Synthesis, in vitro antiproliferative activity and DNA-interaction of benzimidazoquinazoline derivatives as potential anti-tumor agents

Article abstract of DOI:10.1016/S0014-827X(01)01079-5

The synthesis of benzimidazoquinazoline derivatives bearing different alkylamino side chains is reported. All new compounds tested by means of an in vitro assay exhibit antiproliferative activity toward human tumor cell lines. The cytotoxic effect depends on the type of side chain inserted in the planar nucleus and in some cases it is comparable to that of the well-known drug ellipticine. In order to understand the mechanism of action of these compounds, the interaction with DNA has been investigated. Linear flow dichroism measurements allowed us to verify the formation of a molecular complex with DNA and the corresponding geometry of interaction. Intrinsic binding constants have also been evaluated by performing fluorimetric titrations.

Full text of DOI:10.1016/S0014-827X(01)01079-5

www.elsevier.nl/locate/farmac
Il Farmaco 56 (2001) 159–167
Synthesis, in vitro antiproliferative activity and DNA-interaction
of benzimidazoquinazoline derivatives as potential anti-tumor
agents
Lisa Dalla Via a, Ornella Gia a, Sebastiano Marciani Magno a, Antonio Da Settimo b,*,
Anna Maria Marini b, Giampaolo Primofiore b, Federico Da Settimo b, Silvia Salerno b
a Department of Pharmaceutical Sciences, Uni6ersity of Padua, Via Marzolo 5, 35131 Padua, Italy
b Department of Pharmaceutical Sciences, Uni6ersity of Pisa, Via Bonanno 6, 56126 Pisa, Italy
Received 10 November 2000; accepted 8 January 2001
Abstract
The synthesis of benzimidazoquinazoline derivatives bearing different alkylamino side chains is reported. All new compounds
tested by means of an in vitro assay exhibit antiproliferative activity toward human tumor cell lines. The cytotoxic effect depends
on the type of side chain inserted in the planar nucleus and in some cases it is comparable to that of the well-known drug
ellipticine. In order to understand the mechanism of action of these compounds, the interaction with DNA has been investigated.
Linear flow dichroism measurements allowed us to verify the formation of a molecular complex with DNA and the corresponding
geometry of interaction. Intrinsic binding constants have also been evaluated by performing fluorimetric titrations. © 2001
Elsevier Science S.A. All rights reserved.
Keywords: Benzimidazoquinazoline; Antiproliferative activity; DNA binding
induces unwinding, lengthening and stiffening of the
double helix, thus introducing structural changes, which
affect both the condensation of chromatin and its inter-
action with DNA-associated enzymes [1].
In the broad class of intercalating anti-cancer drugs,
the planar heterocyclic moieties play an important role.
Among them, benzimidazole can be considered an in-
teresting tool for the development of compounds exert-
ing biological properties. Many chemical structures
carrying a benzimidazole group indeed, show anti-tu-
mor activity [2–4].
Furthermore, notwithstanding the undoubted impor-
tance of the planar aromatic ring system, interestingly
the addition of groups or side chains which project
from one or both the grooves of the DNA double helix
and which interact like external binder moiety, appears
also to be crucial for drug activity. Indeed, a large
number of anti-tumor compounds have been modified
by linking to a DNA intercalator a suitable side chain,
thus modulating the biological properties and increas-
ing solubility under physiological conditions [5]. In this
connection it is worth noting that the optimization of
1. Introduction
Most anti-tumor drugs bind to DNA suggesting a
direct relationship between their interaction with the
macromolecule and the therapeutic effect. In this con-
nection the DNA has been considered by far the pre-
ferred target for the test of new compounds with
possible anti-tumor activity.
Among the DNA binding compounds one of the
most interesting groups is the intercalators, character-
ized by the presence of a planar aromatic chromophore,
which following the interaction with the nucleic acid,
locates between two consecutive base pairs of the dou-
ble helix.
The complex formation with the macromolecule is
usually considered reversible and van der Waals contact
with the base pairs, hydrophobic interactions and
charge transfer forces were hypothesized to contribute
to bonding energy. The binding of DNA intercalators
* Corresponding author.
E-mail address: fsettimo@farm.unipi.it (A. Da Settimo).
0014-827X/01/$ - see front matter © 2001 Elsevier Science S.A. All rights reserved.
