Applying the designed multiple ligands approach to inhibit dihydrofolate reductase and thioredoxin reductase for anti-proliferative activity

Article abstract of DOI:10.1016/j.ejmech.2016.03.002

The development of multi-targeting drugs is currently being explored as an attractive alternative to combination therapy, especially for the treatment of complex diseases such as cancer. Dihydrofolate reductase (DHFR) and thioredoxin reductase (TrxR) are enzymes belonging to two unrelated cellular pathways that are known to contribute towards cancer cell growth and survival. In order to evaluate whether simultaneous inhibition of DHFR and TrxR by dihydrotriazines (DHFR-targeting) and chalcones (TrxR-targeting) may be beneficial, breast MCF-7 and colorectal HCT116 carcinoma cells were treated with combinations of selected dihydrotriazines and chalcones at a 1:1 M ratio. Two combinations demonstrated synergy at low-to-moderate concentrations. Based on this result, four merged dihydrotriazine-chalcone compounds were designed and synthesized. Two compounds, 11a [DHFR IC₅₀ = 56.4 μM, TrxR IC₅₀ (60 min) = 12.6 μM] and 11b [DHFR IC₅₀ = 2.4 μM, TrxR IC₅₀ (60 min) = 10.1 μM], demonstrated in vitro inhibition of DHFR and TrxR. The compounds showed growth inhibitory activity against MCF-7 and HCT116 cells, but lacked cytotoxicity. Molecular docking experiments showed 11b to possess rational binding modes to both the enzymes. In conclusion, this study has not only identified the dihydrotriazine and chalcone scaffolds as good starting points for the development of dual inhibitors of DHFR and TrxR, but also demonstrated the synthetic feasibility of producing a chemical entity that could result in simultaneous inhibition of DHFR and TrxR. Future efforts to improve the antiproliferative profiles of such compounds will look at alternative ways of integrating the two pharmacophoric scaffolds.

Full text of DOI:10.1016/j.ejmech.2016.03.002

European Journal of Medicinal Chemistry 115 (2016) 63e74
Contents lists available at ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Research paper
Applying the designed multiple ligands approach to inhibit
dihydrofolate reductase and thioredoxin reductase for anti-
proliferative activity
Hui-Li Ng, Shangying Chen, Eng-Hui Chew, Wai-Keung Chui*
Department of Pharmacy, Faculty of Science, 18 Science Drive 4, National University of Singapore, 117543, Singapore
a r t i c l e i n f o
a b s t r a c t
Article history:
The development of multi-targeting drugs is currently being explored as an attractive alternative to
combination therapy, especially for the treatment of complex diseases such as cancer. Dihydrofolate
reductase (DHFR) and thioredoxin reductase (TrxR) are enzymes belonging to two unrelated cellular
pathways that are known to contribute towards cancer cell growth and survival. In order to evaluate
whether simultaneous inhibition of DHFR and TrxR by dihydrotriazines (DHFR-targeting) and chalcones
(TrxR-targeting) may be beneficial, breast MCF-7 and colorectal HCT116 carcinoma cells were treated
with combinations of selected dihydrotriazines and chalcones at a 1:1 M ratio. Two combinations
demonstrated synergy at low-to-moderate concentrations. Based on this result, four merged
dihydrotriazine-chalcone compounds were designed and synthesized. Two compounds, 11a [DHFR
Received 22 November 2015
Received in revised form
29 February 2016
Accepted 2 March 2016
Available online 3 March 2016
Keywords:
Dihydrofolate reductase
Thioredoxin reductase
Dihydrotriazines
Chalcones
IC50
¼
56.4
m
M, TrxR IC50 (60 min)
¼
12.6
m
M] and 11b [DHFR IC50
¼
2.4 mM, TrxR IC50
(60 min) ¼ 10.1
m
M], demonstrated in vitro inhibition of DHFR and TrxR. The compounds showed growth
Antiproliferative activity
inhibitory activity against MCF-7 and HCT116 cells, but lacked cytotoxicity. Molecular docking experi-
ments showed 11b to possess rational binding modes to both the enzymes. In conclusion, this study has
not only identified the dihydrotriazine and chalcone scaffolds as good starting points for the develop-
ment of dual inhibitors of DHFR and TrxR, but also demonstrated the synthetic feasibility of producing a
chemical entity that could result in simultaneous inhibition of DHFR and TrxR. Future efforts to improve
the antiproliferative profiles of such compounds will look at alternative ways of integrating the two
pharmacophoric scaffolds.
