Synthesis, Anticancer, and QSAR Studies of 2-Alkyl(aryl,hetaryl)quinazolin-4(3H)-thione's and [1,2,4]Triazolo[1,5-c]quinazoline-2-thione's Thioderivatives

Article abstract of DOI:10.1002/hlca.201600062

Considering the frightening high level of mortality from cancer, studies of anticancer agents are vital nowadays. The 24 thioderivatives of 2-alkyl(aryl)-quinazolin-4(3H)-thiones and 20 thioderivatives of [1,2,4]triazolo[1,5-c]quinazoline-2-thiones were synthesized and evaluated for preliminary in?vitro anticancer activity with subsequent in silico QSAR analysis. The substance 18 had the best results inhibiting growth of eight cancer cell lines: CCRF-CEM of leukemia; SF-539, SNB-75, and U251 of CNS cancer; 786, RXF393, and UO-31 of renal cancer; and MDA-MB-231/ATCC of breast cancer (?31.50?–?47.41% of cell growth) with low procancer effect. Calculated QSAR-models for CCRF-CEM of leukemia, T-47D and HS 578T of breast cancer, and mean cell growth demonstrated good rate of anticancer activity prediction (r²?=?0.7?–?0.8, Q²_LOO=?0.5?–?0.7).

Full text of DOI:10.1002/hlca.201600062

Helv. Chim. Acta 2016, 99, 1 – 11
1
201600062
FULL PAPER
Synthesis, Anticancer, and QSAR Studies of 2-Alkyl(aryl,hetaryl)quinazolin-4(3H)-
thione’s and [1,2,4]Triazolo[1,5-c]quinazoline-2-thione’s Thioderivatives
by Oleksii M. Antypenkoa), Sergiy I. Kovalenkoa), Oleksandr V. Karpenkob), Vladyslav O. Nikitinb), and Lyudmyla
M. Antypenko*a)
a) Organic and Bioorganic Chemistry Department, Zaporizhzhya State Medical University, 26, Mayakovsky Ave.,
Zaporizhzhya, 69035, Ukraine (phone: +380669658740; e-mail: antypenkol@gmail.com)
b) Enamine Ltd., 78, Chervonotkats’ka str., Kyiv, 02094, Ukraine
Considering the frightening high level of mortality from cancer, studies of anticancer agents are vital nowadays. The 24
thioderivatives of 2-alkyl(aryl)-quinazolin-4(3H)-thiones and 20 thioderivatives of [1,2,4]triazolo[1,5-c]quinazoline-2-thiones
were synthesized and evaluated for preliminary in vitro anticancer activity with subsequent in silico QSAR analysis. The
substance 18 had the best results inhibiting growth of eight cancer cell lines: CCRF-CEM of leukemia; SF-539, SNB-75, and
U251 of CNS cancer; 786, RXF393, and UO-31 of renal cancer; and MDA-MB-231/ATCC of breast cancer (À31.50 – 47.41%
of cell growth) with low procancer effect. Calculated QSAR-models for CCRF-CEM of leukemia, T-47D and HS 578T of
breast cancer, and mean cell growth demonstrated good rate of anticancer activity prediction (r2 = 0.7 – 0.8,
Q2LOO = 0.5 – 0.7).
Keywords: 2-Alkyl(aryl)-quinazolin-4(3H)-thiones, Anticancer, [1,2,4]triazolo[1,5-c]quinazoline-2-thiones, QSAR analysis,
Structure–activity relationship.
quinazolin-2-yl]phenyl}amino)acetamide (F) had GI50 of
45.6 lg/ml against prostate cancer PC3 cell line, as well as
Introduction
Lately, Malvezzi et al. [1] calculated that 1 359 100 Euro-
peans were predicted to die of cancer in 2015, consisting
of 766 200 men and 592 900 women, which projected the
derivatives with electron-withdrawing groups – 4-NO2 (G)
and 4-Cl (H) displayed 55.6 lg/ml (Fig. 1) [8].
Furthermore, quinazoline condensed derivatives have
7.5% decrease in rates for men and a 6% for women, also shown in vitro anticancer activity: 2-(methylsulfonyl)-
compared with data in 2009. Considering these terrifying 4-(4-nitrobenzyl)-[1,2,4]triazolo[1,5-a]quinazolin-5(4H)-one
data, new anticancer agent search is of undoubtedly obli-
gatoriness not only in Europe, but also all over the
world.
(I) had IC50 1.29 lg/ml and 5-chloro-2-(methylsulfanyl)-
[1,2,4]triazolo[1,5-a]quinazoline (J) had 2.93 lg/ml against
medulloblastoma (Daoy) [9]. The latter one also had IC50
The 4-anilinoquinazoline moiety is a known pharma- 3.88 lg/ml against melanoma (SK-MEL283) and 4.52 lg/
cophore, which exhibits potent EGFR tyrosine kinase ml against hepatocellular carcinoma (HepG2), and
inhibitory activity of afatinib [2], erlotinib [3], and saraca- 2-(methylsulfanyl)-[1,2,4]triazolo[1,5-a]quinazoline-5(4H)-
tinib [4] (Fig. 1). Moreover, N-{4-[(3-chlorophenyl)amino] thione (K) had 5.53 lg/ml against medulloblastoma (Daoy),
quinazolin-6-yl}picolinamide showed anticancer properties exceeding dasatinib (Fig. 1).
at 2.9 lM against human c-Src kinases when compared
Considering patented anticancer activity mechanisms
with saracatinib [5]. Benzylideneamino A and 3-phe- of quinazolines, they can be divided into seven major
nylthioureido B derivatives of 4-(3-R-3,4-dihydro-4-oxo- groups: p53 regulators, serine-threonine and tyrosine
quinazolin-2-ylamino)-N-(pyridin-2-yl)benzenesulfonamide
had GI50 of 22.11 and 19.70 lM against human liver cell
line HEPG2, respectively [6]. Moreover, 2-(3-benzyl-3,4-
kinases, dual TK-histone deacetylases, and dehydrofolate
reductase inhibitors, radioligands for imaging, and other
target compounds [10]. For example, afatinib is a tyro-
dihydro-6-iodo-4-oxoquinazolin-2-ylsulfanyl)acetohydrazide sine kinase inhibitor (TKI) that also irreversibly inhibits
substituted with N-propanethione (C), N-benzyloxo (D), epidermal (EGFR) and human epidermal growth factor
and cyclic piperazine-3,6-dione (E) fragment showed receptors (HER2 and HER4) [2]. Another interesting
mean graph midpoint GI50 values of 12.8, 11.3, and
13.8 lM of nine subpanel anticancer cell lines (Fig. 1)
anticancer target for analysis is Ser/Thr protein kinase
CK2 (casein kinase II), a constitutively active serine/
[7]. Besides, N-(4-methoxyphenyl)-2-({4-[4-(m-tolyloxy) threonine kinase, that is involved in a variety of roles
€
© 2016 Wiley-VHCA AG, Zurich
DOI: 10.1002/hlca.201600062

2
Helv. Chim. Acta 2016, 99, 1 – 11
Fig. 1. Derivatives of quinazoline with anticancer properties.
essential for cellular homeostasis. It controls multiple
processes that play important roles in the sensitivity of
(1 – 44). Thus, summary of the already reported and several
novel substances’ synthetic routs were selected and pre-
cancer to EGFR targeting medicines, including PI3K- sented in this article below (Scheme) [13 – 17].
