Estrogen receptor ligands. Part 11: Synthesis and activity of isochromans and isothiochromans

Article abstract of DOI:10.1016/j.bmcl.2004.11.018

The ring oxygen and sulfur analogs of lasofoxifene, 1a and 1b, were synthesized in an attempt to impart ERα selectivity, as found in the closely related dihydrobenzoxathiin compound I, recently discovered in these laboratories. The resulting isochroman an

Full text of DOI:10.1016/j.bmcl.2004.11.018

Bioorganic & Medicinal Chemistry Letters 15 (2005) 715–718
Estrogen receptor ligands. Part 11: Synthesis and activity of
isochromans and isothiochromans
Jian Liu,a,* Elizabeth T. Birzin,b Wanda Chan,b Yi Tien Yang,b Lee-Yuh Pai,b
Carolyn DaSilva,b Edward C. Hayes,b Ralph T. Mosley,a Frank DiNinno,a
Susan P. Rohrer,b James M. Schaefferb and Milton L. Hammonda
aDepartment of Medicinal Chemistry, Merck Research Laboratories, PO Box 2000, Rahway, NJ 07065, USA
bDepartment of Atherosclerosis and Endocrinology, Merck Research Laboratories, PO Box 2000, Rahway, NJ 07065, USA
Received 22 September 2004; revised 5 November 2004; accepted 5 November 2004
Available online 26 November 2004
Abstract—The ring oxygen and sulfur analogs of lasofoxifene, 1a and 1b, were synthesized in an attempt to impart ERa selectivity,
as found in the closely related dihydrobenzoxathiin compound I, recently discovered in these laboratories. The resulting isochroman
and isothiochroman compounds were found to exhibit equipotent binding affinities to the ER isoforms and were less active in the
inhibition of estradiol-triggered uterine growth when compared to I and lasofoxifene.
Ó 2004 Elsevier Ltd. All rights reserved.
Endogenous estrogen interacts with the estrogen recep-
tors, the nuclear hormone receptors that mediate effects
on bone mineral density and regulate the female repro-
ductive, cardiovascular, and central nervous systems.
The selective estrogen receptor modulators (SERMs)1
are compounds with mixed agonist/antagonist proper-
ties, exemplified by the clinically approved tamoxifen,2
for the treatment of breast cancer, and raloxifene,3 for
the treatment and prevention of osteoporosis. Since
the discovery of the estrogen receptor subtype b
(ERb),4 the search for subtype selective SERMs, in par-
ticular ERa selective SERMs or SERAMs (selective
estrogen receptor alpha modulators),5 has attracted
attention in an attempt to achieve an advantage over
nonsubtype selective SERMs. We have previously re-
ported on a new series of dihydrobenzoxathiin-based
SERAMs, exemplified by I (Fig. 1), and we postulated
that the interaction between the bulky sulfur in the ring
and the two discriminating residues in the binding
pocket of the two receptor isoforms (Leu384 for ERa,
Met354 for ERb) was responsible for the observed selec-
Figure 1.
tivity.5b In support of the hypothesis, replacement of the
sulfur atom by the smaller carbon atom resulted in a
Lasofoxifene6 is a potent, nonsubtype selective SERM
based on a tetrahydronaphthalene core structure devel-
oped by Pfizer, which has advanced to late stage clinical
complete loss of a-selectivity.
trials for osteoporosis. Herein, we report our effort to
apply the rationalization for the selectivity in the dihy-
drobenzoxathiin series to the tetrahydronaphthalene
Keywords: Estrogen receptors; Isochroman; Isothiochroman; SERMs.
*
Corresponding author. Tel.: +1 732 594 9600; fax: +1 732 594
3007; e-mail: jian_liu@merck.com
0960-894X/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.11.018

716
J. Liu et al. / Bioorg. Med. Chem. Lett. 15 (2005) 715–718
benzyloxy benzaldehyde provided the corresponding
benzyl alcohols. In turn, TPAP/NMO oxidation af-
forded ketones 3a and 3b, which upon treatment with
Tebbeꢀs reagent and subsequent hydroboration yielded
alcohols 4a and 4b. Then 4a and 4b were converted to
ketones 5a and 5b by sequential oxidation with Dess–
Martin reagent, addition of phenyl lithium and re-
oxidation.