PII: S0014-827X(01)01079-5

160
L. Dalla Via et al. / Il Farmaco 56 (2001) 159–167
macromolecule and the new derivatives. The binding
parameters have been obtained by fluorimetric
titrations.
2. Chemistry
The new target compounds 1–4 were conveniently
prepared from the appropriate 1-alkyl-2-aminobenzimi-
dazoles 6–9 by reaction with 2-bromobenzoic acid, in
Ullmann conditions (Scheme 1). The starting alkylated
derivatives 7–9 have already been described [8], and the
new 1-dimethylaminoethyl-2-aminobenzimidazole (6)
was synthesized following a similar experimental proce-
dure. All products, 1–4, were purified by recrystalliza-
tion and their structures were confirmed by analytical
and spectral data (Table 1).
The synthesis of the desired ethanolamino derivative
5 was carried out using as starting material the hetero-
cyclic compound 10 (Scheme 2), which was obtained
following a literature method [9]. The heterocycle 10
was converted into the intermediate chloroethyl deriva-
tive 11 by reaction with 1-bromo-2-chloroethane in
anhydrous DMF solution, in the presence of sodium
hydride. Then, compound 5 was conveniently obtained
by reaction of 11 with ethanolamine in ethanol reflux-
Fig. 1. Chemical structures of the benzimidazoquinazoline deriva-
tives.
the charged side chains of the derivatives of anthracene-
dione, designed as analogues of the anthracycline an-
tibiotics, led to the well-known anti-tumor agents
ametantrone and mitoxantrone [6,7].
In this paper we report the synthesis of new deriva-
tives 1–5 characterized by the presence of a benzimida-
zoquinazoline nucleus carrying different alkylamino-
substituted side chains. In detail, in the 6 position of
the planar moiety, dialkylamino alkyl side chains have
been inserted in 1–4, and a hydroxyethylaminoethyl
side chain has been inserted in 5 (Fig. 1).
An evaluation of their antiproliferative activity in
vitro was carried out using human tumor cell lines.
Moreover, using spectroscopic techniques the DNA
binding process has been investigated. Linear flow
dichroism measurements were performed to establish
the geometry of the complex formed between the
1
ing solution. IR, H NMR and mass spectral data were
consistent with the proposed 6-alkyl substituted struc-
1
ture. In the H NMR spectrum, the triplets at 4.35 and
2.99 ppm (J=6.20 Hz) assigned, respectively, to the
methylene bound to the 6-position of the benzimidazo-
quinazoline nucleus and to its adjacent methylene
group, showed chemical shift values quite similar to
those of the corresponding methylene groups in the
spectra of derivatives 1–4. Moreover, the assignment of
the position of the side alkyl chain was further con-
firmed by the considerations emerging from the IR
spectra, which, for compound 5, showed a strong CꢀO
absorption band at w=1690 cm−1, superimposable
Scheme 1.

L. Dalla Via et al. / Il Farmaco 56 (2001) 159–167
161
Table 1
Physical and spectral data of compounds 1–5
No.
R
Yield (%) M.p. (°C) 1H-NMR DMSO-d6 (l ppm)
MS m/z (M+
)
Formula
1
(CH2)2N(CH3)2
55
29
160–165
163–165
2.2 (s, 6H, N(CH3)2), 2.73 (t, 2H, CH2CH2N(CH3)2), 306
4.37 (t, 2H, CH2CH2N(CH3)2), 7.29–8.41 (m, 8H,
ArH).
C18H18N4O
C19H20N4O
2
(CH2)3N(CH3)2
1.97 (m, 2H, CH2CH2CH2N(CH3)2), 2.13 (s, 6H,
N(CH3)2), 2.23 (t, 2H, CH2CH2CH2 N(CH3)2), 4.31
(t, 2H, CH2CH2CH2N(CH3)2), 7.31–8.42 (m, 8H,
ArH).
320
3
4
(CH2)2N(CH2CH3)2
(CH2)3N(CH2CH3)2
46
49
140–145
130–135
0.83 (t, 6H, N(CH2CH3)2), 2.4 (q, 4H, N(CH2CH3)2), 334
2.87 (t, 2H, CH2CH2N(C2H5)2), 4.37 (t, 2H,
CH2CH2N(C2H5)2), 7.35–8.45 (m, 8H, ArH).
