Synthesis, anticancer activity and QSAR study of 1,4-naphthoquinone derivatives

Article abstract of DOI:10.1016/j.ejmech.2014.07.024

A series of 2-substituted amino-3-chloro-1,4-naphthoquinone derivatives (3-12) were synthesized as anticancer agents and tested against four cancer cell lines including HepG2, HuCCA-1, A549 and MOLT-3. The most potent cytotoxic activity against the HepG2,

Full text of DOI:10.1016/j.ejmech.2014.07.024

European Journal of Medicinal Chemistry 84 (2014) 247e263
Contents lists available at ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Original article
Synthesis, anticancer activity and QSAR study of 1,4-naphthoquinone
derivatives
Veda Prachayasittikul a, Ratchanok Pingaew b, Apilak Worachartcheewan c,
,
, e, f
Chanin Nantasenamat c, Supaluk Prachayasittikul c *, Somsak Ruchirawat d
,
,
*
Virapong Prachayasittikul a
a Department of Clinical Microbiology and Applied Technology, Faculty of Medical Technology, Mahidol University, Bangkok 10700, Thailand
b Department of Chemistry, Faculty of Science, Srinakharinwirot University, Bangkok 10110, Thailand
c Center of Data Mining and Biomedical Informatics, Faculty of Medical Technology, Mahidol University, Bangkok 10700, Thailand
d Laboratory of Medicinal Chemistry, Chulabhorn Research Institute, Bangkok 10210, Thailand
e Program in Chemical Biology, Chulabhorn Graduate Institute, Bangkok 10210, Thailand
f Center of Excellence on Environmental Health and Toxicology, Commission on Higher Education (CHE), Ministry of Education, Thailand
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of 2-substituted amino-3-chloro-1,4-naphthoquinone derivatives (3e12) were synthesized as
anticancer agents and tested against four cancer cell lines including HepG2, HuCCA-1, A549 and MOLT-3.
The most potent cytotoxic activity against the HepG2, HuCCA-1 and A549 cell lines was found to be m-
Received 11 April 2014
Received in revised form
26 June 2014
Accepted 5 July 2014
Available online 9 July 2014
acetylphenylamino-1,4-naphthoquinone (8) affording IC50 values of 4.758, 2.364 and 12.279
respectively. On the other hand, p-acetylphenylamino-1,4-naphthoquinone (9) exhibited the most potent
cytotoxic activity against the MOLT-3 cell line with an IC50 of 2.118 M. Quantitative structureeactivity
mM,
m
relationship (QSAR) investigations provided good predictive performance as observed from cross-
validated R of 0.9177e0.9753 and RMSE of 0.0614e0.1881. The effects of substituents at the 2-amino
position on the naphthoquinone core structure and its corresponding influence on the cytotoxic activ-
ity were investigated by virtually constructing additional 1,4-naphthoquinone compounds (13e36) for
which cytotoxic activities were predicted using equations obtained from the previously constructed
QSAR models. Interpretation of informative descriptors from QSAR models revealed pertinent knowledge
on physicochemical properties governing the cytotoxic activities of tested cancer cell lines. It is antici-
pated that the QSAR models developed herein could provide guidelines for further development of novel
and potent anticancer agents.
Keywords:
Anticancer
1,4-Naphthoquinone derivatives
QSAR
Structural modification
© 2014 Published by Elsevier Masson SAS.
1. Introduction
bioactivities has been reported for 1,4-naphthoquinone derivatives
to exert anticancer [2,4e8], antimicrobial [9e12], antifungal
There is a continual increase on the incidence of cancer and it is
considered to be the leading cause of morbidity and mortality [1].
Recently, considerable attention has been drawn on the search for
novel anticancer drugs in order to improve survival rates and well-
being.
The quinone scaffold is presented in many currently used anti-
cancer drugs [2,3]. Particularly, 1,4-naphthoquinones are active
quinone derivatives that are widely used as raw materials in
pharmaceuticals and agrochemicals industries. A diverse array of
[10,13e15], antiviral [11,16e18], radical scavenging [19], antiplatelet
[20e22] and trypanocidal [23] activities.
1,4-Naphthoquinone contains two ketone groups as a crucial
chromophore that accounts for its bioactivities owing to their
ability to accept electrons [24]. Structureeactivity relationship
study indicated that cytotoxic activity of 1,4-naphthoquinones are
closely related with their electron accepting capability [25], which
gives rise to reactive oxygen species (ROS) production leading to
DNA damage and cell death [2,25,26]. Furthermore, enhanced
antimicrobial activity was reported in amino derivatives of 1,4-
naphthoquinone [13,27]. The findings indicated that the nitrogen
atom of the amino substituent could improve the redox potential of
the quinone system [25,28]. In addition, insertion of a chlorine
substituent into the amino derivative of 1,4-naphthoquinone
* Corresponding authors.
E-mail addresses: supaluk@swu.ac.th (S. Prachayasittikul), virapong.pra@
mahidol.ac.th (V. Prachayasittikul).
http://dx.doi.org/10.1016/j.ejmech.2014.07.024
0223-5234/© 2014 Published by Elsevier Masson SAS.

248
V. Prachayasittikul et al. / European Journal of Medicinal Chemistry 84 (2014) 247e263
nucleus was found to improve the redox potential of compounds
while not significantly affecting their hydrophobicities [25].
Besides the mechanism of ROS production, cytotoxic effect of
clinically used quinone-based anticancer drugs are closely related
to an inhibition of DNA topoisomerase II enzyme [3,29]. The DNA
topoisomerase II is an essential enzyme required for DNA replica-
tion, chromosome condensation and chromosome segregation
[3,30]. DNA unwinding process is a crucial step in which the coils of
double stranded DNA needs to be uncoiled thereby allowing further
processes of DNA replication [3].
In silico approaches play an important role in drug design and
development. Quantitative structureeactivity relationship (QSAR)
is a computational method for correlating chemical structures of
compounds with their respective biological activities [31e35].
Furthermore, QSAR studies could provide crucial pharmacokinetic
information such as absorption, distribution, metabolism and
excretion [36] which are important parameters in drug design and
development.
The aminoquinones (3e12) were obtained in 26e65% yields as
dark orange, crimson and dark green solid. Structures of the
quinone derivatives were confirmed by spectroscopic methods
comprising of 1H and 13C NMR, IR and HRMS. For example, spectral
data of compound 3 are given. In the HRMSeTOF, the molecular ion
[MþH]þ peak of compound 3 was observed at 264.0783 corre-
sponding to its molecular formula of C14H15ClNO2. The IR spectra of
compound 3 revealed absorption bands of NeH group at 3304 cmꢀ1
and of C]O moiety at 1683 cmꢀ1. The 1H NMR spectra exhibited a
triplet of methyl proton at
methylene protons of amino side chain appeared at
sextet (J ¼ 7.3 Hz), 1.59 ppm as a quintet (J ¼ 7.3 Hz) and
7.51 ppm
d
0.90 ppm with J value of 7.3 Hz. Three
d
1.33 ppm as a
d
d
3.72 ppm as a triplet (J ¼ 7.3 Hz). The broad singlet at
d
was assigned to the NH proton. In addition, four aromatic protons of
naphthoquinone ring displayed at 7.73 (t, J ¼ 7.4 Hz, 1H), 7.82 (t,
J ¼ 7.5 Hz, 1H), 7.97 (d, J ¼ 7.7 Hz, 2H). In the 13C NMR spectra,
characteristic signals of two carbonyl carbons of naphthoquinone
were visible at chemical shift 175.8 and 180.7 ppm.
Herein, a series of 2-amino-1,4-naphthoquinone derivatives
(3e12, Fig. 1) were synthesized using the nucleophilic displace-
ment reaction of 2,3-dichloro-1,4-naphthoquinone (1) by amines
(2) as shown in Scheme 1 and subsequently investigated for their
cytotoxic activity against four human cancer cell lines i.e. HepG2,
HuCCA-1, A549 and MOLT-3. The resulting experimental data were
consequently used for QSAR analysis. In order to elucidate the effect
of 2-substituted amino group on the 1,4-naphthoquinone scaffold,
an additional set of 24 structurally modified compounds with
alkylamino, phenylamino and quinolinylamino substituents
(13e36, Fig. 2) were virtually constructed and their pIC50 values
were predicted from the QSAR models.
2.2. Cytotoxic activity
Cytotoxic activities of the synthesized aminoquinone com-
pounds (3e12) and parent compound (1) were tested against four
human cancer cell lines and normal embryonic lung cells (MRC-5)
using etoposide and doxorubicin as reference drugs (Table 1).
Compounds at 50 mg/mL exhibiting less than 50% inhibition were
considered to be inactive whereas ones exhibiting greater than 50%
inhibition were considered to be active and their IC50 values were
calculated. Compounds 1 and 3e12 were categorized by their
cytotoxic activities as highly active (IC50 < 1
(1 M < IC50 < 10 M), weakly active (IC50 > 10
inactive (IC50 > 50 g/mL) as shown in Table 1.
m
M), moderately active
m
m
mM) [37] and
2. Results and discussion
m
Among all the tested compounds, acetylphenylaminoquinone
compounds (8 and 9) were shown to be the most active compounds
against all cancer cell lines. Compounds 1, 3, 7, 10 and 11 were
inactive to weakly active compounds against all of the tested cell
lines. In addition, compounds 4, 5, 6 and 12 displayed cytotoxic
activity against HuCCA-1 and MOLT-3 cell lines with lower IC50
values as compared to HepG2 and A549 cell lines.
2.1. Chemistry
A series of 2-amino-3-chloro-1,4-naphthoquinones (3e12) were
synthesized via the nucleophilic substitution reaction of 2,3-
dichloro-1,4-naphthoquinone (1) by appropriate primary and sec-
ondary amines (2) in refluxing ethanol (Scheme 1).
The results demonstrated that substitution of the 2-Cl group of
compound 1 with various amino groups afforded compounds with
active and inactive cytotoxic activity. Substitution with alkylamino
(3) and phenylalkylamino (10 and 11) substituents contributed to
reduced cytotoxic activity while quinolinylamino (5) and phenyl-
amino (6, 8 and 9) substituents led to improved cytotoxic activity as
compared to that of the 2,3-dichloroquinone compound (1).
However, these active compounds were found to be less potent
than that of the reference drug (doxorubicin) as deduced by IC50
values. It was noted that compounds 5, 6, 8 and 9 exhibited better
cytotoxic activity against HepG2 cell than the reference drug, eto-
poside. Significantly improved cytotoxic activities were found in
compounds bearing acetylphenylamino substitutions (8 and 9). The
enhanced effect of amino substituents is the following order: ace-
tylphenylamino
>
quinolinylamino
>
alkylamino
>
phenylalkylamino.
Cytotoxicity of compounds 1 and 3e12 (Table 1) toward normal
cell (MRC-5) revealed that quinones 3, 10 and 11 were shown to be
non-cytotoxic. Selectivity index (SI) of these compounds repre-
sented by the ratio of cytotoxicities between normal cell and
different cancer cell lines is summarized in Table 2.
The results revealed that the most active cytotoxic compounds
(8 and 9), particularly, compound 8 had higher SI on HepG2,
HuCCA-1, and A549 with the respective values of 3.26, 6.58 and 1.26
as compared to compound 9 (SI ¼ 1.26, 3.73 and 0.62, respectively).
Inversely, compound 9 had higher SI value (5.79) on MOLT-3 as
Fig. 1. Chemical structures of compounds 1 and 3e12.

V. Prachayasittikul et al. / European Journal of Medicinal Chemistry 84 (2014) 247e263
249
Scheme 1. Synthesis of aminonaphthoquinones derivatives 3e12 by nucleophilic substitution reaction of 2,3-dichloronaphthoquinone (1).
compared to compound 8 (SI ¼ 4.87). It should be noted that
compound 8 showed the highest SI value (6.58) on HuCCA-1 while
the highest SI (5.79) of compound 9 was observed on MOLT-3 cell.
Interestingly, the highest SI (11.46) among the investigated com-
pounds was found to be compound 7 toward MOLT-3 cell.
Cytotoxic activities of the investigated 1,4-quinone analogs (1
and 3e12) against HepG2, HuCCA-1, A549, MOLT-3 and MRC-5 cells
have not been previously reported before. So far, N-phenylamino
(4) and N-methyl-N-phenylamino (6) 1,4-quinone analogs showed
inhibitory effects against cancer cell lines (MCF-7, DU145 and T24)
[25] whereas 2,3-dichloroquinone (1) displayed antiproliferative
effect on cervical cancer (Hela) cells [5].
Cytotoxic effect of 1,4-quinone derivatives acts mainly by
inhibiting the DNA topoisomerase II enzyme [38]. Furthermore, the
quinone moiety can also produce ROS that causes DNA damage
thereby leading to cell death [39,40]. Moreover, cytotoxic activity
could also be explained by the electron accepting capability of
compounds [24] in which the electrophilic center is prone to be
attacked by cellular nucleophiles. On the basis of these principles,
highly potent cytotoxic compounds are likely to require important
molecular properties such as lipophilicity in reaching intracellular
target sites, appropriate size, conformation or geometry for inter-
acting with enzyme binding sites as well as good electron accepting
capability to serve as target sites for cancer cells.
It can be deduced from the experimental results that the elec-
tron accepting capability of the 1,4-quinone compound is increased
when one chlorine atom of the parent compound (1) was
substituted with the phenylamino group as observed in com-
pounds 6, 8 and 9, and with the quinolinylamino group as in
compound 5. Apparently, such substitutions enhanced the elec-
trophilicity of the 1,4-quinone moiety in interacting with cellular
nucleophiles by delocalizing nitrogen lone pair electrons to the
phenyl ring which is directly connected to the amino group. In
strike contrast, substitutions with alkylamino (3) and phenyl-
alkylamino (10 and 11) groups afforded compounds with decreased
cytotoxic activity as well as with non-cytotoxicity toward normal
cell. This may be attributed to the fact that electron-donating ef-
fects of the alkyl group on the N-atom of amino groups could
deactivate the reactivity of 1,4-quinone toward cellular nucleophilic
attack.
2.3. QSAR study of cytotoxic activity
Quantum chemical and molecular descriptors of the 35 com-
Fig. 2. Chemical structures of compounds 13e36.
pounds (1 and 3e36) were calculated using Gaussian [41] and

