Molecular modelling studies, synthesis and biological activity of a series of novel bisnaphthalimides and their development as new DNA topoisomerase II inhibitors

Article abstract of DOI:10.1016/j.bmc.2008.11.024

A series of bisnaphthalimide derivatives were synthesized and evaluated for growth-inhibitory property against HT-29 human colon carcinoma. The N,N′-bis[2-(5-nitro-1,3-dioxo-2,3-dihydro-1H-benz[de]-isoquinolin- 2-yl)]propane-2-ethanediamine (9) and the N,N′-Bis[2-(5-nitro-1,3-dioxo-2,3-dihydro-1H-benz[de]-isoquinolin- 2-yl)]butylaminoethyl]-2-propanediamine (12) derivatives emerged as the most potent compounds of this series. Molecular modelling studies indicated that the high potency of 12, the most cytotoxic compound of the whole series, could be due to larger number of intermolecular interactions and to the best position of the naphthalimido rings, which favours π-π stacking interactions with purine and pyrimidine bases in the DNA active site. Moreover, 12 was designed as a DNA topoisomerase II poison and biochemical studies showed its effect on human DNA topoisomerase II. We then selected the compounds with a significant cytotoxicity for apoptosis assay. Derivative 9 was able to induce significantly apoptosis (40%) at 0.1 μM concentration, and we demonstrated that the effect on apoptosis in HT-29 cells is mediated by caspases activation.

Full text of DOI:10.1016/j.bmc.2008.11.024

Bioorganic & Medicinal Chemistry 17 (2009) 13–24
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Molecular modelling studies, synthesis and biological activity of a series of novel
bisnaphthalimides and their development as new DNA topoisomerase
II inhibitors
a,
Rosanna Filosa a, , Antonella Peduto , Simone Di Micco a, , Paolo de Caprariis a, Michela Festa a,
*
Antonello Petrella a, Giovanni Capranico b, Giuseppe Bifulco a,
*
a Department of Pharmaceutical Science, University of Salerno, via Ponte Don Melillo, variante 11C, 84084 Fisciano, Italy
b Department of Biochemistry G. Moruzzi, University of Bologna, via Irnerio 48, 40126 Bologna, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of bisnaphthalimide derivatives were synthesized and evaluated for growth-inhibitory property
against HT-29 human colon carcinoma. The N,N0-bis[2-(5-nitro-1,3-dioxo-2,3-dihydro-1H-benz[de]-iso-
quinolin-2-yl)]propane-2-ethanediamine (9) and the N,N0-Bis[2-(5-nitro-1,3-dioxo-2,3-dihydro-1H-
benz[de]-isoquinolin-2-yl)]butylaminoethyl]-2-propanediamine (12) derivatives emerged as the most
potent compounds of this series. Molecular modelling studies indicated that the high potency of 12,
the most cytotoxic compound of the whole series, could be due to larger number of intermolecular inter-
Received 28 October 2008
Accepted 11 November 2008
Available online 18 November 2008
Keywords:
Anticancer naphthalimides
DNA topoisomerase II inhibitors
actions and to the best position of the naphthalimido rings, which favours p–p stacking interactions with
purine and pyrimidine bases in the DNA active site. Moreover, 12 was designed as a DNA topoisomerase II
poison and biochemical studies showed its effect on human DNA topoisomerase II. We then selected the
compounds with a significant cytotoxicity for apoptosis assay. Derivative 9 was able to induce signifi-
cantly apoptosis (40%) at 0.1
lM concentration, and we demonstrated that the effect on apoptosis in
HT-29 cells is mediated by caspases activation.
Ó 2008 Elsevier Ltd. All rights reserved.
1. Introduction
N
N
Anticancer naphthalimides constitute an important class of
drugs characterized by a high cytotoxic activity upon a variety of
murine and human tumour cells.1 These drugs perform their bio-
logical activity either by forming a DNA-intercalator-topoisomer-
ase II (topo II) ternary complex or by inhibiting other enzymes
and/or transcription factors that act on DNA. The strong interac-
tions with DNA play a crucial role for their pharmacological prop-
erties. Significant examples include compounds such as
monomeric naphthalimides (1: amonafide 2: mitonafide) and bis-
naphthalimides (3: LU 79553-elinafide, 4: DMP 840-bisnafide).
Nowadays, it is accepted that the planar system of these drugs
can intercalate the DNA base pairs. In the case of amonafide and
mitonafide the antitumour activity is closely related to their ability
to poison human DNA topo II. Moreover, bisnafide analogues kill
eukaryotic cells by stabilizing the cleavage complex of topo II with
O
N
O
O
1
N
O
NH2
NO2
2
O
O
N
N
N
H
N
H
O
O
3
O
O
H
H
N
Me
N
N
NO2
* Corresponding authors. Tel.: +39 089969749; fax: +39 089 962828 (R.F.), Tel.:
+39 089969741 (G.B.).
O2N
N
H
H
Me
O
O
E-mail addresses: rfilosa@unisa.it (R. Filosa), bifulco@unisa.it (G. Bifulco).
Contributed equally to this work.
4
0968-0896/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2008.11.024

14
R. Filosa et al. / Bioorg. Med. Chem. 17 (2009) 13–24
Table 1
The synthesis of compounds 5–10 was accomplished by linking
First series of bisnaphthalimides derivatives
the two heteroaromatic moiety employing a variety of dicationic
linker chains where the length, rigidity, and charge density on
the chain are varied.
In order to gain new detailed molecular insights into their bind-
ing to DNA, we have studied the compounds 5–10 by docking
calculations.
Following the docking results observed for 5–10, we have fo-
cused our attention on designing new topo II inhibitors. We have
rationalized, through virtual screening, the length and chemical
structure of the aminoalkyl linker to improve the interactions with
the biological target.8
X
9
4'
9'
8
O
O
N
O
5'
10
3
3'
6'
7
6
11
2'
1'
11'
10'
1
N
O
Z
2
7'
4
8'
5
X
H
12
15'
13'
N
5: X= NO2; Z=
6: X= NO2; Z=
14'
14
N
H
13
15
12'
2. Chemistry
14
N
12
Polyamines 14a–c from which the bisnaphtalimides 6, 7 and 8
were prepared and then were carried out following the method
outlined in Scheme 1. Thus, the two nitrogens of heterocycles
piperazine, cis-dimethylpiperazine, 3,8-diazabicyclo[3.2.1]octane
(DBO)9 were treated with chloroacetonitrile, in the presence of so-
dium carbonate, to yield the dinitriles 13a–c. Reduction of these
with lithium aluminium hydride led to the corresponding amines
14a–c.
13'
13
13
N
14'
12'
14
N
12
7: X= NO2; Z=
13'
N
14'
12'
The synthesis of ( )-(trans)-N1,N2-bis(2-aminoethyl)cyclopro-
pane-1,2-diamine 19 involved the cyclopropane-1,2-dicarboxylic
acid 15 as the starting material10 (Scheme 2). The conversion of
dicarboxylic acid 15 to diamine moiety 16 was accomplished by
a modified Curtius reaction involving the formation of carboxylic
azides and a final treatment with aqueous HCl.11 The diamine 16
was converted into the corresponding acylated derivative 17, at
each end, with Boc-b-alanine in the presence of carbonyldiimidaz-
ole (CDI). Treatment of 17 with HCl and subsequent borane reduc-
tion afforded in high yield the corresponding polyamine 19.
The rest of amines triethylenetetramine and N,N0-bis(3-amino-
propyl)-ethylenediamine used for synthesis of compounds 5 and
9 were commercially available.
14
N
12
13
8: X= NO2; Z=
9: X= NO2; Z=
13'
14'
N
12'
13'
14'
H
N
15
13
12'
12'
15'
N
12
H
14
H
N
14'
15'
N
10: X= NO2; Z=
13'
H
The preparation of the corresponding bisnaphtalimides 5–10
was achieved following the synthetic protocol by reaction of 3-ni-
tro-1,8-naphthalic anhydride with the corresponding polyamine
(see Scheme 3).
DNA.2 However, this is still controversial as previous studies have
reported that elinafide is not a poison of topo II.3–6
Most of the research has been traditionally focused on the mod-
ification of the naphthalimido ring to enhance anticancer activities
through increased DNA-binding and cleavage.7
3. Structure–activity relationships for cellular growth
inhibition
In order to develop more active DNA-binding antitumour
agents, and to clarify the influence of the aminoalkyl linker chain
in their intricate modes of action, a novel series of bisnaphthali-
mides have been synthesised.
The first series (Table 1) were originally designed and synthe-
sized as bisintercalating agent, starting from previous findings on
the structural requirements for DNA-binding of 3 and 4.
3.1. Cytotoxicity
The in vitro cytotoxicity of all the bisnaphthalmido derivatives
was studied in the human colon adenocarcinoma HT-29 cell line
as described in methods section.
Compounds 5–10 were tested as methanosulfonate salts.
H2N
NH2
b
A
a
A
+
A
Cl
N
N
N
14 a-c
13 a-c
HN
HN
HN
A:
NH
NH
NH
14c
14a
14b
Scheme 1. Reagents and conditions: (a) Na2CO3/EtOH, 18 h,
D; (b) LiAlH4/THF, 4 h, D.