C20H22N4O
C21H24N4O
0.89 (t, 6H, N(CH2CH3)2), 2.02 (m, 2H,
348
CH2CH2CH2N(C2H5)2), 2.29 (q, 4H, N(CH2CH3)2),
2.71 (t, 2H, CH2CH2CH2N(CH2CH3)2), 4.31 (t, 2H,
CH2CH2CH2N(C2H5)2), 7.20–8.50 (m, 8H, ArH).
5
(CH2)2NHCH2CH2OH 72
180–183
2.61 (t, 2H, CH2CH2OH), 2.99 (t, 2H, CH2CH2NH), 322
3.36 (t, 2H, CH2CH2OH), 4.35 (t, 2H, CH2CH2NH),
4.4 (br.t, 1H, OH exch.), 7.21–7.78 (m, 6H, ArH),
8.18–8.50 (m, 2H, ArH) a
C18H18N4O2
a Recorded on a Varian Gemini 200 (200-MHz) spectrometer.
with that of compounds 1–4, similar to that reported in
literature for 6-methylbenzimidazo[2,1-b]quinazolin-
12(6H)-one (w=1684 cm−1) but different from that of
the 5-methyl isomer (w=1724 cm−1) [9].
3.1.1. 1-(2-Dimethylaminoethyl)-2-aminobenzimidazole
(6)
1-Dimethylamino-2-chloroethane hydrochloride (1.17
g, 8.14 mmol) was added to a stirred solution of
2-aminobenzimidazole (1.0 g, 7.5 mmol) and potassium
hydroxide (1.0 g, 18.8 mmol) in ethanol. The mixture
was stirred at room temperature for 5 h and then the
inorganic material (KCl) was filtered off. The solution
was evaporated to dryness and the residue obtained was
purified by recrystallization from acetone to give 0.446
3. Experimental
3.1. Chemistry
1
Melting points were determined using a Reichert
Ko¨fler hot-stage apparatus and are uncorrected. IR
spectra were obtained on a PYE/UNICAM Model PU
g (30% yield) of pure 6; m.p. 140–142°C; H NMR: l
2.21 (s, 6H, N(CH3)2), 2.51 (t, 2H, CH2CH2N(CH3)2),
4.03 (t, 2H, CH2CH2N(CH3)2), 6.39 (s, 2H, NH2 exch.),
6.39–7.16 (m, 4H, ArH); MS: m/z 204 (M+).
1
9561 spectrophotometer as Nujol mulls. H NMR spec-
tra were recorded, if not otherwise indicated, on a
Varian CFT-20 (80 MHz) spectrometer, in DMSO-d6
solution, using TMS as the internal standard. Mass
spectra were obtained on a Hewlett–Packard 5988 A
spectrometer using a direct injection probe and an
electron beam energy of 70 eV. Magnesium sulfate was
always used as the drying agent. Evaporations were
performed in vacuo (rotating evaporator). Analytical
TLC was carried out on Merck 0.2 mm precoated silica
gel aluminium sheets (60 F-254). Elemental analyses (C,
H, N) were performed and agreed with the theoretical
values within 90.4%.
3.1.2. 6-(Dialkylaminoalkyl)benzimidazo[2,1-b]-
quinazolin-12(6H)-ones (1–4)
General procedure: A suspension of the appropriate
1-(dialkylaminoalkyl)-2-aminobenzimidazole (6–9) (7.5
mmol), 2-bromobenzoic acid (1.51 g, 7.5 mmol), anhy-
drous potassium carbonate (1.24 g, 9 mmol), copper-
bronze (0.08 g) and a catalytic amount of KI in 10 ml
of anhydrous DMF was heated at 170°C until the
disappearance of the starting materials (5–7 h, TLC
analysis, chloroform/methanol 7:3 as eluant). After
cooling, the reaction mixture was diluted with water

162
L. Dalla Via et al. / Il Farmaco 56 (2001) 159–167
and left to stand at 4°C overnight. The precipitated
crude products were collected and purified by recrystal-
lization from ethanol (Table 1).
residue was purified by recrystallization from ethanol to
give pure compound 5 (0.156 g) (Table 1).
3.2. Biological acti6ity
3.1.3. 6-(2-Chloroethyl)benzimidazo[2,1-b]quinazolin-
12(6H)-one (11)
3.2.1. Cell cultures
Sodium hydride (50% dispersion in mineral oil, 0.14
g, 2.8 mmol) was added in small portions to a stirred
suspension of 0.60 g (2.56 mmol) of compound 10 in 15
ml of anhydrous DMF, under nitrogen atmosphere.