The cyclization conditions to form the iso- and isothio-
chroman rings were a little different. The preparation of
the isochroman 1a is depicted in Scheme 2. The ketone
5a was treated with TFA in methylene chloride to re-
move the MOM protection, followed by addition of tri-
ethylsilane to affect the dehydrative reduction to form
exclusively the cis-isochroman 6a. The cis-configuration
was confirmed by 1H NMR as judged by the presence of
a small coupling constant (J = 2.5Hz @ d = 5.03 and
4.02ppm) between the C-3 and C-4 protons of the pyran
ring. After removal of the benzyl protecting group by
catalytic hydrogenation, the pyrrolidinoethyloxy side
chain was installed utilizing a Mitsunobu reaction to
form 7a. Unfortunately, no suitable method was found
to convert the methylether to the requisite phenol, as
all conditions concomitantly destroyed the pyran ring.10
Instead, TIPS was used as the protection for the phenol
from the beginning of the synthesis leading to 70a, where
the TIPS was readily removed by TBAF to provide 1a.11
Figure 2. Molecular model of hERa selective I (white) versus model of
1b (pink). HERa is depicted in purple and hERb in green. Residue
numbering is hERa unless otherwise indicated. Arcs represent desired
clash between S and Met354 (hERb) sidechain to achieve selectivity.
skeleton by conversion into isochroman 1a and isothio-
chroman 1b. Based on visual inspection of the molecu-
lar models7 (shown in Fig. 2), it was hoped that either
oxygen or sulfur substitution in the ring might offer
the same discrimination between the receptor subtypes
as is seen in compound I.
In contrast to the reaction of 5a, treatment of 5b with
TFA quickly led to the analogous isothiocoumarin ring
containing a tetrasubstituted double bond, which re-
sisted reduction by either triethyl silane or catalytic
hydrogenation. Hence, as shown in Scheme 3, the
MOM protection in 5b was readily removed by exposure
to silver nitrate to generate the thiol. Subsequent treat-
ment with triethyl silane and boron trifluoride–diethyl
etherate successfully formed only the cis-isothiochro-
man 6b. After debenzylation with boron tribromide at
The method reported by Kim et al.8 was utilized for the
construction of the cis-diaryl iso- and isothiochromans.
Therefore, we started to build ketones 5a and 5b as pre-
cursors for the cyclization–dehydrative–reduction proc-
ess (Scheme 1). Lithiation of the aryl bromides 2a and
2b9 with n-butyl lithium followed by reaction with 4-
Scheme 1. Reagents and conditions: (a) n-butyl lithium, THF, ꢀ78°C,
˚
85%. (b) TPAP, NMO, MS 4A, CH2Cl2, 95%. (c) Tebbe reagent,
THF, 85%. (d) BH3, THF, then H2O2, NaOH, 90%. (e) Dess–Martin
Periodinane, pyridine, CH2Cl2, 96%. (f) Phenyl lithium, THF, 60 %.
(g) Dess–Martin periodinane, pyridine, CH2Cl2, 85%.
Scheme 2. Synthesis of 1a. Reagents and conditions: (a) TFA, CH2Cl2;
(b) Et3SiH, 60%. (c) H2, Pd black, EtOH, 98%. (d) N-(2-hydroxyethyl)
pyrrolidine, PPh3, DIAD, THF, 75%. (e) TBAF, THF, 80%.

J. Liu et al. / Bioorg. Med. Chem. Lett. 15 (2005) 715–718
717
idues (depicted in Fig. 2) can have a significant impact
on selectivity. It is interesting to contrast these results
with those reported16 for the corresponding substituted
nitrogen (tetrahydro-isoquinoline) derivatives lacking
the C-3 phenyl group, in which a modest preference
for ERb was observed amongst substantially less potent
ligands. In spite of the potent binding affinity and potent
antagonism of the estradiol dependent growth of MCF-
7 cells exhibited by both 1a17 and 1b, the uterine weight
assay response contrasted that reported,6c and con-
firmed in our assay, for lasofoxifene. Both 1a and 1b
exhibited substantially less activity in the inhibition of
the estradiol-triggered uterine growth (antagonism)
while effecting significantly more uterine growth
(agonism).