R. Filosa et al. / Bioorg. Med. Chem. 17 (2009) 13–24
15
O
O
a
b
BocHN
NHBoc
N
H
N
HOOC
COOH
H2N
NH2
2
2
H
17
16
15
O
O
c
d
NH2
H2N
3
3
N
N
H
H2N
N
N
NH2
2
2
H
H
H
19
18
Scheme 2. Reagents and conditions: (a) i—SOCl2, benzene, 60–70 °C, 24 h; ii—NaN3/H2O, acetone, 0 °C, 2 h; iii—toluene
alanine, CDI/CH2Cl2 0 °C-rt, 18 h; (c) HCl gas, MeOH, 1 h, 0 °C, 18 h rt; (d) BH3–THF/THF, , 24 h.
D
, 6 h; iv—HCl concd, EtOH 95%,
D
, 5 h; (b) N-Boc-b-
D
O
O
O
a
+
O
H2N
A
NH2
O
n
n
N
A
N
n=2, 3
NO2
O2N
NO2
n
n
O
O
5-10
Scheme 3. Reagents and conditions: toluene/EtOH, 4 h,
D
.
Table 2
Cytotoxic potency of mono- and bisnaphthalimides 5–10 in the human HT-29 cell line
a
Compound
IC50 HT-29
5
6
0.200
2.1
7
6.1
8
7.3
9
10
0.501
3.3
Elinafide
0.038
a
IC50: the drug concentration (
lM) that reduces the cell number of treated
samples to 50% of control cultures.
The activity of the cytotoxic drugs and IC50 values were deter-
mined after 72 h of drug exposures (Table 2). The IC50 values of
the tested compounds range from 0.04 to 6.1 lM.
Figure 1. DNA cleavage stimulated by bisnaphthalimides 5, 9 and 10. Symbols are
as follows: C, control DNA; T, topo II only; m, molecular DNA markers; VM-26,
teniposide; A, Amonafide; ELI, elinafide. Compounds were tested at the indicated
doses with or without topo II
a
a.
Thus, the compounds containing a linear carbon chain (5 and 9)
were more potent than the constrained analogues (6, 7, 8 and 10).
Compound 5 was the most active derivative, in which a linear ami-
noalkyl linker with two methylene groups separate chromophores
from the central ethylendiamine.
The removal of flexibility from the linker chain by incorporating
a piperazine, cis-dimethylpiperazine or DBO (6, 7 and 8) deter-
mined a decrease of cytotoxicity.
II-mediated DNA cleavage was studied using the pBR322 DNA as
a substrate for the recombinant human DNA topo II
a.
The data in Figure 1 show that derivates 5, 9 and 10 were inef-
fective in stimulating DNA cleavage, even though 9 and 10 some-
what showed a slight smear more evident than elinafide, that
may correspond to cleaved DNA fragments at the lowest tested
dose (1 lM). It is well known that topo II poisons with a strong
On the contrary compound 9, with the longer linker chain sep-
aration distance (11.0 Å) showed a good activity at nanomolar
range concentration in HT-29 cells.
DNA-binding activity repress topo II-mediated DNA cleavage at
high concentrations in vitro and in vivo.12
The findings thus show that the tested compounds are margin-
ally active as topo II poisons, suggesting a different mechanism of
action.
When the central ethylendiamine was replaced by a cyclopro-
pane, (10) the cytotoxic activity decreased.
3.2. Mechanism of action studies
3.2.2. Cell cycle effects
To investigate the effects of these compounds in more detail, we
examined the activity on proliferation and cell cycle progression in
HT-29 cells.
The analysed compounds are not able to block cell cycle pro-
gression of HT-29 cells (data not shown).
3.2.1. Topoisomerase inhibition
With respect to the molecular activity the first investigation
dealt with the ability to interfere with DNA functions.
We analysed the possibility that this class of compounds could
poison topo II. The relative drug activity in stimulating topo

16
R. Filosa et al. / Bioorg. Med. Chem. 17 (2009) 13–24
Figure 2. Apoptosis detection by propidium iodide (PI) staining of hypodiploid nuclei. HT-29 cells were incubated with 9 and 10 compounds (0.1–5 mM) for 48 h. Treatments
of 9 at 0.1 M (A) and 10 at 0.5
M (C) induce a significant apoptosis off HT-29 cells (***P < 0.001 vs CTR). To study involvement of caspases in 9 and 10 induced apoptosis a
l
l
general caspases inhibitor (Z-VAD-FMK) was administrated 30 min before both compounds. Z-VAD-FMK (50
l
M) inhibits effect of 9 (B,°°°P < 001 vs 9) and 10 (D,°°°P < 0.001)
on apoptosis of HT-29 cells. Data shown are representative of three experiments performed in triplicate.
3.2.3. Pro-apoptotic effects of compounds 9 and 10 on human
colon carcinoma cells
In order to understand the mechanism involved in cellular
death, compounds 5–10 were tested as pro-apoptotic agents.
Among all compounds analysed, compounds 9 and 10 showed
pro-apoptotic effects. We have checked the activity of compounds
on apoptosis using propidium iodide staining by flow cytometry as
described in methods section. HT-29 cells were incubated with
compounds 9 and 10 at different concentrations (0.1–5 lM) for
48 h. Results in Figure 2 showed a significant increase in apoptotic
cells after treatment with 9 (Fig. 2A) and 10 (Fig. 2C). These effects
are dependent to concentration used and the compound 9 was
Figure 3. Schematic representation of DNA Models (A, B and C) used in the docking
calculations. Sequence and numbering of two 8mers (A and B) and 6mer (C) DNA
duplexes are reported. The arrows indicate the intercalation points.
more able to induce a significant apoptosis (40%) at 0.1 lM concen-
tration. In addition we demonstrated that pro-apoptotic effects of 9
was mediated by caspases activation. Cells were treated with gen-
eral caspases inhibitor Z-VAD-FMK (50 lM) for 30 min before 9
(Fig. 2B) and 10 (Fig. 2D). Z-VAD-FMK inhibits significantly the
pro-apoptotic effects of both compounds to demonstrate involve-
ment of caspases in their apoptotic effect.
For sake of simplicity we will describe in this section only the
detailed docking results obtained with Model B.
The partial charges of the ligands (3 and 5–10) were calculated
at DFT B3LYP level and 6-31G+(d) basis set using the ChelpG14
method for population analysis and were used in the subsequent
docking calculations.
In the theoretical studies, we considered the amine functional-
ity as protonated at physiological pH and, consequently, positively
charged in the calculations.
3.2.4. DNA-binding properties
With the aim to explore the structural determinants responsible
for the activity of this new series of bisnaphthalimides, a molecular
modelling study has been carried out for 5–10.
Docking studies were performed on 5–10 with two DNA models
(Fig. 3, Models A and B), using AutoDock 3.0.5 software.13
The Model A was derived by the NMR solution complex be-
tween the duplex d(CGCTAGGCG)-(GCGATCCGC) and the bis-thia-
zole orange, and the Model B was built from Model A (see
experimental section for details) substituting two central base
pairs.
Both models were used to evaluate the influence of the base se-
quence on the binding affinity.
DNA intercalators, like 3, are reported to show a GC sequence
preference2 and, as it will be shown below, our docking results
are in agreement (Table 3).
Table 3
Values of calculated inhibition constanta associated with the complex between 5–10
and Models A and B
Compound
Ki
Model A
Model B
3
5
6
7
8
9
10
1.51 ꢀ 10ꢁ14
1.99 ꢀ 10ꢁ13
3.37 ꢀ 10ꢁ12
2.24 ꢀ 10ꢁ14
3 ꢀ 10ꢁ14
2.39 ꢀ 10ꢁ15
6.62 ꢀ 10ꢁ14
7.75 ꢀ 10ꢁ13
8.83 ꢀ 10ꢁ15
6.57 ꢀ 10ꢁ15
3.45 ꢀ 10ꢁ15
1.03 ꢀ 10ꢁ15
9.09 ꢀ 10ꢁ15
5.35 ꢀ 10ꢁ15
a
The inhibition constant is expressed as M and it is calculated at 293.15 K.

