Anti-angiogenic activity of basic-type, selective cyclooxygenase (COX)-1 inhibitors

Article abstract of DOI:10.1016/j.bmcl.2006.02.021

Indole- and indoline-type basic COX-1-selective competitive inhibitors, 5-amino-1-(3,5-dimethylbenzoyl)-1H-indole (1) and 5-amino-1-(3,5-dimethylphenyl)-2,3-dihydro-1H-indole (2), were found to possess anti-angiogenic activity estimated as a tube formation-inhibition using human umbilical vein endothelial cells (HUVECs).

Full text of DOI:10.1016/j.bmcl.2006.02.021

Bioorganic & Medicinal Chemistry Letters 16 (2006) 3068–3072
Anti-angiogenic activity of basic-type, selective
cyclooxygenase (COX)-1 inhibitors
Hiroko Sano, Tomomi Noguchi, Atsushi Miyajima,
Yuichi Hashimoto and Hiroyuki Miyachi*
Institute of Molecular and Cellular Biosciences, University of Tokyo, Yayoi, Bunkyo-ku, Tokyo 113-0032, Japan
Received 23 December 2005; revised 26 January 2006; accepted 7 February 2006
Available online 2 March 2006
Abstract—Indole- and indoline-type basic COX-1-selective competitive inhibitors, 5-amino-1-(3,5-dimethylbenzoyl)-1H-indole (1)
and 5-amino-1-(3,5-dimethylphenyl)-2,3-dihydro-1H-indole (2), were found to possess anti-angiogenic activity estimated as a tube
formation-inhibition using human umbilical vein endothelial cells (HUVECs).
ꢀ 2006 Elsevier Ltd. All rights reserved.
Thalidomide is a sedative and/or hypnotic drug, which
was used from the late 1950s to the early 1960s, but
was withdrawn from the market because of severe tera-
togenicity.1 In spite of this, thalidomide research was
continued, due to the serendipitous finding of the drug’s
effectiveness against various intractable diseases, such as
leprosy, AIDS, and certain kinds of cancers.2 Finally,
the United States Food and Drug Administration
(FDA) gave marketing approval to thalidomide for the
treatment of Hansen’s disease in 1998, with special pre-
cautions for usage. At present, several clinical studies of
thalidomide for the treatment of multiple myeloma, co-
lon cancer, prostate cancer, and other conditions are on-
going in the US.3
COX-2 have been well defined. COX-1 is reported to
be constitutively expressed in many organs or tissues,
while COX-2 is inducible with various stimuli. However,
recent molecular-biological studies have indicated that
this simple paradigm has many exceptions. For example,
COX-1 can be regulated during development,6 while
COX-2 is constitutively expressed in the brain7 and in
reproductive tissues.8 Often, both isozymes are involved
in physiological and pathophysiological conditions,
while in some cases, each isozyme plays a distinct role.
COX-2 has been detected in various tumors and its roles
in carcinogenesis and angiogenesis have been well docu-
mented.9 Therefore, COX-2 is thought to be a promising
therapeutic target for cancer.10 Attempts have been
made to apply COX-2 inhibitors, such as celecoxib,
rofecoxib, and sulindac, for chemoprevention of various
cancers, including colon and prostate cancers.11 Howev-
er, current clinical studies of a COX-2-selective inhibi-
tor, rofecoxib (Vioxx), for preventing recurrence of
colorectal polyps in patients with a history of colorectal
adenomas were discontinued and the drug was with-
drawn from the market because its use was associated
with an increased incidence of cardiovascular events,
such as heart attack and stroke.12
The effectiveness of thalidomide for the treatment of cer-
tain kinds of cancers was suggested to be mediated
through inhibition of tumor necrosis factor (TNF)-a pro-
duction- and anti-angiogenic activity.4 However, we sus-
pected that cyclooxygenase (COX) might be another
anticancer-related molecular target of thalidomide.
Prostaglandin and thromboxane biosynthesis involves
the conversion of arachidonic acid to prostaglandin H2
(PGH2), a reaction catalyzed by the sequential actions
of COX and prostaglandin endoperoxidase synthase
(PGHS).5 Three isozymes of COX (COX-1, COX-2,
and COX-3) are known to date, of which COX-1 and
Very recently, experimental results have also indicated a
possible involvement of the other isoform of COX,
COX-1, in angiogenesis, thereby providing the rationale
for the development of selective COX-1 inhibitors.13
However, the effectiveness of COX-1 inhibitors as angi-
ogenesis inhibitors has not yet been well established.
Keywords: Angiogenesis; COX-1; COX-1 inhibitor; HUVEC.
*
Corresponding author. E-mail: miyachi@iam.u-tokyo.ac.jp
0960-894X/$ - see front matter ꢀ 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.02.021

