Thienopyridine and benzofuran derivatives as potent anti-tumor agents possessing different structure-activity relationships

Article abstract of DOI:10.1016/j.bmcl.2004.04.079

(3-Amino-6-thiophen-2-yl-thieno[2,3-b]pyridin-2-yl)phenylmethanone (3) was discovered as a new type of cytotoxic agent selective against a tumorigenic cell line. The molecular structure of a previously reported compound, (4-hydroxy-3-methyl-6-phenylbenzofuran-2-yl)phenylmethanone (2), had remarkably similar bioisosteric substructures to that of compound 3. Although the relationship between the molecular structure and biological activity of each derivative synthesized from these two hit compounds (2 and 3) were studied, unexpectedly no correlation was observed. However, after further synthetic study from 3, one of the most potent derivative (10k) having a different SAR profile from 2, was discovered.

Full text of DOI:10.1016/j.bmcl.2004.04.079

Bioorganic & Medicinal Chemistry Letters 14 (2004) 3411–3414
Thienopyridine and benzofuran derivatives as potent anti-tumor
agents possessing different structure–activity relationships
Ichiro Hayakawa,a Rieko Shioya,a Toshinori Agatsuma,b
Hidehiko Furukawab and Yuichi Suganoa,*
aLead Discovery Research Laboratories, Sankyo Co. Ltd, Shinagawa-ku, Tokyo 140-8710, Japan
bBiomedical Research Laboratories, Sankyo Co. Ltd, Shinagawa-ku, Tokyo 140-8710, Japan
Received 14 April 2004; revised 26 April 2004; accepted 26 April 2004
Available online
Abstract—(3-Amino-6-thiophen-2-yl-thieno[2,3-b]pyridin-2-yl)phenylmethanone (3) was discovered as a new type of cytotoxic agent
selective against a tumorigenic cell line. The molecular structure of a previously reported compound, (4-hydroxy-3-methyl-6-
phenylbenzofuran-2-yl)phenylmethanone (2), had remarkably similar bioisosteric substructures to that of compound 3. Although
the relationship between the molecular structure and biological activity of each derivative synthesized from these two hit compounds
(2 and 3) were studied, unexpectedly no correlation was observed. However, after further synthetic study from 3, one of the most
potent derivative (10k) having a different SAR profile from 2, was discovered.
Ó 2004 Elsevier Ltd. All rights reserved.
1. Introduction
Hit compounds of a screening
As a part of program to develop new drugs, the struc-
ture optimization of lead compounds is the most
important phase in the search for efficient derivatives,
and undesired properties of the lead compounds should
be resolved during this step. What is most problematic,
here, is actually the generation of lead compounds from
hit compounds that were selected by random screening.
Not only must lead compounds have sufficient biologi-
cal potential to be the backbone of a library, they must
contain sufficient structure components for library
construction (Fig. 1).
Lead generation
Library synthesis
Lead compounds
Optimized derivatives
Figure 1. A flow chart from screening hit(s) to library design. If a
screening hit contains sufficient information for the library design, the
compound was simply used after the hit evaluation. Multiple hits were
used to ensure the design of a more reliable library.
Very few hit compounds yield an excellent single lead
compound that is suitable for library construction for
structure optimization. Therefore the structural infor-
mation of multiple hit compounds is often combined to
form a lead compound for an accurate library design.
The structures of the hit compounds are integrated, and
the common substructures are determined to design a
reliable scaffold. The hit compounds that are used in this
approach should interact with the same domain of the
same target molecule.
2. Screening for hit compounds
During our project to develop new anticancer drugs as
small molecules, a cell-based screening system had been
established in our laboratory. Since then, we have found
and reported 4-hydroxy-3-methyl-6-phenylbenzofuran-
2-carboxylic acid ethyl ester (1) to be the most promising
in a series of hit compounds.1;2 This hit compound 1
showed selective and potent cytotoxicity against a
tumorigenic cell line, WI-38 VA-13 subline 2RA
Keywords: Lead compound; SAR; Anticancer.
* Corresponding author. Tel.: +81-3-3492-3131; fax: +81-3-5436-8578;
e-mail: ysugan@sankyo.co.jp
0960-894X/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.04.079