Akt-mTOR signaling, Hsp90 activity, and inhibition of
apoptosis [11].
Moreover, already reported 2-([1,2,4]triazolo[1,5-c] tral data (IR, LC/MS, H-NMR spectra).
The structures of all synthesized compounds were
accurately evaluated by elemental analysis and their spec-
1
quinazolin-2-ylsulfanyl)-N-m-tolylacetamide was already
proven as protein kinase CK2 inhibitor with 82% of
remaining activity in the 33.3 lM concentration as well
Substances 1 and 2 were obtained by treatment of the
quinazoline-4(3H)-one (L) with 2,4-bis(4-methoxyphenyl)-
1,3-dithia-2,4-diphosphatane-2,4-disulfide (Lawesson’s rea-
other derivative – ethyl 4-(2-([1,2,4]triazolo[1,5-c]quina- gent) or P2S5 in dioxane (Scheme) [13]. The corresponding
zolin-2-yl-sulfanyl)acetamido)benzoate (76%) [12].
Therefore, the search of novel anticancer agents
among thioderivatives of 2-alkyl(aryl)-quinazolin-4(3H)- amount of MeOK [13][14].
substituted 3 – 24 were obtained by alkylation of 1 by
proper halogen derivatives in MeOH with an equivalent
thiones and [1,2,4]triazolo[1,5-c]quinazoline-2-thione with
subsequent in silico QSAR analysis definitely seems to be
promising.
For 2-R-4(3H)quinazoline-4-thiones 1 and 2 alkyla-
tion, there were found some peculiarities, when refluxed
with phenacylhalides in MeOH with MeONa for
5 – 10 min for Br derivatives and 10 – 25 min for Cl
derivatives [14] (Scheme), namely, in some cases, ‘sulfide
compression’, described by Eschenmoser [18], has
occurred due to intermolecular attack of enol electron-
Results and Discussion
Among the main aims of this article was SAR analysis of
anticancer activity of the quinazolin-4(3H)-thione’s and deficient thioamide C-atoms, which initially lead to the
[1,2,4]triazolo[1,5-c]quinazoline-2-thione’s
S-derivatives
formation of the corresponding thiirane, and then, after
€
© 2016 Wiley-VHCA AG, Zurich
www.helv.wiley.com

Helv. Chim. Acta 2016, 99, 1 – 11
3
Scheme. Synthetic routes of investigated derivatives.
a) 2,4-Bis(4-methoxyphenyl)-1,3-dithia-2,4-diphosphetane-2,4-disulfide or P2S5, dioxane, reflux, 2 h. b) ClCH2COOH, MeOK, MeOH, reflux,
1 – 5 h; or proper halogen derivative, MeOK, MeOH, reflux, 2 h. c) Proper halogen derivative, MeOK, MeOH, r.t., 5 – 25 min. d) NH2NH2 Á
H2O, dioxane, reflux, 2 h. e) Potassium ethyl xanthogenate, iPrOH, reflux, 3 h. f ) Proper halogen derivative, iPrOH, H2O, reflux, 2 h; KOH,
EtOH, amine Á HCl, reflux, 3 h. g) conc. HCl/H2O/MeOH 1:5:5, reflux, 1 h.
the sulfur elimination, formed the ylidene ethanoles,
whose structure were proved by NMR data [14]. For
Cyclization of 4-hydrazino-3H-quinazoline (M) with
i
potassium ethyl xanthogenate in PrOH lead to the forma-
1
instance, comparing H-NMR spectra of substance 10 and tion of potassium [1,2,4]triazolo[1,5-c]quinazolin-2-ylthiolate
proper 1-(1-benzofuran-2-yl)-2-(quinazolin-4-yl)ethanone
(N), which consequently was alkylated with proper halogen
(10a), the two-H-atom singlet of the last CH2 group was derivatives to get 25 – 42 [15][16]. Alternatively, amides
absent at 5.01 ppm and a one-H-atom singlet appeared at were formed by activation of the corresponding carboxylic
6.99 ppm. Also, the significant increase in electron density
of pyrimidine ring was noticed when the H-atom singlet
in the second position of quinazoline was shifted at
acids with carbonyldiimidazole with further aminolysis.
The nucleophilic degradation of the pyrimidine ring of
the proper compounds 25 – 36 by reflux with soln. of HCl
1 ppm toward the strong field compared with S-containing resulted in 2-(3-(R-sulfanyl)-1H-1,2,4-triazol-5-yl)anilines
derivatives. 43 and 44 [17].
€
© 2016 Wiley-VHCA AG, Zurich
www.helv.wiley.com

4
Helv. Chim. Acta 2016, 99, 1 – 11
to be selective against cell line UO-31 (17.01% of CG) of
renal cancer for 4 or against HOP-62 of nonsmall cell
lung cancer (16.71%) for 3.
Biological Activity
Anticancer Preliminary in vitro Testing. All 44 compounds
were selected by NCI Developmental Therapeutic
Program according to their own internal demands to be
studied in vitro for anticancer activity [19]. The human
tumor cell lines were derived from nine different cancer
types: leukemia, melanoma, lung, colon, CNS, ovarian,
renal, prostate, and breast cancers. A single high dose
was used (1.0 lM) in the full 60-cell panel. Results for
each test agent were reported as the percent growth of
the treated cells when compared with the untreated
control cells (Supplementary Information). The activity of
the compounds was measured according to a value of 100,
which meant no cell growth (CG) inhibition. A value of
30 would mean 70% growth inhibition. A value of 0
meant no net growth over the course of the experiment.
A value of À30 would mean 70% lethality. A value of
À100 meant all cells were dead.
The presence of the MeO group in the Ph fragment was
essential to enhance the inhibition of cancer cell lines pro-
liferation by phenacyls 7 – 18: 1-(3-methoxyphenyl)-2-(qui-
nazolin-4-ylsulfanyl)ethanone (8) had lethal effect against
HOP-62 of nonsmall cell lung cancer (À17.12%) and 786-0
of renal cancer (À52.80%, which was the best result among
all compounds and cell lines), and 1-(2,5-dimethoxyphe-
nyl)-2-(2-morpholinoquinazolin-4-ylsulfanyl)ethanone (18)
– against SNB-75 of CNS cancer (À15.58%) and CCRF-
CEM of leukemia (À31.50%). Also, 2-(3,4-dimethoxyphe-
nyl)quinazoline-4(3H)-thione
2
had strong selective
influence at the last mentioned cell line of leukemia
(18.75%). It is notable that 2-(2-(morpholinomethyl)quina-
zolin-4-ylsulfanyl)-N-phenylacetamide (24) practically had
no anticancer effect, which additionally proved obligatori-
ness of 2,5-dimethoxyphenyl radical presence into the
structure to induct activity.
