxifen akin to a phenotype observed in the case of poorly
metabolizing CYP2D6 alleles. Direct administration of pure
⇑
Corresponding authors.
E-mail address: l.milroy@tue.nl (L.-G. Milroy).
https://doi.org/10.1016/j.bmcl.2018.03.008
0960-894X/Ó 2018 Published by Elsevier Ltd.
Please cite this article in press as: Milroy L.-G., et al. Bioorg. Med. Chem. Lett. (2018), https://doi.org/10.1016/j.bmcl.2018.03.008

2
L.-G. Milroy et al. / Bioorganic & Medicinal Chemistry Letters xxx (2018) xxx–xxx
Figure 1. Chemical structures of tamoxifen and metabolites Z-4 hydroxytamoxifen and Z-endoxifen, and the metabolic relationship between the three compounds.
Z-endoxifen is therefore a promising approach to overcome geno-
typical tamoxifen resistance and undesirable drug interactions.
Besides breast cancer therapy, Z-endoxifen is also being investi-
gated as PKC inhibitors for the treatment of biopolar disorder13,14
and potential bone protective properties in postmenopausal breast
cancer patients.15
of the E-isomer, which was nonetheless successfully separated by
RP-HPLC. For our purposes, we sought a synthetic entry to isomer-
ically pure Z-endoxifen, which was equally short and efficient but
not reliant on RP-HPLC, and which ideally circumvent the use of
highly reactive MeLi, given the need for gram quantities of the
drug. We were also keen to understand more about the adventi-
tious isomerization of Z-endoxifen and sought ways to overcome
this.
Collectively, a short scalable stereoselective synthesis of Z-
endoxifen would potentially be of benefit to the pharmacology
field by improving the tailoring of tamoxifen treatment,16,17 and
enabling wide-scale pharmacological profiling of the drug, ranging
from the mg-quantities needed for biochemical and cellular testing
to the grams needed for animal studies. Two independent synthe-
ses of Z-endoxifen have been reported.14,18 During one route, the
tetra-substituted olefin is generated as a 1/1 mixture of E/Z
isomers, which are separable by RP-HPLC.14 The second route is
stereoselective and founded on an earlier synthesis of Z-4-hydrox-
ytamoxifen published by Gauthier and Labrie et al. (Fig. 2). Here,
the synthesis began with mono-protection of 1 with a pivaloyl
(Piv) group to form 2, followed by an E-selective olefination reac-
tion between 2 and propiophenone under reducing metal McMurry
conditions to generate 3 in a >100/1 E/Z ratio after trituration in
MeOH. At this juncture, Fauq et al. efficiently steered the synthesis
toward Z-endoxifen via first an O-alkylation reaction between 3
and alcohol 4 under Mitsunobu conditions to generate the pro-
tected precursor of Z-endoxifen, 5, importantly with retention of
stereopurity (Fig. 2). The synthesis of Z-endoxifen concluded with
a one-pot deprotection of the pivaloyl and ethyl carbamate pro-
tecting groups using an excess of MeLi at ─78 °C. Some stereoran-
domization of Z-endoxifen was reported to occur during these final
deprotection and isolation steps resulting in accumulation of 30%
Our synthesis efforts began by following the synthetic route to
precursor 5 reported by Fauq et al. (Fig. 2). When we subjected 5 to
the reported MeLi-mediated deprotection conditions, we did not
detect any significant loss of stereopurity of the crude Z-endoxifen
immediately after work-up and prior to silica gel column
chromatography, as judged by 1H NMR in d6-DMSO (circ. 95/5 Z/
E; Figure S3, top panel). However, when we then purified the crude
by silica gel column chromatography using 15% MeOH in CH2Cl2 –
as described in the supporting information provided by Fauq et al.