R. Filosa et al. / Bioorg. Med. Chem. 17 (2009) 13–24
17
4 of a C3, whereas the other nitro group is hydrogen bonded to the
NH2 in 2 of G5 (Fig. 4b).
The linker chain results collocated into the minor groove, with
the piperazine ring arranged in a parallel fashion with respect to
the deoxyribose groove walls of the DNA, providing hydrophobic
contacts. Moreover, in compound 6, the linker amine groups are
not involved in hydrogen bonds, causing a lower predicted binding
affinity compared to 5 (Fig. 4b). For compound 7, as observed for 5,
nitro functionalities do not establish hydrogen bonds; on the other
hand, NH+ 140 of piperazine makes an hydrogen bond with the N3
and anomeric oxygen of deoxyribose of G5 (Fig. 4c), while the
remaining amino group points outside the minor groove. The di-
methyl piperazine ring shows extended hydrophobic contacts, fac-
ing the purine/pyrimidine bases of the minor groove (Fig. 4c).
While the nitro group in 5 of compound 8 presents the same inter-
action observed for 6 (Fig. 4d), the NH+ 140 of piperazine is hydro-
gen bonded to the N3 of guanine 5, as found for 7. NH+ 140 also
faces the external aqueous environment, as observed in 7. More-
over the 3,8-diazabicyclo[3.2.1]octane gives more extensive Van
der Waals interactions with its macromolecular counterpart.
While the linker length is maintained comparable, the struc-
tural difference between 5 and 6–8 is characterized by a linker
tightening due to the insertion of a piperazine ring in the central
part of the chain. Such bulky moiety, especially for 7 and 8, contrib-
utes to complex stability through several Van der Waals interac-
tions with the minor groove. In fact, the compounds 7 and 8
show a better binding affinity for the biological target with respect
to 5 (Table 3).
Figure 4. 3D models of the interactions between 5–10 (respectively, a, b, c, d, e and
f) and the Model B. The DNA is represented by molecular surface and tube (coloured
by atom type: C, grey; polar H, sky blue; N, dark blue; O, red). All ligands are
depicted by sticks (green for 5, light yellow for 6, esmerald for 7, blue for 8, orange
for 9 and yellow for 10) and balls (by atom type). The figure highlights essential
interactions: the naphthalimides are intercalated between base pairs and the linker
interacts with the minor groove. The amino groups establishe hydrogen bonds
(indicated by yellow lines) with nucleotides in the minor groove.
Structure 9 establishes two hydrogen bonds by amino group 15
with the N3 of G13 (Fig. 4e). Compound 9 is similar to 5, but it has a
longer aminoalkyl linker length: 13.8 Å versus 11.1 Å of structure 5
(Fig. 5a).
In 9 there are three methylenes between the polycyclic moiety
and the nitrogen of the linker. This longer chain influences the
binding affinity to the DNA, because it allows a better intercalation
of both aromatic systems compared to the preceding compounds
(Fig. 5a).
In details, both planar portions of the ligand can make more
effective p–p interactions with DNA base pairs, and the aminoalkyl
The analysis of docking results reveals that all ligands share a
similar binding mode (Fig. 4). In details, the common structural
portion, constituted by a polycyclic system, is fundamental for
the line up of the complex with the biological target, through a
base pairs intercalation. The linker of two naphthalimides also
plays a crucial role in the affinity to nucleic acids, contributing to
the complex stability interacting with the minor groove of the
DNA.
chain establishes closer Van der Waals contacts and hydrogen
bonds with the minor groove of the nucleic acid.
The structure 5 establishes p–p interactions with DNA aromatic
Similar results are found for 10, because it presents approxi-
matelythesameaminoalkylchainlength(13.5 Å)of9(Fig. 5b). Asre-
ported in Figure 5b, 9 and 10 showed an overlapped binding mode,
with cyclopropane superimposed to central ethylenes of the amino-
alkyllinkerof9.Compound10establishesbyNH2+ 15thesameinter-
actions found for 9, and a further hydrogen bonds with the anomeric
rings, through intercalated naphthalimides (Fig. 4a) and these
interactions are extended thanks to the presence of the nitro
groups on the aromatic portions. The linker chain establishes Van
der Waals contacts and hydrogen bonds with nucleic acids. In par-
ticular, both amino groups interact with the anomeric oxygen of
deoxyribose and N3 of G5, whereas the NH2 140 forms a further
hydrogen bond with N3 of the same residue (Fig. 4).
Same considerations were applied for structures 6–8. The pla-
nar moiety forms Van der Waals interactions with DNA base pairs,
and the nitro groups are involved in further intermolecular interac-
tions (Fig. 4b, c and d).
+
oxygen of deoxyribose of G13 (Fig. 4f). The NH2 150 only interacts
with anomeric oxygen of deoxyribose of G5 residue. Both nitro
groups, differently from 9, are involved in hydrogen bonds.
In details, the nitro groups 5 and 50 interact with NH2 in 4 of C3
and C11, respectively (Fig. 4f). Moreover, the cyclopropane ring
faces the individual base pairs along the groove floor, contributing
by Van der Waals contacts to the complex stability.
In particular, for 6, the nitro functionality in 5 makes an hydro-
gen bond with NH2 in 4 of a C3, whereas the other nitro group is
hydrogen bonded to the NH2 in 2 of G5 (Fig. 4b).
Thus 10 exerts, compared to 9, a slightly more efficient binding
affinity (Table 3) for the biological target thanks to a larger number
of intermolecular interactions. We have also performed docking
calculations on elinafide (3) as reference structure. Comparing 3
with 10, we observed a similar binding mode. The NH2+ 15 encoun-
ters the same interactions found for 10, whereas the remaining
amino group 150 is hydrogen bonded to the N3 of G5 residue. In
the structure of elinafide the nitro functionalities and cyclopropane
are not present and can not contribute by their interactions to the
binding affinity improvement, causing a lower Ki value for 3.
The theoretical outcomes, obtained docking the ligands on
Model A, are in agreement with data collected with Model B. We
The linker chain results collocated into the minor groove, with
the piperazine ring arranged in a parallel fashion with respect to
the deoxyribose groove walls of the DNA, providing hydrophobic
contacts. Moreover, in compound 6, the linker amine groups are
not involved in hydrogen bonds, causing a lower predicted binding
affinity compared to 5 (Fig. 4b). Same considerations were applied
for structures 6–8. The planar moiety forms Van der Waals interac-
tions with DNA base pairs, and the nitro groups are involved in fur-
ther intermolecular interactions (Fig. 4b, c and d). In particular, for
6, the nitro functionality in 5 makes an hydrogen bond with NH2 in