The reaction mixture was stirred at room temperature
for 1 h, then supplemented dropwise with 0.3 ml (1
mmol) of 1-bromo-2-chloroethane and left at room
temperature with stirring for 16 h. Compound 11 was
recovered from the reaction mixture as insoluble mate-
rial, washed with water and purified by recrystallization
from DMF, obtaining 0.54 g (71% yield) of pure
derivative 11; m.p. 237–238°C; 1H NMR: l 4.12 (t, 2H,
N–CH2), 4.64 (t, 2H, CH2–Cl), 7.33–7.76 (m, 6H,
ArH); 8.17–8.50 (m, 2H, ArH); MS: m/z 297 (M+).
HL-60 (human myeloid leukaemic cells) were grown
in RPMI 1640 (Sigma Chemical Co.) supplemented
with 15% heat-inactivated fetal calf serum (Seromed),
HeLa (human cervix adenocarcinoma cells) were grown
in Nutrient Mixture F-12 [HAM] (Sigma Chemical Co.)
supplemented with 10% heat-inactivated fetal calf
serum (Seromed) and A431 (human squamous car-
cinoma cells) were grown in D-MEM (Sigma Chemical
Co.) supplemented with 10% heat-inactivated fetal calf
serum (Seromed). 100 U/ml penicillin, 100 mg/ml strep-
tomycin and 0.25 mg/ml amphotericin B (Sigma Chemi-
cal Co.) were added to the media. The cells were
cultured at 37°C in a moist atmosphere of 5% carbon
dioxide in air.
3.1.4. 6-(i-Ethanolaminoethyl)benzimidazo[2,1-b]-
quinazolin-12(6H)-one (5)
3.2.2. Inhibition growth assay
0.16 ml (2.70 mmol) of ethanolamine was added to a
suspension of compound 11 (0.20 g, 0.67 mmol) in 30
ml of ethanol. The reaction mixture was refluxed for 40
h and during this time, two amounts of 0.08 ml of
ethanolamine were further added. The resulting suspen-
sion was filtered to recover the unreacted starting
product 11 (0.029 g). The solution was evaporated to
dryness under reduced pressure and the resulting crude
HL-60 cells (3×104) were seeded into each well of a
24-well cell culture plate. After incubation for 24 h,
various concentrations of the test agents were added in
complete medium and incubated for a further 72 h.
HeLa (3×104) and A431 (4×104) cells were seeded
into each well of a 24-well cell culture plate. After
incubation for 24 h, the medium was replaced with an
equal volume of fresh medium and various concentra-
Scheme 2.

L. Dalla Via et al. / Il Farmaco 56 (2001) 159–167
163
Table 2
LD(r)=LD(u)/Aiso(u)
Growth cell inhibition of test compounds 1–5 and ellipticine as
where Aiso(u) is the absorbance of the sample in the
absence of flow. This quantity may be related to an
orientation factor (S) and the angle between the active
transition moment in the chromophore and the DNA
helix axis, h [11,12]:
reference drug
Compounds
IC50 (mM)
HL-60
HeLa
A-431
1
2
3
4
0.590.05
1.490.2
1.790.2
2.290.4
0.690.1
0.6490.02
2.190.3
4.390.7
8.690.5
9.690.9
1.9090.07
0.3190.01
0.890.2
1.290.1
3.590.1
3.690.3
0.5090.09
0.4790.03
LDr=3/2 (3 cos2 h−1)S
Assuming a value of h=90° for the DNA base-pair
chromophore with respect to a local helix axis, it is
possible to evaluate hL for a given ligand:
5
Ellipticine
1/2
hL=arccos[1/3-(LDr)L/3(LDr)DNA
]
tions of the test agents were added. The cells were then
incubated in standard conditions for a further 72 h.
A trypan blue assay was performed to determine cell
viability. Cytotoxicity data were expressed as IC50 val-
ues, i.e. the concentrations of the test agent inducing
50% reduction in cell numbers compared with the con-
trol cultures.
where (LDr)L is the reduced linear dichroism for the
ligand, (LDr)DNA is the reduced LD for DNA and hL
defines the ligand–DNA relative orientation. For the
intercalated system, (LDr)L:(LDr)DNA and hL$90°.
A solution of the salmon testes DNA (1.6×10−3M)
in the ETN buffer containing 10 mM NaCl was used.