– and remeasured the 1H NMR in d6-DMSO, we did observe some
stereorandomization, specifically 22% formation of the E-isomer as
judged by 1H NMR (Figure S2 and S3, bottom panel) and HPLC anal-
ysis (Figure S5, left panel). Our finding suggested that Z-endoxifen
is unstable to the acidic conditions of the silica gel column chro-
matography but, in our hands, stable to the basic cleavage and
mildly acidic work-up conditions. This finding is supported by an
earlier stereoselective synthesis of Z-4-hydroxytamoxifen,19 in
which Gauthier and Labrie use similar basic reaction conditions
(MeLi) to cleave a pivaloyl protecting group from Z-4-hydroxyta-
moxifen. Here, a 66% yield and Z/E-product ratio of >100:1 is
reported after near identical work-up conditions used for this
present synthesis of Z-endoxifen i.e. quench with saturated
Figure 2. Overview of Gauthier and Labrie’s stereoselective synthesis of Z-4-hydroxytamoxifen & Fauq’s stereoselective synthesis of Z-endoxifen.
Please cite this article in press as: Milroy L.-G., et al. Bioorg. Med. Chem. Lett. (2018), https://doi.org/10.1016/j.bmcl.2018.03.008

L.-G. Milroy et al. / Bioorganic & Medicinal Chemistry Letters xxx (2018) xxx–xxx
3
ammonium chloride at ─78 °C, warm to rt, extract aqueous layer
multiple times with EtOAc. Therefore, we were pleased to find that
purification of the crude, stereopure Z-endoxifen via column chro-
matography using neutral alumina resulted in isolation of Z-
endoxifen with a Z/E ratio of 95/5 (1H NMR, Figure S4) or 96/4
(HPLC, Figure S5, right panel). While motivated by this initial find-
ing, we felt that the synthetic route required further optimization
to enable a multi-gram synthesis of Z-endoxifen e.g. avoiding the
use of excess MeLi at the last deprotection step.
(Figure S6–S8). As expected, the Mitsunobu-coupling step for the
formation of 6 was challenging to achieve at full conversion. (Fauq
et al.) Ultimately, the slow addition of multiple equivalents of the
reagents (DIAD and alcohol 7) did lead to full conversion, and after
purification on a short path of neutral alumina the desired
compound 6 could be isolated as a single isomer in 83% yield with
good purity (Figure S10–S12). In this case, the isolated crude 6 was
directly purified by column chromatography because some stereo-
randomization was detectable after aqueous work up. Interest-
ingly, the 2,2,2-trifluoro-N-methylacetamide group present in
compounds 6 and 7 exist as a 2.6:1 rotameric mixture at the amide
bond in CDCl3 as solvent. Evidence for this can be seen in the sin-
glet peaks at 3.11 ppm (minor rotamer) and 3.22 ppm (major rota-
mer) in the 1D 1H NMR spectrum (Figure S10), corresponding to
the N-methyl protons, and distinct resonance peaks at -68.3 ppm
(minor) and -69.9 ppm (major) in the 1D 19F NMR spectrum
(Figure S12), corresponding to the two rotameric forms of the
trifluoroacetyl group. Both sets of signals gave integral values in
the ratio 2.6:1. Further evidence for the existence of two rotameric
forms of the 2,2,2-trifluoro-N-methylacetamide could also be
found in the 1D 13C NMR spectrum (Figure S11), specifically two
A summary of the newly developed optimized route to pure Z-
endoxifen, full details of which can be found in the supporting
information, is shown in Fig. 3. The literature procedure described
by Gauthier et al. for the preparation of compound 2 uses NaH as
base with one equivalent of PivCl. Since a homogeneous mixture
is not formed upon deprotonation, a statistical mixture of products
is not obtained. After purification of this mixture of the starting
material, mono- and bis Piv-adduct, the desired product was iso-
lated in only a 27% yield. In our hands it was more efficient to