In conclusion, we have synthesized the iso- and isothio-
chroman analogs related to lasofoxifene and demon-
strated that the compounds had very high potency in
both an estrogen receptor binding assay and an MCF-
7 inhibition assay. In addition, we demonstrated that
in spite of these similarities to lasofoxifene, the effects
on immature rat uterine tissue were strikingly different.
The conversion of lasofoxifene into an SERAM using
the rationalization for selectivity postulated in the dihy-
drobenzoxathiin series was unsuccessful.
Scheme 3. Synthesis of 1b. Reagents and conditions: (a) AgNO3,
EtOH, 80%. (b) Et3SiH, CH2Cl2, BF3ÆEt2O, 0°C, 60%. (c) BBr3,
CH2Cl2, ꢀ78°C, 80%. (d) N-(2-hydroxyethyl)pyrrolidine, PPh3,
DIAD, THF, 75%. (e) BBr3, CH2Cl2, ꢀ10°C, 55%.
Table 1. Binding affinities (IC50) and in vivo data for compounds 7a,b,
and 1a,b
Compdsa
Binding affinity13 MCF-714
Uterine weight15
References and notes
IC50 (nM)
human
Inhibition % Inhibition/%
control
IC50
(nM)
1. Recent reviews on SERMs: (a) Jordan, V. C. J. Med.
Chem. 2003, 46, 883; (b) Jordan, V. C. J. Med. Chem.
2003, 46, 1081; (c) Meegan, M. J.; Lloyd, D. G. Curr.
Med. Chem. 2003, 10, 181; (d) Miller, C. P. Curr. Pharm.
Des. 2002, 8, 2089; (e) Bryant, H. U. Endocrinol. Metab.
Dis. 2002, 3, 231.
2. (a) Furr, B. J.; Jordan, V. C. Pharmacol. Theraopeut. 1984,
25, 127; (b) Jaiyesimi, I. A.; Buzdar, A. U.; Decker, D. A.;
Hortobagy, G. N. J. Clin. Oncol. 1995, 13, 513; (c)
Osborne, C. K. New. Engl. J. Med. 1998, 339, 1609; (d)
Jordan, V. C. Br. J. Pharmacol. 1993, 3, 495.
3. Grese, T. A.; Pennington, L. D.; Sluka, J. P.; Adrian, M.
D.; Cole, H. W.; Fuson, T. R.; Magee, D. E.; Phillips, D.
L.; Rowley, E. R.; Shetler, P. K.; Short, L. L.; Venugop-
alan, M.; Yang, N. N.; Sato, M.; Glasebrook, A. L.;
Bryant, H. U. J. Med. Chem. 1998, 41, 1272.
ERa/ERb
(fold selective
for ERa)
(antagonism/
agonism)
7a
7b
39/315 (8)
13/110 (8)
2.4/3.6 (2)
0.6/2.1 (4)
0.8/45 (56)
57
164
ND
ND
1a
1b
0.15
24/43
47/49
99/9
0.25
3.0
0.10c
(+) I
Lasofoxifene 1.3/1.8 (1.4)b
74/26
—/100
Estradiol
1.3/1.1
a The isochromans and isothiochromans were a pair of racemic com-
pounds characterized by spectroscopy.
b Data cited from Ref. 16 is 0.5/1.2 (2.4).
c Data cited from Ref. 6a is 0.05.
4. (a) Kuiper, G. C.; Carlsson, B.; Enmark, E.; Pelto-
Huikko, M.; Nilsson, S.; Gustafsson, J. A. Proc. Natl.
Acad. Sci. U.S.A. 1996, 93, 5925; (b) Mosselmann, S.;
Polman, J.; Dijkema, R. FEBS Lett. 1996, 392, 49.
5. We have reported a series of papers on SERAMs: (a)
Chen, H. Y.; Dykstra, K. D.; Birzin, E. T.; Frisch, K.;
Chan, W.; Yang, Y. T.; Mosley, R. T.; DiNinno, F.;
Rohrer, S. P.; Schaeffer, J. M.; Hammond, M. L. Bioorg.
Med. Chem. Lett. 2004, 14, 1417; (b) Kim, S.; Wu, J. Y.;
Birzin, E. T.; Frisch, K.; Chan, W.; Pai, L. Y.; Yang, Y.
