Design, docking, and synthesis of novel indeno[1,2-c]isoquinolines for the development of antitumor agents as topoisomerase I inhibitors

Article abstract of DOI:10.1016/j.bmcl.2007.04.064

An intramolecular radical cyclization reaction of 4-bromo-3-arylisoquinolines 11a-c allowed the efficient synthesis of 11-methylindenoisoquinolines 2a-c. 5-(2-Aminoethylamino)indeno[1,2-c]isoquinolin-11-one 4 was also prepared in the convenient manner. Th

Full text of DOI:10.1016/j.bmcl.2007.04.064

Bioorganic & Medicinal Chemistry Letters 17 (2007) 3531–3534
Design, docking, and synthesis of novel indeno[1,2-c]isoquinolines
for the development of antitumor agents
as topoisomerase I inhibitors
Won-Jea Cho,a,* Quynh Manh Le,a Hue Thi My Van,a Kwang Youl Lee,a
Bok Yun Kang,a Eung-Seok Lee,b Sang Kook Leec and Youngjoo Kwonc
aCollege of Pharmacy and Research Institute of Drug Development, Chonnam National University, Kwangju 500-757,
Republic of Korea
bCollege of Pharmacy, Yeungnam University, Kyongsan 712-749, Republic of Korea
cCollege of Pharmacy, Ewha Womans University, Seoul 120-750, Republic of Korea
Received 8 February 2007; revised 30 March 2007; accepted 20 April 2007
Available online 25 April 2007
Abstract—An intramolecular radical cyclization reaction of 4-bromo-3-arylisoquinolines 11a–c allowed the efficient synthesis of
11-methylindenoisoquinolines 2a–c. 5-(2-Aminoethylamino)indeno[1,2-c]isoquinolin-11-one 4 was also prepared in the convenient
manner. The synthesized compounds were tested in vitro for cytotoxicity and DNA-topoisomerase 1 (top 1) inhibitory activity.
The dramatic enhancement of top 1 inhibitory activity of 4 was explained by a docking study using the FlexX program.
Ó 2007 Elsevier Ltd. All rights reserved.
DNA-topoisomerase 1 (top 1) is considered an impor-
tant enzyme to relax supercoiled DNA for transcription,
replication, and mitosis.1 Elevation of the top 1 levels in
solid tumors compared to normal tissue has made top 1
a promising target for antineoplastic drugs.2 The X-ray
crystal structures of DNA-top 1 have been revealed with
ligands such as topotecan, indenoisoquinoline, and
indolocarbazole molecules.3 Interestingly, the binding
pockets of all of these compounds are almost identical
despite their structural diversity. Furthermore, the
detailed information on the binding sites has made
structure-based drug design possible.
Recently, we reported not only the syntheses of isoquin-
oline alkaloids, but also quantitative structure–activity
relationship studies of 3-arylisoquinolines that showed
top 1 inhibition and cytotoxicity.6–9 The current investi-
gation was undertaken to explore indenoisoquinolines
that are considered constrained forms of 3-arylisoquino-
lines as depicted in Figure 1.
11-Methylindenoisoquinoline 2b was chosen as a first
target compound because of the potent cytotoxicity of
16 and the ethylenediamine moiety was scheduled to
introduce on 5 position of compound 4 due to the strong
top 1 activity of 3.7
Irinotecan and topotecan are camptothecin derivatives,
and they are the only approved drugs for cancer treat-
ment.4 Although these two drugs were developed as
top 1 inhibitors, new compounds need to be found
because of the instability and short infusion times for
maximum activity of these drugs.5
Generally, a rigid molecule has little conformational
entropy but is to fit optimally into the receptor. We
planned to add one carbon atom between the B and C
rings of 3-arylisoquinolines to produce indenoisoquino-
lines. Although the Cushman group have reported pro-
foundly valuable researches on indenoisoquinolines,10
11-methyl or 5-ethylenediamino substituted compounds
have not yet been reported.
Keywords: Topoisomerase 1 inhibitor; Indeno[1,2-c]isoquinolines;
Antitumor agents; Docking study; FlexX; Synthesis of isoquinolines.
The 3-arylisoquinolines 7a–c were synthesized by the
previously reported lithiated toluamide-benzonitrile
cycloaddition method.11 N-methyl-o-toluamide 5a–b
*
Corresponding author. Tel.: +62 530 2933; fax: +62 530 2911; e-mail:
wjcho@jnu.ac.kr
0960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.04.064