© 2016 Elsevier Masson SAS. All rights reserved.
1. Introduction
gruelling issues associated with complex pharmacokinetic and
pharmacodynamic (PK/PD) relationships and potential drugedrug
The rational design of producing a single chemical entity that is
able to act on multiple targets relevant to a disease is an approach
currently gaining popularity. In 2005, the term “designed multiple
ligands” was coined by Morphy and Rankovic to describe such
compounds [1]. In the field of cancer chemotherapy, designed
multiple ligands offer several potential advantages over single
target compounds. Multiple pathways of cell survival could be
targeted simultaneously, leading to increased therapeutic efficacy
[2] and decreased development of cancer drug resistance [3].
Additionally, the use of a single multi-targeted agent instead of
administering multiple agents could help to circumvent the
interactions [1].
Dihydrofolate reductase (DHFR) and thioredoxin reductase
(TrxR) are two enzymes that are implicated in the growth and
survival of cancer cells. As part of the folate pathway, DHFR cata-
lyzes the reduction of dihydrofolic acid to tetrahydrofolic acid,
which is essential for the biosynthesis of nucleic acids and amino
acids. Methotrexate (MTX) is a DHFR inhibitor that has been used in
the clinic as an anticancer agent for more than 60 years. However,
its use has become limited by the emergence of acquired resistance
to MTX by cancer cells, and this has inspired efforts to develop non-
classical DHFR inhibitors that do not contain the glutamate residue
found in MTX [4e6]. Being lipophilic, non-classical inhibitors enter
cells via passive diffusion, thus circumventing drug resistance due
to transport resistance attributed to decreased levels of or muta-
tions in the reduced folate carrier, a putative carrier for active MTX
* Corresponding author.
E-mail address: phacwk@nus.edu.sg (W.-K. Chui).
http://dx.doi.org/10.1016/j.ejmech.2016.03.002
0223-5234/© 2016 Elsevier Masson SAS. All rights reserved.

64
H.-L. Ng et al. / European Journal of Medicinal Chemistry 115 (2016) 63e74
transport. Additionally, since non-classical DHFR inhibitors are not
polyglutamated, they are not affected by impaired poly-
glutamylation resulting from reduced folylpolyglutamate synthe-
tase (FPGS) activity. Examples of non-classical DHFR inhibitors are
the 4,6-diamino-1,2-dihydro-1,3,5-triazines, first synthesized by
E.J. Modest in 1951 [7]. Our laboratory has interest in the design and
development of this class of triazine-based compounds as non-
classical DHFR inhibitors [8e10].
TrxR and its substrate thioredoxin (Trx) make up the Trx system,
which plays an important role in maintaining redox homeostasis
and protecting cells against oxidative damage and mutation [11,12].