250
V. Prachayasittikul et al. / European Journal of Medicinal Chemistry 84 (2014) 247e263
Table 1
Cytotoxic activity of compounds 1 and 3e12 against four cancer cell lines and normal cell (MRC-5).
Compound
IC50
(m
M)
MRC-5
HepG2
HuCCA-1
A549
MOLT-3
1
3
4
5
6
7
8
9
65.187 4.38a
64.462 1.73a
88.118 5.00a
12.935 1.15a
12.528 0.55a
81.38 2.89a
4.758 0.78b
9.731 0.29b
inactivec
19.204 0.51a
61.163 4.20a
8.636 0.35b
7.916 0.64b
5.206 0.07b
50.134 0.66a
2.364 0.53b
3.285 1.03b
inactivec
45.366 0.56a
81.829 1.87a
63.798 5.83a
53.769 12a
36.945 5.66a
inactivec
13.037 0.50a
46.337 2.30a
3.489 0.08b
3.824 0.34b
3.157 0.16b
7.506 1.07b
3.193 0.28b
2.118 0.03b
29.061 0.91a
inactivec
22.99 0.42
non-cytotoxic
16.39 1.28
13.59 1.33
8.60 0.77
86.05 3.04
15.53 0.75
12.28 1.08
non-cytotoxic
non-cytotoxic
20.78 0.68
ed
12.279 1.41a
19.647 1.98a
inactivec
10
11
12
71.730 7.64a
inactivec
inactivec
inactivec
20.389 0.00
0.570 0.13
10.672 0.42a
ed
inactivec
9.178 0.25b
0.051 0.01
ed
Etoposide
Doxorubicin
ed
0.239 0.02
0.258 0.03
non-cytotoxic
a
Weakly active compound.
b
Moderately active compound.
Inactive compound.
Not tested.
c
d
Dragon [42] to obtain 13 and 3224 descriptors, respectively.
Redundant and highly correlated descriptors were removed using
an in-house developed Monte CarloeMultiple Linear Regression
(MCeMLR) method or correlation-based. The remaining set of 10
significant descriptors was comprised of Mor04m, RDF050m,
RDF060m, RDF065m, EEig03d, ESpm13d, R8vþ, R5m, HATS4p and
De. Definitions of these descriptors and the calculated values of
significant descriptors are shown in Tables 3 and 4, respectively.
values of structurally modified compounds (13e36) against the
four tested cell lines are shown in Table 7.
2.6. QSAR analysis and molecular descriptor interpretations
QSAR modeling utilizing the MLR method had successfully been
demonstrated on a wide range of chemical and biological proper-
ties [33e35]. Therefore, experimental IC50 values of eleven tested
compounds (1 and 3e12) were employed for constructing QSAR
models against four data sets corresponding to the four tested cell
lines i.e. HepG2, HuCCA-1, A549 and MOLT-3. For each data set,
compounds with inactive cytotoxic activity were excluded from the
analysis. In order to improve the statistical quality of the model,
outlying compounds were identified using Eq. (2) as described in
the experimental section. Herein, compound 4 was identified as an
outlier for the HuCCA-1 model and was subjected to removal
whereas none of the outliers were identified in HepG2, A549 and
MOLT-3 models.
2.4. Evaluation of predictive performance of QSAR models
Details of the predictive performance of constructed QSAR
models are shown in Table 5 in which the HuCCA-1 model provided
the best performance as observed from RCV and RMSECV values of
0.9753 and 0.1146, respectively. Experimental and predicted pIC50
values of cytotoxic activity of compounds 1 and 3e12 as tested
against four cancer cell lines are summarized in Table 6 and Fig. 3.
Interpretation of molecular descriptors was performed by
analyzing the relationship between the identified significant de-
scriptors and their respective pIC50 values obtained by two data
2.5. Prediction of structurally modified compounds (13e36) using
QSAR models
Predicted pIC50 values of 24 modified compounds 13e36 (Fig. 2)
were calculated according to the QSAR equations (Table 5) as a
function of significant descriptor values (Table 4). Compounds
13e36 were categorized by their predicted cytotoxic activities as
highly active (pIC50 > 0), moderately active (ꢀ1 < pIC50 < 0) and
weakly active to inactive (pIC50 < ꢀ1) [37]. The predicted pIC50
Table 3
Definition of descriptors using for development of QSAR models.
Descriptor Type
Definition
Mor04m 3D-MoRSE descriptors 3D-MoRSEesignal 04/weighted by atomic
masses
RDF050m RDF descriptors
RDF060m RDF descriptors
RDF065m RDF descriptors
Radial distribution function-5.0/weighted
by atomic masses
Radial distribution function-6.0/weighted
by atomic masses
Radial distribution function-6.5/weighted
by atomic masses
Table 2
Selectivity index (SI) of compounds 1 and 3e12.
Compound
SIa
HepG2
HuCCA-1
A549
MOLT-3
EEig03d
Edge adjacency indices Eigenvalue 03 from edge adjacency matrix
weighted by dipole moments
1
3
0.35
ND
1.20
ND
0.51
ND
1.76
ND
ESpm13d Edge adjacency indices Spectral moment 13 from edge adjacency
4
5
6
7
8
9
10
11
12
0.19
1.05
0.69
1.06
3.26
1.26
ND
1.90
1.72
1.65
1.72
6.58
3.73
ND
0.26
0.25
0.23
ND
1.26
0.62
ND
4.70
3.56
2.72
11.46
4.87
5.79
ND
matrix weighted by
dipole moments
GETAWAY descriptors R maximal autocorrelation of lag 8/
weighted by van der Waals volume
GETAWAY descriptors R autocorrelation of lag 5/weighted by
atomic masses
GETAWAY descriptors Leverage-weighted autocorrelation of lag 4/
weighted by atomic polarizabilities
R8vþ
R5m
HATS4p
De
ND
ND
ND
1.95
ND
ND
ND
2.26
WHIM descriptors
D total accessibility index/weighted by
Sanderson electronegativities
a
SI ¼ IC50 normal cell/IC50 cancer cell. ND ¼ not determine.

V. Prachayasittikul et al. / European Journal of Medicinal Chemistry 84 (2014) 247e263
251
Table 4
Values of significant molecular descriptors for predicting the cytotoxic activity of four cell lines.
Compound
Mor04m
R8vþ
RDF065m
RDF050m
RDF060m
EEig03d
ESpm13d
R5m
HATS4p
De
1
3
4
5
6
7
8
9
1.23
0.819
0.611
0.861
1.103
0.248
1.634
1.108
0.441
0.095
0.708
0.492
0.908
0.51
0.598
0.513
0.563
0.251
0.144
ꢀ0.134
0.818
0.336
1.47
1.179
0.833
0.75
1.273
1.016
1.909
1.387
2.486
1.353
1.04
0
5.89
4.217
7.003
8.381
11.549
7.334
9.284
11.011
5.763
7.633
9.37
3.786
3.787
4.245
4.431
5.291
6.251
7.18
7.187
7.822
11.963
9.203
8.325
10.123
11.354
11.804
8.969
13.419
6.348
7.417
10.87
12.992
5.17
8.77
8.205
5.842
5.049
8.1
2.82
17.829
17.699
17.712
17.752
17.827
18.018
17.964
17.961
17.708
17.779
17.73
17.695
17.698
17.699
17.699
17.699
17.699
17.699
17.712
17.967
17.733
17.752
17.752
17.844
17.846
17.846
17.991
17.988
17.801
17.799
19.888
19.888
17.711
17.782
17.731
0.427
0.348
0.653
0.625
0.647
0.685
0.623
0.593
0.339
0.393
0.554
0.400
0.446
0.394
0.321
0.305
0.321
0.290
0.499
0.480
0.619
0.619
0.587
0.689
0.693
0.651
0.612
0.642
0.654
0.601
0.780
0.732
0.380
0.471
0.533
0.194
0.143
0.202
0.208
0.205
0.185
0.193
0.205
0.170
0.144
0.165
0.164
0.125
0.140
0.14
0.134
0.135
0.125
0.169
0.185
0.216
0.205
0.187
0.149
0.143
0.145
0.182
0.216
0.203
0.199
0.188
0.199
0.174
0.155
0.137
0.380
0.404
0.476
0.447
0.491
0.505
0.500
0.512
0.385
0.491
0.490
0.369
0.383
0.391
0.410
0.420
0.425
0.433
0.442
0.435
0.474
0.465
0.525
0.488
0.421
0.382
0.505
0.522
0.431
0.430
0.513
0.536
0.394
0.475
0.438
0.008
0.012
0.026
0.015
0.012
0.024
0.02
2.734
2.805
3.079
3.061
3.398
3.399
3.399
2.761
3.208
2.885
2.725
2.732
2.733
2.734
2.734
2.734
2.734
2.805
3.383
2.958
3.079
3.079
3.097
3.102
3.102
3.399
3.399
2.963
2.955
3.399
3.399
2.81
13.899
20.974
11.461
16.143
18.243
15.136
7.867
10.577
15.387
6.317
6.551
7.008
8.564
8.796
9.095
9.197
7.526
12.114
16.037
16.047
22.359
11.192
13.387
14.038
17.264
14.337
11.652
10.857
18.273
14.278
7.626
7.182
9.176
12.538
12.134
8.449
7.332
9.135
11.233
2.937
3.647
5.091
5.082
5.055
5.236
5.113
9.394
10.71
9.194
8.056
8.878
9.031
9.664
8.27
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
0.016
0.01
0.008
0.008
0.009
0.008
0.012
0.01
0.008
0.006
0.009
0.008
0.028
0.024
0.018
0.013
0.013
0.012
0.014
0.023
0.011
0.01
9.72
8.44
4.969
5.705
16.471
8.729
5.172
6.101
9.658
0.024
0.02
0.014
0.01
0.764
0.871
7.758
9.329
12.165
15.627
3.21
2.914
0.011
sets comprising of tested compounds (1 and 3e12) and virtually
constructed compounds (13e36). As to afford a more uniform
distribution of data values, IC50 were converted to pIC50 by taking
the negative logarithm to the base of 10. Therefore, high pIC50
values represented low IC50 values indicating potent cytotoxic ac-
tivity whereas low pIC50 values represented high IC50 value indi-
cating less potent activity.
The QSAR equation of the HepG2 model is shown below:
pIC50 ¼ 0:5509ðMor04mÞ þ 38:3151ðR8vþÞ ꢀ 2:4803
R8vþ representing the van der Waals volume seems to be the
most important descriptor as deduced from its highest regression
coefficient value of 38.3151. The highest R8vþ (0.024) and high
Mor04m (1.634) values among the tested compounds were both
found in compound 8 (Table 4). Compound 32 (a trifluoroacetyl
analog of compound 8) exhibited the most potent predicted cyto-
toxic activity among the series of modified compounds as shown in
Table 7. Particularly, structural modification of compound 8 by
replacing the hydrogen atom of acetyl group with a fluorine atom
bearing higher atomic number to obtain compound 32 could
significantly improve the Mor04m (2.486) value by 1.52 folds as
compared to its parent compound 8 (Mor04m ¼ 1.634). The results
showed that the predicted pIC50 values of compound 8 (Table 6)
2.6.1. HepG2 and HuCCA-1 models
Similar results were found in both HepG2 and HuCCA-1 models
in which important descriptors concerning atomic masses were
governing the cytotoxic activity. The most potent compound
against each cell line is represented as the highest pIC50 values. In
both models, the highest cytotoxic activity among those investi-
gated was compound 8 while that of the virtual constructs was
observed in compound 32.
Table 5
Summary of QSAR equations and their predictive performances in predicting cytotoxic activity of naphthoquinone derivatives against four cell lines.
Cell line
Equation
N
RTr
RMSETr
RCV
RMSECV
HepG2
HuCCA-1
A549
pIC50 ¼ 0.5509 (Mor04m) þ 38.3151 (R8vþ) ꢀ 2.4803
9
8
7
0.9532
0.9941
0.9952
0.9860
0.1400
0.0526
0.0266
0.0708
0.9177
0.9753
0.9745
0.9624
0.1881
0.1146
0.0614
0.1187
pIC50 ¼ 0.1159 (RDF065m) þ 0.739 (Mor04m) þ 0.029 (RDF050m) ꢀ 3.1356
pIC50 ¼ 0.0386 (RDF060m) þ 0.0599 (EEig03d) þ 1.8865 (ESpm13d) ꢀ 35.6888
pIC50 ¼ ꢀ0.699 (R5m) þ 15.5389 (HATS4p) þ 5.6347 (De) ꢀ 5.9617
MOLT-3
10
pIC50 is the concentration of compound required for 50% inhibition of cell growth.
N is a number of data set.
RTr is a correlation coefficient of the training set.
RMSETr is a root mean square error of the training set.
RCV is a correlation coefficient of leave-one-out cross validation (LOO-CV) of the testing set.
RMSECV is a root mean square error LOO-CV of the testing set.