3532
W.-J. Cho et al. / Bioorg. Med. Chem. Lett. 17 (2007) 3531–3534
reaction with ethylenediamine to give the desired 5-(2-
aminoethylamino)indeno[1,2-c]isoquinolin-11-one 4 in
good yield (Scheme 1).
C
R
A
B
R
NMe
NMe
Cytotoxicity experiments with the synthesized com-
pounds were performed by sulforhodamine B (SRB)
for A549 (lung), Col2 (colon), SNU-638 (stomach),
HT1080 (fibro sarcoma) cell lines and MTT assay for
HL-60 (myeloid leukemic) cells.14
O
O
3-Arylisoquinoline
Indenoisoquinoline
Me
Me
11
Me
Me
To determine top 1 catalytic activity, assays were per-
formed with supercoiled pBR32 DNA as the substrate
according to protocol.7 Semi-quantitative comparisons
of the inhibitory activities are shown in Table 1.
NMe
NMe
O
N
O
1
2b
O
As expected, indenoisoquinolines 2a–c exhibited more
potent cytotoxicity against these five cell lines than 3-
arylisoquinolines 8a–c, 9b–c, 10b–c, and 11b–c (Table
1). Indenoisoquinolines 2a–c displayed 1.70–8.5 lM
activity against the HL 60 cell line, and its activity ran-
ged from 2.58 to 9.0 lM against the other 4 cell lines. All
4-bromo-3-arylisoquinolines 8a–11c exhibited very weak
cytotoxicities. Moreover, none of these compounds had
strong top 1 inhibitory activity, including the indenoiso-
quinolines 2a–c. On the other hand, 5-(2-aminoethyla-
mino)indeno[1,2-c]isoquinolin-11-one 4 exhibited potent
top 1 inhibitory activity and cytotoxicity (Fig. 2).
1
D
C
A
4
B
N
5
NH
NH
4
3
NH2
NH2
Figure 1. The constrained structure of 3-arylisoquinolines to inde-
no[1,2-c]isoquinolines.
In the isoquinolinamine series such as 2a–c and 4, we
found that top 1 inhibitory activity did not correlate
particularly well with cytotoxicity. This discrepancy
indicates that top 1 may not be the sole biological target
for these compounds. Similar results could be found in
other reports15 and explained by the ability of the com-
pounds to penetrate the cell membrane and reach the
target as well as distribution differences within the cell.
Figure 2. Top 1 inhibitory activities of the compounds. lane P:
pBR322; lane T: pBR322 + topoisomerase I; lane C1: pBR322 +
topoisomerase I + camptothecin (0.1 mg/ml); lane C2: pBR322 + topo-
isomerase I + camptothecin (0.01 mg/ml); lanes 1–9 (prepared com-
pound number, 0.1 mg/ml): 1(10b), 2(11b), 3(11c), 4(9b), 5(9c), 6(2a),
7(2b), 8(2c), and 9(4).
Therefore, we used the FlexX docking program, a
molecular modeling tool, to determine the top 1 inhibi-
tory activity of 4. Computational calculations were per-
formed by Sybyl molecular modeling software, version
7.2.5, supplied by Tripos Associates, operating under
Red Hat Linux 4.0 with an IBM computer (Intel Pen-
tium 4, 2.8G CPU, 1G memory).
was treated with n-BuLi to give the anions, which were
then reacted with benzonitrile 612 to afford the 3-aryliso-
quinolines. The obtained compounds were treated with
MeI/60% NaH or PMBCl/K2CO3 to give N-alkylated
compounds 7a–c without detecting O-alkylated ones.
Selective bromination of 7a–c was accomplished with
NBS/ACCN in CCl4 to provide 8a–c in good yield.
P-Methoxybenzyl groups were removed from 8a–c by
DDQ oxidation to give the allyl alcohols 9a–c, which
were then oxidized with PDC to afford the aldehydes
10a–c in excellent yield. In this reaction, p-methoxyben-
zyl group on the N of the amide group was not affected
by DDQ oxidation. Wittig reaction of aldehydes 10a–c
with Ph3PCH3Br in the presence of n-BuLi was carried
out to give the styrenes 11a–c in good yield (62–95%).
Styrenes 11a–c were treated with n-Bu3SnH in the
presence of azobiscyclohexanecarbonitrile (ACCN) to
provide the desired 6,11-dihydro-5H-indeno[1,2-c]iso-
quinolin-5-ones 2a–c in 55–88% yield. This result could
be explained by Baldwin’s rule that the 5-exo-trig path-
way is favored over 6-endo-trig in a general way during
ring closure reactions.13 Commercially available 6-oxa-
benzo[a]fluorine-5,11-dione 12 was efficiently converted
to indenoisoquinoline 13 by treatment with NH4OH;
13 was then reacted with POCl3 followed by substitution
The X-ray crystallographic structure of the indenoiso-
quinoline (MJ238)-DNA-top 1 complex3 (PDB:1SC7)
in Protein Data Bank was selected for the docking study.
The bond to the disconnected phosphoester of G12 in
1SC7 was reconstructed, and the SH of G11 on the scis-
sile strand was changed to OH. The position and orien-
tation of the ligand should be considered to determine
whether it could protect the religation of the G11 hydro-
xyl group to the PTR723-phosphoester bond or not. For
this, we defined the receptor descriptor file based on
indenoisoquinoline molecule (MJ238) intercalated in
the ternary complex. The active site was defined with a
˚
6.5 A radius based on indenoisoquinoline. DNA nucleo-
tides including G12, G11(+1), T10(À1), T9 on the scissile
strand and C112, A113, A114 on the non-scissile strand
were selected as heteroatoms for RDF file preparation.
After running FlexX, 30 docked conformers were dis-
played as a molecular spread sheet ranking the scores.