However, in malignancies, the biological activities of the Trx/TrxR
system contribute to tumor growth and progression. Indeed, it has
been demonstrated in several human cancers that overexpression
of Trx and TrxR are associated with drug resistance and poor pa-
tient prognosis [13e16]. The mammalian TrxR enzyme has a highly
accessible C-terminal active site containing a penultimate seleno-
cysteine (Sec) residue which is easily attacked by agents that
possess an electrophilic center [17e19]. Among such compounds
summarized in Tables 1 and 2 respectively. In vitro efficacy of
enzyme inhibitory activity was expressed as IC50 (50% DHFR/TrxR
inhibition concentration), while antiproliferative activity was
expressed as GI50, which is the concentration of drug at which 50%
of growth inhibition was achieved, and LC50, which is the concen-
tration of drug at which 50% of the cells were killed. As shown in
Table 1, T1 and T2 were moderately potent inhibitors of recombi-
nant human DHFR with IC50 values in the low
also inhibited growth of MCF-7 and HCT116 carcinoma cells with
GI50 values in the low M range, but were not cytotoxic as evident
in the LC50 values that were determined to be above 100 M. This
mM range. T1 and T2
m
m
could be related to their DHFR inhibitory activities that caused cells
to be unable to produce new DNA, RNA and proteins, thus resulting
in growth inhibition but not necessarily cell death. As shown in
Table 2, C1eC3 demonstrated moderately potent inhibition of rat
liver TrxR upon incubation with the enzyme for 60 min, recording
IC50 values in the
mM range. As evident from the obtained GI50 and
LC50 values in the
m
M range, the chalcones displayed moderately
good anti-proliferative activity and exerted cytotoxicity against
MCF-7 and HCT116 cell lines.
that possess electrophilic character, chalcones containing an a,b-
unsaturated carbonyl moiety, also known as a Michael acceptor
moiety, have recently been demonstrated to exert anti-proliferative
effects against human cancer cells through selective and irrevers-
In order to evaluate whether the use of 4,6-diamino-1,2-
dihydro-1,3,5-triazines and chalcones in combination would
result in enhanced anti-proliferative activity, the 3-(4,5-
dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT)
assay was carried out on MCF-7 and HCT116 cells treated with the
individual compounds and the six possible combinations of a 4,6-
diamino-1,2-dihydro-1,3,5-triazine with a chalcone at a 1:1 M ra-
tio. Cell viability after 72 h of incubation with the compounds was
then assessed.
ible inhibition of TrxR [20,21]. The electrophilic
b carbon atom
within the -unsaturated carbonyl moiety of chalcones is pro-
a,b
posed to react irreversibly with the C-terminal Sec residue of the
TrxR enzyme via a Michael addition reaction, thereby forming co-
valent adducts [20].
In the current study, the goal was to harness the combined
benefits of simultaneous inhibition of DHFR and TrxR using com-
pounds possessing the respective pharmacophoric scaffolds to
bring about enhanced anti-cancer effect. To do this, two approaches
were employed and examined. In the first approach, it was hy-
pothesized that the combined use of a 4,6-diamino-1,2-dihydro-
The doseeeffect curves of the 4,6-diamino-1,2-dihydro-1,3,5-
triazines T1-T2 and chalcones C1eC3 were observed to be mark-
edly different from each other. At high doses, the maximum
decrease in cell viability effected by T1-T2 was approximately 50%,
whereas C1eC3 decreased cell viability by 100%. Additionally, the
Hill slope of the doseeeffect curves of T1-T2 was ~1, whereas the
Hill slope of C1eC3 ranged from ~2 to 4. Synergy analysis using
mathematical models such as the Bliss independence model [22],
Loewe additivity model [23] and the median effect principle by
Chou and Talalay [24] was unsuitable, since an underlying
assumption for these models is that the individual drugs will give
rise to equal maximal effects. Modified equations based on the
Loewe additivity model have been derived by Grabovsky and Tal-
larida [25] for cases whereby the two individual drugs produce
different maximal effects. However, these equations assume that
Hill slope ¼ 1 for both drugs. In the absence of an appropriate
mathematical model, we therefore adopted the concept of coop-
erative effect synergy, which is simply defined as an increase in the
effect of the drug combination over the agents alone. In a review of
synergy methodologies, Geary recommended this simpler defini-
tion of synergy because it avoids mathematical difficulties, is easily
applicable, meets FDA criteria for evaluation of combination ther-
apies, and is adequate for basic discovery and clinical research [26].