Spectra were recorded at 25°C at different [DNA]/
[drug] ratios.
3.3. Interaction with DNA
3.3.4. Fluorimetric determinations
3.3.1. Nucleic acid
Fluorescence spectra were recorded on a Perkin–
Elmer LS50B luminescence spectrometer at uex=340
nm. The measurements were carried out at room tem-
perature in ETN buffer in the presence of 0.01 or 0.5 M
NaCl as indicated in Fig. 3.
Salmon testes DNA was purchased from Sigma
Chemical Co. Aqueous solutions of nucleic acid con-
taining Tris 10 mM, EDTA 1 mM (pH 7.0) and NaCl
0.01 M or 0.5 M depending on the desired ionic
strength were used (ETN buffer).
Titration was performed by addition of 0.5×10−6
M test compound to samples containing different con-
centrations of DNA. The buffer background was sub-
tracted from intensities for binding calculations. The
amounts of free and bound ligand were determined as
previously indicated [8].
3.3.2. Spectrophotometric determinations
UV absorption spectra were recorded at room tem-
perature with a Perkin–Elmer model Lambda 5 spec-
trometer. DNA and test compound concentrations were
determined by absorption measurements, using the fol-
lowing extinction coefficients: m=6600 M−1/cm at 260
nm for DNA, m=21 520 M−1/cm at 297 nm for 1,
Experimental binding data were plotted according to
the method of Scatchard [13] and analysed by using the
approach of McGhee and von Hippel [14] to obtain the
intrinsic binding constant (Ki).
m=21 940 M−1/cm at 299 nm for 2, m=21 680 M−1
/
cm at 297 nm for 3, m=23 930 M−1/cm at 299 nm for
4, m=23820 M−1/cm at 300 nm for 5.
4. Results and discussion
3.3.3. Flow linear dichroism
Linear dichroism (LD) measurements were per-
formed on a Jasco J500A circular dichroism spectropo-
larimeter converted for LD and equipped with an IBM
PC and a Jasco J interface.
4.1. Antiproliferati6e acti6ity
The antiproliferative activity of compounds 1–5 has
been evaluated by means of an in vitro test carried out
on three human tumor cell lines to determine the
concentration (mM) that cause the death of 50% of the
cells in comparison with the control cultures. The re-
sults of the experiments, performed as indicated in
Section 3, are reported in Table 2. The well-known
drug ellipticine was used as reference compound.
The results obtained indicate that independent of the
cell line taken into account, a general behavior in
cytotoxic ability can be used as evidence. Indeed, in all
Linear dichroism is defined as:
LD(u)=Aꢀ(u)−AÞ(u)
where Aꢀ and AÞ correspond to the absorbances of the
sample when polarized light is oriented parallel or
perpendicular to the flow direction, respectively. The
orientation is produced by a device designed by Wada
and Kozawa [10] at a shear gradient of 500–700 rpm.
The reduced linear dichroism is defined as:

164
L. Dalla Via et al. / Il Farmaco 56 (2001) 159–167
cases the compounds exerting the higher activity are 5
and 1, whilst the lesser active appears to be compound
4. In particular, it is to be noted that in HL-60 and
A-431 cells the derivatives 1 and 5 exhibited IC50 values
comparable to that of ellipticine.
Due to the fact that all the new derivatives possess
the same chromophore moiety, the differences in behav-
ior above underlined have to be attributed to the
different aminoalkyl side chains (Fig. 1). From the
comparison among the antiproliferative activities of
compounds 1–4 appears that both the presence of three
methylene units linked to the nitrogen belonging to the
benzimidazole nucleus of the planar moiety (derivatives
2 and 4) and the presence of ethyl substituents in the
basic nitrogen of the side chain (derivatives 3 and 4)
concur to decrease the cytotoxic effect. This indicates
that the distance of the basic nitrogen from the planar
chromophore and the steric hindrance around the same
atom play a crucial role in the interactions between the
compounds and the biological target(s) accountable for
the antiproliferative effect. Indeed, the optimal chemi-
cal structure appears to be that of 1, whose side chain
shows the minimal length beside the minimal steric
hindrance around the basic nitrogen. Interestingly, this
behavior is in agreement with a previous work where
the antiproliferative activity of purinoquinazoline
derivatives carrying the aminoalkyl side chains of 1–4
had also been investigated [8]. Altogether these results
underline the capacity of the side chains to modulate
the cellular consequences of a planar aromatic nucleus.