252
V. Prachayasittikul et al. / European Journal of Medicinal Chemistry 84 (2014) 247e263
Table 6
Experimental and predicted cytotoxic activities (pIC50) of compounds 1 and 3e12 against four cancer cell lines.
Compound
HepG2
Exp.
HuCCA-1
Exp.
A549
Exp.
MOLT-3
Exp.
Pred.
Pred.
Pred.
Pred.
1
3
4
5
6
7
8
9
ꢀ1.814
ꢀ1.809
ꢀ1.945
ꢀ1.112
ꢀ1.098
ꢀ1.911
ꢀ0.677
ꢀ0.988
ea
ꢀ1.775
ꢀ1.703
ꢀ1.64
ꢀ0.942
ꢀ1.332
ꢀ1.872
ꢀ0.643
ꢀ1.13
ea
ꢀ1.283
ꢀ1.786
outlier
ꢀ0.899
ꢀ0.717
ꢀ1.700
ꢀ0.374
ꢀ0.517
ea
ꢀ1.217
ꢀ1.597
outlier
ꢀ0.802
ꢀ0.523
ꢀ1.286
ꢀ0.305
ꢀ0.619
ea
ꢀ1.657
ꢀ1.913
ꢀ1.805
ꢀ1.731
ꢀ1.568
ea
ꢀ1.669
ꢀ1.961
ꢀ1.783
ꢀ1.751
ꢀ1.506
ea
ꢀ1.115
ꢀ1.666
ꢀ0.543
ꢀ0.583
ꢀ0.499
ꢀ0.875
ꢀ0.504
ꢀ0.326
ꢀ1.463
ea
ꢀ1.094
ꢀ1.762
ꢀ0.614
ꢀ0.688
ꢀ0.453
ꢀ0.608
ꢀ0.598
ꢀ0.247
ꢀ1.333
ea
ꢀ1.089
ꢀ1.293
ea
ea
ea
ꢀ1.214
ꢀ1.232
ea
ea
ea
10
11
12
ꢀ1.856
ꢀ2.182
ea
ea
ea
ea
ꢀ1.028
ꢀ1.001
ꢀ0.963
ꢀ1.059
a
e ¼ the compound was experimentally inactive and was excluded from the data set of QSAR analysis.
and compound 32 (Table 7) towards HepG2 were ꢀ0.643
and ꢀ0.191, respectively. This led to marked improvements (3.39
folds) in the cytotoxic activity of compound 32 when compared to
compound 8 as represented by the higher pIC50 value of the analog
32.
The worst predicted cytotoxic activity against HepG2 cell line
was found in compound 21 in which the negative Mor04m value
of ꢀ0.134 (Table 4) was found in concomitant with low pIC50 value
of ꢀ2.248 (Table 7). Chemical structure of compound 21 was
modified from compound 4 by replacing the phenylamino moiety
with the 1-adamantyl amino group. The loss of cytotoxic activity is
plausibly due to steric hindrance of the bulky 1-adamantyl group
that affects its binding to the active site of DNA topoisomerase II.
The QSAR equation of the HuCCA-1 model is shown below:
It can be seen that the Mor04m descriptor is primarily respon-
sible for prediction of the cytotoxic activity of HuCCA-1 cell line as
deduced from its high regression coefficient value of 0.739. Simi-
larly, the trifluoroacetyl analog 32 (Mor04m ¼ 2.486) could
significantly increase the value of the atomic mass descriptor
Mor04m by 1.52 folds when compared to the acetyl compound 8
(Mor04m ¼ 1.634) thereby leading to markedly improved cytotoxic
activity. However, the cytotoxic activity of compound 32
(pIC50 ¼ 0.491) against HuCCA-1 was 1.26 folds less potent than the
reference drug, doxorubicin, showing pIC50 value of 0.621 (Table 7).
On the other hand, the worst predicted cytotoxic activity against
HuCCA-1 cell line (Table 7) was found to be compound 13
(pIC50 ¼ ꢀ2.150) which also had the lowest atomic masses
weighted descriptor values of 3.786 for RDF065m and 6.317 for
RDF050m as shown in Table 4. It could be explained that compound
13 has the methylamino group on the 1,4-quinone core structure in
which methyl is the smallest group on the amino moiety among the
investigated series.
pIC50 ¼ 0:1159ðRDF065mÞ þ 0:739ðMor04mÞ
þ 0:029ðRDF050mÞ ꢀ 3:1356
Fig. 3. Plots of experimental versus predicted pIC50 values of cytotoxic activities against four cell lines (A: HepG2, B: HuCCA-1, C: A549, D: MOLT-3) generated by QSAR models
(training set: compounds are represented by closed hex and regression line is shown as a solid line, leave-one-out validated testing set: compounds are represented by open
diamond and regression line is shown as a dotted line).

V. Prachayasittikul et al. / European Journal of Medicinal Chemistry 84 (2014) 247e263
253
Table 7
interacting with the target site as compared to p-compound 33.
Similar findings were found for m-acetyl compound 8 in which its
higher RDF060m value (12.134) afforded better cytotoxic activity
when compared to p-acetyl compound 9 with RDF060m value of
8.449. Apparently, compounds with m-acetylphenylamino (8) or
m-trifluoroacetylphenylamino (32) substituents exhibited more
potent cytotoxic activity than the corresponding para-substituents
(9 and 33). Similar effects of m- and p-acetyl substituents on the
phenyl ring (8, 9, 32 and 33) were noted for the prediction of
cytotoxic activity against HepG2 and HuCCA-1 cell lines.
It could be assumed from these findings that the ortho, meta
and para position of substituents could affect the cytotoxic activity
in which only particular isomers can act as potent anticancer
agents.
Predicted cytotoxic activity (pIC50) of modified compounds (13e36) and experi-
mental cytotoxic activity of reference drugs.
Compound
pIC50
HepG2
HuCCA-1
A549
MOLT-3
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
ꢀ1.903a
ꢀ1.635a
ꢀ1.893a
ꢀ1.691a
ꢀ1.815a
ꢀ1.864a
ꢀ2.112a
ꢀ2.056a
ꢀ2.248a
ꢀ0.957b
ꢀ1.376a
ꢀ0.981b
ꢀ1.333a
ꢀ1.523a
ꢀ1.607a
ꢀ1.243a
ꢀ1.039a
ꢀ1.007a
ꢀ1.333a
ꢀ0.191b,c
ꢀ0.969b
ꢀ1.371a
ꢀ1.676a
ꢀ1.579a
ꢀ1.309
ꢀ2.150a
ꢀ1.836a
ꢀ2.063a
ꢀ1.932a
ꢀ1.888a
ꢀ1.731a
ꢀ1.851a
ꢀ1.978a
ꢀ1.977a
ꢀ0.680b
ꢀ1.355a
ꢀ0.436b
ꢀ0.767b
ꢀ0.816b
ꢀ0.806b
ꢀ0.655b
ꢀ0.414b
ꢀ0.651b
ꢀ0.936b
0.491c,d
ꢀ0.216b
ꢀ1.545a
ꢀ1.319a
ꢀ0.958b
eg
ꢀ2.031a
ꢀ1.997a
ꢀ1.939a
ꢀ1.940a
ꢀ1.941a
ꢀ1.934a
ꢀ1.939a
ꢀ1.744a
ꢀ1.178a
ꢀ1.703a
ꢀ1.704a
ꢀ1.673a
ꢀ1.492a
ꢀ1.463a
ꢀ1.517a
ꢀ1.170a
ꢀ1.225a
ꢀ1.738a
ꢀ1.714a
2.669c,d,e
2.370d,e
ꢀ1.909a
ꢀ1.715a
ꢀ1.692a
eg
ꢀ1.614a
ꢀ2.173a
ꢀ1.858a
ꢀ1.700a
ꢀ1.726a
ꢀ1.694a
ꢀ1.782a
ꢀ1.194a
ꢀ0.971b
ꢀ0.367b
ꢀ0.589b
ꢀ0.508b
ꢀ1.378a
ꢀ1.852a
ꢀ2.011a
ꢀ0.716b
ꢀ0.113b,c
ꢀ0.836b
ꢀ0.867b
ꢀ0.695b
ꢀ0.361b
ꢀ1.303a
ꢀ1.206a
ꢀ1.737a
1.292
2.6.3. MOLT-3 model
Distinct from other cell lines, the cytotoxic activity of com-
pounds against the MOLT-3 cell line is dependent on atomic
polarizability, electronegativity and atomic mass as shown in the
equation below:
pIC50 ¼ ꢀ0:699ðR5mÞ þ 15:5389ðHATS4pÞ þ 5:6347ðDeÞ
ꢀ 5:9617
It can be noted that the atomic polarizability-weighted
descriptor (HATS4p) was the most influential descriptor for pre-
dicting the cytotoxic activity owing to the fact that it afforded the
Etoposidef
Doxorubicinf
0.244
0.621
0.588
eg
a
highest
regression
coefficient
value
(15.5389).
The
Weakly active to inactive compound.
Moderately active compound.
The most potent compound against each cancer cell line.
Highly active compound.
Compound with more potent cytotoxic activity compared to reference drugs.
Experimental pIC50 values.
Not tested.
b
c
d
e
f
electronegativity-weighted descriptor (De) ranked the second most
important one affording regression coefficient of 5.6347 while
atomic mass-weighted descriptor (R5m) plays its minimal role with
negative regression coefficient of ꢀ0.699. According to the QSAR
equation, high HATS4p and high De values together with low R5m
value are likely required for potent cytotoxic activity. It was found
that p-amino benzoic acid analog 29 was predicted as the most
potent anticancer agent against MOLT-3 cell line as represented by
the highest predicted pIC50 value of ꢀ0.113 (Table 7). However, such
compound exhibited approximately 11.4 folds less potent activity
than the reference drug, etoposide (pIC50 ¼ 1.292, Table 7). The
potent cytotoxic activity of compound 29 could possibly be gov-
erned by its high HATS4p and De values resulting from high elec-
tronegativity O-atom of CO2H group that can be polarized and
withdraw lone pair electron on N-atom of amino group, giving rise
to a significantly improved redox potential of the compound.
Within the same series of isomeric amino benzoic acid analogs (7,
28 and 29), the p-amino benzoic acid analog (29) provided the
highest HAPS4p value (0.216, Table 4) as compared to ortho (7) and
meta (28) analogs with HAPS4p values of 0.185 and 0.182, respec-
tively (Table 4). The analog 29 had approximately 5 and 6 folds
greater activity than ortho (7) and meta (28) analogs in which their
predicted pIC50 values were ꢀ0.608 (Table 6) and ꢀ0.716 (Table 7),
respectively. In addition, isomeric effects of COCH3 group at m- and
p-positions can significantly affect polarizability of the compounds
by decreasing HATS4p values from p- to m-analogs as found in
compounds 9 and 8 with the values of 0.205 and 0.193, respectively
(Table 4). In case of COCF3 analogs 32 and 33, their HATS4p values
were decreased as compared to acetyl analogs 8 and 9, and shown
to be 0.188 and 0.199, respectively. Apparently, the reduced polar-
izabilities values of 32 and 33 could be due to the lipophilic effect of
trifluoroacetyl group. These results revealed that substitution of
acetyl (9), trifluoroacetyl (33) and carboxyl (29) groups at para
position on phenylamino moiety enhances cytotoxic activity of
compounds against MOLT-3 cell line while substitution at meta and
ortho positions aggravates the activity. In contrast, substitution of
these functional groups at meta position seems to be the
g
2.6.2. A549 model
Besides atomic masses, dipole moment properties of com-
pounds are highly involved in cytotoxic activity against the
A549 cell line as shown by the QSAR equation below:
pIC50 ¼ 0:0386ðRDF060mÞ þ 0:0599ðEEig03dÞ
þ 1:8865ðESpm13dÞ ꢀ 35:6888
It is noted that high regression coefficients were observed for
dipole moment weighted descriptors, particularly ESpm13d, which
is shown to be more important in predicting the cytotoxic activity
as compared to atomic mass weighted descriptor, RDF060m.
Interestingly,
isomeric
trifluoroacetyl
compounds
32
(pIC50 ¼ 2.669) and 33 (pIC50 ¼ 2.370), which are derivatives of the
corresponding acetyl compounds 8 and 9, exhibited significantly
more potent cytotoxic effect against A549 cell line when compared
to doxorubicin (pIC50 ¼ 0.588). This indicated that the respective
pIC50 values of m- (32) and p- (33) compounds were approximately
4.5 folds and 4 folds higher than that of the reference drug. These
results showed that the fluorine atom may improve the lip-
ophilicity of such compounds thereby allowing them to reach
intracellular target sites. As shown in Table 4, both compounds (32
and 33) possessed exactly the same values for dipole moment
weighted descriptors, EEig03d (3.399) and ESpm13d (19.888),
while bearing different values for the atomic mass weighted
descriptor, RDF060m (16.471 and 8.729, respectively). Moreover,
compound 32 had RDF060m value of 2 folds greater than that of
compound 33. Such high RDF060m value of compound 32 could
plausibly be attributed to the effect of substituent on the m-posi-
tion that makes the compound with more appropriate shape in

254
V. Prachayasittikul et al. / European Journal of Medicinal Chemistry 84 (2014) 247e263
appropriate strategy for developing anticancer agents against
HepG2, HuCCA-1 and A549 cell lines.
Regarding the effect of fluorine substitution deteriorating cytotoxic
activity against MOLT-3 cell line by increasing atomic mass-weight
descriptor value (R5m), it could be hypothesized that besides the
lipophilicity, the appropriate atomic polarizabilities and masses or
size of the molecules are crucial for exerting potent cytotoxic ac-
tivity against MOLT-3 cell line in terms of cell entry and interaction
at intracellular target sites.
The effect of modifications to chemical structures on descriptor
values and cytotoxic activity against the four investigated human
cancer cell lines are summarized in Tables 8e11. Particularly, the
experimental pIC50 of tested compounds 1 and 3e12 and the pre-
dicted pIC50 of virtual constructs 13e36 along with their significant
descriptor values were provided. It can be deduced that sub-
stituents (R) on the 2-amino position of 1,4-naphthoquinone core
structures (Figs. 1 and 2) may affect the cytotoxic activity of com-
pounds by improving or aggravating it, which depends on molec-
ular properties of compounds that are influenced by their
substituents. Apparently, compounds which R ¼ acetylphenyl
groups (8 and 9) were found to be the most potent anticancer
agents. Moreover, fluorine substituents (R ¼ trifluoroacetylphenyl)
of virtually constructed compounds (32 and 33) exhibited markedly
improved cytotoxic activity against every cancer cell lines except
for MOLT-3 cell. Furthermore, some particular isomers seemed to
Previously, QSAR study of antiproliferative and cytotoxic activ-
ities of 1,4-naphthoquinone derivatives against the L1210 cell line
demonstrated that the hydrophobic effect is an important param-
eter for defining the bioactivities [43]. Herein, the trifluoroacetyl
compound 33 exhibited the second most potent activity against
MOLT-3 cell line among all of structural modified compounds
affording the predicted pIC50 value of ꢀ0.361 (Table 7). However,
compound 33 is less potent than its parent compound (9). This
could be due to the fact that replacement of acetyl group in com-
pounds 8 and 9 with trifluoroacetyl group can significantly increase
atomic mass-weighted descriptor (R5m) value leading to higher
lipophilicity (or lower polarizabilities) that decreased cytotoxic
activity as demonstrated in compounds 32 and 33. The R5m values
of compounds 8, 9, 32 and 33 are 0.623, 0.593, 0.780 and 0.732,
respectively (Table 4). Such results indicate HAPS4p as the most
influential descriptor governing the cytotoxic activity toward
MOLT-3 cell line. Interestingly, these findings are different from
those of HepG2, HuCCA-1 and A549 cell lines in which fluorine
analogs (32 and 33) exhibited more potent activity as compared to
their parents compounds (8 and 9) by improving lipophilicity of the
compounds thereby enhancing cell entry to reach their target sites.
Table 8
The effect of phenyl and cycloalkane rings on values of significant descriptors and cytotoxic activity against four cell lines.
Effect of substituent
HepG2
HuCCA-1
A549
MOLT-3
Most potent compound
32
32
32a
29
Worst potent compound
21
13
13
14
YMor04mb
Yactivityb
YMor04mb
[RDF065mb
[RDF050mb
[activityb
YESpm13db
Yactivityb
[all descriptorsb
[activityb
YMor04mb
Yactivityb
YMor04mb
Yactivityb
YESpm13db
Yactivityb
YHATS4pb
[R5mb
Yactivityb
YMor04mb
Yactivityb
YMor04mb
Yactivityb
[all descriptorsb
[activityb
YHATS4pb
[R5mb
[Deb
[activityb
[Mor04mb
[activityb
[Mor04mb
[activityb
YESpm13db
Yactivityb
[all descriptorsb
[activityb
[Mor04mb
[activityb
[Mor04mb
[activityb
YESpm13db
Yactivityb
[R5mb
[Deb
[activityb
Fluorine substituents
(32 & 33)
[all descriptorsb
[activityb
[all descriptorsb
[activityb
[ESpm13db
[RDF060mb
[activityb
[R5mb
[Deb
[[activityb
a
More potent than reference drug, doxorubicin.
Comparison of descriptor values and cytotoxic activity of considered compound and compound 1.
b

