Synthesis, antiproliferative, and pharmacokinetic properties of 3- and 17-double-modified analogs of 2-methoxyestradiol

Article abstract of DOI:10.1016/j.bmcl.2009.08.020

The syntheses of 21 analogs of 2-methoxyestradiol are presented, including ENMD-1198 which was selected for advancement into Phase 1 clinical trials in oncology. These analogs were evaluated for antiproliferative activity using breast tumor MDA-MB-231 cells, for antiangiogenic activity in HUVEC proliferation assays, and for estrogenic activity in MCF-7 cell proliferation. The most active analogs were evaluated for iv and oral pharmacokinetic properties via cassette dosing in rat and in mice pharmacokinetic models.

Full text of DOI:10.1016/j.bmcl.2009.08.020

Bioorganic & Medicinal Chemistry Letters 19 (2009) 6241–6244
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Synthesis, antiproliferative, and pharmacokinetic properties of 3-
and 17-double-modified analogs of 2-methoxyestradiol
Gregory E. Agoston , Jamshed H. Shah, Lita Suwandi, Arthur D. Hanson, Xiaoguo Zhan, Theresa M. LaVallee,
*
Victor Pribluda, Anthony M. Treston
EntreMed, Inc., 9640 Medical Center Drive, Rockville, MD 20850, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
The syntheses of 21 analogs of 2-methoxyestradiol are presented, including ENMD-1198 which was
selected for advancement into Phase 1 clinical trials in oncology. These analogs were evaluated for anti-
proliferative activity using breast tumor MDA-MB-231 cells, for antiangiogenic activity in HUVEC prolif-
eration assays, and for estrogenic activity in MCF-7 cell proliferation. The most active analogs were
evaluated for iv and oral pharmacokinetic properties via cassette dosing in rat and in mice pharmacoki-
netic models.
Received 25 June 2009
Revised 4 August 2009
Accepted 5 August 2009
Available online 8 August 2009
Keywords:
Antitumor
Ó 2009 Elsevier Ltd. All rights reserved.
Antiproliferative
2-Methoxyestradiol analogs
2-Methoxyestradiol (2ME2, PanzemÒ) was originally identified
as an endogenous metabolite of estradiol, and has antiproliferative
and antiangiogenic activity when dosed at pharmacological levels
in in vitro and in vivo oncology models.1 The primary mechanisms
of action for 2ME2 are thought to be inhibition of tubulin polymer-
ization by binding to the colchicine binding site on tubulin and
selected were based on our previous work from a series of analogs
modified individually at the 3 or 17 positions of 2ME2.5,6 The novel
double-modified analogs were evaluated in vitro for inhibition of
tumor cell and angiogenic proliferation, and for estrogenicity. Ana-
logs with acceptable in vitro profiles were assessed for pharmaco-
kinetic (PK) properties by cassette dosing in rats or in CD1 mice.
These studies ultimately lead to selection of several lead candi-
dates for pharmacological evaluation, including ENMD-1198,
which is currently in Phase 1 oncology clinical trials (Fig. 1).7
The synthetic strategy used in this study is shown in Schemes 1
and 2. 2-Methoxyestrone (1)8 was modified at position 17 (Scheme
1) by Wittig coupling to give 2 (2 a R1 = H 75% yield, 2b R1 = Me
84% yield).9 The methylene olefin can be reduced to the 17-b-
methyl isomer (3, quantitative) by catalytic hydrogenation.10
Alternatively, 1 can be reduced to 17-deoxy analog (4, 90%) by a
Wolf–Kishner reduction.11 The 16,17-olefin analog 5 can be pre-
pared by Shapiro reduction of 1 (90%).12 Further modification of
2, 3, 4 or 5 at position 3 as depicted in Scheme 2 gave the dou-
ble-modified analogs used in this study.