W.-J. Cho et al. / Bioorg. Med. Chem. Lett. 17 (2007) 3531–3534
3533
PMBO
PMBO
NC
R
Me
NHMe
1) n-BuLi, THF
R1
a: R1=H, R2= Me
b: R1=Me, R2=Me
c: R1=Me, R2=PMB
+
2) MeI, NaH or
PMBCl,K2CO3
N
R2
7
6
O
5
O
a: R=H
b: R=Me
H
O
PMBO
Br
HO
Br
Br
R1
R1
NBS
PDC
R1
ACCN
DDQ
N
O
R2
N
CH2Cl2
R2
H2O
CH2Cl2
CCl4
hv
N
R2
O
10
8
9
O
a: (73%), b; (90%), c: (73%)
a: (73%), b: (69%), c: (81%)
a: (90%), b: (88%), c: (89%)
Me
Br
R1
n-Bu3SnH
ACCN
R1
Ph3P+CH3Br-
n-BuLi,THF
N
R2
Toluene
N
R2
O
O
2
11
a: (95%), b: (62%), c: (89%)
a: (88%), b: (39%), c: (29%)
O
O
O
O
POCl3
y. 73%
NH4OH
ethylenediamine
K2CO3/ DMF
y. 68%
N
DMF
O
N
NH
HCl
y. 77%
HN
O
Cl
O
NH2
12
13
4
14
Scheme 1. The synthesis of indenoisoquinolines.
Table 1. IC50 Cytotoxicity (lM) and top 1 activity of the compounds
No
Compound
R1
R2
A549
Col2
SNU-638
HL60
HT1080
Top 1a
1
2
8a
8c
H
Me
PMB
Me
PMB
Me
Me
PMB
Me
PMB
Me
Me
PMB
À
77.64
>100
25.06
43.37
60.67
37.58
18.46
49.35
84.27
2.58
>100
>100
99.3
23.4
67.4
33.8
69.8
61.1
16.8
4.90
7.4
>100
>100
18.7
16.4
16.4
22.0
19.8
14.0
14.5
7.40
9.0
76.7
>100
>100
>100
6.0
>100
>100
70.2
31.7
11.4
46.5
17.8
38.6
42.2
5.80
5.9
À
Me
Me
Me
H
À
3
9b
9c
À
4
À
5
10a
10b
À
6
Me
Me
Me
Me
H
17.6
>100
18.3
>100
1.70
7.1
++
À
7
10c
11b
8
++
À
9
11c
2a
10
11
12
13
14
15
16
À
2b
2c
Me
Me
À
2.98
3.84
À
4.40
>100
1.43
2.3
6.60
>100
1.68
2.2
8.50
>100
10.4
5.8
4.20
40.1
9.6
À
13
4
>100
0.85
1.9
+
À
À
++++
ntb
++++
Ellipticine
Camptothecin
À
À
4.3
0.080
À
À
0.069
0.045
0.098
0.018
a Activity is expressed semi-quantitatively as follows: À, very weak activity; ++, weak activity; ++++, similar activity as camptothecin.
bnt, not tested.
We selected the conformer with the best total score
(À33.56) and speculated the detailed binding patterns
in the cavity. In our model, the ligand intercalated well
between the À1 and +1 bases, parallel to the plane of
the base pairs. In a previous study, camptothecin was
found stacked between the pyrimidine ring of T10 and