3412
I. Hayakawa et al. / Bioorg. Med. Chem. Lett. 14 (2004) 3411–3414
OMe
OH
Me
Me
O
O
a
1
4: EC50 > 4000 ng/mL
5: EC50 = 420 ng/mL
6: EC50 > 4000 ng/mL
O
O
O
OEt
Me
O
HN
S
1: EC50 = 40 ng/mL
Figure 2. Hit compound 1.
S
S
N
N
Me
b
Me
N
OH
Me
3
NH2
O
O
O
S
S
S
O
N
3: EC50 = 90 ng/mL
2: EC50 = 600 ng/mL
Scheme 1. Synthesis and cytotoxicity of methylated hit compounds.
(a) 1.5 equiv NaH (55% dispersion in mineral oil), then 2.0 equiv MeI
in DCM, 25 °C, 1.0 h (98%); (b) 1.1 equiv ButOK and 1.1 equiv MeI in
DMF, 25 °C, 1.0 h (5: 79%, 6: trace).
Figure 3. Lead compound 2 derived from hit compound 1, and the
second hit compound 3 discovered in the same screening.
gen, had completely disappeared. These results suggest
that the functional groups contributed largely to the
biological activity, and that one might replace the other
if these two hit compounds bound to the same target
molecule.
(VA-13) (EC50 ¼ 40 ng/mL), despite no cytotoxicity was
observed against the normal parental cell line, WI-38
(EC50 > 4000 ng/mL) (Fig. 2).
After further synthetic study, (4-hydroxy-3-methyl-
6-phenylbenzofuran-2-yl)phenylmethanone (2) was
derived as a lead compound that did not contain a
biologically unstable ester group.1;2 After further screen-
ing, (3-amino-6-thiophen-2-yl-thieno[2,3-b]pyridin-2-yl)-
phenylmethanone (3) was also discovered as a hit comp-
ound in the same screening (Fig. 3).
Although a thorough investigation of the domain com-
prising the central fused ring and its substituents would
be required for further structure exploitation of the hit
compounds, the synthetic modification of this domain
was considered challenging. The phenyl ketone moiety
of compound 3 was substituted instead to investigate the
relationship between the structure and cytotoxic activity
of the derivatives. The biological activity was compared
with that of benzofuran (2) derivatives, which have al-
ready been described in our previous report.1;2 The
derivatives of compound 3 were readily synthesized
from acetylthiophene (7) as shown in Scheme 2.3;4
The chemical structures of these two compounds (2 and
3) can be considered to have bioisoteric substructures
except functional groups on central fused rings. A ben-
zene ring is usually biologically equivalent to a thio-
phene or a pyridine ring, and the dimension of furan
ring is almost equal to thiophene ring. Although these
hits were discovered by a cell-based screening and the
target molecules of these compounds were unidentified,
lead generation based on these two hit compounds was
attempted.
4. Structure–activity relationship
The cytotoxicities of the synthesized derivatives are lis-
ted in Table 1. While phenyl ketone (11a [2]) without a
3. Evaluation of different functional groups
N
The major difference between hit compound 2 and 3 is
the functional groups connected to the central fused ring
and their positions. Compound 3 possesses an amino
group instead of the methyl group in compound 2. Since
an amino group and not a hydroxy group was present in
3, this may explain the hydrophilic properties of those
functional groups are interchangeable.
S
O
S
O
8
a
7
b
NH2
CN
S
O
R
S
In order to evaluate these functional groups in terms of
biological activity, derivatives whose functional group
was masked with methyl groups were synthesized and
the cytotoxicity of the resulting compounds was tested
(Scheme 1). Methyl ether 4, which was prepared from
the original hit compound 1, completely lost its cyto-
toxicity. Monomethylamine 5 derived from 3 still
showed weak activity, and the cytotoxicity of dimeth-
ylamine 6, which had no hydrogen atom on the nitro-
N
S
c
S
N
H
10
9
Scheme 2. Synthesis of thieno[2,3-b]pyridine derivatives. (a) 4.0 equiv
Me2NCH(OMe)2 neat, reflux, 7.0 h (60%); (b) 1.0 equiv NCCH2-
CSNH2 and 2.2 equiv ButOK in DMF, 50 °C, 3.0 h (93%); (c) 1.0 equiv
BrCH2COR, 1.0 equiv aqueous 10% KOH in DMF 25 °C, and then
1.0 equiv aqueous 10% KOH 25 °C, 1.0 h (10b:8 67%; 10c: 61%; 10d:
55%; 10f: 78%; 10g: 82%; 10h: 77%; 10i: 63%; see also Table 2).