V. Prachayasittikul et al. / European Journal of Medicinal Chemistry 84 (2014) 247e263
255
Table 9
The effect of substituted phenyl on values of significant descriptors and cytotoxic activity against four cell lines.
Effect of substituent
HepG2
HuCCA-1
A549
MOLT-3
Mor04ma: 28 > 29 > 7
activitya: 29 > 28 > 7
activityb: 7Y, 28[, 29[
para > meta > ortho
Mor04ma: 28 > 29 > 7
activitya: 29 > 28 > 7
activityb: 7Y, 28[, 29[
para > meta > ortho
ESpm13da: 28 > 29 > 7
activitya: 28 > 29 > 7
activityb: 7Y, 28[, 29[
meta > para > ortho
HATS4pa: 29 > 7 > 28
activitya: 29 > 7 > 28
activityb: 7[, 28[, 29[[
para > ortho > meta
Mor04ma: 36 > 12
Mor04ma: 36 > 12
RDF050ma: 36 > 12
activitya: 36 > 12
activityb: 12[, 36[
meta > para
EEig03da: 36 > 12
RDF060ma: 12 > 36
activitya: 36 > 12
activityb: 12Y, 36Y
meta > para
HATS4pa: 12 > 36
Dea: 12 > 36
a
R8vþ : 36 > 12
activitya: 36 > 12
activityb: 12Y, 36[
meta > para
activitya: 12 > 36
activityb: 12[, 36Y
para > meta
Mor04ma: 8 > 9
activitya: 8 > 9
activityb: 8[, 9[
meta > para
Mor04ma: 8 > 9
activitya: 8 > 9
activityb: 8[, 9[
meta > para
RDF060ma: 8 > 9
activitya: 8 > 9
activityb: 8[, 9[
meta > para
HATS4pa: 9 > 8
R5ma: 8 > 9
activitya: 9 > 8
activityb: 8[, 9[
para > meta
Mor04ma: 32 > 33
activitya: 32 > 33
activityb: 32[, 33[
meta > para
Mor04ma: 32 > 33
activitya: 32 > 33
activityb: 32[, 33[
meta > para
RDF060ma: 32 > 33
activitya: 32c > 33c
activityb: 32[, 33[
meta > para
HATS4p & Dea: 33 > 32
R5ma: 32 > 33
activitya: 33 > 32
activityb: 32[, 33[
para >> meta
a
Comparison within the same series.
Comparison of descriptor values and cytotoxic activity of considered compound and compound 1.
More potent than reference drug, doxorubicin.
b
c
be more cytotoxic against particular cell lines as shown by the ef-
normal cellular functions [45] and it can be altered by exogenous
compounds such as anticancer drugs [3]. As the concentration of
enzymeeDNA cleavage complex becomes too low, it gives rise to
mitotic failure, slow cell growth and cell death [46e49]. Whereas
when the concentration becomes too high, this induces the con-
version of transient DNA breaks to permanent DNA damages [48]
thereby leading to cell death [50]. In other words, the essential
topoisomerase II enzyme is converted to cellular toxin known as
topoisomerase poison [3]. The alteration of these steps in catalytic
cycle can increase the concentration of enzymeeDNA cleavage
complex. Such anticancer drugs exert their cytotoxic activity as
topoisomerase poison by inhibiting religation of transient DNA-
strand breaks thereby maintaining the enzymeeDNA cleavage
complexes. As a result, the high level of enzymeeDNA cleavage
complexes inhibits DNA synthesis subsequently leading to cell
death [3,29,30].
Topoisomerase II poison can be classified by its mechanism of
action as non-intercalative and intercalative types. Both mecha-
nisms of action increase the concentration of enzymeeDNA cleav-
age complex resulting in cell death. However, non-intercalative
type interacts at enzymeeDNA interfaces whereas the intercalative
type intercalates DNA strands to cause structural changes of DNA
[3].
fects caused by variations of substituents at the ortho, meta and
para-positions as well as length of alkyl chain and position of qui-
nolinyl substituents.
2.7. Molecular descriptor interpretations based on mechanisms of
action of the reference drugs
The steps of DNA unwinding is a crucial process for DNA syn-
thesis whereby it allows the separation of complementary DNA
strands to facilitate initiation and elongation processes of DNA
replication [3]. The DNA unwinding process is catalyzed by the
homodimeric enzyme known as topoisomerase II [3]. Topoisom-
erase II is the essential enzyme that regulates the DNA winding and
unwinding processes. DNA topoisomerase II functions in a cyclic
fashion [3] and ATP is required as a cofactor for driving the catalytic
cycle as well as for enzyme recycling [44]. The catalytic cycle ini-
tiates when topoisomerase II binds double-stranded DNA to form
the enzymeeDNA cleavage complex [29]. Topoisomerase II then
cleaves one of the DNA strands creating a transient strand break
thereby allowing the passage of another intact DNA region [29].
After the DNA passage is completed, the transient DNA strand break
is religated and the enzyme is released from the cleavage complex
[29] to be reused in further catalytic cycles [44]. Thus far, topo-
isomerase II is one of the promising targets for the treatment of
cancers [3,30] owing to its potential to be genotoxic [3] as it creates
transient DNA damage each time it functions [3]. An appropriate
concentration of enzymeeDNA cleavage complex is essential for
Both reference drugs employed in this study are clinically used
topoisomerase II poison. Therefore, their mechanism of actions
could be reasonably applied for describing the structureeactivity
relationships through the utilization of identified significant de-
scriptors and predicted cytotoxic activities.

256
V. Prachayasittikul et al. / European Journal of Medicinal Chemistry 84 (2014) 247e263
Table 10
The effect of quinolinyl ring on values of significant descriptors and cytotoxic activity against four cell lines.
Effect of substituent
HepG2
HuCCA-1
A549
MOLT-3
Quinolinylamino position
activitya
activitya
activitya
activitya
5 > 22 > 24 > 23
24 > 22 > 5 > 23
24 > 22 z 23 > 5
22 > 24 > 23 > 5
b
↑R8vþ , [activityb
↑RDF050mb
↑RDF065mb
[activityb
↑EEig03db
YESpm13db
Yactivityb
[all descriptorsb
[activityb
b
↑R8vþ , [activityb
↑RDF050mb
↑RDF065mb
[activityb
YESpm13db
Yactivityb
[all descriptorsb
[activityb
b
↑R8vþ , [activityb
↑RDF050mb
↑RDF065mb
Yactivityb
YESpm13db
Yactivityb
[all descriptorsb
[activityb
b
↑R8vþ , [activityb
↑all descriptorsb
[activityb
YESpm13db
Yactivityb
YHATS4pb
[R5mb
[Deb
[activityb
a
Comparison within the same series.
Comparison of descriptor values and cytotoxic activity of considered compound and compound 1.
b
Doxorubicin is an intercalative topoisomerase II poison con-
taining the 1,4-quinone moiety as shown in Fig. 4. The cytotoxic
effect of doxorubicin arises from its dual action as topoisomerase II
catalytic inhibitor and topoisomerase II poison [3,29,30]. Firstly,
doxorubicin acts as a topoisomerase inhibitor as to inhibit the
topoisomerase II catalytic cycle. The planar structure of doxorubicin
enhances its intercalation between DNA base pairs [29] giving rise
to DNA structural changes [30] thereby prevents the binding of
topoisomerase II with DNA strands in the initial step of the catalytic
cycle [29] as well as creates permanent DNA damages [29]. Sec-
ondly, doxorubicin acts as a topoisomerase II poison by increasing
and stabilizing enzymeeDNA cleavage complexes [29]. Moreover,
the 1,4-quinone moiety of doxorubicin plays an important role in
the production of ROS leading to oxidative damages to the cells
[30].
Etoposide (Fig. 4) is a non-intercalative topoisomerase II poison
that acts by inhibiting the religation of transient DNA breaks
[29,30]. Etoposide metabolites such as semiquinone and phenoxyl
radical serve as ROS sources thereby leading to cellular damages
[30]. Moreover, semiquinone metabolite is a potent inhibitor of the
topoisomerase II-DNA cleavage complex [51,52] that directly binds
to the DNA content of topoisomerase IIeDNA cleavage complex [29]
and finally resulting in the inhibition of DNA religation. As a
consequence, the increasing level of topoisomerase IIeDNA cleav-
age complex eventually causes cell death [3,29,30]. Such ROS
including superoxide (Oꢁ2ꢀ) and hydroxyl radical (OHꢁ), which can
Table 11
The effect of hydrocarbon chain length on values of significant descriptors and cytotoxic activity against four cell lines.
Effect of substituent
Length of alkyl chain
HepG2
HuCCA-1
A549
MOLT-3
14a
3a
18a
13a
n ¼ 1 (2C length)
activityb: 14[
highest Mor04mc
n ¼ 3 (4C length)
n ¼ 6 (7C length)
n ¼ 0 (1C length)
activityb: 3Y
activityb: 18Y
activityb:13Y
6a
6a
26a
6a
n ¼ 0 (1C length)
n ¼ 0 (1C length)
n ¼ 2 (3C length)
activityb: 26[
highest RDF060mc
n ¼ 0 (1C length)
activityb: 6[
lowest R5mc
highest HATS4p & Dec
activityb: 6[
activityb: 6[
a
The most appropriate chain length that gives the highest activity within the same series.
Comparison of descriptor values and cytotoxic activity of considered compound and compound 1.
Comparison of descriptor values within the same series.
b
c

V. Prachayasittikul et al. / European Journal of Medicinal Chemistry 84 (2014) 247e263
257
substituents on the quinone ring [59]. For example, electron-
donating substituents increase electron density but decrease elec-
trophilicity on the ring thereby impaired redox potential and ROS
producing ability. In contrast, electron-withdrawing substituents
affect vice versa [59]. In addition, it should be considered that the
cytotoxic activity of quinones is influenced by multiple mecha-
nisms and factors including alkylation, ROS production and phar-
macokinetic properties.
It could be deduced from the QSAR analysis that chemical
properties such as high atomic masses, high van der Waal volume,
high electronegativity, high dipole moments and high polarizabil-
ities are partially responsible for potent cytotoxic activity of com-
pounds as shown in Table 12.
Fig. 4. Chemical structures of reference drugs.
These chemical properties are associated with pharmacokinetic
profiles, which affect the activities of compounds. Particularly, cell
entry is a crucial step that allows compounds to reach the intra-
cellular target sites of actions. Appropriate size, shape and lip-
ophilicity of compounds are also important factors facilitating cell
entry. In addition, particular conformations or isomers are required
for their interaction at target sites. The QSAR analysis demonstrated
that values of atomic masses-weighted, dipole moment-weighted
and polarizability-weighted descriptors related to the type, size,
shape, hydrocarbon chain length and position of substituents.
In summary, atomic masses, dipole moments and polarizabil-
ities play integral roles in determining cell entry capability of
compounds. That is the appropriate atomic mass refers to the
appropriate size and shape of compounds that are suitable for cell
entry. Dipole moments and polarizabilities may account for the
charge distribution and net charge of molecules pertaining to the
lipophilicity as well as the ROS production ability. Therefore, the
importance of chemical and pharmacokinetic properties (as
described by the selected significant descriptors) in concomitant
with the mechanism of actions of reference drugs could be used in
explaining the bioactivities of the investigated cell lines. Herein, the
mechanism of action of doxorubicin could be applied for the
explanation of QSAR analysis of HepG2, HuCCA-1 and A549 cell
lines while etoposide could be used for the MOLT-3 cell line.
Doxorubicin exerted its anticancer activity via intercalation be-
tween DNA base pairs [3], stabilization of the topoisomerase II-DNA
cleavage complex [29] as well as production of ROS [29,30]. Shape,
size and charge distribution of compounds are important factors
facilitating the two former mechanism of actions whereas the 1,4-
quinone moiety accounts for the latter mechanism that is the ROS
production. According to QSAR analysis, the shape and size of
compounds are likely to play predominant roles in the cytotoxic
activity against HepG2 and HuCCA-1 cell lines as represented by
atomic masses descriptors (Mor04m, RDF050m and RDF065m) and
atomic van der Waals volume (R8vþ) as shown in Table 5. The
highest value of atomic mass-weighted descriptor (Mor04m), ex-
erts the highest cytotoxic activity against HepG2 and HuCCA-1 cell
lines, as observed in trifluoroacetyl analog 32. In addition, the high
value of atomic van der Waals volume (R8vþ) was found in analog
32 exhibiting the highest cytotoxic activity against the HepG2 cell
line. It is notable that the lipophilic trifluoroacetylphenylamino
moiety and the planar quinone structure of analog 32 facilitated cell
entry and enhanced DNA intercalation thereby resulting in more
potent activity as compared to acetyl compound 8. It could be
explained that replacement of hydrogen atoms of acetyl group (8)
by fluorine atoms gives rise to high atomic masses and high atomic
van der Waals volume as well as increased redox potential of the
trifluoroacetyl analog 32. In contrast, 1-adamantyl amino com-
pound 21 exerted the lowest activity against the HepG2 cell line.
This may be ascribed to the effect of non-planar bulky structure of
the 1-adamantyl moiety that makes compound 21 unsuitable for
DNA intercalation.
be generated from quinone by enzymatic one-electron reduction,
and/or direct reduction (non-enzymatically) via NADPH or NADH
present in the cells (Fig. 5) [8,53,54].
QSAR results demonstrated that cytotoxic activities of com-
pounds 1 and 3e36 against the investigated cell lines were influ-
enced by distinct molecular descriptors. This indicates that such
distinct chemical properties are crucial for the cytotoxic activity
against particular cell lines. Particularly, atomic masses, van der
Waals volume and dipole moments seemed to be important de-
scriptors for the cytotoxic activity against cell lines that use doxo-
rubicin as the reference drug such as HepG2, HuCCA-1 and A549
whereas the atomic polarizability was shown to be crucial for
MOLT-3 that uses etoposide as the reference drug. Therefore, it
could be hypothesized that the mechanism of cytotoxic activity of
quinone compounds against HepG2, HuCCA-1 and A549 cells may
be similar whereas that of MOLT-3 cell may act through a different
mechanism of action.
Alkylation of biomolecules and ROS production have been pro-
posed to be the main mechanisms of quinones cytotoxicity
[24,56,57]. The ability of quinones to act as electrophiles contrib-
utes to alkylation mechanism whereas their ability to act as oxi-
dants gives rise to the ROS production [57]. Quinones are highly
redox active molecules and they are capable of generating ROS via
redox cycles (Fig. 5) [57,59]. The ease of reduction of such com-
pounds is expressed by their redox potentials [59]. The redox po-
tential of quinones is correlated to their ROS production ability as
well as biological activities [6,60,61] and can be modulated by
Fig. 5. Possible anticancer effect of 1,4-quinone (modified from Refs. [55e58]).