H. Sano et al. / Bioorg. Med. Chem. Lett. 16 (2006) 3068–3072
3069
We have recently demonstrated that thalidomide direct-
ly inhibits COX-1/COX-2 with efficacy comparable to
that of the representative drug, aspirin.14 Our earlier
work on the COX-inhibiting activity of thalidomide
afforded a new structural type of COX-1-selective inhib-
itors, such as compounds 1 and 2; the IC50s of 1 for
COX-1 and COX-2 are 1.9 lM and undetermined
(13.8 % inhibition at 30 lM), respectively, and those
of 2 are 1.8 and 55 lM, respectively.15 These compounds
are structurally unique, because most of the previously
reported COX-1-selective and non-selective COX inhib-
itors are acidic or neutral compounds.
(hEGF, VEGF, hFGF-b, and R3-IGF-1, as well as
FBS) in the presence or absence of various concentra-
tions of the test compounds and incubated at 37 ꢁC un-
der 5% CO2 for 6 h. After incubation, each well was
photographed using a · 5 objective to analyze tube for-
mation. The corresponding area was measured as the
number of pixels using Meta Morph software (Universal
Imaging, Downingtown, PA). Experiments were repeat-
ed at least three times.
To assess the cytotoxicity of the test compounds, HU-
VECs were treated with various concentrations of these
compounds at 37 ꢁC for 6 h. After the incubation, the
viability of the treated cells was measured by direct
counting under a microscope. The LD50 values of these
compounds were more than 100 lM.
We have also reported the anti-angiogenic activity of
thalidomide and its metabolites, using a human umbili-
cal vein endothelial cell (HUVEC) assay.16 The study
clearly indicated that thalidomide, as well as its main
metabolites 5-hydroxythalidomide and N-hydroxytha-
lidomide, exhibits moderate anti-angiogenic activity.
Although the exact mechanism(s) of anti-angiogenic
activity elicited by thalidomide and its metabolites is un-
known, we hypothesized that a COX-1-mediated path-
way might be involved in angiogenesis, and these
compounds might act by inhibiting this pathway. There-
fore, we expected that compounds, 1 and 2, possess
more potent anti-angiogenic activity than thalidomide.
In this paper, we would like to report the anti-angiogen-
ic activity of compounds 1 and 2.
We have previously demonstrated that the N-substituted
phenylindoline and N-substituted phenylindole skele-
tons are useful scaffolds for the development of COX
inhibitors other than the clinically used diaryl heterocy-
clic COX inhibitors, such as celecoxib.15 We have also
demonstrated that strongly electron-donating groups,
such as an unsubstituted amino group at the 5- or 4-po-
sition of the indoline and/or indole skeleton, favor po-
tent and COX-1-selective inhibitory activity.15 In the
previous study, the substituent on the N-phenyl, and/
or N-benzoyl group, was fixed to 3,5-dimethyl group,
considering the early SAR of the series of compounds,
but the substituent was not optimized. Therefore, we
synthesized 3,5-disubstituted-benzoyl derivatives and as-
sayed their COX-inhibiting activity. Results are summa-
rized in Table 1. Roughly speaking, in the case of the
5-nitroindole and 5-nitroindoline series, all compounds
which have CF3, C(CH3)3, and Cl as a 3,5-substituent
did not show apparent COX-1-inhibiting activity (4, 5,
6, 11, and 12), or weak inhibitory activity (13). As for
COX-2, these compounds also show no activity at the
concentration of 100 lM (11–13), or showed very weak
inhibitory activity (4–6). In the case of 5-aminoindole
and 5-aminoindoline series, all the compounds show
moderate to high COX-1 inhibitory activity, except 8.
As for COX-2, all these compounds show weak inhibito-
ry activity at the concentration of 100 lM. As a whole,
none of the compounds show superior activity and selec-
tivity than 3,5-dimethyl derivative (1). Therefore, we
selected compounds 1 and 2 for further study (Fig. 1).
The preparation of 1, 2, 10, 14, 18 (2,3-dihydro-1-(3,5-
dimethylphenyl)-1H-indole), and 19 (2,3-dihydro-5-
methoxy-1-(3,5-dimethylphenyl)-1H-indole) was described
previously.15 The N-benzoyl-substituted indoles and
the N-benzoyl-substituted indolines were prepared
according to the preparation of 1. Briefly, 5-nitroindole
sodium (or 5-nitroindoline sodium) was prepared from
5-nitroindole (or 5-nitroindoline) and NaH, with 3,5-di-
substituted-benzoyl chloride in N,N-dimethylformamide
to afford 4–6 and 11–13. These nitro-derivatives were re-
duced with 10% Pd–C to afford amino derivatives 7–9
and 15–17. The structures of the synthesized compounds
1
were confirmed by H NMR, mass spectroscopy, and
elemental analysis. SC-560 (SC), DUP-697 (DUP), and
Ibuprofen (IBU) were purchased from Funakoshi. Co.
Ltd (Japan).
A kinetic study of the inhibition of COX-1-mediated
oxidation of arachidonic acid was performed with an en-
zyme immunoassay-based COX inhibitor screening
assay kit purchased from Cayman Chemical (Ann Ar-
bor, MI, Catalog No. 560101) according to the suppli-
er’s protocol.
First, we analyzed the inhibitory mode of COX-1 inhib-
itors 1 and 2, because the screening method that we
previously used17 could not completely exclude false-po-
sitive results from simple anti-oxidant compounds. The
Lineweaver–Burk plot depicted in Figure 2 clearly indi-
cated that compounds 1 and 2 are true competitive
inhibitors of COX-1. Thus, we confirmed that these
compounds inhibit COX-1 by competing directly with
the natural substrate, arachidonic acid, at the ligand
binding site.
HUVEC tube formation assay16 was performed as fol-
lows: human umbilical vein endothelial cells (HUVECs)
were plated on Matrigel and treated with the test com-
pounds for 6 h, then tube formation was measured as
previously reported.16 Briefly, 6-well plates were coated
with 1.5 mL of the Matrigel basement membrane matrix
(Becton Dickinson) and allowed to gel at 37 ꢁC under an
atmosphere of 5% CO2 in air for 30 min. Then, HU-
VECs were plated at 5 · 105 cells/well in DMEM con-
taining the vehicle (0.5% DMSO) and growth factors
We then investigated the anti-angiogenic activity of
representative COX inhibitors ((SC (COX-1-selective
inhibitor), DUP (COX-2-selective inhibitor), IBU (non-
selective COX inhibitor)), and our COX-1-selective