18
R. Filosa et al. / Bioorg. Med. Chem. 17 (2009) 13–24
mechanism catalysed by this protein. The topo II makes two breaks
on both DNA strands with a four base pairs overhang and the en-
zyme becomes covalently linked to the 50 ends of the break sites.8
This transient covalent complex is the target of different drugs,
that stabilize the ternary complex through a base pairs intercala-
tion mechanism. As the choosen X-ray structural complex presents
the intercalation points separated by 4 base pairs, it is consistent
with our purpose to design topo II poisons.
The partial charges of the ligands are calculated by the same
theoretical level employed for 5–10 and are used in the docking
calculations.
Initially we have docked structure 3 as starting point for the
development of new drugs. The docking pose of 3 shows only
one chromophore portion intercalated between the DNA base
pairs, because the aminoalkyl linker does not cover a 4 base pairs
length. This finding has suggested us to lengthen the chain separat-
ing the two chromophore portions. We have tried various aminoal-
kyl linker (Table 4), and selected the structure 11.
The aminoalkyl chain of 11 differs from the other elongated
linkers for the number of amino groups. We have inserted two fur-
ther amino functionalities to favour the hydrogen bonds forma-
tions with nucleotides. The central amino groups are separated
by three carbon atoms as in 3. The choice of the amino functional-
ities 16 and 160 position along the chain, instead, is based on the
distance between two consecutive base pairs. The theoretical stud-
ies showed that the linker length permitted the contemporary
intercalation of both aromatic moieties of 11 and all amino groups
are involved in hydrogen bonds with their macromolecular coun-
terparts, gaining a better binding affinity (Ki = 5.22 ꢀ 10ꢁ18). The
aminoalkyl chain perfectly collocates in the minor groove. It fol-
lows the helical turn covering the 4 base pairs distance and estab-
lishes significant hydrophobic contacts (Fig. 6a).
+
The NH2 16 and 160 interact with O(@C) in 2 of C11 and C5,
respectively, whereas the functionality 19 with N3 of A4. The ami-
no group 190 establishes hydrogen bonds with carbonilyc oxygen of
T9, N3 of A10 and NH2 of G8.
The analysis of docking studies on 5–10 has showed the nitro
groups contribution to the complex stability by their interactions
with purine/pyrimidine bases. Thus we have designed structure
12 with the common polycycle system of 5–10.
Figure 5. Compounds 5 and 9 (a), 9 and 10 (b), superimpositions in the DNA-
binding site. The figure shows the influence of linker length on the contemporary
intercalation of chromophore moieties. The DNA is represented by tube (coloured
by atom type: C, grey; polar H, sky blue; N, dark blue; O, red). All ligands are
depicted by sticks (green for 5, orange for 9 and yellow for 10) and balls (by atom
type).
+
The aminoalkyl chain is involved in hydrogen bonds. The NH2
16 and 19 are hydrogen bonded to anomeric oxygen of deoxyribose
+
and N3 of A4 residue, and the NH2 16 makes further interactions
with NH2 of G2 and O(@C) of T3 (Fig. 6b).
The amino functionalities 190 and 160 establish hydrogen bonds
with anomeric oxygen of deoxyribose of C5 and A10 residues.
observed, in fact, a similar trend for ligands binding affinity. More-
over the comparison of docking results collected from Models A
and B, shows a GC sequence preference in accordance with data re-
ported in literature.2 We observed for 5–10 docking poses on Mod-
el B the amidic CO group of chromophore moiety facing the NH2 in
+
Moreover the NH2 160 makes an hydrogen bond with O(@C) of
T9 and with NH2 of G8. The nitro groups extend p–p interactions
with DNA base pairs, giving a slightly improvement of binding
affinity (Ki = 4.32 ꢀ 10ꢁ18). Compared to 5–10, compound 12
established more Van der Waals contacts and hydrogen bonds by
its longer linker chain reaching a better complex stability. The
experimental trials of growth-inhibitory properties, in fact, shows
for 12 the most interesting IC50 value, in accordance with the the-
oretical data (Table 5). Moreover the theoretical outcomes are in
agreement with the biological assays evaluating the Topoisomer-
ase II inhibition. Compounds 5, 9 and 10 do not inhibit the enzyme,
whereas 12 shows an inhibition of the activity (see Scheme 4).
2 of guanine residue. This spatial arrangement extends the
p–p
interactions and probably gives raise to the little preference for
GC sequences.
The predicted Ki are not perfectly in line with IC50 values
describing the growth-inhibitory properties (Table 2). This is prob-
ably due to the possibility that each compound could interfere
with more than a single biological process.
3.2.5. Design of novel topoisomerase II inhibitors
A crystal structure of a covalent topo II–DNA complex has not
been solved, thus we have used a simplified model of the biological
target in our docking calculations (Fig. 3, Model C).
It consists in a DNA duplex built from the X-ray structural
complex between the double strand d(CG(Br-U)ACG)-(GCATCG)
and two N-(2-(dimethylamino)ethyl)acridine-4-carboxamide. The
choice of this experimental structure is dictated from the reaction
3.2.6. Chemistry
In order to synthesize compounds 11 and 12 in which more het-
eroatoms in the linker chain were introduced, N-alkylation reac-
tion was chosen according to a modified method which was
recently reported.15
The common intermediate for the synthesis of two derivatives
is N1,N3,N7,N10-tetramesytil-1,3-bis(2-aminoethyl)-1,3-propanedi-

R. Filosa et al. / Bioorg. Med. Chem. 17 (2009) 13–24
19
Table 4
Second series of bisnaphthalimides derivatives
X
9
4
4'
8
O
O
N
O
5'
10
3
3'
6'
7
6
11
2'
1'
11'
10'
1
N
Z
2
7'
O
8'
9'
5
X
16
16'
H
N
H
N
14
14
18'
14'
18
18
12'
12
12
11: X= H; Z=
N
H
N
17
17
13
13
17'
17'
13'
15
15
15'
15'
H
19
19'
16
H
16'
H
18'
14'
12'
N
N
12: X= NO2; Z=
N
H
N
H
13'
19
19'
H
N
H
N
25: X= H; Z=
26: X= H; Z=
H
N
N
H
H
N
27: X= H; Z=
28: X= H; Z=
N
H
H
N
N
H
Table 5
10
lM (Fig. 7). Thus, our findings indicate that further structural
Growth-inhibitory properties for Bisnaphthalimides 11 and 12 compared to elinafide
and compound 5
modifications of the bisnaphthalimides series may lead to more
effective inhibition of human Top2a.
a
Compound
IC50 HT-29
5
11
0.200
6.1
4. Conclusion
12
Elinafide
0.160
0.038
The biological activity of naphthalimides is likely based on
strong interactions with DNA in the target cells. Thus, we have here
performed docking studies to rationalize the binding mode to nu-
cleic acid evaluating the structural determinants of the binding
affinities to the biological target. The analysis of docking results
on 5–10 indicates all ligands as bis intercalating agents. The com-
mon planar moieties of the ligands intercalate between base pairs
a
IC50: concentration of drug (
l
M) to reduce cell number to 50% of control
cultures.
amine 23. This was prepared reacting tetramine with mesitylene
chloride in pyridine at room temperature. N-alkylation of the latter
compound with O-tosylbutylnaphthalimides 22a–b, with caesium
carbonate in anhydrous DMF afforded the fully protected bis-
naphthalimidobutyl derivatives 24a–b which upon deprotection
with hydrobromic acid/glacial acetic acid in dichloromethane
afforded compounds 11 and 12. O-Tosylbutylnaphthalimides
22a–b were prepared by first reacting 1,8-naphthalic anhydride
with aminobutanol to give N-(4-hydroxybutyl)naphthalimides
21a–b which upon reaction with tosyl chloride gave 22a–b. It is
important to note that during the tosylation reaction, the best con-
dition to obtain a maximum yield is to use four times excess of to-
syl chloride in a small volume of solvent.16 Under other conditions
(e.g., with equimolar or 2 molar quantities of tosyl chloride), a mix-
ture of O-tosylbutylnaphthalimide and N-(4-chlorobutyl)naphthal-
imide derivatives is always formed thus reducing the overall yield
of the reaction and also, renders purification very difficult.
establishing p–p interactions with DNA aromatic rings, with nitro
groups extending these intermolecular interactions or forming
hydrogen bonds with nucleotides. The linker of two chromophores
also contributes to complex stability because it can form, along the
minor groove of the DNA, hydrogen bonds and Van der Waals con-
tacts (especially where rings are present as structural elements of
the linker chain).
The theoretical outcomes highlight that the linker chain length
plays a crucial role in the binding affinity to DNA. Increasing the
distance between two naphthalimides (in the series 5–10), we
reached more effective p–p interactions with base pairs and closer
Van der Waals contacts and hydrogen bonds with the minor
groove.
The docking results, obtained from 5–10, have been the starting
points to design new topo II poisons. Based on topoisomerase II
catalysed mechanism, by a virtual screening we have designed
the length and chemical structure (11 and 12) of the aminoalkyl
linker to improve the interactions with the biological target.
The docking results suggest that to stabilize the ternary com-
plex topo II–DNA-drug through a bisintercalator, it is fundamental
a linker chain covering the four base pairs overhang. Such linker
length causes a double intercalation in a position adjacent to the
protein–DNA covalent bond.
Among the latter compounds, derivative 12 showed the highest
cytotoxicity against HT-29 cells with a IC50 value of 0.16 lM. The
compound 11, in which the nitro group was removed, was also
cytotoxic but significantly less. Then, compound 12 was selected
for the DNA cleavage assay in the presence of recombinant human
topo IIa. The results reported in Figure 7 show that 12 is a topo II
poison with a potency intermediate between amonafide and elina-
fide. 12 stimulated topo II-mediated breakage of pBR322 DNA at