258
V. Prachayasittikul et al. / European Journal of Medicinal Chemistry 84 (2014) 247e263
Table 12
Summary of relationships between significant descriptors and possible mechanism of actions of compounds 1 and 3e36 based on mechanism of actions of reference drugs.
Cell line
Reference drug and
mechanism of
action
Significant descriptors and their effects on
chemical properties of compounds
Significant effects
related with cytotoxic
activity of compounds
Requirements of significant
values of descriptor for
potent cytotoxic activity
HepG2
HuCCA-1
Doxorubicin : DNA
intercalation
ROS production
Doxorubicin : DNA
intercalation
van der Waals volume (HepG2)
Atomic mass-weighted (HepG2 & HuCCA-1)
ꢁ Size & Shape
Cell entry
DNA intercalation
High values
A549
Dipole moment weighted
ꢁ Polarity/charge
Cell entry
ROS production
High values
ROS production
DNA intercalation
ROS production
High values
High values
Atomic mass-weighted
ꢁ Size & Shape
MOLT-3
Etoposide:
Semiquinone and
ROS productions
Polarizability-weighted
Electronegativity-weighted
ꢁ Polarity/charge
Cell entry
Low values
Atomic mass-weighted
ꢁ Size & Shape
Furthermore, dipole moment-weighted descriptors (EEig03d
and ESpm13d) play crucial roles in governing the cytotoxic effect
against the A549 cell line. Compounds with high dipole moment
could induce positive and negative charges on the molecule giving
rise to potent cytotoxic activity as seen in trifluoroacetyl com-
pounds 32 and 33 in which analog 32 was shown to be the most
potent compound. Such potent activity of compound 32 could be
attributed to the high electronegativity of fluorine atoms in the
trifluoroacetyl group thereby inducing high charge distribution on
the molecule as well as increasing the electrophilicity of the 1,4-
quinone core structure in interacting with nucleophilic target
cells. Moreover, the ROS production of 1,4-quinone compounds
arises from the generated semiquinone that ultimately produced
superoxide and/or hydroxyl radicals culminating in oxidative cell
damage and cell death.
On the other hand, etoposide is a non-intercalative topoisom-
erase II poison that interacts within the enzyme active site at the
DNAeenzyme interface [48,49,52]. The cytotoxic activity of eto-
poside is highly dependent on the redox reaction, which accounts
for the production of phenoxyl radical and the semiquinone toxic
metabolite [29,52]. All of the studied compounds contained the 1,4-
quinone moiety as the core structure (Figs. 1 and 2), which is the
key pharmacophore for topoisomerase II poison activity. Therefore,
it could be hypothesized that the tested compounds possibly have
the same capability of inhibiting DNA unwinding via the production
of semiquinone (Fig. 5). The QSAR analysis revealed that polariz-
ability and electronegativity of the compound are closely related
with potent cytotoxic activity against MOLT-3 cell line, as shown in
QSAR equation (Table 5), in which high polarizability and high
electronegativity give rise to high redox potential of the com-
pounds thereby enhance semiquinone radical production. Partic-
ularly, the most potent p-amino benzoic acid analog (29) possessed
the highest value of polarizability-weighted descriptor (HATS4p) as
well as high value of electronegativity-weighted descriptor (De) as
shown in Table 4. Similarly, the second most potent p-tri-
fluoroacetylphenylamino compound (33) possessed the highest
value of De and high value of HATS4p (Table 4). Furthermore,
fluorine substitution on compounds 32 and 33 significantly gives
rise to an increase of atomic masses-weighted descriptor (R5m) or
lipophilicity which in turn weakens the cytotoxic activity (MOLT-3)
as compared to their parents (8 and 9) as shown in Table 4. Owing
to higher polarized molecules afforded higher activity, therefore, it
is possible that the molecules with higher lipophilicity may impair
cell entry to reach their intracellular target sites. Of the series of p-
substituents (9, 12, 29 and 33) on phenylamino moiety, their
cytotoxic activity (MOLT-3) are governed by the polarized func-
tional groups as the following: 29 (p-CO2H) > 9 (p-COCH3) > 33 (p-
COCF3) > 12 (p-NH-C6H5).
At this point, the findings from QSAR analyses indicate the sig-
nificance of diverse substituents at specific positions that may
affect the overall chemical properties as well as the pharmacoki-
netic profiles of compounds. A summary of the relationships be-
tween the selected significant descriptors and the possible
underlying mechanisms of cytotoxic activity based on reference
drugs are presented in Table 12.
3. Conclusion
A series of 1,4-naphthoquinone derivatives (3e12) were syn-
thesized and investigated for their cytotoxic activity against
HepG2, HuCCA-1, A459 and MOLT-3 cell lines. Novel anticancer
agents such as compound 8 exhibited the most potent cytotoxic
activity against HepG2, HuCCA-1 and A549 cell lines whereas
compound 9 exerted the most potent activity against the MOLT-
3 cell line. Interestingly, quinones 5, 6, 8, and 9 displayed higher
cytotoxic activity against HepG2 cell than the reference drug,
etoposide. QSAR analysis demonstrated the significance of atomic
masses-weighted descriptors (Mor04m, RDF050m, RDF060m and
RDF065m) in the prediction of cytotoxic activity against HepG2,
HuCCA-1 and A549 cell lines. Moreover, the atomic van der Waals
volume (R8vþ) was also deduced to be involved with the cytotoxic
activity against the HepG2 cell line. In addition, dipole moments-
weighted descriptors (EEig03d and ESpm13d) were also shown to
be important in prediction of the cytotoxic activity against the
A549 cell line. In contrast, polarizabilities-weighted descriptor
(HATS4p) and electronegativities-weighted descriptor (De) in-
fluence the cytotoxic activity against the MOLT-3 cell line. The
results indicated that the cytotoxic activity of such compounds
against particular cell lines might arise from distinct mechanism
of actions. The selected significant descriptors indicated that
certain chemical properties might be relevant to pharmacokinetic
profiles (i.e. cell entry and interaction at the target sites). It could
be deduced that chemical properties such as size and shape, lip-
ophilicity, polarity and charge distribution are important factors
related to pharmacokinetic profiles that may enhance the cyto-
toxic activity. Significantly, the findings revealed the rational
design of trifluoroacetyl analog (32) as a promising potential
anticancer agent toward HepG2, HuCCA-1 and A549 cell lines
while the p-amino benzoic acid analog (29) was the most potent
compound against the MOLT-3 cell line.

V. Prachayasittikul et al. / European Journal of Medicinal Chemistry 84 (2014) 247e263
259
4. Experimental section
4.2.3. 2-Chloro-3-(quinolin-8-ylamino)naphthalene-1,4-dione (5)
Yield 43%. mp 301e303ꢂC. IR (UATR) cmꢀ1: 3299, 1670, 1593. 1H
NMR (300 MHz, dmso-d6) 7.44e7.57 (m, 2H, ArH), 7.63 (t,
4.1. General
d
J ¼ 7.6 Hz, 1H, ArH), 7.72e7.90 (m, 4H, ArH), 8.16 (d, J ¼ 7.5 Hz, 1H,
Column chromatography was carried out using silica gel 60
ArH), 8.40 (d, J ¼ 8.3 Hz, 1H, ArH), 8.69 (d, J ¼ 4.1 Hz, 1H, ArH), 9.45
(70e230 mesh ASTM). Analytical thin-layer chromatography (TLC)
(s, 1H, NH). 13C NMR (75 MHz, dmso-d6)
d 121.8, 122.3, 124.1, 126.5,
was performed with silica gel 60 F254 aluminum sheets. 1H and 13
C
126.7, 127.0, 128.8, 131.2, 132.2, 133.8, 135.0, 136.4, 137.1, 140.8, 146.1,
NMR spectra were recorded on a Bruker AVANCE 300 NMR spec-
trophotometer (operating at 300 MHz for 1H and 75 MHz for 13C).
FTIR spectra were obtained using a universal attenuated total
reflectance attached to a PerkineElmer Spectrum One spectro-
photometer. High-resolution mass spectra (HRMS) were recorded
on a Bruker Daltonics (microTOF). Melting points were determined
using a Griffin melting point apparatus and were uncorrected. HPLC
was carried out using Dionex, Ultimate 3000 equipped with
149.7, 176.9, 180.2. HRMSeTOF: m/z [MþH]þ calcd for
C19H12ClN2O2: 335.0582; found: 335.0589. Purity 96.04%.
4.2.4. 2-(N-methyl-N-phenylamino)-3-chloronaphthalene-1,4-
dione (6)
Yield 56%. mp 100e102ꢂC. IR (UATR) cmꢀ1: 1671, 1590. 1H NMR
(300 MHz, dmso-d6)
ArH), 6.95 (d, J ¼ 7.8 Hz, 2H, ArH), 7.23 (t, J ¼ 7.3 Hz, 2H, ArH),
7.85e8.10 (m, 4H, ArH). 13C NMR (75 MHz, dmso-d6)
39.7
d
3.34 (s, 3H, NCH3), 6.89 (t, J ¼ 7.3 Hz, 1H,
UVevisible detector at 254 nm, column (Acclaim TM120, C18, 3
mm,
d
120ꢂA, 150 mm ꢃ 2.1 mm), mobile phases: 0.1% formic acid in H2O
(A) and acetonitrile (B), condition: isocratic at 95% B, flow rate
0.1 mL/min, running time 10 min, samples were prepared in
acetonitrile or methanol.
(superimposed with dmso-d6 peaks), 117.1, 120.9, 126.8, 127.2,
129.3, 131.9, 132.0, 134.6, 134.7, 136.2, 146.7, 148.7, 178.4, 180.8.
HRMSeTOF: m/z [MþH]þ calcd for C17H13ClNO2: 298.0629; found:
298.0625. Purity 100.00%.
Chemicals purchased from commercial suppliers were used
without further purification. All biological tested compounds were
>95% purity (96e100%) as determined by HPLC, except for com-
pound 7 that showed 91% purity, 1 ꢄ 98.0% (Fluka).
4.2.5. 2-(2-Chloro-1,4-dihydro-1,4-dioxonaphthalen-3-ylamino)
benzoic acid (7)
Yield 26%. mp 240e242 ꢂC. IR (UATR) cmꢀ1: 3302, 2925, 1677,
Reagents using in cell culture and assay were obtained from the
following sources. Ham's/F12 (nutrient mixture F-12), DMEM
(Dulbecco's Modified Eagle's Medium), RPMI-1640 (Rosewell Park
Memorial Institute medium), FBS (fetal bovine serum) from
Hyclone laboratories, USA; MTT (3(4,5-dimethylthiazole-2-yl)-2,5-
diphenyltetrazolium bromide), XTT (2,3-bis-(2-methoxy-4-nitro-5-
1653,1605,1567.1H NMR (300 MHz, dmso-d6)
d
7.00 (d, J ¼ 8.2 Hz,1H,
ArH), 7.16 (t, J ¼ 7.7 Hz,1H, ArH), 7.55 (t, J ¼ 7.2 Hz,1H, ArH), 7.81e8.09
(m, 5H, ArH), 10.05 (s, 1H, NH). 13C NMR (75 MHz, dmso-d6)
118.5,
d
120.0, 123.2,123.6, 126.8, 127.1, 130.8, 131.1, 132.3, 132.7, 134.0, 135.3,
140.0, 142.5, 169.3, 177.3, 180.5. HRMSeTOF: m/z [MþH]þ calcd for
C17H11ClNO4: 328.0371; found: 328.0384. Purity 91.07%.
sulfophenyl)-2H-tetrazolium-5-carboxanilide),
PMS
(5-
methylphenazin-5-ium methyl sulfate),
icillinestreptomycin, sodium pyruvate and glucose from Sigma,
DMSO (dimethyl sulfoxide) from Merck.
L
-glutamine, pen-
4.2.6. 2-(3-Acetylphenylamino)-3-chloronaphthalene-1,4-dione (8)
Yield 60%. mp 205e206 ꢂC (207e209 ꢂC) [62]. IR (UATR) cmꢀ1
:
3283, 1673, 1593. 1H NMR (300 MHz, dmso-d6)
d 2.56 (s, 3H,
COCH3), 7.38 (d, J ¼ 7.9 Hz, 1H, ArH), 7.46 (t, J ¼ 7.6 Hz, 1H, ArH), 7.69
(s, 1H, ArH), 7.71 (d, J ¼ 8.6 Hz, 1H, ArH), 7.82 (dt, J ¼ 7.4, 1.4 Hz, 1H,
ArH), 7.88 (dt, J ¼ 7.4, 1.4 Hz, 1H, ArH), 8.05 (dd, J ¼ 7.4, 1.4 Hz, 2H,
4.2. Chemistry: general procedure
A
mixture of appropriate amine
2
(5 mmol) and 2,3-
ArH), 9.44 (s, 1H, NH). 13C NMR (75 MHz, dmso-d6)
d 27.2, 123.4,
dichloronaphthoquinone 1 (5 mmol) in absolute ethanol (20 mL)
was stirred under reflux with stirring until completion of the re-
action as monitored by TLC and then concentrated under reduced
pressure. Water (20 mL) was added and extracted with EtOAc
(3 ꢃ 20 mL). The organic layer was dried over anhydrous sodium
sulfate, filtered and concentrated. The crude product was purified
using silica gel column chromatography to afford the pure product.
124.5, 126.6, 127.0, 128.5, 128.7, 130.8, 132.4, 133.8, 135.2, 137.1,
139.8, 143.5, 177.3, 180.5, 198.1. HRMSeTOF: m/z [MþH]þ calcd for
C18H13ClNO3: 326.0578; found: 326.0571. Purity 97.38%.
4.2.7. 2-(4-Acetylphenylamino)-3-chloronaphthalene-1,4-dione (9)
Yield 61%. mp 237e238 ꢂC. IR (UATR) cmꢀ1: 3236, 1672, 1594. 1H
NMR (300 MHz, dmso-d6)
2H, ArH), 7.80e7.92 (m, 4H, ArH), 8.06 (d, J ¼ 7.4 Hz, 2H, ArH), 9.55
(s, 1H, NH). 13C NMR (75 MHz, dmso-d6)
27.0, 122.3, 126.7, 127.1,
d
2.59 (s, 3H, COCH3), 7.17 (d, J ¼ 8.5 Hz,
4.2.1. 2-(n-Butylamino)-3-chloronaphthalene-1,4-dione (3)
d
Yield 51%. mp 108e110 ꢂC (110e112 ꢂC) [62]. IR (UATR) cmꢀ1
:
128.9, 131.0, 132.1, 132.2, 134.0, 135.2, 143.1, 144.2, 177.5, 180.4, 197.1.
HRMSeTOF: m/z [MþH]þ calcd for C18H13ClNO3: 326.0578; found:
326.0581. Purity 100.00%.
3304, 1683, 1598. 1H NMR (300 MHz, dmso-d6)
d
0.90 (t, J ¼ 7.3 Hz,
3H, CH3),1.33 (sex, J ¼ 7.3 Hz, 2H, CH2CH3), 1.59 (quin, J ¼ 7.3 Hz, 2H,
CH2CH2CH2), 3.72 (t, J ¼ 7.3 Hz, 2H, NHCH2CH2), 7.51 (br s, 1H, NH),
7.73 (t, J ¼ 7.4 Hz, 1H, ArH), 7.82 (t, J ¼ 7.5 Hz, 1H, ArH), 7.97 (d,
4.2.8. 2-Chloro-3-(phenethylamino)naphthalene-1,4-dione (10)
Yield 62%. mp 109e111 ꢂC. IR (UATR) cmꢀ1: 3326, 1674, 1599. 1H
J ¼ 7.7 Hz, 2H, ArH). 13C NMR (75 MHz, dmso-d6)
d 14.2, 19.8, 33.4,
44.1, 126.3, 127.0, 130.3, 132.6, 133.1, 135.4, 145.7, 175.8, 180.7.
HRMSeTOF: m/z [MþH]þ calcd for C14H15ClNO2: 264.0786; found:
264.0783. Purity 98.11%.
NMR (300 MHz, dmso-d6)
d
2.92 (t, J ¼ 7.3 Hz, 2H, CH2Ph), 3.96 (q,
J ¼ 7.1 Hz, 2H, NHCH2), 7.16e7.32 (m, 5H, ArH), 7.50 (br s, 1H, NH),
7.73 (t, J ¼ 7.5 Hz, 1H, ArH), 7.83 (t, J ¼ 7.5 Hz, 1H, ArH), 7.97 (d,
J ¼ 7.5 Hz, 2H, ArH). 13C NMR (75 MHz, dmso-d6)
d 37.4, 45.9, 126.3,
4.2.2. 2-Chloro-3-(phenylamino)naphthalene-1,4-dione (4)
126.8, 127.0, 128.9, 129.2, 130.4, 132.4, 133.1, 135.4, 139.0, 145.8,
175.8, 180.6. HRMSeTOF: m/z [MþH]þ calcd for C18H15ClNO2:
312.0786; found: 312.0785. Purity 98.79%.
Yield 45%. mp 218e220ꢂC (210 ꢂC) [10]. IR (UATR) cmꢀ1: 3243,
1674, 1591. 1H NMR (300 MHz, dmso-d6)
d 7.10e7.35 (m, 5H, ArH),
7.81 (dt, J ¼ 7.6, 1.5 Hz, 1H, ArH), 7.87 (dt, J ¼ 7.6, 1.5 Hz, 1H, ArH),
8.03 (dd, J ¼ 7.3, 0.9 Hz, 2H, ArH), 9.31 (s,1H, NH). 13C NMR (75 MHz,
4.2.9. 2-(3,4-Dimethoxyphenethylamino)-3-chloronaphthalene-
1,4-dione (11)
dmso-d6) d 124.4, 124.9,126.6, 127.0, 128.4, 130.7, 132.4, 133.7, 135.3,
139.3, 143.6, 177.2, 180.6. HRMSeTOF: m/z [MþH]þ calcd for
Yield 65%. mp 148e149 ꢂC. IR (UATR) cmꢀ1: 3327, 1676, 1601. 1H
C
16H11ClNO2: 284.0473; found: 284.0472. Purity 100.00%.
NMR (300 MHz, dmso-d6)
d
2.84 (t, J ¼ 7.1 Hz, 2H, CH2Ph), 3.65, 3.71