inhibition of HIF-1a 2
. Phase 1 and Phase 2 oncology clinical trials
of 2ME2 showed that it, like other estranes, undergoes extensive
metabolism involving conjugation (glucuronidation, sulfation) at
positions 3 and 17 and oxidation to 2-methoxyestrone.3
Most prior structure–activity relationship (SAR) studies for
2ME2 analogs focused on making single point modifications to
the A, B or D rings of the steroid to increase inhibition of tubulin
polymerization or antiproliferative activity.4 Our data indicated
that the principal liability of 2ME2 was its rapid metabolism,
rather than insufficient activity per se. This suggested that analogs
designed to reduce metabolism could be better drug candidates
than analogs selected solely on the basis of in vitro tubulin or anti-
proliferative activity. Additionally, our previous SAR studies indi-
cated that single modifications at the 3, 16, or 17 positions were
not sufficient to provide the desired improvement in metabolic
profile and/or pharmacological activity.4–6
Conjugation
and
Oxidation
OH
In this Letter, we report simultaneous modification of positions
3- and 17- of 2ME2 using substituents that increase metabolic sta-
bility, and increase or maintain in vitro potency. The substituents
17
16
MeO
2ME2
HO3
* Corresponding author.
Conjugation
E-mail address: tonyt@entremed.com (A.M. Treston).
Principal author.
Figure 1. Structure of 2-methoxyestradiol showing oxidation and conjugation sites.
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.08.020

6242
G. E. Agoston et al. / Bioorg. Med. Chem. Lett. 19 (2009) 6241–6244
directly from 6 via a CO insertion reaction using HMDS as the nitro-
O
gen source to give 9 (50–75%).15 Triflate 6 was converted to amine
10 by Pd-mediated benzophenone imine formation and subse-
quent acid hydrolysis (70–77%, two steps).16 Amine 10 was con-
verted to the formamide (11, 78–88%) by in situ preparation of
formyl imidazolidine from formic acid and carbonyl diimidazole
in THF followed by amide coupling. Acetamide (12, 85%) was pre-
pared from acetic anhydride and 10 under basic conditions. Urea
13 was prepared from amine 10 using sodium cyanate under acidic
conditions (50%).17 Carbamate 14 was prepared from 10 using
methyl chloroformate with triethylamine (60%).18 Aminonitrile
15 (40%) was prepared using cyanogen bromide from 10.19 Carbon-
ate 16 was prepared directly from A using methyl chloroformate
with DMAP (60%).
The in vitro antitumor and antiangiogenic activities are
summarized in Table 1. All compounds were screened in HUVEC
proliferation assays7 as a surrogate for antiangiogenic activity, on
MDA-MB-231 tumor cells7 for antitumor activity, and for ability
to sustain proliferation of estrogen-dependant MCF-7 cells (estrog-
enicity). Desirable characteristics of the analogs were maintained
or improved antiproliferative activities (decreased IC50 values)
and decreased estrogenicity (stimulation index <1.0 at any concen-
tration compared to 2ME2).
MeO
HO
d
1
a
R1
MeO
HO
5
MeO
HO
2a R1=H
2b R1=Me
b
c
MeO
HO
R2
4
MeO
HO
3
Scheme 1. Modification of 17-position of 2ME2. (a) BuLi, (Ph)3PR, toluene
(R1 = H(2a) or Me(2b)); (b) H2, Pd/C EtOAc (R2 = CH3(3)); (c) NH2NH2, NaOH; (d)
(1) p-tosyl sulfonamide, (2) BuLi.
The 3- and 17-moieties used to design this series of doubly
modified 2ME2 analogs were chosen based on in vitro activity
R3
R3
and/or PK characteristics as individual substituents from our prior
5,6
g
studies.
In general, all activities presented in the discussion
MeO
O
MeO
HO
were compared to 2ME2 as a benchmark.