3534
W.-J. Cho et al. / Bioorg. Med. Chem. Lett. 17 (2007) 3531–3534
lenediamine group of 4 had H-bond with the oxygen in
the G11 ribofuranose, which could strongly block the
religation process. At this stage, we should mention that
the carbonyl group on the C ring is essential for H-bond
with Arg 364, and the flat plane of the ligand is advanta-
geous for intercalation with base pairs.
In summary, we have presented a novel radical cyclization
that allows efficient synthesis of indenoisoquinolines. The
dramatic enhancement of top 1 inhibitory activity of the
synthesized compounds could be explained by the FlexX
docking program. In order for ligands to show top 1 relax-
ation activity, they should stabilize the ternary DNA-top
1 cleavable complex through intercalations with base
pairs, as well as H-bond with the enzyme in the cleavage
site. The ethylenediamine group of 4 had a strong H-bond
with the G11 furanose in the scissile strand. This result
suggests that blocking the religation of the G11 hydroxyl
group could be the main design point for novel top 1
inhibitors. Further study on the binding modes is now
being carried out, the details of which will be reported in
due course.
Figure 3. Superimposition of 4 (white), MJ238 of docked model
(green) and MJ238 of X-ray structure (red).
Acknowledgment
Financial support from the Korea Research Foundation
(KRF-2003-041-E00336) of Korea is greatly appreciated.
Figure 4. Docking model of 4.
References and notes
O-
O
1. Redinbo, M. R.; Stewart, L.; Kuhn, P.; Champoux, J. J.;
Hol, W. G. J. Science 1998, 279, 1504.
2. Rasheed, Z. A.; Rubin, E. H. Oncogene 2003, 22, 7296.
3. Staker, B. L.; Feese, M. D.; Cushman, M.; Pommier, Y.;
Zembower, D.; Stewart, L.; Burgin, A. B. J. Med. Chem.
2005, 48, 2336, and references cited therein.
4. Hardman, W. E.; Moyer, M. P.; Cameron, I. L. Antican-
cer Res. 1999, 19, 2269.
O
N
O
5. Laco, G. S.; Collins, J. R.; Luke, B. T.; Kroth, H.; Sayer,
J. M.; Jerina, D. M.; Pommier, Y. Biochemistry 2002, 41,
1428.
MJ-238
Figure 5. Structure of MJ-238.
6. Cho, W. J.; Park, M. J.; Chung, B. H.; Lee, C. O. Bioorg.
Med. Chem. 1998, 8, 41.
7. Cho, W. J.; Kim, E. K.; Park, I. Y.; Jeong, E. Y.; Kim, T.
S.; Le, T. N.; Kim, D. D.; Lee, E. S. Bioorg. Med. Chem.
2002, 10, 2953.
8. Cho, W. J.; Min, S. Y.; Le, T. N.; Kim, T. S. Bioorg. Med.
Chem. Lett. 2003, 13, 4451.
9. Le, T. N.; Gang, S. G.; Cho, W. J. J. Org. Chem. 2004, 69,
2768.
10. Morrell, A.; Antony, S.; Kohlhagen, G.; Pommier, Y.;
Cushman, M. J. Med. Chem. 2006, 49, 7740.
11. The structures of all synthesized compounds were con-
firmed by a spectral data.
12. Cho, W. J.; Kim, I. J.; Park, S. J. Bull. Korean Chem. Soc.
2000, 21, 1035.
the purine ring of G11 in the minor groove. To validate
our docking model, the newly sketched and minimized
MJ238 was first docked using the defined RDF file and
then compared to the original complex structure of
MJ238 in 1SC7 as depicted in Figure 3. Our docked mod-
el (green) was almost identical to MJ238 (red) between
the À1 and +1 bases. The distances of carboxylates and
ketones between our docked structure and MJ238 were
˚
2.063 and 0.754 A, respectively, and their H-bonds with
Asn 352 and Arg 364 were also detected. Thus, com-
pound 4 was docked using the same confirmed RDF file
to obtain the reasonable model shown in Figure 3 (white)
and Figure 4. Compound 4 intercalated at almost the
same position as MJ238 between À1 and +1 bases with
positioning of the D ring toward the minor groove
(Fig. 5). We observed the p–p stacking interaction be-
tween the G11 purine and the B and C rings of ligand
4, and the T10 pyrimidine and A and B rings of 4. As
shown in the X-ray complex of MJ238, the ketone group
of 4 had H-bond with Arg 364. Interestingly, the ethy-
13. Yamamoto, Y.; Ohno, M.; Eguchi, S. J. Org. Chem. 1996,
61, 9264.
14. Rubinstein, L. V.; Shoemaker, R. H.; Paull, K. D.; Simon,
R. M.; Tosini, S.; Skehan, P.; Scudiero, D. A.; Monks, A.;
Boyd, M. R. J. Nat. Cancer Inst. 1990, 82, 1113.
15. Nagarajan, M.; Morrell, A.; Ioanoviciu, A.; Antony, S.;
Kohlhagen, G.; Agama, K.; Hollingshead, M.; Pommier,
Y.; Cushman, M. J. Med. Chem. 2006, 49, 6283.