260
V. Prachayasittikul et al. / European Journal of Medicinal Chemistry 84 (2014) 247e263
(2s, 6H, 2 ꢃ OCH3), 3.95 (q, J ¼ 7.1 Hz, 2H, NHCH2), 6.72 (d,
J ¼ 8.5 Hz, 1H, ArH), 6.80 (s, 1H, ArH), 6.81 (d, J ¼ 7.6 Hz, 1H, ArH),
7.37 (br t, 1H, NH), 7.73 (t, J ¼ 7.4 Hz, 1H, ArH), 7.82 (t, J ¼ 7.4 Hz, 1H,
compounds exhibited IC50 > 50 mg/mL were considered as non-
cytotoxic.
ArH), 7.93 (d, J ¼ 7.4 Hz, 1H, ArH), 7.96 (d, J ¼ 7.4 Hz, 1H, ArH). 13
C
4.5. QSAR analysis
NMR (75 MHz, dmso-d6)
d 36.9, 46.0, 55.8, 55.9, 112.5, 113.2, 121.3,
126.2, 126.9,130.4,131.4, 132.4, 133.1, 135.3, 147.9, 149.1, 175.9, 180.7.
HRMSeTOF: m/z [MþNa]þ calcd for C20H18ClNNaO4: 394.0817;
found: 394.0823. Purity 99.51%.
Conceptually, molecular structures of 35 compounds encom-
passing the tested compounds (1 and 3e12) and the virtually
constructed compounds (13e36) were drawn, geometrically opti-
mized and subjected to descriptor calculation in order to obtain
values of quantum chemical and molecular descriptors. The X
(descriptors) and Y (cytotoxic activity) block of data from the tested
compounds (1 and 3e12) were further used in the development of
QSAR models. Finally, the cytotoxic activities of virtual constructs
(13e36) were calculated from the QSAR equations.
4.2.10. 2-Chloro-3-[[4-(phenylamino)phenyl]amino]-naphthalene-
1,4-dione (12)
Yield 37%. mp 196e198 ꢂC. IR (UATR) cmꢀ1: 3342, 3295, 1670,
1595. 1H NMR (300 MHz, dmso-d6)
7.01e7.09 (m, 6H, ArH), 7.24 (t, J ¼ 7.5 Hz, 2H, ArH), 7.79 (t,
J ¼ 7.4 Hz, 1H, ArH), 7.86 (t, J ¼ 7.3 Hz, 1H, ArH), 8.03 (d, J ¼ 7.4 Hz,
2H, ArH), 8.21, 9.24 (2s, 2H, NH). 13C NMR (75 MHz, dmso-d6)
d
6.82 (t, J ¼ 7.3 Hz, 1H, ArH),
4.5.1. Data set
Eleven compounds (1 and 3e12) and their experimental IC50
values were used in the development of QSAR models. IC50 values
were converted to pIC50 values by taking the negative logarithm to
the base of 10 (ꢀlog IC50) as to obtain distribution of data points.
Herein, four data sets were constructed pertaining to the cytotoxic
activity against four cancer cell lines (HepG2, HuCCA-1, A549 and
MOLT-3). QSAR models of the cytotoxic activity against each cell
line were developed separately.
d
116.6, 117.1, 120.1, 126.2, 126.5, 127.0, 129.7, 130.6, 131.2, 132.6,
133.4, 135.3, 141.1, 143.7, 143.9, 176.8, 180.6. HRMSeTOF: m/z
[MþH]þ calcd for C22H16ClN2O2: 375.0895; found: 375.0895. Purity
99.92%.
4.3. Cytotoxic assay
The cytotoxic activity of compounds (1 and 3e12) against the
four tested types of human cancer cell lines was performed using
etoposide and doxorubicin as reference drugs. Lung carcinoma
(A549) and cholangiocarcinoma (HuCCA-1) cell lines were grown in
Hamm's/F12 containing 2 mM L-glutamine supplemented with
100 U/mL penicillinestreptomycin and 10% FBS. Lymphoblastic
4.5.2. Molecular structure optimization and descriptor calculation
Chemical structures of a set of 35 compounds were drawn
using the GaussView software [41] as shown in Figs. 1 and 2. All
molecular structures were initially subjected to geometrical
optimization using Gaussian 09 [67] at the semi-empirical Austin
Model 1 (AM1) level followed by a subsequent density functional
theory (DFT) calculation using the Becke's three-parameter hybrid
method with the LeeeYangeParr correlation functional (B3LYP)
together with the 6e31g(d) level. Geometrical optimization pro-
vided low-energy conformers in which their output files were
consequently used for extracting quantum chemical descriptors
using an in-house developed script. Quantum chemical de-
scriptors consisted of the mean absolute atomic charge (Qm), total
leukemia (MOLT-3) cell line was cultured in RPMI-1640 medium
containing 2 mM L-glutamine, 100 U/mL penicillinestreptomycin,
sodium pyruvate, glucose and 10% FBS. Hepatocellular carcinoma
(HepG2) cell line was cultured in DMEM medium containing 100 U/
mL penicillinestreptomycin and 10% FBS. Briefly, cell lines sus-
pended in RPMI-1640 containing 10% FBS were seeded at a density
of 5 ꢃ103e2 ꢃ 104 cells per well in a 96-well plate (Costar No. 3599,
USA) and then incubated at 37 ꢂC under a humidified atmosphere
with 95% air and 5% CO2 for 24 h. An equal volume of additional
medium containing either the serial dilutions of the tested com-
pounds, positive control (etoposide and/or doxorubicin) or negative
control (DMSO) was added to the desired final concentrations. The
microtiter plates were further incubated for 48 h. Cell viability in
each well was determined by staining with MTT assay [63e65] (for
adherent cells: A549, HuCCA-1 and HepG2 cells) or XTT assay [66]
(for suspended cells: MOLT-3 cells). The plates were read on a
microplate reader (Molecular Devices, USA). The absorbance was
recorded at 550 nm. Finally, IC50 values were determined as the
required concentration in which tested compounds afforded 50%
inhibition of cell growth.
energy (Etotal), total dipole moment (m), highest occupied molec-
ular orbital energy (EHOMO), lowest unoccupied molecular orbital
energy (ELUMO), energy difference of HOMO and LUMO (HOMO-
eLUMOGap), electron affinity (EA), ionization potential (IP), Mul-
liken electronegativity (c), hardness (h), softness (S), electrophilic
index (ui) and electrophilicity (
u
). Dragon software (version 5.5)
[42] was used for further calculation of an additional set of 3224
molecular descriptors using the aforementioned low-energy mo-
lecular structures from B3LYP/6-31g(d) calculation as input data.
Dragon descriptors can be categorized into the following 22
classes as follows: Constitutional descriptors, Topological de-
scriptors, Walk and path counts, Connectivity indices, Information
indices, 2D autocorrelation, Edge adjacency indices, Burden ei-
genvalues, Topological charge indices, Eigenvalue-based indices,
Randic molecular profiles, Geometrical descriptors, RDF de-
scriptors, 3D-MoRSE descriptors, WHIM descriptors, GETAWAY
descriptors, Functional group counts, Atom-centered fragments,
Charge descriptors, Molecular properties, 2D binary fingerprints
and 2D frequency fingerprints.
4.4. Cytotoxicity assay
Cytotoxicity was tested using normal embryonic lung cells
(MRC-5). The cells were grown in DMEM medium supplemented
with 100 U/mL penicillinestreptomycin and 10% FBS. The assay was
performed as previously described for human cancer cells (Section
4.3). The cell viability was determined by MTT assay [63e65]. The
MTT solution (10
mL/100
mL medium) was added to all wells of an
4.5.3. Descriptor selection
assay and plates were incubated as described above for 2e4 h.
Subsequently, DMSO was added to dissolve the resulting formazan
by sonication. The plates were read on a microplate reader (Mo-
lecular Devices, USA) using a test wavelength of 550 nm and a
reference wavelength of 650 nm. The IC50 value was determined as
the compound concentration that inhibited cell growth by 50%. The
All quantum chemical and molecular descriptors of the experi-
mentally tested compounds (1 and 3e12) were used as input data
for descriptor selection using Monte Carlo or correlation-based
approaches.
The starting set of descriptors consisted of 13 quantum chemical
descriptors and 3224 molecular descriptors and important