A
Doubly modified 3-amines 10a–10e showed a moderate to
significant decrease in in vitro activity. Amines 10b and 10c had
a slight decrease in antiproliferative activity but were equivalent
in HUVEC proliferative activity. Amines 10a, 10d and 10e had
significantly less activity in both antiproliferative and antiangio-
genic activities. 3-Aminonitrile 15 was slightly more active in
in vitro proliferation assays.
In the carboxamide series 9, the order of potency for the 17
substituent in MDA-MB-231 antiproliferative activity was:
17-methylene (9b) > 16,17-olefin (9d) > 17-deoxy (9c) > 17-oxo
O
MeO
16
R3
a
b
R3
MeO
MeO
d
7
6
TfO
R3
c
R3
MeO
H2N
MeO
O
10
e
R4HN
8 or 9
R3
f
Table 1
R3
In vitro activities of double-modified analogs of 2ME2
MeO
#
3-Subst.
17-Subst.
MDA-MB-231 IC50
M)
HUVEC IC50
M)
O
MeO
(
l
(l
N
R5
11, 12, 13 or 14
NC
2ME2
7
8
9a
9b
–OH
–OH
–H
–H
0.79 0.14
1.86 0.22
0.68 0.15
0.58 0.04
H
HN
–CH@CH2
–CONH(Et)
–CONH2
–CONH2
–CONH2
–CONH2
–NH2
–NH2
–NH2
–NH2
–NH2
–NHCOH
–NHCOH
–NHCOH
–NHCOH
–NHCOH
–NHCOMe
–NHCONH2
–NHCOOMe
–NHCN
15
53.01 9.67
2.20 0.43
0.12 0.06
1.86 0.86
0.20 0.1
46.91 8.15
2.11 0.57
1.94 0.27
7.38 1.66
7.31 0.58
1.22 0.60
0.30 0.08
0.56 0.02
8.29 1.64
0.61 0.00
9.25 0.66
0.71b
37.98 2.37
5.46 1.57
0.25 0.01
1.19 0.33
0.13 0.05
23.42 4.14
1.00 0.19
0.72 0.10
4.72 0.49
2.65 0.77
0.53 0.22
0.22 0.03
0.58 0.00
5.80 0.73
0.59 0.00
1.68 0.32
0.61b
Scheme 2. Modification of 3-position of 17-substituted-2-methoxyestranes. R3
substituents are shown in Table 1. (a) Tf2O, pyridine, 0 °C to rt; (b) (Bu)3SnCHCH2,
Pd(Cl)2PPh3, 2,6-di-tert-butyl methyl phenol, LiCl; (c) Pd(OAc)2, dppf, CO, KOAc,
DMSO, 55 °C then SOCl2 at reflux, then EtNH2, THF R4 = Et (8) or Pd(Cl)2, dppp, CO,
HMDS, DMF, 100 °C, MeOH, then acidic workup R4 = H (9); (d) Pd(OAc)2 rac-BINAP.
Cs2CO3, benzophenone imine, toluene, reflux, then 2 M HCl, THF; (e) formic acid,
N,N-carbonyldiimidazole, THF (R5 = H, 11) or Ac2O, 10 M NaOH 0 °C to rt (R5 = CH3,
12) or NaOCN, H2O, AcOH (R5 = NH2, 13) or methyl chloroformate, Et3N, DCM
(R5 = OCH3, 14); (f) BrCN, Et2O, DCM 0 °C to rt; (g) methyl chloroformate, DMAP,
DCM.
@O
@CH2
–H
9c
9d
–Ha
10a
10b
10c
10d
10e
11a
11b
11c
11d
11e
12
@O
@CH2
@CHCH3
–H
–Me
@O
@CH2
@CHCH3
–H
–Me
–H
3-Position modifications are shown in Scheme 2. Initially, for
analogs of general formula A (1, 2a, 2b, 3, 4 or 5 from Scheme 1)
the 3-hydroxy group was converted to the 3-triflate (6, 90%) using
pyridine and triflic anhydride. The resulting triflate can undergo
Pd-catalyzed vinylation using vinyl tributyl tin to give 7 (30%).13
Similarly, 6 can undergo Pd-mediated CO insertion to give a car-
boxylic acid14 which can be converted to a secondary amide (8,
50%, two steps). Alternatively, the primary amide was generated
13
14
15
16
@O
@O
@O
@O
5.60 1.42
0.20 0.12
24.99b
8.22 2.79
0.25 0.02
32.87 11.43
–OCOOMe
a
16, 17 Olefin.
b
Standard deviation not available where results are reported from a single
experiment.