The rationale which drove to the synthesis of com-
pound 5 was based on the properties of the well-known
drug mitoxantrone. In particular, a hydroxyethyl-
aminoethyl group has been inserted in the position 6 of
the planar chromophore. The results obtained confirm
the important role played by this type of side chain,
indeed the antiproliferative effect exerted by 5 is signifi-
cantly high and comparable to that of compound 1 in
all the cell lines taken into account.
4.2. Flow linear dichroism measurements
Biochemical evidence suggests that deoxyribonucleic
acid is among the principal cell targets for many anti-
tumor agents and that the biological activities of these
drugs appear to be related to their specific affinities and
to their particular mode of binding [15].
To investigate the interaction of compounds 1–5
with DNA, flow linear dichroism (LD) spectra were
performed as reported in Section 3. In Fig. 2 the flow
LD of 5 for a [DNA]/[drug]=12.5 was reported as
representative spectrum (middle panel). The upper
panel and the lower panel depicted the light absorption
and the reduced linear dichroism (LDr) spectra, respec-
tively, evaluated for the same sample.
The LD spectra of DNA solutions containing the test
compounds display a strong negative signal at 260 nm,
as expected for this macromolecule. Furthermore, a
significant negative signal appears in the ligand absorp-
tion region (320–380 nm). Since these small molecules
cannot themselves become oriented in the flow field, the
appearance of a dichroic signal in this region is at-
tributable to the induced orientation of the ligand
chromophore upon interaction with the macromolecule,
so indicating the formation of a molecular complex
with DNA. Moreover, the negative sign of the spec-
trum in this region can indicate an orientation of the
molecular plane of the ligand chromophore preferen-
tially parallel to the plane of the DNA bases [8].
More specific information about the geometry of the
complexation between 1–5 and DNA can be obtained
by (LDr) estimation and the subsequent calculations of
the values of the average orientation angle hL by means
Fig. 2. Absorbance (upper panel), linear dichroism (LD, middle
panel) and reduced linear dichroism (LDr, lower panel) spectra of
compound 5 in the presence of DNA (1.6×10−3 M) at [DNA]/
[drug]=12.5.

L. Dalla Via et al. / Il Farmaco 56 (2001) 159–167
165
Table 3
of the equations reported in Section 3. In Table 3 the
calculated values of hL for the test compounds are
reported.
Calculated values of the average orientation angle hL for the test
compounds 1–5 ([DNA]=1.6×10−3 M, [DNA]/[drug]=12.5.)
The hL values calculated for compounds 1, 2 and 3
are 90, 80 and 82°, respectively. Large values of hL
(80–90°) are retained consistent with an intercalative
mode of binding [16], and this means that in the
complex between the above indicated compounds and
DNA the planar aromatic moiety belonging to the
drugs assumes a perpendicular geometry with respect to
the helix axis of the macromolecule.
Compounds
hL (°)
1
2
3
4
5
90
80
82
74
73
With regard to compounds 4 and 5, the calculated hL
values, i.e. 73 and 74°, respectively (see Table 3), are
lower in comparison with previous ones. An hL value
smaller than 80–90° obtained for the ligand absorption
band indicates that the corresponding ligand transition
deviates from co-planarity with the DNA base plane.
For example, if the drug was bound in the minor
groove an hL value around 45° would be expected [17].
According to previous results achieved for the interac-
tion with DNA of analogues of ellipticine having hete-
rocyclic ring system with three rings and
a
dimethylaminoethyl side chain [18], the hL values ob-
served for 4 and 5 can be retained indicative for the
occurrence of more than one type of binding mode. In
particular, the resulting hL values suggest a binding
mode where both intercalation (for which hL=90° is
expected) and groove binding (for which hL should be
45°) take place.
By analysing the chemical structures of 1–5 with the
calculated hL values, it is conceivable to ascribe to the
different side chains a crucial role in determining the
geometry of complexation with the double helix of
DNA. In detail, comparing the behavior shown by 4
with that of its congeners 1–3, the increase of the
lengthening beside the presence of steric hindrance in
the terminal part of the side chain seem to promote the
capacity to give rise to external binding which, at least
in part, competes with the intercalation ability of the
planar moiety. In the case of 5 which is characterized
by having a hydroxyethylaminoethylamino side chain,
it is reasonable to state that external interactions occur
due to its particular side chain which is probably ar-
ranged in an elongated manner with respect to the
planar moiety so that interaction (e.g. hydrogen bonds)
between the terminal hydroxyl group and DNA base
atoms located externally to the double helix can take
place.