3070
H. Sano et al. / Bioorg. Med. Chem. Lett. 16 (2006) 3068–3072
H2N
H2N
Me
Me
N
N
O
Me
Me
2
1
Figure 1. Chemical structures of the basic-type selective COX-1
inhibitors 1 and 2.
Figure 2. Lineweaver–Burk plots for compounds 1 and 2. AA means
arachidonic acid.
inhibitors 1 and 2 at the concentration of 100 lM (Fig. 3).
Compounds 18 and 19, both of which structurally resem-
ble 1 and 2, but which lack COX-inhibitory activity, were
also assayed.
As can be seen from Figure 3, the reference compounds
SC, DUP, and IBU exhibited anti-angiogenic activity.
SC, a COX-1-selective inhibitor, exhibited moderate
anti-angiogenic activity (ca. 50% inhibition at
100 lM). IBU, which is a well-known non-selective
COX inhibitor, also exhibited moderate anti-angiogenic
activity, comparable to that of SC (ca. 35% inhibition at
100
75
50
25
0
Figure 3. Anti-angiogenic effects of representative COX inhibitors
(SC, DUP, and IBU) and our derivatives.

H. Sano et al. / Bioorg. Med. Chem. Lett. 16 (2006) 3068–3072
3071
100 lM). DUP,
a
COX-2-selective inhibitor, also
known indole and/or indoline structure-dependent
pathway.
showed anti-angiogenic activity. However, the potency
of DUP (ca. 25% inhibition at 100 lM) is weaker than
that of SC. It is reported that HUVEC apparently does
not express COX-2 in the absence of endogenous stimu-
li, such as IL-1b,18 so DUP might not show potent activ-
ity in these cells. In the case of our compounds, both 1
and 2 exhibited moderate anti-angiogenic activity, com-
parable or superior to that of SC (ca. 50% and 75% inhi-
bition at 100 lM, respectively), whereas 18 and 19,
which have similar structures to 1 and 2, but do not
show apparent COX-inhibitory activity, lacked anti-an-
giogenic activity. These results indicate that the anti-an-
giogenic activity of 1 and 2 is mediated through the
COX-1 pathway, at least in part, but not via an un-
Figures 4 and 5 show the dose-dependency of the anti-
angiogenic effects of COX inhibitors. All the compounds
tested showed dose-dependent anti-angiogenic activity,
and the rank order of inhibitory activity is as follows;
2 > SC > 1 > IBU > DUP (none of these compounds
showed cytotoxicity at the concentration of 300 lM).
Compound 2 was found to possess highly potent anti-
angiogenic activity, comparable with that of the well-
known COX-1-selective antagonist, SC.
In conclusion, we have discovered that the novel basic-
type selective COX-1 inhibitors, 1 and 2, exhibit apparent
anti-angiogenic activity in a HUVEC tube formation as-
say. Compound 2 was the most potent among the pre-
pared compounds. Angiogenesis has recently become a
primary target of anticancer therapy. A recent report
has indicated that COX-1 is over-expressed in a signifi-
cant number of ovarian cancers.19 Further, ovarian can-
cer is known to be highly vascular and is a primary