T.; Mosley, R. T.; Fitzgerald, P. M. D.; Sharma, N.;
Dahllund, J.; Thorsell, A. G.; DiNinno, F.; Rohrer, S. P.;
Schaeffer, J. M.; Hammond, M. L. J. Med. Chem. 2004,
47, 2171; (c) Chen, H. Y.; Kirn, S.; Wu, J. Y.; Birzin, E. T.;
Chan, W.; Yang, Y. T.; Dahllund, J.; DiNinno, F.;
Rohrer, S. P.; Schaeffer, J. M.; Hammond, M. L. Bioorg.
Med. Chem. Lett. 2004, 14, 2551; (d) Kim, S.; Wu, J.;
Chen, H. Y.; Birzin, E. T.; Chan, W. D.; Yang, Y. T.;
Colwell, L.; Li, S.; Dahllund, J.; DiNinno, F.; Rohrer, S.
P.; Schaeffer, J. M.; Hammond, M. L. Bioorg. Med. Chem.
ꢀ78°C,12 the pyrrolidinoethyloxy side chain was ap-
pended to form 3b. Again in contrast to isochroman
7a, treatment of 7b with boron tribromide at ꢀ10°C
successfully removed the methyl protection on the phe-
nol without degrading the isothiochroman ring to pro-
vide 1b.11
As indicated in Table 1, four compounds were evaluated
in the estrogen receptor binding assay, the MCF-7 cell
proliferation assay and the immature rat uterine weight
assay. Consistent with prior observations, methylethers
7a,b exhibited much lower binding affinity than the phe-
nolic counterparts 1a and 1b, the latter of which rivaled
the potency of lasofoxifene. While in agreement with the
calculated energetics,7 the lack of desired selectivity sug-
gests that even small differences in the position of the
sulfur in compound I and 1b relative to the receptor res-

718
J. Liu et al. / Bioorg. Med. Chem. Lett. 15 (2005) 715–718
Lett. 2004, 14, 2741; (e) Tan, Q.; Birzin, E. T.; Chan, W.; All of the conditions led to the degradation of the
Yang, Y. T.; Pai, L. Y.; Hayes, E. C.; DaSilva, C. A.;
DiNinno, F.; Rohrer, S. P.; Schaeffer, J. M.; Hammond,
M. L. Bioorg. Med. Chem. Lett. 2004, 14, 3747; (f) Tan,
Q.; Birzin, E. T.; Chan, W.; Yang, Y. T.; Pai, L. Y.;
Hayes, E. C.; DaSilva, C. A.; DiNinno, F.; Rohrer, S. P.;
Schaeffer, J. M.; Hammond, M. L. Bioorg. Med. Chem.
Lett. 2004, 14, 3753; (g) Blizzard, T. A.; DiNinno, F.;
Morgan, J. D.; Wu, J. Y.; Chen, H. Y.; Kim, S.; Chan, W.;
Birzin, E. T.; Yang, Y. T.; Pai, L.-Y.; Zhang, Z.; Hayes, E.
C.; DaSilva, C. A.; Tang, W.; Rohrer, S. P.; Schaeffer, J.
M.; Hammond, M. L. Bioorg. Med. Chem. Lett. 2004, 14,
3865; (h) Blizzard, T. A.; DiNinno, F.; Morgan, J. D.; Wu,
J. Y.; Gude, C.; Kim, S.; Chan, W.; Birzin, E. T.; Yang, Y.
T.; Pai, L.-Y.; Zhang, Z.; Hayes, E. C.; DaSilva, C. A.;
Tang, W.; Rohrer, S. P.; Schaeffer, J. M.; Hammond, M.
L. Bioorg. Med. Chem. Lett. 2004, 14, 3861.
benzopyran ring.
11. Owing to the lack of selectivity in binding to the estrogen
receptor isoforms described in Table 1, the chiral separa-
tion of 1a and 1b was not undertaken.
12. The use of catalytic hydrogenation with a palladium
catalyst to remove the benzyl protecting group was
unsuccessful, presumably owing to the poisoning of the
catalyst by sulfur.