I. Hayakawa et al. / Bioorg. Med. Chem. Lett. 14 (2004) 3411–3414
3413
substituent on the benzene ring showed relatively weak
cytotoxicity, its derivatives, which had a methoxy group
at the ortho position (11b), were moderately active. The
methoxy group at the para position (11c) also improved
the biological activity to some extent, and 11d, which
has two methoxy groups at the ortho and para positions,
was the most potent benzofuran derivative. Ethyl ester
(11f [1]) and methyl ketone (11g) were another class of
highly potent derivatives.1;2
differentiated. The benzofuran derivatives (11) have a
benzene substituent at the sixth position, and thieno-
pyridine derivatives (10) were made to have a thiophene
ring instead. Although these substituents are usually
recognized to be biologically equivalent and inter-
changeable, the steric hindrance of the ortho hydrogen
atoms in the thiophene ring and pyridine of 10 was
estimated to be very small because of the lack of two
hydrogen atoms (Fig. 4). The semi-empirical AM1 cal-
culations for energy minimization indicated that the
torsional angle of thienopyridine and thiophene rings
of 10g was almost flat, the corresponding angle of
benzofuran derivative 11g, however, was approximately
41.8°.5 This result was consistent with that of simple
biphenyl.6
In addition, the simple phenyl ketone (10a [3]) was rel-
atively potent. Although a derivative with a methoxy
group at the ortho position (10b8) showed cytotoxicity,
derivatives with a para methoxy group (10c and d)
considerably lost their biological activity. Although the
dimethoxy phenyl group was recognized to be the best
substructure of benzofuran derivative (11d), it did not
perform as a substructure of thienopyridine derivatives
(10d). In contrast with 11f [1] and 11g, neither ethyl ester
(10f) nor methyl ketone (10g) showed cytotoxicity.
As this calculation suggests that the structure difference
between thiophene and benzene may not be negligible as
a bioisosteric structure in this case, the cytotoxicity of
derivatives that have a benzene ring instead of a thio-
phene ring (12 and 13) was tested.7 To our surprise,
these compounds were revealed to have no biological
activity at all, and this result suggested that the core
structure of 10 and 11 were not bioisosteric (Fig. 5).
5. Conformational differences of the hit compounds
As the structure–activity relationship of the derivatives
with two similar core structures (10 and 11) did not
show the same profile, the core structures were further
In order to discover the most potent substituent of the
thienopyridine derivatives (10), another investigation
was carried out independently from the structure–
activity relationship studies of benzofuran derivatives
(11) (Table 2). While the bioactivity of the benzofuran
derivative was improved by placing a methoxy group in
the ortho and/or a para position (11b and c), placing it in
the meta position decreased activity.1 In contrast to the
benzofuran derivatives, some of the thienopyridine
derivatives with a meta substituent (10j10 and k11)
showed increased cytotoxicity. The most potent deriva-
tive was the compound with a meta fluoro substituent
(10k11), which was more potent than all the synthesized
benzofuran derivatives. Introduction of a para sub-
stituent (10n, o, p, and q) weakened the biological
activity.
Table 1. The structure–activity relationship of benzofuran derivatives
(11) and thienopyridine derivatives (10)
OH
NH2
Me
O
R
O
R
S
S
O
N
10
11
Compd
R
Cytotoxicity of 11 Cytotoxicity of 10
EC50, ng/mLa;b EC50, ng/mLa
a
600 (compound 2) 90 (compound 3)
b8
90
280
40
59
OH
NH2
Me
H
MeO
H
O
R
O
R
H
S
S
O
N
c
770
520
79
H
11
H
10
H
OMe
Figure 4. The steric hindrance of the protons at ortho position of two
benzene rings (11) twists the connecting bond to an approximately
41.8° torsional angle. The thiophene and pyridine ring (10), however,
produces almost no steric interaction.
d
MeO
MeO
OMe
OMe
e9
90
NH2
NH2
f
OEt
Me
40 (compound 1)
>4000
2700
890
O
O
g
h
i
80
270
410
S
S
OEt
N
N
Et
tert-Bu
570
12
13
a Selectivity against a tumor cell line was confirmed.
b The compounds and corresponding biological data are reported in
Ref. 1.
Figure 5. Derivatives of compound 10 whose thiophene ring is
replaced with bioisosteric benzene ring.