V. Prachayasittikul et al. / European Journal of Medicinal Chemistry 84 (2014) 247e263
261
descriptors were identified by a Monte Carlo approach [68] coupled
to Multiple Linear Regression (MCeMLR) using an in-house
developed script.
4.5.8. Prediction of structurally modified compounds (13e36) using
QSAR models
Equations obtained from QSAR analysis in concomitant with
molecular descriptors obtained from Gaussian and Dragon calcu-
lated were used in computing the predicted pIC50 values of 24
virtually constructed compounds (13e36) against the four tested
cancer cell lines.
Prior to performing MCeMLR, constant and insignificant de-
scriptors were initially removed whose standard deviation is less
than 0.1. Remaining descriptors were standardized and correlation
coefficients between descriptor pairs were calculated. Any pair of
descriptor whose correlation coefficient is greater than 0.9 were
considered as highly correlated and was removed. The same pro-
cess was repeated until there were no redundant or highly corre-
lated descriptors.
Correlation-based feature selection was initially performed by
calculating the pair-correlation of each descriptor value and
bioactivity (pIC50). Threshold value of correlation coefficient (r) of
0.6 was used as the cut off. Descriptors representing jrj < 0.6 were
considered as low correlated descriptor and were discarded from
this study. The remaining descriptors along with the pIC50 value
were used as input data for further feature selection by stepwise
MLR regression as implemented in SPSS statistics 18.0 (SPSS Inc.,
USA).
Acknowledgment
This project is supported by the Office of the Higher Education
Commission, Mahidol University under the National Research
Universities Initiative and Annual Government Grant under Mahi-
dol University (2556e2558 B.E.). We appreciate Miss Sasiwimon
Lalitmanat and Dr. Thummaruk Suksrichavalit for performing HPLC
analysis of the compounds. We gratefully acknowledge The Chu-
labhorn Research Institute for cytotoxic testings.
Appendix A. Supplementary data
Supplementary data related to this article can be found online at
http://dx.doi.org/10.1016/j.ejmech.2014.07.024. These data include
MOL files and InChiKeys of the compounds described in this article.
4.5.4. Multivariate analysis using multiple linear regression
Selected descriptors were used as independent variables and
pIC50 values were used as the dependent variable for multivariate
analysis by Waikato Environment for Knowledge Analysis (WEKA)
version 3.4.5 [69] using MLR.
References
[1] T. Vos, A.D. Flaxman, M. Naghavi, R. Lozano, C. Michaud, M. Ezzati, K. Shibuya,
J.A. Salomon, S. Abdalla, V. Aboyans, J. Abraham, I. Ackerman, R. Aggarwal,
S.Y. Ahn, M.K. Ali, M. Alvarado, H.R. Anderson, L.M. Anderson, K.G. Andrews,
C. Atkinson, L.M. Baddour, A.N. Bahalim, S. Barker-Collo, L.H. Barrero,
D.H. Bartels, M.G. Basanez, A. Baxter, M.L. Bell, E.J. Benjamin, D. Bennett,
E. Bernabe, K. Bhalla, B. Bhandari, B. Bikbov, A. Bin Abdulhak, G. Birbeck,
J.A. Black, H. Blencowe, J.D. Blore, F. Blyth, I. Bolliger, A. Bonaventure,
S. Boufous, R. Bourne, M. Boussinesq, T. Braithwaite, C. Brayne, L. Bridgett,
S. Brooker, P. Brooks, T.S. Brugha, C. Bryan-Hancock, C. Bucello, R. Buchbinder,
G. Buckle, C.M. Budke, M. Burch, P. Burney, R. Burstein, B. Calabria,
B. Campbell, C.E. Canter, H. Carabin, J. Carapetis, L. Carmona, C. Cella,
F. Charlson, H. Chen, A.T. Cheng, D. Chou, S.S. Chugh, L.E. Coffeng, S.D. Colan,
S. Colquhoun, K.E. Colson, J. Condon, M.D. Connor, L.T. Cooper, M. Corriere,
M. Cortinovis, K.C. de Vaccaro, W. Couser, B.C. Cowie, M.H. Criqui, M. Cross,
K.C. Dabhadkar, M. Dahiya, N. Dahodwala, J. Damsere-Derry, G. Danaei,
A. Davis, D. De Leo, L. Degenhardt, R. Dellavalle, A. Delossantos, J. Denenberg,
S. Derrett, D.C. Des Jarlais, S.D. Dharmaratne, M. Dherani, C. Diaz-Torne,
H. Dolk, E.R. Dorsey, T. Driscoll, H. Duber, B. Ebel, K. Edmond, A. Elbaz, S.E. Ali,
H. Erskine, P.J. Erwin, P. Espindola, S.E. Ewoigbokhan, F. Farzadfar, V. Feigin,
D.T. Felson, A. Ferrari, C.P. Ferri, E.M. Fevre, M.M. Finucane, S. Flaxman,
L. Flood, K. Foreman, M.H. Forouzanfar, F.G. Fowkes, R. Franklin, M. Fransen,
M.K. Freeman, B.J. Gabbe, S.E. Gabriel, E. Gakidou, H.A. Ganatra, B. Garcia,
F. Gaspari, R.F. Gillum, G. Gmel, R. Gosselin, R. Grainger, J. Groeger,
F. Guillemin, D. Gunnell, R. Gupta, J. Haagsma, H. Hagan, Y.A. Halasa, W. Hall,
D. Haring, J.M. Haro, J.E. Harrison, R. Havmoeller, R.J. Hay, H. Higashi, C. Hill,
B. Hoen, H. Hoffman, P.J. Hotez, D. Hoy, J.J. Huang, S.E. Ibeanusi, K.H. Jacobsen,
S.L. James, D. Jarvis, R. Jasrasaria, S. Jayaraman, N. Johns, J.B. Jonas,
G. Karthikeyan, N. Kassebaum, N. Kawakami, A. Keren, J.P. Khoo, C.H. King,
L.M. Knowlton, O. Kobusingye, A. Koranteng, R. Krishnamurthi, R. Lalloo,
L.L. Laslett, T. Lathlean, J.L. Leasher, Y.Y. Lee, J. Leigh, S.S. Lim, E. Limb, J.K. Lin,
M. Lipnick, S.E. Lipshultz, W. Liu, M. Loane, S.L. Ohno, R. Lyons, J. Ma,
J. Mabweijano, M.F. MacIntyre, R. Malekzadeh, L. Mallinger, S. Manivannan,
W. Marcenes, L. March, D.J. Margolis, G.B. Marks, R. Marks, A. Matsumori,
R. Matzopoulos, B.M. Mayosi, J.H. McAnulty, M.M. McDermott, N. McGill,
J. McGrath, M.E. Medina-Mora, M. Meltzer, G.A. Mensah, T.R. Merriman,
A.C. Meyer, V. Miglioli, M. Miller, T.R. Miller, P.B. Mitchell, A.O. Mocumbi,
T.E. Moffitt, A.A. Mokdad, L. Monasta, M. Montico, M. Moradi-Lakeh, A. Moran,
L. Morawska, R. Mori, M.E. Murdoch, M.K. Mwaniki, K. Naidoo, M.N. Nair,
L. Naldi, K.M. Narayan, P.K. Nelson, R.G. Nelson, M.C. Nevitt, C.R. Newton,
S. Nolte, P. Norman, R. Norman, M. O'Donnell, S. O'Hanlon, C. Olives, S.B. Omer,
K. Ortblad, R. Osborne, D. Ozgediz, A. Page, B. Pahari, J.D. Pandian, A.P. Rivero,
S.B. Patten, N. Pearce, R.P. Padilla, F. Perez-Ruiz, N. Perico, K. Pesudovs,
D. Phillips, M.R. Phillips, K. Pierce, S. Pion, G.V. Polanczyk, S. Polinder,
C.A. Pope 3rd, S. Popova, E. Porrini, F. Pourmalek, M. Prince, R.L. Pullan,
K.D. Ramaiah, D. Ranganathan, H. Razavi, M. Regan, J.T. Rehm, D.B. Rein,
G. Remuzzi, K. Richardson, F.P. Rivara, T. Roberts, C. Robinson, F.R. De Leon,
L. Ronfani, R. Room, L.C. Rosenfeld, L. Rushton, R.L. Sacco, S. Saha, U. Sampson,
L. Sanchez-Riera, E. Sanman, D.C. Schwebel, J.G. Scott, M. Segui-Gomez,
S. Shahraz, D.S. Shepard, H. Shin, R. Shivakoti, D. Singh, G.M. Singh, J.A. Singh,
J. Singleton, D.A. Sleet, K. Sliwa, E. Smith, J.L. Smith, N.J. Stapelberg, A. Steer,
T. Steiner, W.A. Stolk, L.J. Stovner, C. Sudfeld, S. Syed, G. Tamburlini,
Constructed MLR models are summarized according to the
following equation:
X
Y ¼ B0 þ
BnXn
(1)
where Y is the pIC50 values of compounds, B0 is the intercept and Bn
are the regression coefficient of descriptors Xn.
4.5.5. Data sampling
The data set was divided into 2 subsets comprising of the
training set and the testing set and such data divisions were per-
formed by means of leave-one-out cross validation (LOO-CV). In
brief, one sample was excluded and used as the testing set while
the remaining N ꢀ 1 samples were used as the training set. The
process was repeated until every sample in the data set was used as
the testing set to predict the Y variable.
4.5.6. Outlier detection
Outliers were detected when the standardized differences be-
tween actual and predicted values (i.e. the residual) are greater
than 2. Standardization of residuals was calculated as following
equation:
,
,
N
X
ꢀ
ꢁ
ꢀ
ꢁ2
stn
ij
X
¼ Xij ꢀ Xj
xij ꢀ Xj
N
(2)
i¼1
stn
ij
where X is the standardized value, is the value of each sample,
is the mean of each descriptors and N is the sample size of the
data set. Identified outliers were subjected to removal from the
model.
4.5.7. Evaluating the performance of QSAR models
The performance of QSAR models was evaluated by statistical
parameters comprising of correlation coefficient (R) and the root
mean square error (RMSE) which represented the predictive per-
formance and predictive error of the models, respectively.