In order to determine whether the drug combinations showed a
better effect than their individual components, their doseeeffect
curves were plotted and compared. Dose-effect curves on MCF-7
and HCT116 cells are shown in Figs. 2 and 3 respectively. In the
dihydrotriazine-chalcone combinations T1 þ C1 and T1 þ C3, when
1,3,5-triazine and
a
chalcone would result in enhanced
antiproliferative profiles against cancer cells in comparison to the
anti-tumor activities of the individual compounds. As such, MCF-7
breast and HCT116 colorectal carcinoma cells were treated with a
combination of 4,6-diamino-1,2-dihydro-1,3,5-triazines and chal-
cones at a 1:1 M ratio. In the second approach, it was hypothesized
that the structural combination of the 4,6-diamino-1,2-dihydro-
1,3,5-triazine and chalcone scaffolds would produce new molecules
that possessed potent antiproliferative profiles correlated to strong
DHFR and TrxR inhibitory activities. To test this hypothesis, the 4-6-
diamino-1,2-dihydro-1,3,5-triazine and chalcone scaffolds were
incorporated synthetically into a single chemical entity; in a pre-
liminary attempt, four merged dihydrotriazine-chalcone com-
pounds were synthesized and evaluated for anti-proliferative
activities and possible in vitro inhibitory effects on DHFR and TrxR.
In summary, findings obtained in this proof-of-concept study have
highlighted the feasibility of producing a single chemical entity to
bring about simultaneous inhibition of DHFR and TrxR, albeit the
chemical approach to merge dihydrotriazine and chalcone phar-
macophores to obtain dual-targeted compounds possessing well-
correlated enzyme (DHFR and TrxR) inhibitory and anti-
proliferative activities would need further optimization.
2. Results and discussion
used at low-to-moderate concentrations of 1.5e15 mM, both 4,6-
diamino-1,2-dihydro-1,3,5-triazine and chalcone components in
the combination contributed to a stronger anti-proliferative effect
that resulted in a greater reduction in cell viability in comparison to
the decrease in cell viability brought about by treatment with the
individual components (Figs. 2A, C, 3A, C). Hence, at these con-
centrations, a small cooperative effect was observed between T1
and C1 or C3. However, at higher concentrations, the doseeeffect
2.1. Combination treatment
Two 4,6-diamino-1,2-dihydro-1,3,5-triazines, T1 and T2, and
three chalcones, C1eC3, were selected for this study. The structures
of the compounds are shown in Fig. 1, and the enzyme inhibitory
activities and antiproliferative activities of T1-T2 and C1eC3 are

H.-L. Ng et al. / European Journal of Medicinal Chemistry 115 (2016) 63e74
65
Fig. 1. Structures of the 4,6-diamino-1,2-dihydro-1,3,5-triazines T1-T2 and the chalcones C1eC3 used in the combination study.
Table 1
in enhanced anti-proliferative activities at low-to-moderate con-
centrations of up to ~10 M. Thus, it should be noted that
Recombinant human DHFR inhibitory activity and anti-proliferative activity of
dihydrotriazines T1 and T2 on MCF-7 and HCT116 cells.
m
dihydrotriazine-chalcone combinations were still advantageous
when compared to treatment with 4,6-diamino-1,2-dihydro-1,3,5-
triazine or chalcone alone. In general, at lower concentrations, the
combinations had the advantage of greater growth inhibitory ac-
tivity when compared to the individual chalcones. At higher con-
centrations, the combinations demonstrated cell kill that was not
observed with the use of 4,6-diamino-1,2-dihydro-1,3,5-triazines
alone. Hence, these observations have laid the basis that supports
Compound DHFR IC50
(
m
M) Anti-proliferative activity
MCF-7 cells HCT116 cells
GI50 M) LC50 M) GI50 M)
6.6 1.8
(
m
(
m
(
m
LC50
(mM)
T1
T2
9.36 0.70
0.43 0.03
12.7 2.6 >100
2.5 0.9 >100
>100
0.24 0.11 >100
IC50, GI50 and LC50 values are presented as means
experiments.