In the result, there were found 16 compounds possess-
ing anticancer activity: 2 – 4, 8 – 13, 16, 18, 19, 27, 29, 32,
and 37 (Tables 1 and 2), with subsequent structure–activ-
ity relationship analysis.
The colon, melanoma, and prostate cancers were practi-
cally insensitive to the synthesized substances. Only 1-(2,5-
difluorophenyl)-2-(2-styrylquinazolin-4-ylsulfanyl)ethanone
(16) inhibited growth of MDA-MB-435 of melanoma and
HCT-116 of colon cancer, with 2,3-dihydrobenzo[b][1,4]-
dioxin-6-yl derivative 11, that also acted against the last
cancer cell line.
Thus, anticancer data of the 3-[(2-methylquinazolin-4-
yl)sulfanyl]acetic acid (3) and 2-{[2-(trifluoromethyl)-
quinazolin-4-yl]sulfanyl}acetic acid (4) showed that the
presence of 2-CF3 in the latter instead of 2-Me was vital
Table 1. Cancer cells growth percentage [%] after addition of the most active compoundsa)
Compoundc)
Typeb)
CNS
Cell line
3
8
11
12
13
16
18
27
32
U251
SNB–75
SF-295
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
42.07
À15.58
–
–
–
44.78
–
–
–
–
SF-539
HOP-62
–
16.71
–
39.65
–
39.92
16.34
–
23.41
–
23.47
–
3.37
NSCL
R
À17.12
786-0
A498
–
–
–
–
–
–
–
À52.8
47.98
–
–
–
–
–
–
–
–
47.89
49.69
51.55
–
22.28
–
46.20
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
À42.37
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
44.03
–
–
–
–
–
–
–
RXF 393
UO-31
47.41
–
–
–
42.54
B
L
MDA-MB-468
T-47D
MCF7
18.03
41.38
–
43.93
MDA-MB-231/ATCC
SR
–
40.46
–
25.84
–
32.79
–
–
–
HL-60(TB)
RPMI-8226
CCRF-CEM
IGROV1
NCI/ADR-RES
MDA-MB-435
HCT-16
–
–
–
–
–
–
–
–
23.47
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
32.95
–
43.18
–31.50
O
–
–
–
–
–
–
–
–
–
–
–
M
Col
33.12
35.06
a) The results are given for the most active compounds inhibiting the growth in more than 52%. b) CNS, central nervous system cancer; R,
renal cancer; NSCL, nonsmall cell lung cancer; Col, colon cancer; B, breast cancer; L, leukemia; O, ovarian cancer; M, melanoma. c) All other
data are presented in the Supplementary Material. Compounds with strong inhibition against only one line were 2 – 18.75% (CCRF-CEM of
L), 4 – 17.01% (OU-31 of R), 9 – 26.91% (SR of L), 10 – 49.25% (T-47D of B), 19 – 46.04% (MDA-MB-468 of B), 29 – 36.43% (A-498 of
R) and 37 – 42.82% (RPMI-8226 of L).
€
© 2016 Wiley-VHCA AG, Zurich
www.helv.wiley.com

Helv. Chim. Acta 2016, 99, 1 – 11
5
Table 2. The best QSAR models’ statistical characteristics
No.
Tr.
Cell linea)
L, CCRF-CEM
BC, T-47D
BC, HS 578T
Mean CG
nb)
r2
RMSE
34
35
30
36
0.802
12.3146
13.333
29.311
0.698
9
0.787
7.325
8.047
21.475
0.701
9
0.701
9.886
10.794
15.270
0.517
8
0.817
3.290
3.604
26.853
0.764
8
s
F
Q2LOO
Pr.
n
r2
RMSE
0.699
20.802
0.7026
17.865
0.519
23.586
0.767
11.936
a) L, leukemia; BC, breast cancer. b) n, number of investigated compounds in set; RMSE, root mean square error; s, standard error; F, vari-
ance ratio, Fisher coefficient; Q2LOO, weighted correlation coefficient by leave-one out method.
Condensation with triazole ring greatly weakened sub- substituent was more preferable than in the third position,
stances anticancer potential. Among alkylamides 25 – 27, for cell line A498 of renal cancer inhibition indicating that
t
branching from Bu in 25 to Bu for 26 lead to decrease in even minimum differences in molecule mattered a lot.
anticancer properties, and shortening to diethyl amide
allowed substance 27 to inhibit HOP-62 of nonsmall cell
Synthesis of the aliphatic alkylamine 39, alicyclic
amine 38, alcohol 40, aliphatic amide 42, and phenacyl 41,
lung cancer and SR of leukemia at the high level – as well as nucleophilic degradation to 2-(3-(R-sulfanyl)-
23.47% and 32.79% of CG, correspondingly. 1H-1,2,4-triazol-5-yl)anilines 43 and 44 caused no posses-
Furthermore, one of the strongest effects was achieved sion of strong anticancer properties, but just stimulation
by 2-([1,2,4]triazolo[1,5-c]quinazolin-2-ylsulfanyl)-N-(4- of some cancer cell lines proliferation.
chlorophenyl)acetamide (32) with À42.37% rate of cell
death against 786-0 of renal cancer and 96.63% inhibi-
tion of HOP-62 of nonsmall cell lung cancer. But aceta-
mide 31, which differed only by 3-chlorophenyl radical,
practically, had no influence at the same cell lines.
Thus, not only the presence of the chlorine in the Ph
fragment, but also its exact position in the ring was essen-
tial for activity. Comparing the results of acetamide 32
Summing up, the widest range of moderate activity
was exhibited by 2-(2-R-quinazolin-4-ylsulfanyl)-1-pheny-
lethanones 11 – 13, 16, and 18, showing the importance of
the second position substitution of their quinazolin-4(3H)-
thione’s ring by relatively small substituents or aliphatic
heterocycle and the presence of electron-donating groups
in the Ph residue.
Although, condensation of the triazole with quinazo-
(amide of acetic acid) and 35 (amide of 2-propionic acid), line ring to [1,2,4]triazolo[1,5-c]quinazoline-2-thiones
the great decreasing effect of the small Me group was
observed on the growth and proliferation of the cancer
(25 – 42) lessened their anticancer activity range and
level, surprisingly, acetamide 32 had lethal effect against
cells. The growth range for 32 was from À42.37 to line 786-0 of renal cancer. This fact once again reaffirmed
150.88% and for 35, it decreased to 67.83 – 113.43%.
similar correlation was detected for alicyclic
piperidines: the absence of Me branching of 37 opposite
that only in vitro results could show the real picture of
biological activities.