bis-protect the dihydroxybenzophenone first with PivCl and subse-
quently deprotect one Piv-group with one equivalent of LiOH. In
this way, a mixture of starting material, mono Piv-adduct and bis
Piv-adduct is obtained with a ratio of 7:85:7. Unfortunately upon
purification of the material on silica a part of the desired product
is deprotected further leading to a lower than expected yield
(47%) – Figure S6. While not investigated on this occasion, recrys-
tallization may prove to be a more efficient technique to isolate 2
on a large scale, given the high conversion to the mono-Piv adduct
in the crude material. The McMurry coupling was conducted as
described in literature,19 though some changes concerning the
workup on a large scale were made. To obtain very isomerically
pure compound 5, a substantial amount of material was sacrificed
during trituration (Figure S9). The biggest improvement for obtain-
ing highly pure Z-endoxifen in the final stage was achieved
through use of a different protecting group for the hydroxylamine
tail. We anticipated that a trifluoroacetyl protection of the nitrogen
would be adequate for the Mitsunobu reaction, and could be
deprotected along with the Piv-group under mild conditions at
the same stage. Hydroxylamine 7 was therefore prepared through
treatment of 2-(methylamino)ethanol with ethyl trifluoroacetate
2
sets of overlapping quartets – 116.44 (quartet, JC-F = 285 Hz –
2
minor rotamer), 116.41 (quartet, JC-F = 285 Hz – major rotamer)
3
and 157,40 (quartet, JC-F = 35.3 Hz – minor rotamer), 157,55
3
(quartet, JC-F = 35.3 Hz – major rotamer), the first set logically
assigned to -C(O)CF3 of each rotamer, the second set to -C(O)CF3.
The use of the trifluoroacetyl protecting group meant that both
protecting groups present on 6 could be simultaneously cleaved
on a 57 g scale by treatment with methanolic LiOH, thus improving
the safety and practicability of the large scale synthesis. We did not
observe any formation of the E-isomer under these basic reaction
conditions and the product was stable during aqueous workup.
However, over time, we did observe partial isomerization of 6
and Z-endoxifen in the acidic CDCl3. Finally, any trace amounts of
the E-isomer were removed by trituration of the crude material,
to afford 37 g of Z-endoxifen (83% yield) with purity >97% (two
separate qNMR experiments) and >99/1 Z/E (1D 1H NMR) – Fig. 4.
It is worth emphasizing that a column chromatography step was
not needed to deliver the final pure compound. The overall yield
Figure 3. Overview of large-scale synthesis of Z-endoxifen.
Please cite this article in press as: Milroy L.-G., et al. Bioorg. Med. Chem. Lett. (2018), https://doi.org/10.1016/j.bmcl.2018.03.008

4
L.-G. Milroy et al. / Bioorganic & Medicinal Chemistry Letters xxx (2018) xxx–xxx
c
a
e
d
b
Figure 4. Molecular characterization of Z-endoxifen: 1D 1H NMR (top), 1H NMR (300 MHz, CDCl3): d 0.91 (t, J = 6 Hz, 3H), 2.48 (q, J = 6 Hz, 2H), 2.50 (s, 3H), 2.93 (t, J = 6 Hz,
2H), 3.96 (t, J = 6 Hz, 2H), 6.49 (d, J = 6 Hz, 2H), 6.71–6.75 (m, 4H), 7.02 (d, J = 9 Hz, 2H), 7.08–7.18 (m, 5H).; and analytical LC (bottom), Peak #1 = unknown impurity, Peak #2
= Z-endoxifen, Peak #3 = unknown impurity.
of the synthesis was thus increased from 3% to 18% judging on
Gauthier, Labrie and Fauq’s combined synthetic route. Since the
route still contains two steps with moderate yield, further opti-
mization of the synthetic route is recommended.
4-hydroxytamoxifen, and Dr. Abdul H. Fauq for his previous work
on the stereoselective synthesis of Z-endoxifen. The authors also
wish to thank the Dutch Ministry of Education, Culture, and
Science (Gravity Program 024.001.035) for funding.