20
R. Filosa et al. / Bioorg. Med. Chem. 17 (2009) 13–24
X
O
OH
+
O
O
H2N
a
X=NO2, H
O
OH
X
N
H
N
H
O
N
H2N
NH2
21a X=H
21b X=NO2
c
b
O
OTosyl
Mts
N
Mts
N
X
N
Mts
Mts
+
N
N
H
O
H
23
22a X=H
22b X=NO2
d
O
Mts
N
Mts
N
O
N
O
N
O
X
N
Mts
N
Mts
X
24a X=H
24b X=NO2
e
O
O
N
O
H
N
H
N
N
X
N
H
N
X
H
O
4HBr
11 X=H
12 X=NO2
Scheme 4. Reagents and conditions: (a) H2O,
D, 40 °C; (b) tosyl-Cl, Et3N/CH2Cl2, D,
overnight; (c) Mts-Cl, Pyr, rt, 4 h; (d) CsCO3/DMF, 80 °C, 18 h; (e) HBr/gꢂCH3COOH,
CH2Cl2, rt, 12 h.
Figure 6. 3D models of the interactions between 11–12 (a and b, respectively) and
the Model C. The DNA is represented by molecular surface and tube (coloured by
atom type: C, grey; polar H, sky blue; N, dark blue; O, red). Two ligands are depicted
by sticks (yellow for 11, orange for 12) and balls (by atom type). The figure
highlights essential interactions: the naphthalimides are intercalated between base
pairs and the linker interacts with the minor groove and allow
a good bis
intercalation. The amino groups establish hydrogen bonds (indicated by yellow
lines) with nucleotides in the minor groove.
Figure 7. DNA cleavage stimulated by bisnaphthalimides 5, 9, 10 and 12. Symbols
The guidelines used to design new topo II poisons have been
corroborated by biological assays. In details, compounds 5, 9 and
10 do not inhibit the enzyme, whereas compound 12 shows a topo
II inhibition. Moreover, the described agreement between theoret-
ical and experimental data validates the reliability of our simplified
model used to project new topo II poisons.
To design new anticancer compounds bisintercalating the DNA,
the theoretical studies should suggest the use of a chromophore lin-
ker with suitable length to guarantee an optimum double intercala-
tion between base pairs. The introductions of amino functionalities
along the aliphatic chain and of donor/acceptor groups of hydrogen
bonds on the planar portions, should also contribute to the binding
affinity through a network of hydrogen bonds with the nucleotides.
are as follows: C, control DNA; T, topoisomerase II
markers; VM-26, teniposide; A, amonafide; ELI, elinafide. Compounds were tested
at 1 M and 10 M with or without enzyme.
a only; m, molecular DNA
l
l
TLC was carried out using plates coated with silica gel 60F
254 nm purchased from Merck (Darmstadt, Germany). 1H and 13C
NMR spectra were registered on a Brucker AC 300. Chemical shifts
are reported in ppm. The abbreviations used are follows: s, singlet;
d, doublet; dd double doublet; bs, broad signal. MS spectrometry
analysis ESI-MS was carried out on a Finnigan LCQ Deca ion trap
instrument. Elemental analysis was performed by Desert Analytics
(Tucson, AZ).
5.2. General procedure for synthesis of 13a–c
5. Experimental
5.1. Chemistry
To a solution of piperazine, 2,6-cis-dimethylpiperazine or DBO
(1 g) and chloroacetonitrile (2.4 equiv) in absolute ethanol
(50 mL), anhydrous sodium carbonate was added (4 equiv).
The reaction mixture was stirred at reflux for 4 h. The hot mix-
ture was filtered and the insoluble part washed with hot ethanol
All reagents were analytical grade and purchased from Sigma–
Aldrich (Milano, Italy). Flash chromatography was performed on
Carlo Erba silica gel 60 (230–400 mesh; CarloErba, Milan, Italy).

R. Filosa et al. / Bioorg. Med. Chem. 17 (2009) 13–24
21
and filtered again. The combined filtrate was evaporated to give a
residue which was submitted to flash chromatography. Elution
with CHCl3/MeOH 98/2 afforded desired compounds:
vent was evaporated, and the residue was washed with diethyl
ether and dried, to give the title compound 18 (98%) as a white
solid.
1,4-Bis-(cyanomethyl)piperazine 13a (71%) 1H NMR (CDCl3):
2.71 (s, 8H); 3.57 (s, 4H); 1,4-bis-(cyanomethyl)-2,6-cis-dimethyl
piperazine 13b with (61%) 1H NMR (CDCl3): 1.09 (s, 3H); 1.11 (s,
3H); 1.95 (d, 2H, J = 7.78 Hz); 2.73 (d, 2H J = 7.78 Hz); 2.75–2.77
(m, 2H); 3.50 (s, 2H); 3.77 (s, 2H); 3,8-bis-(cyanomethyl)-3,8-
diazabicyclo[3.2.1]octane 13c (63%) 1H NMR (CDCl3): 1.87–2.02
(m, 4H); 2.58 (dd, 2H, J = 2.19H, 7.78 Hz); 2.73 (d, 2H
J = 7.78 Hz); 3.36 (s, 2H); 3.72 (s, 2H); 3.49 (s, 2H).
1H NMR (D2O) 1.42 (t, 2H, J = 7.45 Hz); 2.60 (t, 4H, J = 7.02 Hz);
3.04 (t, 2H J = 7.45 Hz); 3.20 (t, 4H, J = 7.02 Hz).
5.8. ( )-(trans)-N,N0-bis(2-aminopropyl)cyclopropane-1,2-
diamine tetrahydrochloride (19)
A solution of 18 (450 mg, 1.3 mmol) in 15 mL of anhydrous tet-
rahydrofuran was added to 28.6 mL of 1 M borane–THF complex
solution. This mixture was refluxed for 24 h and cooled to room
temperature, and then 20 mL of methanol was added slowly and
refluxed for an additional 15 h. After filtration, the solvents were
evaporated and the residue dissolved in 5 mL of methanol. After
the mixture cooled in an ice bath, 1 mL of concentrated hydrochlo-
ric acid was added. This mixture was treated with diethyl ether,
and the solid formed was filtered to give the title compound 19
(57%) as a hygroscopic solid.
5.3. General procedure for synthesis of 14a–c
A solution of 13a–c (500 mg) in THF (20 mL) was added drop-
wise to a suspension of LiAlH4 (4.6 equiv) in THF (20 mL). The reac-
tion mixture was heated under reflux for 4 h and then cooled in ice,
and water (0.47 mL), 15% aqueous NaOH (0.47 mL), and water
(1.41 mL) were added sequentially. The resulting granular precipi-
tate was filtered and washed with Et2O. The combined filtrates
were evaporated under reduced pressure, and the residue was
chromatographed on silica gel (eluting with MeOH/CH2Cl2/NH4OH,
30:60:10) to give:
1H NMR (D2O): 1.41 (t, 2H, J = 7.45 Hz); 1.63 (m, 4H); 2.67 (m,
8H); 3.05 (t, 2H, J = 7.45 Hz).
5.9. General procedures for synthesis of compounds 5–10
1,4-Bis(aminoethyl)piperazine 14a (87%) 1H NMR (CDCl3): 1.6
(s, 2H); 2.45 (t, 4H, J = 7.45 Hz); 2.51 (br, 8H); 2.81 (t, 4H,
J = 7.45 Hz);
To a suspension of 3-nitro-1,8-naphthalic anhydride (2 equiv)
in toluene (5 mL), a solution of corresponding polyamine in etha-
nol was added drop wise.
The reaction mixture was refluxed for 4 h and then the precip-
itated solid was filtered and washed with ether and AcOEt to give
desired products.
1,4-Bis(aminoethyl)-2,6-cis-dimethyl piperazine 14b (73%) 1H
NMR (CDCl3): 1.10 (s, 3H); 1.12 (s, 3H); 1.51 (br, 8H); 1.90 (t, 2H,
J = 10.12 Hz); 2.40 (t, 2H, J = 7.45 Hz); 2.63–2.82 (m, 10H);
1,4-Bis(aminoethyl)-3,8-diazabicyclo[3.2.1]octane 14c (53%) 1H
NMR (CDCl3): 1.77–1.83 (m, 4H); 2.29 (d, 2H, J = 7.78 Hz); 2.61 (dd,
2H, J = 2.19 H, 7.78 Hz); 3.13 (s, 2H).
Compound 5: Yield (75%): 1H NMR (CDCl3): 3.42 (t, 4H,
J = 7.02 Hz); 3.58 (s, 4H); 4.29 (t, 4H, J = 7.02 Hz); 7.93 (t, 2H,
J = 7.23 Hz); 8.43 (d, 2H J = 7.23 Hz); 8.73 (d, 2H J = 7.23 Hz); 9.15
(d, 2H J = 2.19 Hz); 9.32 (d, 2H J = 2.19 Hz); 13C (DMSO) as meth-
anesulfonate salt: 38.2, 46.1; 122.8; 123.2; 125.4; 130.1; 130.2;
130.3; 132.4; 135.3; 147.1; 146.4; 164.1; 165.3. MS (ESI) m/z:
597.3. Mp (2 CH3SO3H salt) 251.4 °C; Anal. Calcd for C30H24N6O8:
C, 60.40; H, 4.06; N, 14.09; O, 21.46. Found: C, 60.47; H, 4.15; N,
14.22; O, 21.39.
5.4. ( )-trans-l,2-Cyclopropanedicarboxylic acid (15)
Compound 15 was prepared according to the method of Payne10
affording crystals, mp 176–177 (lit.17 mp 177–177.5 °C).
5.5. ( )-trans-l,2-Cyclopropanediamine dihydrochloride (16)
Compound 6: (Yield 73%): 1H NMR (CDCl3) 2.64 (s, 8H); 2.74 (t,
4H, J = 6.80 Hz); 4.39 (t, 4H, J = 6.80 Hz); 7.97 (t, 2H, J = 7.23 Hz);
8.44 (d, 2H, J = 7.23 Hz); 8.86 (d, 2H, J = 7.23 Hz); 9.16 (d, 2H,
J = 2.19 Hz); 9.34 (d, 2H, J = 2.19 Hz). 13C (DMSO) as methanesulfo-
nate salt: 37.2; 47.2; 50.7; 55.2; 122.7; 123.2; 124.7; 130.1; 130.2;
130.3; 131.3; 134.9; 137.4; 146.2; 164.4; 165.2. MS (ESI) m/z:
622.7. Mp (2 CH3SO3H salt) 298 °C dec. Anal. Calcd for
C32H26N6O8: C, 61.73; H, 4.21; N, 13.50; O, 20.56. Found: C,
61.69; H, 4.12; N, 14.72; O, 20.60.
Compound 16 was prepared according to the method of
Witiak11 affording diamine dihydrochloride 16 as white crystals,
mp 210 °C dec.
5.6. ( )-(trans)-N,N0-Bis[[(tert-butoxycarbonyl)amino]
propionyl]-ciclopropan-1,2-diamine (17)
A
solution of N-(tert-butoxycarbonyl)-b-alanine (0,1 g,
0,57 mmol) in 5 mL of dichloromethane was treated at 0 °C with
1,10-carbonyldiimidazole (93 mg, 0,57 mmol). After the mixture
was stirred in an ice bath for 1.5 h, 16 (40 mg, 0,28 mmol) in 2 mL
of dichloromethane was added and stirred for an additional 1 h at
0 °C. After keeping at room temperature overnight, the solution
was washed with a saturated solution of Na2CO3 followed by water.
This was dried on anhydrous magnesium sulfate and filtered, the
solvent was evaporated and the residue was chromatographed on
silica gel eluting with CH2Cl2/MeOH 95:5 to give 17 (70%).
Compound 7: (Yield 82%): 1H NMR (CDCl3) 1.34 (s, 3H); 1.36 (s,
3H); 2.03 (t, 2H, J = 9.87 Hz); 2.70 (t, 2H, J = 7.13 Hz); 2.76–2.88 (m,
2H); 2.92–3.05 (m, 4H); 4.24–4.34 (m, 2H); 4.38 (t, 2H, J = 7.13 Hz);
7.97 (t, 2H, J = 7.23 Hz); 8.44 (d, 2H J = 7.23 Hz); 8.86 (d, 2H,
J = 7.23 Hz); 9.16 (d, 2H, J = 2.19 Hz); 9.34 (d, 2H, J = 2.19 Hz). 13C
(DMSO) as methanesulfonate salt: 14.3; 37.2; 46.2; 50.1; 55.2;
59.3; 64.6; 122.8; 123.2; 125.4; 126.8, 130.1; 130.2; 130.3;
132.3; 132.4; 135.4; 137.1; 146.3; 164.2; 165.3. MS (ESI) m/z:
651.3. Mp (2 CH3SO3H salt) 233.0 °C. Anal. Calcd for C34H30N6O8:
C, 62.76; H, 4.65; N, 12.92; O, 19.67. Found: C, 62.82; H, 4.79; N,
13.09; O, 19.59.
1H NMR (CDCl3): 1.40 (t, 2H, J = 7.45 Hz); 1.43 (s, 18H); 2.42 (t,
4H, J = 7.02 Hz); 3.03 (t, 2H J = 7.45 Hz); 3.24 (t, 4H, J = 7.02 Hz).
Compound 8: (Yield 83%): 1H NMR (CDCl3) 2.31 (d, 2H,
J = 9.98 Hz); 2.61–2.67 (m, 8H); 2.72 (d, 2H, J = 9.98 Hz); 3.23 (s,
2H); 4.28–4.34 (m, 4H); 7.91 (t, 2H, J = 7.13 Hz); 8.34 (d, 2H,
J = 7.13 Hz); 8.78 (d, 2H, J = 7.13 Hz); 9.14 (d, 2H, J = 2.19 Hz);
9.32 (d, 2H, J = 2.19 Hz).
5.7. ( )-(trans)-N,N0-Bis(aminopropionyl)-ciclopropan-1,2-
diamine tetrahydrochloride (18)
A solution of 17 (100 mg, 0.26 mmol) in 8 mL of dry methanol,
at 0 °C, was saturated with hydrochloric acid (gas). After 1 h at 0 °C,
the mixture was stirred at room temperature overnight. The sol-
13C (DMSO) as methanesulfonate salt: 22.2; 25.1; 37.2; 50.1;
55.2; 59.3; 60.1; 122.8; 123.2; 125.4; 126.8, 130.1; 130.2; 130.3;