4.3. Binding parameters
The new derivatives 1–5 show a significant fluores-
cence signal which is dramatically quenched upon addi-
tion of the macromolecule. These properties allow us to
determine the thermodynamic parameters for the DNA
binding process by using fluorimetric titration and the
Fig. 3. Scatchard plots for the interaction of compound 1 with DNA
in the presence of 0.01 M (A) and 0.5 M (B) ionic strength. B is the
ratio between bound drug and moles of DNA base pairs and F is the
free drug concentration.

166
L. Dalla Via et al. / Il Farmaco 56 (2001) 159–167
methodological approach reported in Section 3. The
binding parameters obtained for compound 1 and rep-
resented as Scatchard plots, are reported in Fig. 3 as an
example. The ionic strengths of 0.01 M (Fig. 3A) and
0.5 M (Fig. 3B) have been taken into account.
The intrinsic binding constant (Ki) obtained in the
above mentioned conditions for compounds 1–5 are
reported in Table 4. The higher values obtained are
those corresponding to 1 and 5, while lower and com-
parable values have been found for 2–4. Taking into
account that 1 and 5 are the most active compounds in
inducing cytotoxicity (Table 2), it appears that for these
derivatives a high affinity in the formation of a molecu-
lar complex with the macromolecule could be funda-
mental to promote biological events which lead to
antiproliferative effects.
tionship seems to exist (see Tables 2 and 3). Actually,
the cytotoxicity exhibited by 1–4 appears to be corre-
lated with the capacity to form an intercalative complex
being lower for the compound 4 for which another
mode of binding occurs, at least in part. Interestingly,
unlike the behavior shown by 4, for 5 the ability to
form an intercalative complex with the macromolecule
does not correlate with the cytotoxic capacity. Indeed,
cytotoxicity appears to be comparable to that observed
for 1, which behaves as an intercalator, and is clearly
higher than that of 4, which has a similar value of hL.
For compound 5 it is thus reasonable to assume that
the external interactions due to the hydroxyethy-
laminoethylamino side chain play an important role in
modulating cytotoxic activity promoting the formation
of a complex with DNA able to induce marked antipro-
liferative effects.
Furthermore, it is interesting to observe that for all
compounds the increase in ionic strength induces a
clear decrease in the binding affinity, so confirming the
involvement of charged groups in the complex forma-
tion with DNA.
Moreover, higher values of intrinsic binding con-
stants toward the macromolecule accompany higher
cytotoxic activity confirming that the ability to interact
with the macromolecule is fundamental to exert the
cellular effect.
These results allow us to confirm the importance of
the intercalation in the cytotoxic ability, but also to
affirm the crucial role carried out by the presence of
suitable side chains and in this framework these struc-
tures can constitute an useful tool for the development
of rationales devoted to the synthesis of new drugs
endowed with better specificity of action towards DNA.
5. Conclusions
The new benzimidazoquinazoline derivatives 1–5
show significant antiproliferative activity toward hu-
man tumor cell lines and in particular 1 and 5 exert an
effect comparable to that of the well-known drug ellip-
ticine in HL-60 and A-431 cells.
Linear flow dichroism studies reveal that these struc-
tures are able to form a complex with DNA mainly by
intercalation. From the calculation of hL it can be
hypothesized that for compound 4 and 5 beside the
intercalation a concurrent external mode of binding can
take place. It seems reasonable that the presence in the
side chain of the basic amino nitrogen at a suitable
distance from the planar moiety (compound 4) or the
presence of a terminal hydroxy group (compound 5)
results in the possibility to give rise to an external
binding ascribable to electrostatic interactions or hy-
drogen-bond formation.
Acknowledgements
This work was supported by grants from the Min-
istry of University and Scientific and Technological
Research (MURST) (Research fund 60%).
References
[1] B.C. Baguley, DNA intercalating anti-tumor agents, Anti-Can-
cer Drug Des. 6 (1991) 1–35.
[2] W.A. Denny, G.W. Rewcastle, B.C. Baguley, Potential antitu-
mor agents. 59. Structure–activity relationships for 2-phenylben-
Comparing complexation geometry with the antipro-
liferative activity, it is interesting to note that a rela-
zimidazole-4-carboxamides,
a
new class of minimal
DNA-intercalating agents which may not act via topoisomerase
II, J. Med. Chem. 33 (1990) 814–819.