G. E. Agoston et al. / Bioorg. Med. Chem. Lett. 19 (2009) 6241–6244
6243
Table 2
(9a) and ranged from sixfold more to threefold less potent than
In vivo rat and mouse PK summary
2ME2. For HUVEC proliferation assay the rank order was: 17-meth-
ylene (9b) > 17-deoxy (9c) > 16,17-olefin (9d) > 17-oxo (9a) and
were fivefold more to eightfold less potent compared to 2ME2. In
this series, the 17-methylene and 16,17-olefin had greatest
in vitro potency and the deoxy showed a modest drop in activity.
In the formamide series, the order of antitumor potency was:
17-methylene (11b) > 17-ethylene (11c) ꢀ 17-methyl (11e) > 17-
oxo (11a) > 17-deoxy (11d) and ranged from twofold more to
10-fold less potent in tumor cell antiproliferative activity. HUVEC
antiproliferative activity was ranked: 17-methylene (11b) > 17-oxo
(11a) > 17-ethylene (11c) ꢀ 17-methyl (11e) > 17-deoxy (11d)
and were threefold more to eightfold less potent compared to
2ME2. Substituents with sp2 hybridization such as the methylene,
ethylene and oxo were the most active analogs in this series. Addi-
tion of steric bulk to the exocyclic olefin brought a modest drop in
activity (compare 11b to 11c). In this series, the deoxy substitution
had the greatest dropoff in activity.
Since some of the most potent analogs (9 and 11) in this series
contained carbonyl substituents at position 3, we chose to investi-
gate the SAR of related substituents on the doubly substituted
scaffold. From this investigation, we found that as steric bulk
increased, antiproliferative activity decreased. For example, ethyl
carboxamide (8), 3-acetamide (12), 3-methylcarbamate (14), and
3-methyl carbonate (16) all showed decreases in MDA-MB-231
and HUVEC antiproliferative activity compared to 9 and 11.
Conversely, when steric bulk did not increase at position 3, we
maintained in vitro potency as illustrated by the 3-vinyl (7) and
3-urea (13).
#
Rat cassette dosinga
Mouse PKa,b
PO Cmax
(nM)
PO AUC
(nM h)
F
(%)
PO Cmax
(nM)
PO AU C0–1
(nM h)
t1/2
(h)
2ME2 BLOQc
9a
9b
9c
BLOQc
61.5
467
624
320
1273
2146
1703
659
—
294
142
nd
3196
2527
3359
nd
nd
nd
nd
nd
0.3
nd
4.8
1.1
2.1
nd
nd
nd
nd
nd
nd
nd
47
62
130
42
150
190
127
83
13.6 nd
22.8 334
14.9 504
14.3 1427
19.3 nd
9d
10b
10c
11b
11c
11d
11e
15
27
nd
39.5 nd
24.7 nd
24.7 nd
20.6 nd
53
73
508.2
673.2
BLOQc
nd
nd
BLOQc
—
nd
a
Pharmacokinetic variables for compounds were determined using liquid ion
chromatography–tandem mass spectrometric analysis. Pharmacokinetic data were
modeled using noncompartmental analysis. Oral bioavailability (%F) was calculated
from the ratio of oral (PO) and intravenous area under the curve (AUC) values.
All compounds were dosed at 45 mg/m2.
b
c
At the low dose used for cassette dosing, values for these analogs were below
the limit of quantization.