Some of the proposed correlations were in accordance
A
to 42 initiated 42.28% of RPMI-8226 CG of leukemia, with those observed by Patel and co-authors that among
which was the better result, than 75.80%. N-Ph substituted quinazoline thioacetamide derivatives,
The N-phenyl-2-(2-methylquinazolin-4-ylsulfanyl)acet- MeO, NO2, and Cl substituents had positive impact in the
amide 21 and 2-propionamide 23 had similar principle in anticancer properties against human PC3 cells [8]. Still,
cancer cell lines inhibition, because the presence of the
Me group in the chain of the carboxylic acid residue also
decreased the minimum CG for the first amide, by small
amount, from 52.08 to 64.41%, as well as the absence of
the Me radical in the second position of the quinazoline
core in 19 improved its activity to 46.04%.
the majority of substances had the best result against dif-
ferent cancer cell lines and, unfortunately, even promoted
its growth (Supplementary Information). The strongest
promotion of the cancer cell growth was detected for sub-
stances 44 (160%, presence of COOH group), 10 (164%,
benzofuran-2-yl), 4 (170%, CF3), 31 (171%, 3-ClPh), and
Even for acetamide 33 and 2-propionamide 36, the 36 (197%, 4-NO2Ph). Thus, these substances were not
same correlation was observed: the first promoted growth
at 112.98% of melanoma SK-MEL-28, but the second
promoted its CG at 197.19%, which was the highest rate
among all investigated substances.
By analyzing activities of substances 29 and 30, it was
found that the Me group in the second position of the Ph
chosen to be investigated in five doses.
QSAR Studies. Molecular descriptors are digital values,
which received the quantification of diverse structural and
physicochemical characteristics of the molecule [20]. They
could estimate that these attributes to establish the
€
© 2016 Wiley-VHCA AG, Zurich
www.helv.wiley.com

6
Helv. Chim. Acta 2016, 99, 1 – 11
molecule biological activities, namely, anticancer properties
in this case. The program QSARINS 2.2 calculated the best
regression using descriptors obtained by Dragon 5.5 and
MOPAC2012 [21 – 23]. Due to their very high quantity,
genetic algorithm (GA) and multiple linear regression
analysis (MLRA) were used.
Graphical representations of the experimental and
calculated values of growth percent for all reported
QSAR models are shown in Fig. 2.
Comparing in vitro experimental data with calculated
data using models, the best predicted ones were found for
substances 28 and 40 against CCRF-CEM cell line of leu-
kemia (Table 4). The values differed only in 0.01 points
for the last substance.
Hence, all cancer cell lines were analyzed and upon
their descriptors, 60 QSAR models were calculated. But
only the four best equations were reported hereby:
CCRF-CEM of leukemia, T-47D and HS 578T of breast
cancer, and mean cell growth, relying on their highest
squared regression (r2) and weighted correlation coeffi-
cients by leave-one out method (Q2LOO; Tables 2 and 3).
When mean growth percent was used as single end
point, the best r2 (0.817) was found, still very close to
Also for cell line T-47D of breast cancer, the best pre-
diction was achieved for compound 9 (distinction is 0.2
points). For cell line HS 578T of breast cancer, there were
found more substances with closest results: 1, 3, 14, and
25 with accuracy of the prediction in < 0.5 odds.
To propose the direction of the future modifications
of the synthesized compounds, for example, against
all results. Then, simple (up to 4 – 5 descriptors in T-47D cell line of breast cancer, it was decided to create
model), transparent, and thereby realistic QSAR models novel structures’ base and calculate their anticancer activ-
were proposed to avoid errors arising from overfitting ity using proposed QSAR model.
of data (i.e., inclusion of greater number of descriptor;
Table 3).
Thus, 177 compounds were drawn and their activity
was predicted to establish structure–activity correlation
(Supplementary Information, Table 2, WinRAR file).
In the result, considerable amount of proposed sub-
Analyzing the QSAR models, among the essential
descriptors were: RDF065e – weighted by atomic Sander-
son electronegativities; RDF075m, RDF90m, RDF105m stances with better anticancer properties were derivatives
and RDF140m – weighted by mass, RDF125p – weighted of quinazolin-4(3H)-thione and condensation with triazole
by polarizability; piPC10 – molecular multiple path count ring proved to worsen the cancer cells growth inhibition.
of order 10; RSIpw3 and RSIpw5 – randic shape index;
The summarized SAR to enhance or to reduce anti-
H7v – H autocorrelation of lag 7/weighted by Van der cancer properties is shown in Fig. 3.
Waals volume; BEHe2 – highest eigenvalue n. 2 of Bur-
den matrix/weighted by atomic Sanderson electronegativi-
ties; HATS4v – leverage-weighted autocorrelation of lag
Therefore, it can be concluded that proposed QSAR
models could be used for prediction of the anticancer
properties of the consequently designed and synthesized
4/weighted by Van der Waals volume; C-032 – atom-cen- compounds in the series of quinazolin-4(3H)-thione’s and
tered fragments X-CX-X; F07[O-S] – frequency of O-S at [1,2,4]triazolo[1,5-c]quinazoline-2-thione’s derivatives to
topological distance 7; D/Dr10 – distance/detour ring improve effectiveness of anticancer agents search.
index of order 10; T(N..N) – sum of topological distances
between N..N; MATS7v – Moran autocorrelation – lag 7/
weighted by atomic Van der Waals volumes; STD(O,O)-
sum of topological distances between O..O [23].
Conclusions
In the result of NCI anticancer preliminary screening
of 44 2-alkyl(aryl,hetaryl)quinazolin-4(3H)-thione’s and
[1,2,4]triazolo[1,5-c]quinazoline-2-thione’s thioderivatives,
16 substances were found to possess mild to strong anti-
cancer activities. Among 60 studied types of cancer cell
lines, 21 were sensitive to synthesized compounds. The
The performance of all selected models had a little
positive bias, due to conducted randomization according
to the best r2. The prediction of the anticancer activity
by the calculated QSAR models showed their high
accuracy. It was decided to add prediction of the mean
cell growth data into the Supplementary Information cell line 786-0 of renal cancer had the highest rate of cell
death: À52.8% for 1-(3-methoxyphenyl)-2-(quinazolin-4-
ylsulfanyl)ethanone (8) and À42.37% for 2-([1,2,4]triazolo
and focused in the text at the exact cancer cell lines
(Table 4).
Table 3. Calculated QSAR models for leukemia (L) and breast cancer (BC)
Equation of growth inhibition
Cell line
L, CCRF-CEM
17.782 (Æ 8.4423)*piPC10-2.8039 (Æ 0.6886)*RDF065e-297.024 (Æ 130.0034)*H7v-615.574
(Æ 399.6342)*path/walk3-Randicshapeindex(RSIpw3)+208.5102 (Æ 120.4788)
BC, T-47D
À77.8571 (Æ 30.7625)*BEHe2 + 2.2147 (Æ 1.5242)*RDF090m-3.4268 (Æ 1.9111)*RDF105 m-94.1573
(Æ 90.3526)*HATS4v+12.7357 (Æ 6.3241)*C-032 + 389.8157* (Æ 111.3616)
BC, HS 578T
Mean cell growth
11.9385 (Æ 4.0578)*RDF140m-5.6197 (Æ 3.4452)*RDF125p-16.4186 (Æ 12.3636)*F07[O-S]-1801.38
(Æ 783.495)*path/walk5-Randicshapeindex (RSIpw5)+290.8168 (Æ 78.6734)
À0.1802 (Æ 0.0456)*D/Dr10-0.0957 (Æ 0.0831)*T(N..N)-9.7387 (Æ 4.4643)*MATS7v+1.7553
(Æ 0.5924)*RDF075m-0.4213 (Æ 0.162)*STD(O,O)+103.9585 (Æ 4.3738)
€
© 2016 Wiley-VHCA AG, Zurich
www.helv.wiley.com

Helv. Chim. Acta 2016, 99, 1 – 11
7
Table 4. Experimental data of substances in vitro anticancer activity in comparison to QSAR models [%]
Compound
Leukemia, CCRF-CEM
Breast cancer, T-47D
Breast cancer, HS 578T
Statusa)
Exp.