SD from 3 independent
Table 2
Rat liver TrxR inhibitory activity and anti-proliferative activity of chalcones C1eC3 on MCF-7 and HCT-116 cells.
Compound
TrxR IC50
(
m
M) (60 min incubation)
Anti-proliferative activity
MCF-7 cells
HCT116 cells
GI50 M)
GI50
(m
M)
LC50
(
m
M)
(
m
LC50 (mM)
C1
C2
C3
10.8 0.3
2.6 0.1
24.5 5.5
12.4 2.0
5.0 0.4
11.2 2.9
30.7 5.1
17.0 2.5
42.7 7.3
15.7 2.2
3.8 0.7
13.5 0.6
22.5 1.5
6.5 0.3
26.6 4.7
IC50, GI50 and LC50 values are presented as means SD from 3 independent experiments.
curve of the combination overlapped with that of the individual
chalcone (C1, C3), suggesting that the antiproliferative effect of the
combinations was predominantly due to the chalcone component.
As for the combination of T1 þ C2, the doseeeffect curve was
observed to overlap with that of C2 across all concentrations
(Figs. 2B and 3B), which indicated that the antiproliferative activity
of the T1 þ C2 combination was predominantly due to C2 that
possessed markedly stronger anti-proliferative profile than that of
T1. Moving on to the dihydrotriazine-chalcone combinations
T2 þ C1, T2 þ C2 and T2 þ C3, at low-to-moderate concentrations
the notion that incorporation of the 4-6-diamino-1,2-dihydro-
1,3,5-triazine and chalcone scaffolds into a single chemical entity
would potentially result in novel compounds possessing stronger
antiproliferative profiles than their parent (dihydrotriazine/chal-
cone) compounds.
2.2. Merged dihydrotriazine-chalcone compounds
2.2.1. Design and synthesis of merged dihydrotriazine-chalcone
compounds
(up to ~10
mM), the doseeeffect curves of the combinations over-
2,2-Dimethyl-1-phenyl-1,2-dihydro-1,3,5-triazine-4,6-diamine
HCl T1 and unsubstituted chalcone C1 were used as starting point
for the design of novel merged dihydrotriazine-chalcone com-
pounds. T1 and C1 were chosen due to their (1) potent inhibitory
activities against DHFR and TrxR respectively, (2) strong
lapped with that of T2, while at higher concentrations (>10
mM),
the doseeeffect curves overlapped with that of the individual
chalcones (C1, C2, C3) (Figs. 2DeF, 3DeF). This had indicated that at
low-to-moderate concentrations, the combinations were no better
than T2 alone while at higher concentrations, the combinations
were no better than the respective chalcones alone. Therefore,
taken together, results obtained had indicated that treatment with
4,6-diamino-1,2-dihydro-1,3,5-triazine-chalcone combinations at a
1:1 M ratio did not produce strong synergy.
Nonetheless, despite the general lack of synergistic
antiproliferative activities of the 4,6-diamino-1,2-dihydro-1,3,5-
triazine-chalcone combination, it was notable that the combina-
tion of T1 and C1/C3 displayed cooperative effects that culminated
antiproliferative activity in the mM range of concentrations, and (3)
cooperative effect observed when used together as combination
treatment. Additionally, T1 and C1 have low molecular weights of
253.7 and 208.3 respectively. Considering that designed multiple
ligands are generally bulkier and more lipophilic than other drugs,
thereby resulting in a poorer pharmacokinetic profile3, it is thus
advantageous that the individual scaffolds are small in size. There
are three ways of joining different pharmacophores to obtain multi-
targeted compounds: by conjugation, fusion and merging of the

66
H.-L. Ng et al. / European Journal of Medicinal Chemistry 115 (2016) 63e74
Fig. 2. Dose-dependent effects of 4,6-diamino-1,2-dihydro-1,3,5-triazines T1/T2, chalcones C1/C2/C3 or their combination on viability of MCF-7 cells. MCF-7 cells were incubated
with (A) T1, C1 and their combination at a 1:1 M ratio, (B) T1, C2 and their combination at a 1:1 M ratio, (C) T1, C3 and their combination at a 1:1 M ratio, (D) T2, C1 and their
combination at a 1:1 M ratio, (E) T2, C2 and their combination at a 1:1 M ratio, and (F) T2, C3 and their combination at a 1:1 M ratio for 72 h and cell viabilities were determined
using MTT assay. Concentration was plotted on a logarithmic scale. Each data point represents the average value from three independent experiments.