22
R. Filosa et al. / Bioorg. Med. Chem. 17 (2009) 13–24
132.3; 132.4; 135.4; 137.1; 146.3; 164.2; 165.3.MS (ESI) m/z:
648.95. Mp (2 CH3SO3H salt) 252.8 °C. Anal. Calcd for C34H28
N6O8: C, 62.96; H, 4.35; N, 12.96; O, 19.73. Found: C, 62.84; H,
4.73; N, 13.10; O, 19.55.
5.14. N1,N3,N7,N10-tetramesytil-1,3-bis(2-aminoethyl)-1,3-
propanediamine (23)
N,N0-Bis(2-aminoethyl)-1,3propanediamine (1.0 g, 6.05 mmol)
was dissolved in anhydrous pyridine (22 mL) followed by the addi-
tion of mesitylene chloride (5.4 g, 24.8 mmol). The resulting solu-
tion was stirred at room temperature for 4 h. Pyridine was
evaporated and the crude product was recrystallized from hot eth-
anol as a yellow powder (59%).
Compound 9: Yield (75%): 1H NMR (CDCl3): 1.96 (m, 4H); 2.67 (t,
8H, J = 7.02 Hz); 4.31 (t, 2H, J = 7.02 Hz);7.97 (t, 2H J = 7.23 Hz); 8.44
(d, 2H, J = 7.23 Hz); 8.80 (d, 2H J = 7.23 Hz); 9.15 (d, 2H, J = 2.19 Hz);
9.32 (d, 2H, J = 2.19 Hz). 13C (DMSO) as methanesulfonate salt: 24.3;
37.2; 38.2, 46.1; 122.8; 123.2; 125.4; 130.1; 130.2; 130.3; 132.4;
135.3; 137.1; 146.4; 164.1; 165.3. MS (ESI) m/z: 625.3. Mp (2
CH3SO3H salt) 223.0 °C. Anal. Calcd for C32H28N6O8: C, 61.53; H,
4.52; N, 13.46; O, 20.49. Found: C, 61.44; H, 4.39; N, 13.29; O, 20.55.
Compound 10: (Yield 68%): 1H NMR (CDCl3): 1.42 (t, 2H
J = 7.45 Hz); 1.80 (m, 4H); 2.15 (t, 4H, J = 7.02 Hz); 3.05 (t, 2H,
J = 7.45 Hz); 4.31 (t, 4H, J = 7.02 Hz); 7.97 (t, 2H, J = 7.23 Hz); 8.44
(d, 2H, J = 7.23 Hz); 8.80 (d, 2H, J = 7.23 Hz); 9.15 (d, 2H,
J = 2.19 Hz); 9.32 (d, 2H, J = 2.19 Hz). 13C (DMSO) as methanesulfo-
nate salt: 15.5; 28.3; 35.9; 37.2; 37.9; 42.5; 122.6; 123.1; 124.5;
130.1; 130.2; 130.3; 131.3; 134.9; 137.4; 146.2; 164.4; 165.2 MS
(ESI) m/z: 636.72. Mp (2 CH3SO3H salt) 243.8 °C. Anal. Calcd for
C33H28N6O8: C, 62.26; H, 4.43; N, 13.20; O, 20.11. Found:, 62.22;
H, 4.55; N, 13.09; O, 20.07.
1H NMR (CDCl3): 1.75–1.90 (m, 2H); 2.35 (s, 12H); 2.55 (s, 12H);
2.60 (s, 12H); 2.95 (q, 4H, J = 6.14 Hz); 3.15 (t, 4H, J = 7.23 Hz); 3.35
(t, 4H, J = 6.14 Hz); 6.90 (s, 8H).
5.15. General procedures for N-alkylation reaction (24a–b)
N1,N3,N7,N10-Tetramesytil-1,3-bis(2-aminoethyl)-1,3-propane-
diamine (23) (0.651 mmol) was dissolved in anhydrous DMF
(13.5 mL) followed by the addition of 22a–b (0.13 mmol) and cae-
sium carbonate (1.06 g). The solution was left at 80 °C. Completion
of the reaction was monitored by thin-layer chromatography. DMF
was removed under vacuo, and the residue was poured into cold
water and the resulted precipitate filtered and washed thoroughly
with water. After drying, the crude product was recrystallised
from ethanol to give the fully protected pure product in high
yield.
5.10. Naphthalimidobutanol (21a)
Compound 24a. (Yield 85%) 1H NMR (CDCl3): 1.30–1.70 (m,
10H); 2.15 (s, 12H); 2.35 (s, 12H); 2.60 (s, 12H); 2.55 (s, 12H);
2.95 (q, 4H, J = 6.14 Hz); 3.15 (t, 4H, J = 6.14 Hz); 3.25 (s, 8H);
4.00 (t, 4H, J = 7.45 Hz); 6.88 (s, 4H); 7.02 (s, 4H); 7.75 (t, 4H,
J = 7.78 Hz); 8.25 (d, 4H, J = 7.78 Hz); 8.60 (d, 4H, J = 7.78 Hz).
Compound 24b. (Yield 95%) 1H NMR (CDCl3): 1.30–1.70 (m,
10H); 2.15 (s, 12H); 2.35 (s, 12H); 2.60 (s, 12H); 2.55 (s, 12H);
2.95 (q, 4H, J = 6.14 Hz); 3.15 (t, 4H, J = 6.14 Hz); 3.25 (s, 8H);
4.00 (t, 4H, J = 7.45 Hz); 6.88 (s, 4H); 7.02 (s, 4H); 8.00 (t, 2H,
J = 7.78 Hz); 8.55 (d, 2H, J = 8.33 Hz); 8.80 (d, 2H, J = 7.23 Hz);
9.15 (d, 2H, J = 2.19 Hz); 9.35 (d, 2H, J = 2.19 Hz).
A mixture 1,8-napthalic anhydride (2.5 g, 0.013 mol), and 4-
amino-1-butanol (3.6 mL, 0.04 mmol) in water (10 mL) was re-
fluxed for 30 min and then cooled at 4 °C. Then product 7a formed
was filtered, washed with ice cold water and crude product was
recrystallized from acetone and dried over P2O5 to give the com-
pound 21a as white crystalline solid (89%).
1H NMR (CDCl3) 1.65–1.95 (m, 4H); 3.75 (t, 2H, J = 5.92 Hz);
4.25 (t, 2H, J = 7.45 Hz); 7.75 (t, 2H, J = 7.67 Hz); 8.20 (d, 2H,
J = 8.11 Hz); 8.6 (d, 2H, J = 7.13 Hz).
5.11. 3-Nitro-naphthalimidobutanol (21b)
5.16. General procedures for deprotection reaction (11–12)
Compound 21b was prepared using 3-nitro-1,8-naphthalic
anhydride following the above-described procedure.
The fully protected polyamine derivatives (24a–b) (0.222 mmol)
were dissolved in anhydrous dichloromethane (10 mL) followed by
the addition of hydrobromic acid/glacial acetic acid (1 mL). The solu-
tion was left stirring at room temperature for 24 h. The yellow precip-
itate formed was filtered off and washed with dichloromethane, ethyl
acetate and ether.
Yield 86%. 1H NMR (CDCl3) 1.65–1.95 (m, 4H); 3.75 (t, 2H,
J = 6.40 Hz); 4.30 (t, 2H, J = 7.45 Hz); 7.95 (t, 1H, J = 7.78 Hz); 8.45
(d, 1H, J = 8.33 Hz); 8.8 (d, 1H, J = 7.23 Hz); 9.15 (d, 1H,
J = 2.19 Hz); 9.35 (d, 1H, J = 2.19 Hz).
Compound 11. (Yield 97%) 1H NMR (D2O) 1.50–1.90 (m, 8H); 2.15–
2.28 (m, 2H); 3.15–3.32 (m, 8H); 3.56 (br, 8H); 3.9 (t, 4H, J = 7.45 Hz)
7.05 (t, 4H, J = 7.78 Hz); 7.90 (d, 4H, J = 7.78 Hz); 8.15 (d, 4H,
J = 7.78 Hz). 1H NMR (DMSO): 23.3; 24.2; 25.6; 43.3; 43.4; 45.0;
47.7; 122.8; 128.2; 131.7; 132.2; 135.3; 164.4. MS (ESI) m/z: 633.43.
Mp (4 HBr salt) 235.6 °C Anal. Calcd for C38H44N6O4: C, 70.35; H,
6.84; N, 12.95; O, 9.86. Found: C, 70.30; H, 6.72; N, 12.80; O, 9.80.
Compound 12. (Yield 93%) 1H NMR (D2O) 1.50–1.90 (m, 8H);
2.15–2.28 (m, 2H); 3.15–3.32 (m, 8H); 3.56 (br, 8H); 3.9 (t, 4H,
J = 7.45 Hz); 8.07 (t, 2H, J = 7.78 Hz); 8.57 (d, 2H, J = 8.33 Hz);
8.73 (d, 2H, J = 7.23 Hz); 9.18 (d, 2H, J = 2.19 Hz); 9.24 (d, 2H,
J = 2.19 Hz). 1H NMR (DMSO): 23.2; 24.4; 25.3; 43.3; 43.5; 45.2;
47.7; 122.8; 123.2; 125.4; 130.1; 130.2; 130.3; 132.4; 135.3;
147.1; 146.4; 164.1; 165.3. MS (ESI) m/z: 753.45. Mp (4 HBr salt)
244.6 °C Anal. Calcd for C34H42N8O8: C, 61.78; H, 5.73; N, 15.17;
O, 17.33. Found: C, 61.70; H, 5.60; N, 15.03; O, 17.39.
5.12. O-tosylbutyloxynaphthalimide (22a)
Naphthalimidobutanol (21a) (1.0 g, 3.7 mmol) was dissolved in
anhydrous CH2Cl2 (10 mL) followed by the addition of tosyl chlo-
ride (2.89 g, 14.85 mmol) and triethylamine (2.8 mL, 20.4 mmol).
The solution was left overnight at 50 °C. After cooling the CH2Cl2
solution was washed with water and saturated bicarbonate solu-
tion. After drying and the removal of the solvent, a dark residue
was obtained which was then recrystallized from hot ethanol as
a white powder (55%).
1H NMR (CDCl3) 1.60–1.80 (m, 4H); 2.40 (s, 3H); 4.10 (br, 2H);
4.20 (br, 2H); 7.34 (d, 2H, J = 8.00 Hz); 7.70–7.90 (m, 4H); 8.25
(d, 2H, J = 8.22 Hz); 8.6 (d, 2H, J = 7.23 Hz).
5.13. O-Tosylbutyloxy-3-nitronaphthalimide (22b)
Compound 22b was prepared following the above described
procedure starting from 21b. Yellow powder, yield 46%. 1H NMR
(CDCl3) 1.70–1.85 (m, 4H); 2.50 (s, 3H); 4.10 (br, 2H); 4.20 (br,
2H); 7.35 (d, 2H, J = 8.00 Hz); 7.80 (d, 2H, J = 8.00 Hz); 8.00 (t, 1H,
J = 7.78 Hz); 8.55 (d, 1H, J = 8.33 Hz); 8.80 (d, 1H, J = 7.23 Hz);
9.15 (d, 1H, J = 2.19 Hz); 9.35 (d, 1H, J = 2.19 Hz).
5.17. Biological tests
5.17.1. Cells culture
Human colon adenocarcinoma HT-29 cells were cultured in
DMEM supplemented with 2 mM L-glutammine, 10% heat-inacti-