Pr.
93.89
Status
Exp.
Pr.
Status
Exp.
Pr.
1
2
Pr.
Tr.
Tr.
Tr.
Tr.
Pr.
Tr.
Tr.
Pr.
Tr.
Tr.
Tr.
Tr.
Tr.
Tr.
Tr.
Tr.
Tr.
Tr.
Pr.
Pr.
Tr.
Tr.
Pr.
Tr.
Pr.
Tr.
Tr.
Pr.
Tr.
Tr.
Tr.
Tr.
Tr.
–c)
Pr.
Tr.
Tr.
Tr.
Tr.
Tr.
Tr.
–
102.64
18.75
85.20
102.67b)
95.86
97.58
99.83
109.17
86.70
78.61
51.55
77.69
81.55
78.08
92.34
70.27
69.03
À31.5
56.83
93.09
72.54
82.06
67.74
94.56
94.97
99.68
82.61
102.05
93.12
96.19
68.96
111.89
92.28
88.74
–
Tr.
Tr.
Tr.
Pr.
Tr.
Tr.
Tr.
Tr.
Tr.
Tr.
Tr.
Pr.
Tr.
Tr.
Tr.
Pr.
Tr.
Tr.
Tr.
Pr.
Tr.
Tr.
Tr.
Pr.
Tr.
Tr.
Tr.
Tr.
Tr.
Tr.
Pr.
Tr.
Pr.
Tr.
–
102.96
92.97
118.67
107.69
101.89
104.21
84.12
78.29
77.2
100.75
89.97
113.84
114.59
86.37
103.37
78.02
83.04
77.00
51.60
71.90
81.40
80.74
74.15
84.22
77.51
84.70
99.81
71.62
81.44
71.50
67.52
70.95
72.75
113.13
92.36
108.05
85.28
96.53
82.07
71.25
90.84
94.34
74.09
100.45
85.68
98.36
104.65
96.03
105.54
92.65
95.60
103.88
98.11
Tr.
Tr.
Tr.
Tr.
Pr.
Tr.
Tr.
Tr.
Tr.
Tr.
Tr.
Tr.
Tr.
Tr.
Tr.
Tr.
Tr.
Tr.
Tr.
Tr.
Tr.
Tr.
Pr.
Tr.
Tr.
Tr.
Tr.
Tr.
–c)
Tr.
Pr.
Tr.
Tr.
Pr.
Tr.
Pr.
–
111.23
85.79
116.00
109.87
103.99
107.62
108.38
121.92
116.21
103.72
71.74
81.03
89.09
102.81
113.26
97.53
110.69
70.86
100.26
104.2
114.05
89.08
105.28
82.25
105.38
106.78
103.58
92.55
–
169.10
93.76
116.00
116.31
101.41
84.60
122.60
–
–
–
–
–
111.45
79.64
26.20
98.56
103.82
94.52
94.82
85.01
76.48
71.42
67.37
43.18
72.34
75.09
71.90
82.59
79.38
84.10
À20.68
81.42
85.31
81.49
72.60
75.77
61.66
85.67
127.94
96.32
101.61
81.54
94.93
85.31
99.19
84.48
79.68
98.17
89.16
106.65
111.76
94.82
89.23
97.79
96.90
97.19
108.52
3
116.22
108.15
104.67
114.87
111.88
111.94
109.86
88.61
4
5
6
7
8
9
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
37
38
39
40
41
42
43
44
49.25
59.52
76.2
80.54
103.50
109.16
102.68
98.17
72.32
64.14
90.99
52.53
83.35
94.99
76.18
95.83
67.20
71.25
65.92
109.75
91.71
87.11
110.68
84.07
91.78
90.29
89.9
93.66
114.27
74.56
107.04
104.86
110.43
86.24
77.34
104.23
115.98
103.93
106.71
84.52
104.25
157.59
89.48
99.55
84.91
79.04
–
82.27
95.75
108.53
100.03
110.36
109.8
90.13
103.29
109.86
98.13
109.01
101.05
89.37
80.94
79.81
129.37
92.41
89.24
89.64
96.63
–
Tr.
Tr.
Tr.
Tr.
Pr.
Tr.
Tr.
Tr.
Pr.
68.49
101.26
118.01
96.08
–
–
–
–
–
Tr.
Pr.
95.10
89.83
–
94.47
111.27
119.54
116.57
130.27
Tr.
99.25
a) Tr., Training set; Pr., prediction set. b) Bold values are the closest predicted results with < 1.5 digits difference. c) Substances have no
experimental data at this cell line.
[1,5-c]quinazolin-2-ylsulfanyl)-N-(4-chlorophenyl)acetamide ATCC of breast cancer (from À31.50 to 47.41%)). It is
(32) influence. The cell line HOP-62 of nonsmall cell lung
cancer had the widest range of effective substances
worth to mention its lowest cancer growth stimulation
effect (105.44%). Besides, the widest range of moderate
against it: 3, 8, 11, 16, 27, and 32. 1-(2,5-Dimethoxyphe- activity was shown by 1-(2,3-dihydro-1,4-benzodioxin-6-yl)-
nyl)-2-(2-morpholinoquinazolin-4-ylsulfanyl)ethanone (18) 2-(quinazolin-4-ylsulfanyl)ethanone (11; 22.28 – 79.59%).
was the best one among all investigated compounds
according to the results of the lethal effect and strong
inhibition against eight cancer cell lines (CCRF-CEM of
leukemia, SF-539, SNB-75, and U251 of CNS cancer; 786,
RXF393, and UO-31 of renal cancer; and MDA-MB-231/
SAR revealed that: i) introduction of MeO groups into
the thioacetophenone substituent in the fourth position
and morpholyl fragment into the second position of the
quinazoline ring amplified its anticancer properties of ser-
ies 1; ii) replacement of small substituents in the second
€
© 2016 Wiley-VHCA AG, Zurich
www.helv.wiley.com

8
Helv. Chim. Acta 2016, 99, 1 – 11
Fig. 2. Correlation graphs of predicted vs. experimental growth percentage for model equations.
position of the quinazoline ring was more effective, than inhibition; iv) branching of the thioalkyl residue lead to
changing radicals in the Ph ring for substances 7 – 18; iii)
condensation of quinazoline with 1,2,4-triazole ring for
series 2 highly narrowed the types and level of cell lines
procancer effect; and v) cell line HOP-62 of nonsmall cell
lung cancer, 786-0 of renal cancer, and SR, RPMI-8226,
and CCRF-CEM of leukemia were the most sensitive to
synthesized compounds. The calculated four best QSAR
models for CCRF-CEM of leukemia, T-47D and HS
578T of breast cancer, and mean cell growth predicted
anticancer data with satisfactory accuracy. Hence, the
directions of the purposeful modification in the series of
2-alkyl(aryl)-quinazolin-4(3H)-thione’s and [1,2,4]tria-
zolo[1,5-c]quinazoline-2-thione’s thioderivatives were
found, and would be consequently done along with the
multitarget QSAR method to derive predictive anti-
cancer models as an alternative to represent selection
among triazolo-, tetrazolo-, triazino, and uncondensed
quinazolines.