13. The single IC50 values were generated in an estrogen
receptor ligand binding assay. This scintillation proximity
assay was conducted in NEN basic flashplates using
tritiated estradiol and full length recombinant human ERa
and ERb proteins with a 3h incubation time. This assay
provides IC50 values that are reproducible to within a
factor of 2–3.
14. In the MCF-7 proliferation assay, estrogen depleted
MCF-7 cells were plated into 96-well cell culture plates
at a density of 1000cells/well. To determine the antagonist
activity of a compound, the test compounds and 3pmol
estradiol were applied to the cells on days 1 and 4. The
assay was terminated between days 8 and 10 and the
cellular protein content/well used to determine the IC50.
15. Twenty-day old intact Sprague–Dawley rats were treated
(sc) with the tested compounds for 3 days at 1mpk. The
anti-estrogenic activity of compounds was determined by
co-administration of the compound with a subcutaneous
injection of 17-b-estradiol one hour after compound at
4lg/kg dose and reported as % inhibition. The estrogenic
activity (partial agonism) of the compounds was deter-
mined by administering the test compound without
estradiol and reported as % control.
6. (a) Rosati, R. L.; Da Silva-Jardine, P.; Cameron, K. O.;
Thompson, D. D.; Ke, H. Z.; Toler, S.; Brown, T. A.; Pan,
L. C.; Ebbinghaus, C. F.; Reinhold, A. R.; Elliot, N. C.;
Newhouse, B. N.; Tjoa, C. M.; Sweetnam, P. M.; Cole, M.
J.; Arriola, M. W.; Gauthier, J. W.; Crawford, D. T.;
Nickerson, D. F.; Pirie, C. M.; Qi, H.; Simmons, H. A.;
Tkalcevic, G. T. J. Med. Chem. 1998, 41, 2928; (b)
MacLean, D. B.; Thompson, D. D. Eur. Patent 0792641;
(c) Ke, H. Z.; Paralkar, V. M.; Grasser, W. A.; Crawford,
D. T.; Qi, H.; Simmons, H. A.; Pirie, C. M.; Chidsey-
Frink, K. L.; Owen, T. A.; Smock, S. L.; Chen, H. K.; Jee,
W. S. S.; Cameron, K. O.; Rosati, R. L.; Brown, T. A.;
Dasilva-Jardine, P.; Thompson, D. D. Endocrinology
1998, 139, 2068.
7. In work to be published elsewhere, the interaction energies
for lasofoxifene, 1a and 1b with the two receptor subtypes
fall within 1kcal/mol of each other compared within each
subtypes. Despite the visual similarity of placement of the
S in compound I and compound 1b, these results suggest
that the selectivity profile of 1a and 1b should be similar to
that of lasofoxifene.
16. Chesworth, R.; Zawistoski, M. P.; Lefker, B. A.; Cam-
eron, K. O.; Day, R. F.; Mangano, F. M.; Rosati, R. L.;
Colella, S.; Petersen, D. N.; Brajlt, A.; Lu, B.; Pan, L. C.;
Perry, P.; Ng, O.; Castleberry, T. A.; Owen, T. A.; Brown,
T. A.; Thompson, D. D.; DaSilva-Jardine, P. Bioorg. Med.
Chem. Lett. 2004, 14, 2729.
8. Kim, S.; Wu, J. Y.; Chen, H. Y.; DiNinno, F. Org. Lett.
2003, 5, 685.
9. Compound 2a was prepared from commercially available
methyl 2-bromo-5-methoxybenzoate in two steps, while 2b
was synthesized from 2-bromo-5-methoxytoluene in three
steps.
17. Bury, P. S.; Christiansen, L. B.; Jacobsen, P.; Jorgensen,
A. S.; Kanstrup, A. K.; Naerum, L.; Bain, S.; Fledelius,
C.; Gissel, B.; Hansen, B. S.; Korsgaard, N.; Thorpe, S.
M.; Wassermann, K. Bioorg. Med. Chem. 2002, 10, 125,
describes the C-1 oxygen analog of lasofoxifene, NNC 45-
0781, which appears to possess a very similar SERM
profile of activity as lasofoxifene, however, the receptor
affinity for ERb was not measured.
10. Several conditions were tried: (a) BBr3, CH2Cl2, ꢀ78 to
ꢀ20°C; (b) AlCl3, 3h, 0°C; (c) LiI, Collidine, reflux, 10h.