2ME2, respectively. Compound 9b had the longest t1/2 for this set
of analogs (4.8 h), while 9d and 9c had a t1/2 of 2.1 h and 1.1 h,
respectively, compared to 0.3 h for 2ME2. Based on favorable PK
and in vitro properties compared to 2ME2, carboxamides 9b, 9c
and 9d were selected as leads for further development. The results
of these studies have been communicated elsewhere.7
It was desirable to maintain the significantly reduced estroge-
nicity of 2ME2 compared to estradiol. The substituents evaluated
in this study had been selected partially on a basis of low degree
of estrogenicity as single substituents.5,6 All analogs with signifi-
cant in vitro activities were evaluated using the estrogen-depen-
In conclusion, twenty-one 3- and 17-double-modified analogs
of 2ME2 were synthesized and evaluated for antiproliferative
activity using MBA-MB-231 cells, and antiangiogenic activity using
HUVEC proliferation. The most potent carboxamide analogs 9b and
9d were three- to sixfold more potent in antiproliferative activity
and three to fivefold more potent in antiangiogenic activity com-
pared to 2ME2. Representatives of the most potent analog families
were tested via cassette dosing for PK properties. The formamide,
amine and carboxamide analogs all had greater bioavailability
compared to 2ME2 when dosed orally. Based on potential toxicity
liabilities and intrinsic activities, three carboxamides were selected
for further PK evaluation in an oral single-dose CD1 mouse study.
dent MCF-7 in vitro proliferation assay as
a surrogate for
estrogenicity.7 As expected from our reported data on individual
substitutions,5,6 all double-modified analogs tested had equivalent
or less estrogenic activity than 2ME2 (data not shown).
Rat cassette dosing was used as an initial screen for PK evalua-
tion in our discovery program. An advantage of this technique is
that several analogs can be evaluated simultaneously in a single
animal.20 Compounds 9a, 9b, 9c, 9d, 10b, 10c, 11b, 11c, 11d, 11e,
and 15 were evaluated and results are presented in Table 2. Amino-
nitrile 15 had no oral bioavailability. Amines 10c and 10b had the
greatest oral Cmax. Formamide 11d and carboxamide 9d had
approximately equal oral (PO) Cmax PO. Most analogs had similar
PO AUC with the exceptions being carboxamide 9a, amines 10b
and 10c and formamide 11b. All analogs had significantly higher
oral bioavailability (%F) compared to 2ME2.
All carboxamides had significantly better PK properties (Cmax
,
AUC and t1/2) compared to 2ME2 in this model. Based on pharma-
cological evaluation presented elsewhere, carboxamide 9d is cur-
rently in Phase 1 clinical trials for patients with solid tumors
under the name ENMD-1198.
References and notes
Based on these data, formamides (11b,c,d,e), amines (10b,c)
and carboxamides (9a,b,c,d) had acceptable PK properties in cas-
sette dosing analysis. However, the amines were not as active
in vitro as the best formamides or carboxamides. Furthermore,
primary aromatic amine toxicity is well known.21 The lead forma-
mides have a risk of aniline-like toxicity profile, since they could
act as prodrugs for anilines via cleavage of the formamide. Carbox-
amides will not generate the same metabolite and were considered
to have lower risks of toxicity. This, along with their in vitro pro-
files, resulted in several of the carboxamides being selected for
evaluation in mouse single-compound PK in preparation for mouse
tumor models.
Carboxamides 9b, 9c and 9d were tested by oral administration
in CD1 mice and compared to 2ME2. Data are presented in Table 2.
All three double-modified analogs had much higher AUC and long-
er t1/2 than 2ME2. Compound 9d has the highest measured Cmax
and AUC of the analogs tested—fivefold and 23-fold higher than
1. Pribluda, V.; Gubish, E.; LaVallee, T.; Treston, A.; Swartz, G.; Green, S. Cancer
Metastat. Rev. 2000, 19, 173.