The authors gratefully acknowledge team of the Drug
Synthesis and Chemistry Branch, National Cancer
Institute, Bethesda, MD, USA, for in vitro evaluation of
anticancer activity.
Supporting Information
Fig. 3. Necessary modifications need to be done to improve anti-
cancer activity against T-47D cell line of breast cancer for the investi-
gated compounds according to the proposed QSAR model.
Supporting Information for this article is available on the
WWW under http://dx.doi.org/10.1002/hlca.201600062.
€
© 2016 Wiley-VHCA AG, Zurich
www.helv.wiley.com

Helv. Chim. Acta 2016, 99, 1 – 11
9
J = 8.2, 1 H, H–C(5)); 8.22 – 8.04 (m, 4 H, H–C(7), C(8),
Ph(2,6); 7.92 (t, J = 7.1, 1 H, H–C(3)); 7.69 (t, J = 7.2, 1
H, Ph(5)); 7.58 (t, J = 7.6, 2 H, H–C(6), Ph(4); 5.07 (s, 2
H, SCH2). LC/MS: 349 ([M + H]+). Anal. calc. for
Experimental Part
Chemistry
M.p.: determined in open capillary tubes in a Thiele’s appa- C17H11F3N2OS: C 58.62, H 3.18, N 8.04, O 4.59; found: C
ratus (Staffordshire, UK); uncorrected. UV Spectra: Ana- 58.66, H 3.14,
lytic Jena UV/VIS spectrophotometer Specord 200 (Jena, N 8.08, O 4.53.
Germany). IR Spectra (4000 – 600 cmÀ1): Bruker ALPHA 1-Phenyl-2-{[2-(2-phenylethyl)quinazolin-4-yl]sulfanyl}eth-
(Ettingen, Germany) FT-IR spectrometer using a module
for measuring attenuated total reflection (ATR);
anone (14). Yield: 68%. M.p. 92 – 94°. IR: 3060, 3025,
2978, 2903, 1697, 1613, 1596, 1561, 1548, 1483, 1449, 1372,
1341, 1322, 1291, 1252, 1202, 1151, 1111, 1074, 1028, 990,
À1
m in cm
.
1H- and 13C-NMR spectra (400 MHz and
~
500 MHz): Varian-Mercury 400 and Bruker Avance DRX- 936, 905, 879, 867, 844, 760, 739, 690, 646, 620. LC/MS:
500 spectrometers in (D6)DMSO soln.; d in ppm rel. to
385 ([M + H]+). Anal. calc. for C24H20N2OS: C 74.97, H
Me4Si as internal standard, J in Hz. LC/MS: chromatogra- 5.24, N 7.29, O 4.16; found: C 74.99, H 5.21, N 7.33, O
phy/mass spectrometric system which consists of HPLC
4.12.
Agilent 1100 Series (Palo Alto, CA, USA) equipped with 1-(2,4-Dichlorophenyl)-2-{[2-(trifluoromethyl)quinazolin-4-
diode-matrix and mass-selective detector Agilent LC/MSD yl]sulfanyl}ethanone (17). Yield: 78%. M.p. 164 – 166°.
SL (atmospheric pressure chemical ionization – APCI; Palo IR: 2995, 2943, 2886, 1709, 1584, 1566, 1552, 1487, 1463,
Alto, CA, USA). ESI-MS Spectra: Varian 1200l instrument 1380, 1349, 1287, 1262, 1246, 1185, 1163, 1145, 1109, 1061,
at 70 eV. The purity of all obtained compounds was
checked by 1H-NMR and LC/MS.
1002, 983, 958, 882, 870, 850, 819, 800, 773, 755, 732, 705,
689, 627. LC/MS: 418 ([M + H]+). Anal. calc. for
Starting substances (L, M, N), 1 – 5 [13], 7 – 12, 15, C17H9Cl2F3N2OS: C 48.94, H 2.17, N 6.71, O 3.83; found:
16, 19 – 24 [14], 43, 44 [15], 25 – 37, 42 [16], and 40, 41 C 48.91, H 2.19, N 6.68, O 3.86.
[17] were already synthesized and reported. Other start- 1-(2,5-Dimethoxyphenyl)-2-{[2-(morpholin-4-yl)quinazolin-
ing materials and solvents were obtained from comm-
ercially available sources and used without additional
purification.
4-yl]sulfanyl}ethanone (18). Yield: 75%. M.p. 172 – 174°.
IR: 3011, 2987, 2944, 2896, 2859, 1658, 1613, 1554, 1497,
1478, 1452, 1427, 1412, 1368, 1349, 1327, 1291, 1279, 1255,
1240, 1224, 1162, 1114, 1070, 1048, 1029, 1014, 995, 979,
928, 888, 863, 842, 816, 786, 753, 726, 713, 653, 638. 1H-
NMR (400 MHz, (D6)DMSO + CCl4): 7.88 (d, J = 7.9, 1
H, H–C(5)); 7.70 (t, J = 7.0, 1 H, H–C(7)); 7.46 (d,
J = 8.1, 1 H, H–C(8)); 7.26 (t, J = 7.4, 1 H, H–C(6));
7.17 – 7.23 (m, 2 H, Ph(3,4)); 7.13 (s, 1 H, Ph(2)); 4.77 (s,
2 H, SCH2); 3.89 – 3.94 (m, 2 H, OCH2); 3.70 – 3.75 (m,
2 H, OCH2); 3.60 (s, Me); 3.52 (s, Me); 3.30 – 3.36 (m, 4
Procedure to Obtain Substances 6, 13, 14, 17, and 18. To
a soln. of 0.23 g (0.01 mol) of metallic Na in 25 ml of
MeOH was added 0.01 mol of 2-R-quinazolin-4-(3H)-
thione. After the dissolution, the mixture was cooled to
5 – 15°, and 0.01 mol of appropriate halogen derivative
was added, heated until a neutral pH. A small amount of
H2O was poured and cooled. The precipitate was filtered
i
out, washed with H2O, and recrystallized from PrOH/ H, N(CH2)2). LC/MS: 426 ([M + H]+). Anal. calc. for
H2O 3:1 soln.
C22H23N3O4S: C 62.10, H 5.45, N 9.88, O 15.04; found: C
62.13, H 5.42, N 9.91, O 15.01.