cancer in which current anti-angiogenic therapies are
being tested.19 Therefore, our present results may yield
new candidate drugs for the treatment of ovarian cancer.
100
75
50
25
0
Acknowledgments
100
DMSO
10 100 10 100 10 100 10 100 10 100
We thank Dr. Hideki Nohara (University of Tokyo) for
helpful discussions and technical assistance. The work
described here was partly supported by Grants-in-Aid
for Scientific Research from The Ministry of Education,
Culture, Sports, Science and Technology, Japan, and the
Japan Society for the Promotion of Science.
SC
DUP
IBU
1
2
conc. (μM)
Figure 4. Dose–response relationship of representative COX inhibitors
(SC, DUP, and IBU) and our derivatives (1, 2).
Figure 5. Photographic representation of the dose-dependent inhibition of 2 on the HUVEC tube-formation. (A) Control; (B) 10 lM 2; (C) 100 lM
2; (D) 300 lM 2.

3072
H. Sano et al. / Bioorg. Med. Chem. Lett. 16 (2006) 3068–3072
12. Onn, A.; Tseng, J. E.; Herbst, R. S. Clin. Cancer Res.
2001, 7, 3311.
References and notes
13. Kitamura, T.; Itoh, M.; Noda, T.; Matsuura, M.; Waka-
bayashi, K. Int. J. Cancer 2004, 109, 576.
14. Noguchi, T.; Shimazawa, R.; Nagasawa, K.; Hashimoto,
Y. Bioorg. Med. Chem. Lett. 2002, 12, 1043.
1. Hales, B. F. Nat. Med. 1999, 5, 489.
2. Calabrese, L.; Fleisher, A. B. Am. J. Med. 2000, 108, 487.
3. Richardson, P.; Hideshima, T.; Anderson, K. Annu. Rev.
Med. 2002, 53, 629.
15. Sano, H.; Noguchi, T.; Tanatani, A.; Hashimoto, Y.;
Miyachi, H. Bioorg. Med. Chem. 2005, 13, 3079.
16. Ng, S. S. W.; Gutschow, M.; Weiss, M.; Hauschildt, S.;
Teubert, U.; Hecker, T. K.; Luzzio, F. A.; Kruger, E. A.;
Eger, K.; Figg, W. D. Cancer Res. 2003, 63, 189.
17. The first screening was performed with a colorimetric
COX (ovine) inhibitor screening assay kit purchased from
Cayman Chemical (Ann Arbor, MI, Catalog No. 760111)
according to the supplier’s protocol.
4. D’Amato, R. J.; Loughnan, M. S.; Flynn, E.; Folkman, J.
Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 4082.
5. Hawkey, C. J. Lancet 1999, 353, 307.
6. Smith, W. L.; Langenbach, R. J. Clin. Invest. 2001, 12,
1491.
7. Yamagata, K.; Andreasson, K. I.; Kaufmann, W. E.;
Barnes, C. A.; Worley, P. F. Neuron 1993, 11, 371.
8. Kniss, D. A. J. Soc. Gynecol. Investig. 1999, 6, 285.
9. Vane, J. Nature (London) 1994, 367, 215.
10. Reddy, B. S.; Hirose, Y.; Lubet, R.; Steele, V.; Kelloff, G.;
Paulson, S.; Seibert, K.; Rao, C. V. Cancer Res. 2000, 60,
293.
18. Imaizumi, T.; Itaya, H.; Fujita, K.; Kudoh, D.; Kudoh, S.;
Mori, K.; Fujimoto, K.; Matsumiya, T.; Yoshida, H.;
Satoh, K. Thromb. Haemost. 2000, 83, 949.
19. Rajnish, A. G.; Lovella, V. T.; Beverly, J. T.; Sanjoy, K.
D.; Jason, D. M.; Sudhansu, K. D.; Raymond, N. D.
Cancer Res. 2003, 63, 906.
11. Hsu, A. L.; Ching, T. T.; Wang, D. S.; Song, X.;
Rangnekar, V. M.; Chen, C. S. J. Biol. Chem. 2000, 275,
1397.