[3] W.F. Bran˜a, J.M. Castellano, G. Keilhauer, A. Machuca, Y.
Mart´ın, C. Redondo, E. Schlick, N. Walker, Benzimidazo[1,2-
c]quinazolines: a new class of antitumor compounds, Anti-Can-
cer Drug Des. 9 (1994) 527–538.
Table 4
Intrinsic binding constants (Ki) of complexes between test compounds
1–5 and DNA
Comp. Ki×10−5 (M−1) (ionic
Ki×10−5 (M−1) (ionic
strength 0.5 M)
strength 0.01 M)
[4] L. Garuti, M. Roberti, T. Rossi, C. Cermelli, M. Portolani, M.
Malagoli, M. Castelli, Synthesis, antiviral and antiproliferative
activity of some N-benzenesulphonyl-2-(2- or 3-pyridylethyl)-
benzimidazoles, Anti-Cancer Drug Des. 13 (1998) 397–406.
[5] L.M. Werbel, M. Angelo, D.W. Fry, D.F. Worth, Basically
substituted ellipticine analogues as potential antitumor agents, J.
Med. Chem. 29 (1986) 1321–1322.
1
2
3
4
5
2.390.3
1.0990.14
1.1190.09
1.1490.12
2.590.3
0.08590.013
0.2090.03
0.1490.02
0.1890.02
0.07990.012

L. Dalla Via et al. / Il Farmaco 56 (2001) 159–167
167
[6] R.J. White, F.E. Durr, Development of mitoxantrone, Investig.
New Drugs 3 (1985) 85–93.
[7] J. Koeller, M. Eble, Mitoxantrone: a novel anthracycline deriva-
tive, Clin. Pharmacol. 7 (1988) 574–581.
[8] A. Da Settimo, F. Da Settimo, A.M. Marini, G. Primofiore, S.
Salerno, G. Viola, L. Dalla Via, S. Marciani Magno, Synthesis,
DNA binding and in vitro antiproliferative activity of purino-
quinazoline, pyridopyrimidopurine and pyridopyrimidobenzimi-
dazole derivatives as potential antitumor agents, Eur. J. Med.
Chem. 33 (1998) 685–696.
[9] W.H.W. Lunn, R.W. Harper, Methods for preparing benzimi-
dazo[2,1-b]quinazolin-12-ones and related compounds, J. Hete-
rocycl. Chem. 8 (1971) 141–147.
[13] G. Scatchard, The attractions of proteins for small molecules
and ions, Ann. N. Y. Acad. Sci. 51 (1949) 660–672.
[14] J.D. McGhee, P.H. von Hippel, Theoretical aspects of DNA–
protein interaction: co-operative and non co-operative binding of
large ligands to one dimensional homogeneous lattice, J. Mol.
Biol. 86 (1974) 469–489.
[15] M.J. Waring, DNA modification and cancer, Annu. Rev.
Biochem. 50 (1981) 159–192.
[16] U. Sehlstedt, S.K. Kim, P. Carter, J. Goodisman, J.F. Vollano,
B. Norde´n, J.C. Dabrowiak, Interaction of cationic porphyrins
with DNA, Biochemistry 33 (1994) 417–426.
[17] B. Norde´n, T. Kurucsev, Analysing DNA complexes by cir-
cular and linear dichroism, J. Mol. Recognit. 7 (1994) 141–
156.
[18] G. Behravan, M. Leijon, U. Sehlstedt, B. Norde´n, H. Vallberg,
J. Bergman, A. Gra¨slund, The interaction of ellipticine deriva-
tives with nucleic acids studied by optical and 1H-NMR spec-
troscopy: effect of size of the heterocyclic ring system,
Biopolymers 34 (1994) 599–609.
[10] A. Wada, S. Kozawa, Instrument for the studies of differential
flow dichroism of polymer solutions, J. Polym. Sci., Part A 2
(1964) 853–864.
[11] B. Norde´n, Applications of linear dichroism spectroscopy, Appl.
Spectrosc. Rev. 14 (1978) 157–248.
[12] B. Norde´n, M. Kubista, T. Kurucsev, Linear dichroism spec-
troscopy of nucleic acid, Q. Rev. Biophys. 25 (1992) 51–170.
.