2-Acetamido-3-{[2-(trifluoromethyl)quinazolin-4-yl]sulfa-
nyl}propanoic Acid (6). Yield: 73%. M.p. 178 – 180°. IR:
3389, 2992, 2886, 2828, 1914, 1721, 1606, 1566, 1529, 1487,
1438, 1412, 1390, 1351, 1308, 1290, 1260, 1246, 1226, 1205,
1165, 1136, 1110, 1045, 1017, 999, 983, 931, 881, 849, 837,
Procedure to Obtain Substances 38 and 39. To a soln. of
0.34 g (0.006 mol) KOH in 15 ml of EtOH was added
proper amine hydrochloride (0.006 mol) and 1.22 g
794, 766, 755, 734, 693, 631. 1H-NMR (400 MHz, (D6) (0.005 mol) of potassium [1,2,4]triazolo[1,5-c]quinazoline-
DMSO + CCl4): 12.96 (br. s, COOH); 8.48 (d, J = 8.2, 1
2-thiolate. The mixture was refluxed for 3 h and then
H, H–C(5q)); 8.21 (t, J = 8.3, 1 H, H–C(7q)); 8.11 (d, cooled down. Then, 30 ml of H2O was added, oil was
J = 8.3, 1 H, H–C(8q)); 7.89 (t, J = 5.6, 1 H, H–C(6q)); extracted with CHCl3, and recrystallized from iPrOH/
4.70 (q, J = 8.5, 1 H, CH2CH); 4.05 (dd, J = 13.6, 4.4, 1
H2O 3:1 soln.
H, NH); 3.52 (dd, J = 13.4, 9.0, 2 H, SCH2); 1.81 (s, Me). 2-{[2-(Morpholin-4-yl)ethyl]sulfanyl}[1,2,4]triazolo[1,5-c]
LC/MS: 360 ([M + H]+). Anal. calc. for C14H12F3N3O3S: quinazoline (38). Yield: 41%. M.p. 126 – 128°. IR: 3396,
C 46.80, H 3.37, N 11.69, O 13.36; found: C 46.84, H 3.34,
N 11.73, O 13.33.
1-Phenyl-2-{[2-(trifluoromethyl)quinazolin-4-yl]sulfanyl}-
3061, 2966, 2924, 2871, 2850, 2792, 1621, 1602, 1516, 1475,
1454, 1393, 1383, 1369, 1350, 1300, 1264, 1242, 1198, 1158,
1105, 1088, 1066, 1008, 982, 959, 912, 898, 869, 805, 768,
ethanone (13). Yield: 76%. M.p. 132 – 134°. IR: 3062, 745, 714, 641.1H-NMR: 9.50 (s, 1 H, H–C(5)); 8.41 (d, 1
3040, 2988, 2934, 1746, 1697, 1613, 1593, 1565, 1552, 1486,
1450, 1388, 1348, 1322, 1291, 1262, 1247, 1196, 1140, 1108,
1022, 999, 985, 938, 881, 866, 847, 778, 764, 757, 747, 730,
H, J = 8.0, H–C(10)); 8.06 (d, 1 H, J = 7.8, H–C(7)); 7.94
(t, 1 H, J = 8.0, H–C(8)); 7.82 (t, 1 H, J = 8.0, H–C(9));
3.42 (t, 2 H, J = 7.0, SCH2); 3.04 – 2.98 (m, 4 H, O
688, 644, 627. 1H-NMR (500 MHz, CDCl3): 8.33 (d, (CH2)2); 2.65 (q, 2 H, J = 7.8, SCH2–CH2); 1.32 – 1.27
€
© 2016 Wiley-VHCA AG, Zurich
www.helv.wiley.com

10
Helv. Chim. Acta 2016, 99, 1 – 11
(m, 4 H, (CH2)2N). LC/MS: 316 ([M + H]+). Anal. calc.
for C15H17N5OS: C 57.12, H 5.43, N 22.21, O 5.07; found: rized elsewhere [21 – 23]. Optimized structures were also
descriptors and the calculation procedures were summa-
C 57.16, H 5.41, N 22.23, O 5.04.
N-(1-methylethyl)-N-[2-([1,2,4]triazolo[1,5-c]quinazolin-2-
ylthio)ethyl] propan-2-amine (39). Yield: 69%. M.p.
used for calculation of additional important quantum-
chemical parameters (final heat of formation, total energy,
electronic energy, core–core repulsion, ionization poten-
82 – 84°. IR: 3055, 2957, 2868, 2836, 1622, 1608, 510, tial, HOMO, and LUMO), that were also used as descrip-
1475, 1453, 1385, 1361, 1305, 1288, 1251, 1234, 1212, 1196,
1159, 1122, 1107, 1068, 1034, 972, 958, 925, 900, 867, 782,
tors. MOPAC2012 was used to do such computation
(MOPAC2009 for CAChe, Stewart Computational Chem-
1
766, 726, 714, 683, 640. H-NMR: 9.53 (s, 1 H, H–C(5)); istry, Oregon, LLC, USA). The correlation coefficients for
8.40 (d, J = 8.0, 1 H, H–C(10)); 8.04 (d, J = 7.8, 1 H, H–
C(7)); 7.95 (t, J = 8.0, 1 H, H–C(8)); 7.82 (t, J = 8.0, 1 H,
all pair of descriptor variables used in the models were
evaluated to identify highly correlated descriptors in order
H–C(9)); 3.43 (t, J = 8.0, 2 H, SCH2); 2.88 – 2.84 (m, 2 H, to detect redundancy in the data set. Hence, descriptors
NCH2); 0.97 – 1.10 (m, 4 Me). LC/MS: 330 ([M + H]+). with constant variables and near-constant variables were
Anal. calc. for C17H23N4S: C 61.97, H 7.04, N 21.26, S
9.73; found: C 61.95, H 7.06, N 21.24, S 9.75.
excluded from the further consideration (r ≥ 0.95).
The genetic algorithm (GA) and multiple linear regres-
sion analysis (MLRA) were used to select the descriptors
and to generate the correlation models that relate the
structural features to the cell growth percent of different
Anticancer Activity
Primary anticancer assay was performed against human cancer cell lines. The combination of the GA-MLRA tech-
tumor cell lines panel derived from nine neoplastic dis-
eases, in accordance with the protocol of the Drug Evalu-
ation Branch, National Cancer Institute, Bethesda [19].
The human tumor cell lines of the cancer screening panel
were grown in RPMI 1640 medium containing 5% fetal
bovine serum and 2 mM L-glutamine. For a typical screen-
ing experiment, cells were inoculated in 96-well microtiter
plates in 100 ml assay volume, at plating densities ranging
from 5000 to 40 000 cell/well. After cell inoculation, the
microtiter plates were incubated at 37°, under an atmo-
nique was applied to obtain the best QSAR models using
the QSARINS 2.2. The latter program split compounds
data as the following: random selection of 20% of com-
pounds for prediction set and 80% for training set. For
each obtained model, such random selection was different.