262
V. Prachayasittikul et al. / European Journal of Medicinal Chemistry 84 (2014) 247e263
M. Tavakkoli, H.R. Taylor, J.A. Taylor, W.J. Taylor, B. Thomas, W.M. Thomson, synthesized 1,4-naphthoquinone derivative, Biochem. Pharmacol. 60 (2000)
G.D. Thurston, I.M. Tleyjeh, M. Tonelli, J.A. Towbin, T. Truelsen,
M.K. Tsilimbaris, C. Ubeda, E.A. Undurraga, M.J. van der Werf, J. van Os,
M.S. Vavilala, N. Venketasubramanian, M. Wang, W. Wang, K. Watt,
D.J. Weatherall, M.A. Weinstock, R. Weintraub, M.G. Weisskopf,
M.M. Weissman, R.A. White, H. Whiteford, S.T. Wiersma, J.D. Wilkinson,
H.C. Williams, S.R. Williams, E. Witt, F. Wolfe, A.D. Woolf, S. Wulf, P.H. Yeh,
A.K. Zaidi, Z.J. Zheng, D. Zonies, A.D. Lopez, C.J. Murray, M.A. AlMazroa,
Z.A. Memish, Years lived with disability (YLDs) for 1160 sequelae of 289
diseases and injuries 1990e2010: a systematic analysis for the global burden
of disease study 2010, Lancet 380 (2012) 2163e2196.
1001e1008.
[23] L. Salmon-Chemin, E. Buisine, V. Yardley, S. Kohler, M.A. Debreu, V. Landry,
C. Sergheraert, S.L. Croft, R.L. Krauth-Siegel, E. Davioud-Charvet, 2- and 3-
Substituted 1,4-naphthoquinone derivatives as subversive substrates of try-
panothione reductase and lipoamide dehydrogenase from Trypanosoma cruzi:
synthesis and correlation between redox cycling activities and in vitro cyto-
toxicity, J. Med. Chem. 44 (2001) 548e565.
[24] P.J. O'Brien, Molecular mechanisms of quinone cytotoxicity, Chem. Biol.
Interact. 80 (1991) 1e41.
[25] J. Benites, J.A. Valderrama, K. Bettega, R.C. Pedrosa, P.B. Calderon, J. Verrax,
Biological evaluation of donor-acceptor aminonaphthoquinones as antitumor
agents, Eur. J. Med. Chem. 45 (2010) 6052e6057.
[2] R.P. Verma, Anti-cancer activities of 1,4-naphthoquinones: a QSAR study,
Anticancer. Agents. Med. Chem. 6 (2006) 489e499.
[3] A.K. McClendon, N. Osheroff, DNA topoisomerase II, genotoxicity, and cancer,
Mutat. Res. Fundam. Mol. Mech. Mutagen 623 (2007) 83e97.
[4] H.R. Lawrence, A. Kazi, Y. Luo, R. Kendig, Y. Ge, S. Jain, K. Daniel, D. Santiago,
W.C. Guida, S.M. Sebti, Synthesis and biological evaluation of naphthoquinone
analogs as a novel class of proteasome inhibitors, Bioorg. Med. Chem. 18
(2010) 5576e5592.
[5] V.K. Tandon, H.K. Maurya, A. Tripathi, G.B. ShivaKeshava, P.K. Shukla,
P. Srivastava, D. Panda, 2,3-Disubstituted-1,4-naphthoquinones, 12H-benzo[b]
phenothiazine-6,11-diones and related compounds: synthesis and biological
evaluation as potential antiproliferative and antifungal agents, Eur. J. Med.
Chem. 44 (2009) 1086e1092.
[6] E.H. da Cruz, C.M. Hussene, G.G. Dias, E.B. Diogo, I.M. de Melo, B.L. Rodrigues,
M.G. da Silva, W.O. Valenca, C.A. Camara, R.N. de Oliveira, Y.G. de Paiva,
M.O. Goulart, B.C. Cavalcanti, C. Pessoa, E.N. da Silva Junior, 1,2,3-Triazole-,
arylamino- and thio-substituted 1,4-naphthoquinones: potent antitumor ac-
tivity, electrochemical aspects, and bioisosteric replacement of C-ring-modi-
fied lapachones, Bioorg. Med. Chem. 22 (2014) 1608e1619.
[7] H.R. Nasiri, M.G. Madej, R. Panisch, M. Lafontaine, J.W. Bats, C.R. Lancaster,
H. Schwalbe, Design, synthesis, and biological testing of novel naph-
thoquinones as substrate-based inhibitors of the quinol/fumarate reductase
from Wolinella succinogenes, J. Med. Chem. 56 (2013) 9530e9541.
[8] D. Bhasin, S.N. Chettiar, J.P. Etter, M. Mok, P.K. Li, Anticancer activity and SAR
studies of substituted 1,4-naphthoquinones, Bioorg. Med. Chem. 21 (2013)
4662e4669.
[9] C.-K. Ryu, D.-H. Kim, The synthesis and antimicrobial activities of some 1,4-
naphthoquinones (II), Arch. Pharm. Res. 15 (1992) 263e268.
[10] V.K. Tandon, H.K. Maurya, N.N. Mishra, P.K. Shukla, Design, synthesis and
biological evaluation of novel nitrogen and sulfur containing hetero-1,4-
naphthoquinones as potent antifungal and antibacterial agents, Eur. J. Med.
Chem. 44 (2009) 3130e3137.
[26] T. Schneider, Y. Muthukumar, B. Hinkelmann, R. Franke, M. Doring, C. Jacob,
F. Sasse, Deciphering intracellular targets of organochalcogen based redox
catalysts, Med. Chem. Commun. 3 (2012) 784e787.
[27] V.K. Tandon, D.B. Yadav, R.V. Singh, M. Vaish, A.K. Chaturvedi, P.K. Shukla,
Synthesis and biological evaluation of novel 1,4-naphthoquinone derivatives
as antibacterial and antiviral agents, Bioorg. Med. Chem. Lett. 15 (2005)
3463e3466.
[28] S. Salunke-Gawali, O. Pawar, M. Nikalje, R. Patil, T. Weyhermüller,
V.G. Puranik, V. Badireenath Konkimalla, Synthesis, characterization and
molecular structures of homologated analogs of 2-bromo-3-(n-alkylamino)-
1,4-napthoquinone, J. Mol. Struct. 1056e1057 (2014) 97e103.
~
ꢀ
[29] C. Avendano, J.C. Menendez, DNA intercalators and topoisomerase inhibitors,
in: Medicinal Chemistry of Anticancer Drugs, Elsevier, Oxford, 2008, pp.
199e228.
[30] K.R. Hande, Topoisomerase II inhibitors, Update Cancer Ther. 3 (2008) 13e26.
[31] C. Nantasenamat, C. Isarankura-Na-Ayudhya, V. Prachayasittikul, Advances in
computational methods to predict the biological activity of compounds,
Expert Opin. Drug Discov. 5 (2010) 633e654.
[32] C. Nantasenamat, C. Isarankura-Na-Ayudhya, T. Naenna, V. Prachayasittikul,
A practical overview of quantitative structure-activity relationship, Excli J. 8
(2009) 74e88.
[33] A. Worachartcheewan, C. Nantasenamat, C. Isarankura-Na-Ayudhya,
V. Prachayasittikul, QSAR study of amidino bis-benzimidazole derivatives as
potent anti-malarial agents against Plasmodium falciparum, Chem. Zvesti 67
(2013) 1462e1473.
[34] A. Worachartcheewan, C. Nantasenamat, C. Isarankura-Na-Ayudhya,
V. Prachayasittikul, Predicting antimicrobial activities of benzimidazole de-
rivatives, Med. Chem. Res. (2013) 1e13.
[35] A. Worachartcheewan, S. Prachayasittikul, R. Pingaew, C. Nantasenamat,
T. Tantimongcolwat, S. Ruchirawat, V. Prachayasittikul, Antioxidant, cytotox-
icity, and QSAR study of 1-adamantylthio derivatives of 3-picoline and phe-
nylpyridines, Med. Chem. Res. 21 (2012) 3514e3522.
[36] C. Hansch, A. Leo, S.B. Mekapati, A. Kurup, QSAR and ADME, Bioorg. Med.
Chem. 12 (2004) 3391e3400.
[11] J.J. Inbaraj, C.F. Chignell, Cytotoxic action of juglone and plumbagin:
a
mechanistic study using HaCaT keratinocytes, Chem. Res. Toxicol. 17 (2004)
55e62.
[12] S.T. Huang, H.S. Kuo, C.L. Hsiao, Y.L. Lin, Efficient synthesis of 'redox-switched'
naphthoquinone thiol-crown ethers and their biological activity evaluation,
Bioorg. Med. Chem. 10 (2002) 1947e1952.
[13] V.K. Tandon, R.B. Chhor, R.V. Singh, S. Rai, D.B. Yadav, Design, synthesis and
evaluation of novel 1,4-naphthoquinone derivatives as antifungal and anti-
cancer agents, Bioorg. Med. Chem. Lett. 14 (2004) 1079e1083.
[14] K. Sasaki, H. Abe, F. Yoshizaki, In vitro antifungal activity of naphthoquinone
derivatives, Biol. Pharm. Bull. 25 (2002) 669e670.
ꢀ
~
ꢀ
[37] E. Perez-Sacau, R.G. Díaz-Penate, A. Estevez-Braun, A.G. Ravelo, J.M. García-
Castellano, L. Pardo, M. Campillo, Synthesis and pharmacophore modeling of
naphthoquinone derivatives with cytotoxic activity in human promyelocytic
leukemia HL-60 cell line, J. Med. Chem. 50 (2007) 696e706.
[38] W.R. Leopold, J.L. Shillis, A.E. Mertus, J.M. Nelson, B.J. Roberts, R.C. Jackson,
Anticancer activity of the structurally novel antibiotic CI-920 and its ana-
logues, Cancer Res. 44 (1984) 1928e1932.
[15] V.K. Tandon, D.B. Yadav, H.K. Maurya, A.K. Chaturvedi, P.K. Shukla, Design,
synthesis, and biological evaluation of 1,2,3-trisubstituted-1,4-dihydrobenzo
[g]quinoxaline-5,10-diones and related compounds as antifungal and anti-
bacterial agents, Bioorg. Med. Chem. 14 (2006) 6120e6126.
[16] V.K. Tandon, R.V. Singh, S. Rai, R.B. Chhor, Z.K. Khan, Synthesis and pharma-
cological studies of some 2-t-amino and 2,3-di-t-amino substituted 1,4-
naphthoquinones and related compounds, Boll. Chim. Farm. 141 (2002)
304e310.
[17] T.V. Ilina, E.A. Semenova, T.R. Pronyaeva, A.G. Pokrovskii, I.V. Nechepurenko,
E.E. Shults, O.I. Andreeva, S.N. Kochetkov, G.A. Tolstikov, Inhibition of HIV-1
reverse transcriptase by aryl-substituted naphtho- and anthraquinones,
Dokl. Biochem. Biophys. 382 (2002) 56e59.
[39] K.M. Tewey, G.L. Chen, E.M. Nelson, L.F. Liu, Intercalative antitumor drugs
interfere with the breakage-reunion reaction of mammalian DNA topoisom-
erase II, J. Biol. Chem. 259 (1984) 9182e9187.
[40] J.W. Lown, S.K. Sim, K.C. Majumdar, R.Y. Chang, Strand scission of DNA by
bound adriamycin and daunorubicin in the presence of reducing agents,
Biochem. Biophys. Res. Commun. 76 (1977) 705e710.
[41] I.I.R. Denning ton, T. Keith, J. Millam, K. Eppinnett, W.L. Hovell, R. Gilliland,
GaussView, Version 3.09, Semichem Shawnee Mission, KS, USA, 2003.
[42] Talete, Dragon for Windows (Software for Molecular Descriptor Calculations),
Version 5.5, Milan, Italy, 2007.
[43] R.P. Verma, C. Hansch, Elucidation of structure-activity relationships for 2- or
6-substituted-5,8-dimethoxy-1,4-naphthoquinones, Bioorg. Med. Chem. 12
(2004) 5997e6009.
[44] N. Osheroff, Eukaryotic topoisomerase II. Characterization of enzyme turn-
over, J. Biol. Chem. 261 (1986) 9944e9950.
[45] J.C. Wang, DNA topoisomerases, Annu. Rev. Biochem. 65 (1996) 635e692.
[46] J.C. Wang, Cellular roles of DNA topoisomerases: a molecular perspective, Nat.
Rev. Mol. Cell Biol. 3 (2002) 430e440.
[47] J.J. Champoux, DNA topoisomerases: structure, function, and mechanism,
Annu. Rev. Biochem. 70 (2001) 369e413.
[48] J.M. Fortune, N. Osheroff, Topoisomerase II as a target for anticancer drugs:
when enzymes stop being nice, Prog. Nucleic Acid Res. Mol. Biol. 64 (2000)
221e253.
[18] A.J.M. Da Silva, C.D. Buarque, F.V. Brito, L. Aurelian, L.F. Macedo, L.H. Malkas,
€
R.J. Hickey, D.V.S. Lopes, F. Noel, Y.L.B. Murakami, N.M.V. Silva, P.A. Melo,
R.R.B. Caruso, N.G. Castro, P.R.R. Costa, Synthesis and preliminary pharmaco-
logical evaluation of new ( ) 1,4-naphthoquinones structurally related to
lapachol, Bioorg. Med. Chem. 10 (2002) 2731e2738.
[19] G.Y. Song, Y. Kim, Y.J. You, H. Cho, S.H. Kim, D.E. Sok, B.Z. Ahn, Naphthazarin
derivatives (VI): synthesis, inhibitory effect on DNA topoisomerase I and
antiproliferative activity of 2- or 6-(1-oxyiminoalkyl)-5,8-dimethoxy-1,4-
naphthoquinones, Arch. Pharm. 333 (2000) 87e92.
[20] J.C. Lien, L.J. Huang, C.M. Teng, J.P. Wang, S.C. Kuo, Synthesis of 2-alkoxy 1,4-
naphthoquinone derivatives as antiplatelet, antiinflammatory, and anti-
allergic agents, Chem. Pharm. Bull. 50 (2002) 672e674.
[21] Y.R. Jin, C.K. Ryu, C.K. Moon, M.R. Cho, Y.P. Yun, Inhibitory effects of J78, a
newly synthesized 1,4-naphthoquinone derivative, on experimental throm-
bosis and platelet aggregation, Pharmacology 70 (2004) 195e200.
[22] D.Y. Yuk, C.K. Ryu, J.T. Hong, K.H. Chung, W.S. Kang, Y. Kim, H.S. Yoo, M.K. Lee,
C.K. Lee, Y.P. Yun, Antithrombotic and antiplatelet activities of 2-chloro-3-
[49] A.M. Wilstermann, N. Osheroff, Stabilization of eukaryotic topoisomerase II-
DNA cleavage complexes, Curr. Top. Med. Chem. 3 (2003) 321e338.
[50] S.H. Kaufmann, Cell death induced by topoisomerase-targeted drugs: more
questions than answers, Biochim. Biophys. Acta. 1400 (1998) 195e211.
[51] T.G. Gantchev, D.J. Hunting, The ortho-quinone metabolite of the anticancer
drug etoposide (VP-16) is a potent inhibitor of the topoisomerase II/DNA
cleavable complex, Mol. Pharmacol. 53 (1998) 422e428.
[4-(ethylcarboxy)-phenyl]-amino-1,4-naphthoquinone (NQ12),
a
newly

V. Prachayasittikul et al. / European Journal of Medicinal Chemistry 84 (2014) 247e263
263
[52] D.A. Jacob, S.L. Mercer, N. Osheroff, J.E. Deweese, Etoposide quinone is a
redox-dependent topoisomerase II poison, Biochemistry 50 (2011)
5660e5667.
[53] A. Hakura, H. Mochida, Y. Tsutsui, K. Yamatsu, Mutagenicity and cytotoxicity
of naphthoquinones for Ames Salmonella tester strains, Chem. Res. Toxicol. 7
(1994) 559e567.
[54] H. Soedjak, J. Hajdu, J.D. Raffetto, R. Cano, B.L. Bales, L.S. Prasad, L.D. Kispert,
The structure of streptonigrin semiquinone in solution, Biochim. Biophys.
Acta. 1335 (1997) 73e90.
[62] R. Parr, J. Reiss, N-Substituted enaminones from 3-aminobutenoic acids and
1,4-naphthoquinones, Aust. J. Chem. 37 (1984) 389e394.
[63] T. Mosmann, Rapid colorimetric assay for cellular growth and survival:
application to proliferation and cytotoxicity assays, J. Immunol. Methods. 65
(1983) 55e63.
[64] H. Tominaga, M. Ishiyama, F. Ohseto, K. Sasamoto, T. Hamamoto, K. Suzuki,
M. Watanabe, A water-soluble tetrazolium salt useful for colorimetric cell
viability assay, Anal. Commun. 36 (1999) 47e50.
[65] J. Carmichael, W.G. DeGraff, A.F. Gazdar, J.D. Minna, J.B. Mitchell, Evaluation of
[55] S.L. de Castro, F.S. Emery, E.N. da Silva Júnior, Synthesis of quinoidal mole-
cules: strategies towards bioactive compounds with an emphasis on lapa-
chones, Eur. J. Med. Chem. 69 (2013) 678e700.
[56] E.A. Hillard, F.C. de Abreu, D.C. Ferreira, G. Jaouen, M.O. Goulart, C. Amatore,
Electrochemical parameters and techniques in drug development, with an
emphasis on quinones and related compounds, Chem. Commun. (Camb.)
(2008) 2612e2628.
[57] J.L. Bolton, M.A. Trush, T.M. Penning, G. Dryhurst, T.J. Monks, Role of quinones
in toxicology, Chem. Res. Toxicol. 13 (2000) 135e160.
[58] T.J. Monks, R.P. Hanzlik, G.M. Cohen, D. Ross, D.G. Graham, Quinone chemistry
and toxicity, Toxicol. Appl. Pharmacol. 112 (1992) 2e16.
[59] Y. Song, G.R. Buettner, Thermodynamic and kinetic considerations for the
reaction of semiquinone radicals to form superoxide and hydrogen peroxide,
Free Radic. Biol. Med. 49 (2010) 919e962.
[60] M.O.F. Goulart, C.L. Zani, J. Tonholo, L.R. Freitas, F.C. De Abreu, A.B. Oliveira,
D.S. Raslan, S. Starling, E. Chiari, Trypanocidal activity and redox potential of
heterocyclic- and 2-hydroxy-naphthoquinones, Bioorg. Med. Chem. Lett. 7
(1997) 2043e2048.
[61] L. Bouffier, I. Gosse, M. Demeunynck, P. Mailley, Electrochemistry and bioac-
tivity relationship of 6-substituted-4H-pyrido[4,3,2-kl]acridin-4-one anti-
tumor drug candidates, Bioelectrochemistry. 88 (2012) 103e109.
a tetrazolium-based semiautomated colorimetric assay: assessment of
radiosensitivity, Cancer Res. 47 (1987) 936e942.
[66] A. Doyly, J.B. Griffiths, Mammalian Cell Culture-essential Techniques, John
Wiley and Sons Chichester, New York, Weinheim, Brisbane, Toronto, 1997.
[67] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb,
J.R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G.A. Petersson,
H. Nakatsuji, M. Caricato, X. Li, H.P. Hratchian, A.F. Izmaylov, J. Bloino,
G. Zheng, J.L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda,
J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven,
J.A. M. Jr., J.E. Peralta, F. Ogliaro, M. Bearpark, J.J. Heyd, E. Brothers, K.N. Kudin,
V.N. Staroverov, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell,
J.C. Burant, S.S. Iyengar, J. Tomasi, M. Cossi, N. Rega, J.M. Millam, M. Klene,
J.E. Knox, J.B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts,
R.E. Stratmann, O. Yazyev, A.J. Austin, R. Cammi, C. Pomelli, J.W. Ochterski,
R.L. Martin, K. Morokuma, V.G. Zakrzewski, G.A. Voth, P. Salvador,
€
J.J. Dannenberg, S. Dapprich, A.D. Daniels, O Farkas, J.B. Foresman, J.V. Ortiz,
J. Cioslowski, D.J. Fox, Gaussian Inc., Wallingford CT, 2009.
[68] M. Hall, Correlation Based Feature Selection for Machine Learning (Doctoral
dissertation), Department of Computer Science, University of Waikato, 1999.
[69] I.H. Witten, E. Frank, M.A. Hall, Data Mining: Practical Machine Learning Tools
and Techniques, second ed., Morgan Kaufmann, San Francisco, 2011.