R. Filosa et al. / Bioorg. Med. Chem. 17 (2009) 13–24
23
vated foetal bovine serum (FBS), 1% penicillin/streptomycin (all
from Cambrex Bioscience, Verviers, Belgium) at 37 °C in an atmo-
sphere of 95% O2 and 5% CO2. The cells were plated at density of
5 ꢀ 105 in 6-cm cell culture dishes the day before treatment with
compounds 9 and 10 at different concentrations. Z-VAD-FMK
(50 lM) (BD Bioscience, USA) was administrated to cells 30 min
before compounds analysed. At the end of the incubation times
cells were processed for detection of apoptosis by FACS analysis.
ing the ligand structure (Macromodel 8.5 Software Package).20
From Model A we have built the Model B changing the T4 and
T12, A5 and A13 in C and G residues, respectively. The same steps
are made for the Model C using as template the X-ray structure
367D (PDB archive code). We have eliminated the ligands and
water molecules, and converted the Br-U3 and Br-U9 in T residues.
The 3 and 5–12 geometries are optimized at the hybrid DFT
B3LYP level using the 6-31G(d) basis set (Gaussian 03 Software
Package).21 Torsion restraints have been employed to avoid the for-
mation of intramolecular hydrogen bonds affecting the atomic
charge values. Subsequently, the charges of 3 and 5–12 are calcu-
lated by the ChelpG method21 at the B3LYP/6-31 + G(d) level. The
above calculated charges are used for the docking calculations.
For all the docking calculations of 3 and 5–10, a grid box size of
52 ꢀ 52 ꢀ 60 with spacing of 0.375 Å between the grid points, cen-
tred between central base pairs and covering the minor and major
groove surface of DNA is used. The same criteria are used for all the
docking calculations of 11 and 12, but through a grid box size of
58 ꢀ 68 ꢀ 58.
5.17.2. MTT assay
HT-29 cells (1 ꢀ 105 cells/ml in 96-well culture plates) were
incubated for 72 h with different concentrations of compounds (
5–12) dissolved in DMSO. After treatment, MTT solution (5 mg/
ml in PBS) was added to each well. After 3 h incubation lysis buffer
(200 g/L SDS, 50% Formamide, pH 4.7) was added to each well to
dissolve formazan. The absorbance was measured at 620 nm with
a Microplate reader. All experiments were performed at least three
times and average of the percentage absorbance was plotted
against concentration. The results were expressed as percentage
relative to untreated control and IC50 value was calculated for each
compounds.
In order to achieve a representative conformational space dur-
ing the docking calculations on 3 and 5–10, seven calculations con-
sisting of 256 runs are performed, obtaining 1792 structures
(256 ꢀ 7). The Lamarkian genetic algorithm is used for dockings.
An initial population of 150 randomly placed individuals, a maxi-
mum number of 2.5 ꢀ 105 energy evaluations, and a maximum
number of 2.7 ꢀ 104 generations are taken into account. A muta-
tion rate of 0.02, a crossover rate of 0.8 and local search frequency
of 0.06 are used.
5.17.3. Analysis of apoptosis
Hypodiploid DNA was analysed using propidium iodide staining
(PI) by flow cytometry as described.18 Briefly, cells were washed in
phosphate buffered saline (PBS) and resuspended in 500
solution containing 0.1% sodium citrate, 0.1% Triton X-100 and
50 g/ml PI (Sigma–Aldrich). After incubation at 4 °C for 30 min
ll of a
l
in the dark, cell nuclei were analysed with Becton Dickinson FAC-
Scan flow cytometer using CellQuest program. Cellular debris was
excluded from analysis by raising the forward scatter threshold,
and the DNA content of the nuclei was registered on a logarithmic
scale. The percentage of cells in the hypodiploid region was
calculated.
The docking calculation parameters on 11 and 12 are incre-
mented for the larger number of active torsions. An initial popula-
tion of 600 randomly placed individuals, a maximum number of
5 ꢀ 106 energy evaluations, and a maximum number of 6 ꢀ 106
generations are taken into account. A mutation rate of 0.02, a
crossover rate of 0.8 and local search frequency of 0.26 are used.
Results (3 and 5–12) differing by less than 2 Å in positional
root-mean-square deviation (RMSD) are clustered together and
represented by the result with the most favourable free energy of
binding.
5.17.4. Statistical analysis
All results are shown as means SEM of three experiments per-
formed in triplicate. Statistical comparison between groups were
made using parametric Bonferroni test. P values <0.05 were consid-
ered significant.
All the 3D models are depicted using the Phyton software:22
molecular surfaces are rendered using Maximal Speed Molecular
Surface (MSMS).23
5.17.5. DNA cleavage assays
DNA cleavage reactions were performed with recombinant hu-
Acknowledgment
man DNA topo IIa
, purified as described.19 pBR322 plasmid (5 ng)
was incubated with equal amounts of the enzyme in the presence
of different drug concentrations in 10 mM Tris–Cl, pH 7.5, 150 mM
KCl, 5 mM EDTA, 10 mM MgCl2, 0.01 mg/mL BSA, 10 mM DTT and
1 mM ATP. The reactions were incubated for 30 min at 37 °C, and
then stopped with 0.1% SDS and 0.5 mg/mL proteinase K (Sigma)
for 30 min at 45 °C. DNA samples were then run in 1% agarose gels
containing 0.05% SDS, and after extensive washing the gel was
stained with ethidium bromide. Then, DNA fragments were trans-
ferred to nitrocellulose membranes and blocked with UV irradia-
tion. pBR322 DNA was labelled with 32P-dCTP with Ready-to-go
DNA labelling Kit (GE Healthcare) and used as a probe to hybridize
the DNA fragments onto the membrane. Hybridization reaction
was performed in 0.5 M Na-phosphate, pH 7.2, 1 mM EDTA, 7%
SDS and 1% BSA at 65 °C. After final washes in 1x SSC and 0.1%
SDS at 65 °C, membranes were exposed for 1–24 h, and images
were then acquired with a Molecular Dynamics PhosphorImager.
Giovanni Capranico acknowledges financial support from Asso-
ciazione Italiana per la Ricerca sul Cancro (AIRC), Milan, Italy.
References and notes
1. Braña, M. F.; Ramos, A. Curr. Med. Chem. Anticancer Agents 2001, 1, 237.
2. Nitiss, J. L.; Zhou, J.; Rose, A.; Hsiung, Y.; Gale, K. C.; Osheroff, N. Biochemistry
1998, 37, 3078.
3. Malviya, V. K.; Liu, P. Y.; Alberts, D. S.; Surwit, E. A.; Craig, J. B.; Hanningan, E. V.
Am. J. Clin. Oncol. 1992, 15, 41.
4. Bousquet, P. F.; Braña, M. F.; Conlon, D.; Fitzgerald, K. M.; Perron, D.; Cocchiaro,
C.; Miller, R.; Moran, M.; George, J.; Qian, X.-D.; Keilhauer, G.; Romerdahl, C.
Cancer Res. 1995, 55, 1176.
5. Bailly, C.; Braña, M. F.; Waring, J. Eur. J. Biochem. 1996, 240, 195.
6. Gamage, S. A.; Spicer, J. A.; Finlay, G. J.; Stewart, A. J.; Charlton, P.; Baguley, B. C.;
Denny, W. A. J. Med. Chem. 2001, 44, 1407.
7. Cholody, W. C.; Kosakowska-Cholody, T.; Hollingshead, M. G.; Hariprakasha, H.
K.; Michejda, C. J. J. Med. Chem. 2005, 48, 4474.
8. Holden, J. A. Curr. Med. Chem. Anticancer Agents 2001, 1, 1.
9. Filosa, R.; Peduto, A.; de Caprariis, P.; Festa, M.; Petrella, A.; Pau, A.; Pinna, G. A.;
La Colla, P.; Busonera, B.; Loddo, R. Eur. J. Med. Chem. 2007, 42, 293.
10. Payne, G. B. J. Org. Chem. 1967, 32, 3351.
11. Witiak, D. T.; Lee, H. J.; Hart, R. W.; Gibson, R. E. J. Med. Chem. 1977, 20, 630.
12. Capranico, G.; Binaschi, M.; Borgnetto, M. E.; Zunino, F.; Palumbo, M. Trends
Pharmacol. Sci. 1997, 18, 323.
13. Morris, G. M.; Goodsell, D. S.; Halliday, R. S.; Huey, R.; Hart, W. E.; Belew, R. K.;
Olson, A. J. J. Comp. Chem. 1998, 19, 1639.
5.17.6. Computational details
Autodock 3.0.510 is used for all docking calculations. The NMR
structure of the covalent complex between d(CGCTAGGCG)-
(GCGATCCGC) and the bis-thiazole orange (PDB archive code
108D), is used as a template to build the homology Model A delet-