In conclusion, a short stereoselective synthesis of Z-endoxifen
has been achieved, which delivers the drug in >97% purity and
>99/1 Z/E isomeric ratio in a multi-gram-scale quantity, as demon-
strated by the 37 g scale synthesis described herein. The key con-
tribution of our work is that Z-endoxifen can be deprotected
under basic conditions without significant loss of stereopurity –
acidic silica gel chromatography should be avoided – and purified
by straightforward trituration instead of reverse-phase HPLC at the
last step. Building on the previous important contributions by Gau-
thier, Labrie and Fauq our hope is that this current streamlined
synthesis of Z-endoxifen broadens the accessibility of this clinically
important drug to many research groups.
A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at https://doi.org/10.1016/j.bmcl.2018.03.008.
These data include MOL files and InChiKeys of the most important
compounds described in this article.
References
1. Ahmad A, Ali SM, Ahmad MU, Sheikh S, Ahmad I. Breast Cancer Res Treat.
2010;122:579–584.
2. Jordan VC, McDaniel R, Fadeke Agboke, Maximov PY. Steroids. 2014;90:3–12.
3. Mürdter T, Schroth W, Bacchus-Gerybadze L, et al. Clin Pharmacol Ther.
2011;89:708–717.
4. Teunissen SF, Rosing H, Seoane MD, et al.
2011;55:518–526.
5. Johnson MD, Zuo H, Lee K-H, et al. Breast Cancer Res Treat. 2004;85:151–159.
6. Huang P, Chandra V, Rastinejad F. Annu Rev Physiol. 2010;72:247–272.
Acknowledgments
J Pharm Biomed Anal.
The authors wish to thank Dr. Sylvain Gauthier and Dr. Fernand
Labrie for their earlier work on the stereoselective synthesis of Z-
Please cite this article in press as: Milroy L.-G., et al. Bioorg. Med. Chem. Lett. (2018), https://doi.org/10.1016/j.bmcl.2018.03.008

L.-G. Milroy et al. / Bioorganic & Medicinal Chemistry Letters xxx (2018) xxx–xxx
5
7. Jager NGL, Rosing H, Schellens JHM, Linn SC, Beijnen JH. Breast Cancer Res Treat.
2014;143:477–483.
8. Reid JM, Goetz MP, Buhrow SA, et al. Cancer Chemother Pharmacol.
2014;74:1271–1278.
9. de Schultink AHM, Alexi V, Werkhoven X, et al. Breast Cancer Res Treat.
2017;161:567–574.
13. Ahmad A, Sheikh S, Shah T, et al. Clin Transl Sci. 2016;9:252–259.
14. Ali SM, Ahmad A, Shahabuddin S, et al. Bioorg Med Chem Lett.
2010;20:2665–2667.
15. Gingery A, Subramaniam M, Pitel KS, et al. PLoS ONE. 2014;9:e98219.
16. Jager NGL, Linn SC, Schellens JHM, Beijnen JH. Clin Breast Cancer.
2015;15:241–244.
10. Goetz MP, Suman VJ, Reid JM, et al. J Clin Oncol. 2017;35:3391–3400.
11. Rondón-Lagos M, Villegas VE, Rangel N, Sánchez MC, Zaphiropoulos PG. Int J
Mol Sci. 2016;17:1357.
17. Krou S de, Rosing H, Nuijen B, Schellens JHM, Beijnen JH. Ther Drug Monit.
2017;39:132–137.
18. Fauq AH, Maharvi GM, Sinha D. Bioorg Med Chem Lett. 2010;20:3036–3038.
19. Gauthier S, Mailhot J, Labrie F. J Org Chem. 1996;61:3890–3893.
12. Hoskins JM, Carey LA, McLeod HL. Nat Rev Cancer. 2009;9:576.
Please cite this article in press as: Milroy L.-G., et al. Bioorg. Med. Chem. Lett. (2018), https://doi.org/10.1016/j.bmcl.2018.03.008