3414
I. Hayakawa et al. / Bioorg. Med. Chem. Lett. 14 (2004) 3411–3414
Table 2. Biological activities of thienopyridine derivatives (10)
Acknowledgements
NH2
O
The authors would like to express their gratitude to Dr.
Mihiro Sunose for performing the synthetic study, and
Ms. Youko Oda for her technical support in the cyto-
toxicity test.
S
S
R
N
10
Compd
R
Yield (%)
Cytotoxicity of 10
from 9 to 10a EC50, ng/mLb
References and notes
a
93
59
90
79
1. Hayakawa, I.; Shioya, R.; Agatsuma, T.; Furukawa, H.;
Naruto, S.; Sugano, Y. Bioorg. Med. Chem. Lett. 2004, 14,
455.
OMe
e9
2. US Patent Application US60/306116, July 17, 2001.
3. Moryashova, S. I.; Salamandra, L. K.; Fedorov, A. E.;
Rodinovskaya, L. A.; Shestopalov, A. M.; Semenov, V. V.
Russ. Chem. Bull. 1998, 47, 357.
4. Moryashova, S. I.; Salamandra, L. K.; Fedorov, A. E.;
Rodinovskaya, L. A.; Shestopalov, A. M.; Semenov, V. V.
Izv. Akad. Nauk Ser. Khim. 1998, 2, 365.
5. SYBYL 6.6 (Tripos, Inc.) was used as a front-end.
6. Streitwieser, A.; Heathcook, C. H.; Kosower, E. M.
Introduction to Organic Chemistry. 4th ed.; Macmillan
Publishing Co.: New York, NY, 1992, pp 1051–
1052.
MeO
OMe
j10
k11
l
63
78
68
30
25
F
NO2
Br
150
7. The tested compounds were purchased from ChemBridge
Corporation, San Diego, CA, USA.
m
n
75
64
500
8. Compound 10b: orange powder. Mp 203–204 °C, 1H
NMR (270 MHz, DMSO-d6) d 8.63 (d, 1H, J ¼ 8:1 Hz),
8.21 (br s, 2H), 8.03 (d, 1H, J ¼ 8:4 Hz), 7.97 (d, 1H,
J ¼ 3:8 Hz), 7.74 (d, 1H, J ¼ 5:1 Hz), 7.49 (t, 1H,
J ¼ 8:1 Hz), 7.32 (d, 1H, J ¼ 7:6 Hz), 7.23–7.15 (m, 2H),
7.05 (t, 1H, J ¼ 7:6 Hz), 3.78 (s, 3H); MS(APCI, m/z) 367
(M+1)þ.
1800
F
F
F
9. Compound 10e: yellow powder. Mp 107–109 °C, 1H NMR
(270 MHz, DMSO-d6) d 8.63 (d, 1H, J ¼ 8:6 Hz), 8.22 (br
s, 2H), 8.03 (d, 1H, J ¼ 8:9 Hz), 7.97 (dd, 1H, J1 ¼ 3:5 Hz,
J2 ¼ 1:1 Hz), 7.75 (dd, 1H, J1 ¼ 5:1 Hz, J2 ¼ 1:1 Hz), 7.22
(dd, 1H, J1 ¼ 5:1 Hz, J2 ¼ 1:1 Hz), 7.12–7.02 (m, 2H), 6.89
(d, 1H, J ¼ 2:7 Hz), 3.75 (s, 3H), 3.73 (s, 3H); MS(APCI,
m/z) 397 (M+1)þ.
o
58
>4000
p
q
45
46
2300
2800
F
F
10. Compound 10j: yellow powder. Mp 222–223 °C, 1H NMR
(270 MHz, DMSO-d6) d 8.67 (d, 1H, J ¼ 8:6 Hz), 8.43 (br
s, 2H), 8.06 (d, 1H, J ¼ 8:6 Hz), 7.99 (d, 1H, J ¼ 3:2 Hz),
7.77 (d, 1H, J ¼ 3:2 Hz), 7.47 (t, 1H, J ¼ 7:6 Hz), 7.37 (d,
1H, J ¼ 7:6 Hz), 7.30 (s, 1H), 7.22 (dd, 1H, J1 ¼ 4:9 Hz,
J2 ¼ 3:8 Hz), 7.16 (dd, 1H, J1 ¼ 8:1 Hz, J2 ¼ 2:7 Hz), 3.84
(s, 3H); MS(APCI, m=z) 367 (M+1)þ.
Br
a Yields were calculated based on 9.
b Selectivity against a tumor cell line was confirmed.
1
11. Compound 10k: yellow solid. Mp 221–222 °C, H NMR
In conclusion, although the lead generation from two hit
compounds 2 and 3 was unsuccessful, biologically
potent derivatives from each hit compound were dis-
covered. These compounds will be evaluated in the next
stage of biological tests.
(270 MHz, DMSO-d6) d 8.68 (d, 1H, J ¼ 8:6 Hz), 8.49 (br
s, 2H), 8.07 (d, 1H, J ¼ 8:6 Hz), 7.99 (dd, 1H, J1 ¼ 3:8 Hz,
J2 ¼ 1:1 Hz), 7.77 (dd, 1H, J1 ¼ 4:8 Hz, J2 ¼ 1:1 Hz), 7.68–
7.54 (m, 3H), 7.49–7.44 (m, 1H), 7.22 (dd, 1H, J1 ¼ 4:8 Hz,
J2 ¼ 1:1 Hz); MS(APCI, m=z) 355 (M+1)þ.