24
R. Filosa et al. / Bioorg. Med. Chem. 17 (2009) 13–24
14. Breneman, C. M.; Wiberg, K. B. J. Comp. Chem. 1990, 11, 361.
15. Oliveira, J.; Ralton, L.; Tavares, J.; Codeiro-da-Silva, A.; Bestwick, C.
S.; McPherson, A.; Lin, P. K. T. Bioorg. Med. Chem. Lett. 2007, 10,
541.
16. Lin, P. K. T.; Pavlov, V. A. Bioorg. Med. Chem. Lett. 2000, 10, 1609.
17. McCoy, L. L. J. Am. Chem. Soc. 1958, 80, 6568.
18. Nicoletti, I.; Migliorati, G.; Pagliacci, M. C.; Grignani, F.; Riccardi, C. J. Immunol.
Methods 1991, 139, 271.
19. Guano, F.; Pourquier, P.; Tinelli, S.; Binaschi, M.; Bigioni, M.; Animati, F.;
Manzini, S.; Zunino, F.; Kohlhagen, G.; Pommier, Y.; Capranico, G. Mol.
Pharmacol. 1999, 56, 77.
Scalmani, G.; Rega, N.; Petersson, G. A.; Nakatsuji, H. Hada, M.; Ehara, M.;
Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao,
O.; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Adamo,
C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi,
R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G. A.;
Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.;
Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.;
Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.;
Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Martin, R. L.;
Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe,
M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A.
Gaussian 03, revision B.05, Gaussian, Inc., Pittsburgh, PA, 2003.
20. MacroModel, version 8.5, Schrödinger LLC, New York, NY, 2003.
21. Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.;
Cheeseman, J. R.; Montgomery J. A.; Jr, Vreven, T.; Kudin, K. N.; Burant, J. C.;
Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.;
22. Sanner, M. F. J. Mol. Graph. Model. 1999, 17, 57–61.
23. Sanner, M. F.; Olson, A. J.; Spehner, J. C. Biopolymers 1996, 38, 305–
320.