2. Mabjeesh, N.; Escuin, D.; LaVallee, T.; Pribluda, V.; Swartz, G.; Johnson, M.;
Willard, M.; Hong, H.; Simons, J.; Giannakakou, P. Cancer Cell 2003, 3, 363.
3. Sweeney, C.; Liu, G.; Yiannoutsos, C.; Kolesar, J.; Horvath, D.; Staab, M. J.; Fife,
K.; Armstrong, V.; Treston, A.; Sidor, C.; Wilding, G. Clin. Cancer Res. 2005, 11,
6625.
4. Agoston, G.; Shah, J.; LaVallee, T.; Zhan, X.; Pribluda, V.; Treston, A. Bioorg. Med.
Chem. 2007, 15, 7524.
5. Suwandi, L.; Agoston, G.; Shah, J.; Hanson, A.; Zhan, X.; LaVallee, T.; Pribluda,
V.; Treston, A. Bioorg Med. Chem. Lett., in press. Available from: <http://
dx.doi.org/10.1016/j.bmcl.2009.09.009>.
6. Shah, J.; Agoston, G.; Suwandi, L.; Zhan, X.; Swartz, G.; LaVallee, T.; Treston, A.
Bioorg. Med. Chem., in press. Available from: <http://dx.doi.org/10.1016/
j.bmc.2009.08.038>.
7. LaVallee, T.; Burke, P.; Swartz, G.; Hamel, E.; Agoston, G.; Shah, J.; Suwandi, L.;
Hanson, A.; Fogler, W.; Sidor, C.; Treston, A. Mol. Cancer Ther. 2008, 7, 1472.
8. Edsall, A.; Mohanakrishnan, A.; Yang, D.; Fanwick, P.; Hamel, E.; Hanson, A.;
Agoston, G.; Cushman, M. J. Med. Chem. 2004, 47, 5126.
9. Schow, R.; McMorris, C. J. Org. Chem. 1979, 44, 3760.
10. Peters, R.; Crowe, D.; Avery, M.; Chong, W.; Tanabe, M. J. Med. Chem. 1989, 32,
1642.

6244
G. E. Agoston et al. / Bioorg. Med. Chem. Lett. 19 (2009) 6241–6244
11. Leese, M.; Leblond, B.; Newman, S.; Purhit, A.; Reed, M.; Potter, B. J. Steroid
Biochem. Mol. Biol. 2005, 94, 239.
12. Berliner, D.; Adams, N.; Jennings-White, C. U.S. Patent 578,357,1, 1998.
13. Shi, Y.; Koh, J. J. Am. Chem. Soc. 2002, 124, 6921.
14. Shi, Y.; Koh, J. Chem. Biol. 2001, 8, 501.
15. Morera, E.; Ortar, G. Tetrahedron Lett. 1998, 39, 2835.
16. Wolfe, J.; Ahman, J.; Sadighi, J.; Singer, R.; Buchwald, S. Tetrahedron Lett. 1997,
38, 6367.
17. Kurzer, F.. In Organic Synthesis; Rajohn, N., Ed.; John Wiley & Sons: NY, 1963;
Vol. IV, pp 49–54.
18. Abdelaal, S.; Kong, S.; Bauer, L. J. Heterocycl. Chem. 1992, 29, 1069.
19. Niwa, R.; Kamada, H.; Shitara, E.; Horiuchi, J.; Kibushi, N.; Kato, T. Chem. Pharm.
Bull. 1996, 44, 2314.
20. Six rats were implanted with jugular vein canulae. The compounds (3–5/
cassette) were administered to three rats intravenously and three rats orally.
Blood was taken at ten timepoints for intravenous and eight time points for the
oral route. Compounds were extracted by plasma protein precipitation and
analyzed by LC–MSMS. PK parameters were calculated using WinNonlin
(Pharsight Inc.).
21. Benigni, R.; Passerini, L. Mutat. Res. 2002, 511, 191.