Calculation of QSAR models was conducted sepa-
rately for each of 60 cell lines. Growth percent according
to the NCI protocol was not converted to any other value;
it was used in original version to built models. Some cell
lines were given the value of À999, which means, that
sphere of 5:95 CO2/air (v/v) at 100% relative humidity, they were not tested.
for 24 h prior to addition of drugs under assessment. Fol-
lowing drug addition (1 lM), the plates were incubated
for an additional 48 h, under the same conditions. Sul-
The amount of generation algorithm setup was set
until four descriptors, and generation per size was estab-
lished to the value of 3000, and the division into training
forhodamine B (SRB) soln. (100 ll, 0 – 4% w/v in 1% and test sets was performed automatically at a ratio of
aq. AcOH) was added to each well and plates were incu-
bated for 10 min at r.t. The percentage growth was evalu-
ated spectrophotometrically vs. controls not treated with
test agents.
80 – 20 percent relatively. Models, which showed statisti-
cal significance according to the parameters at a higher
level (r2 ≥ 0.5), were selected for a more thorough ren-
dering. For these lines, the following options were given:
the amount of generation algorithm setup was set until
seven descriptors, and generation per size was established
to the value of 10 000. Substances were spited into train-
ing and test sets and the division, was made such, as to
QSAR Modeling
First, all structures were built by MarvinSketch 6.3.0 (Mar-
vinSketch version: 6.3.0, ChemAxon, Cambridge, MA, establish equal distribution of substances of high and
USA). Then, they were preliminary optimized by program
HyperChem 8.0.8 using molecular mechanical MM+ algo-
rithm combined with semiempirical PM3 molecular mod-
eling method with a maximum number of cycles and
Polak-Ribiere (Conjugate Gradient) algorithm. Molecular
mechanics has been used to produce more realistic geome-
try values for the majority of org. molecules owing to the
fact of being highly parameterized. The next step was a
reoptimization of the MM+ optimized structures by apply-
ing semiempirical PM3 molecular modeling method.
Obtained files were further used for calculations.
moderate percentage of inhibition of cell growth.
REFERENCES
[1] M. Malvezzi, P. Bertuccio, T. Rosso, M. Rota, F. Levi, C. La
Vecchia, E. Negri, Ann. Oncol. 2015, 26, 779.
[2] J. Bliesath, N. Huser, M. Omori, D. Bunag, C. Proffitt, N. Strei-
ner, C. Ho, A. Siddiqui-Jain, S. E. O’Brien, J. K. C. Lim, D. M.
Ryckman, K. Anderes, W. G. Rice, D. Drygin, Cancer Lett.
2012, 322, 113.
[3] A. Aleksic, V. MacLaren, Lung Cancer 2014, 83, S9.
[4] M. A. Gubens, M. Burns, S. M. Perkins, M. S. Pedro-Salcedo,
S. K. Althouse, P. J. Loehrer, H. A. Wakelee, Lung Cancer
2015, 89, 57.
Descriptors were calculated using Dragon 5.5 (> 1500
descriptors) (DRAGON 5.5 for Windows, TALETE SRL,
Milano, Italy). The definition of all used molecular
[5] F. Fang, D. D. Li, J. R. Li, J. Sun, Q. R. Du, H. B. Gong, H. L.
Zhu, RSC Adv. 2013, 3, 26230.
€
© 2016 Wiley-VHCA AG, Zurich
www.helv.wiley.com

Helv. Chim. Acta 2016, 99, 1 – 11
11
[6] M. M. Ghorab, F. A. Ragab, H. I. Heiba, A. A. Bayomi,
J. Heterocycl. Chem. 2014, 51, E255.
[16] L. N. Antipenko, A. V. Karpenko, S. I. Kovalenko, A. M. Kat-
sev, E. Z. Komarovska-Porokhnyavets, V. P. Novikov, Arch.
Pharm. (Weinheim, Ger.) 2009, 342, 651.
[17] L. M. Antypenko, S. I. Kovalenko, O. V. Karpenko, A. M. Kat-
sev, V. P. Novikov, N. S. Fedyunina, Curr. Comput. Aided Drug
Des. 2015, 12, 29.
[7] A. A. Khalil, S. G. A. Hamide, A. M. Al-Obaid, H. I. El-Sub-
bagh, Arch. Pharm. (Weinheim, Ger.) 2003, 336, 95.
[8] A. B. Patel, K. H. Chikhalia, P. Kumari, Res. Chem. Intermed.
2015, 41, 4439.
€
[18] M. Roth, P. Dubs, E. Gotschi, A. Eschenmoser, Helv. Chim.
Acta 1971, 54, 710.
[9] R. Al-Salahi, M. Marzouk, A. E. Ashour, I. Alswaidan, Asian J.
Chem. 2014, 26, 2173.
[10] G. Marzaro, A. Guiotto, A. Chilin, Exp. Opin. Ther. Pat. 2012,
22, 223.
[19] M. R. Boyd, ‘The NCI In Vitro Anticancer Drug Discovery
Screen’, Anticancer Drug Development Guide Humana Press,
1997, p. 23.
[11] S. Kumar, R. Agrawal, Recent Pat. Anticancer Drug Discov.
2014, 9, 382.
€
[20] J. Kirchmair, A. H. Goller, D. Lang, J. Kunze, B. Testa, I. D.
[12] S. A. Starosyla, M. V. Protopopov, O. I. Voskoboinik, I. S.
Nosulenko, L. M. Antypenko, O. M. Antypenko, I. M. Fesun,
S. I. Kovalenko, S. M. Yarmoluk, Ukrain. Bioorg. Acta 2014, 1,
12. (http://www.bioorganica.org.ua/UBAdenovo/vol_12_1.htm)
[13] L. Antypenko, S. Kovalenko, Y. Posylkina, V. Nikitin, N.
Fedyunina, V. Ivchuk, J. Enzyme Inhibit. Med. Chem. 2015, 1.
[14] V. A. Nikitin, S. I. Kovalenko, I. F. Belenichev, Y. M. Maksi-
mov, N. O. Vrynchanu, Pharmacology 2006, 1, 122. (http://far-
macomua.narod.ru/Arx1_2_2006/1_2_2006.htm)
Wilson, R. C. Glen, G. Schneider, Nat. Rev. Drug Discov. 2015,
14, 387.
[21] P. Gramatica, N. Chirico, E. Papa, S. Cassani, S. Kovarich,
J. Comput. Chem. 2013, 34, 2121.
[22] P. Gramatica, S. Cassani, N. Chirico, J. Comput. Chem. 2014,
35, 1036.
[23] R. Todeschini, V. Consonni, ‘Handbook of Molecular Descrip-
tors’, Eds. R. Mannhold, H. Kubinyi, H. Timmerman, Wiley-
VCH, Weinheim, New York, 2000.
[15] L. Antipenko, A. Karpenko, S. Kovalenko, A. Katsev, E.
Komarovska-Porokhnyavets, V. Novikov, A. Chekotilo, Chem.
Pharm. Bull. 2009, 57, 580.
Received March 12, 2016
Accepted May 18, 2016
€
© 2016 Wiley-VHCA AG, Zurich
www.helv.wiley.com