chemical frameworks containing the individual pharmacophores.1
Briefly, “conjugated” compounds contain a distinct linker sepa-
rating the frameworks containing the pharmacophores for each
biological target whereas compounds are “fused” when the
frameworks are directly connected. Finally, overlap between the
pharmacophoric elements results in “merged” compounds. A
common structural feature present in both T1 and C1 is the phenyl
ring. As shown in Fig. 4, in T1, the phenyl ring is attached to the N1
atom of the 1,2-dihydro-1,3,5-triazine ring and C1 contains two
phenyl rings that are denoted as ring A and ring B. The phenyl rings
are located in tolerant regions of both scaffolds, meaning that
substitutions can be made on the phenyl rings without destroying
enzyme inhibitory activity and anti-proliferative activity. Hence,
the structures of T1 and C1 were “merged” in our effort to syn-
thesize novel dual-target compounds against DHFR and TrxR. The
chemical structures of T1, C1 and the resulting merged
dihydrotriazine-chalcone compounds are shown in Fig. 4: the
phenyl ring attached to N1 of T1 was merged with either ring A or
ring B of C1 at the meta- or para-position.
The synthetic route for compounds 6a and 6b is outlined in
Scheme 1. Nitro-substituted chalcones (3a-b) (81e86% yield) were
synthesized via the ClaiseneSchmidt condensation reaction be-
tween acetophenone and p-nitrobenzaldehyde (1) or m-nitro-
benzaldehyde (2). The nitro group was then reduced by refluxing
with iron powder and ammonium chloride in an aqueous ethanolic
solution to obtain amino-substituted chalcones (4a-b) (55e63%
yield), which were further reacted with cyanoguanidine in the
presence of concentrated HCl to form compounds 5a-b (72e75%
yield). Finally, ring closure of these compounds with acetone in the
presence of concentrated HCl afforded the target compounds 6a
and 6b (40e55% yield). Ring closure to obtain 6a was achieved at
room temperature, whereas ring closure to obtain 6b required
microwave-assisted synthesis at 90 ꢀC for 1 h. The final products
were purified by recrystallization from ethanol/water.
The synthetic route for compounds 11a and 11b is outlined in
Scheme 2. Attempts to synthesize 3’-nitrochalcone and 4’-

H.-L. Ng et al. / European Journal of Medicinal Chemistry 115 (2016) 63e74
67
Fig. 3. Dose-dependent effects of 4,6-diamino-1,2-dihydro-1,3,5-triazines T1/T2, chalcones C1/C2/C3 or their combination on viability of HCT116 cells. HCT116 cells were incubated
with (A) T1, C1 and their combination at a 1:1 M ratio, (B) T1, C2 and their combination at a 1:1 M ratio, (C) T1, C3 and their combination at a 1:1 M ratio, (D) T2, C1 and their
combination at a 1:1 M ratio, (E) T2, C2 and their combination at a 1:1 M ratio, and (F) T2, C3 and their combination at a 1:1 M ratio for 72 h and cell viabilities were determined
using MTT assay. Concentration was plotted on a logarithmic scale. Each data point represents the average value from three independent experiments.
nitrochalcone by sodium hydroxide-catalyzed condensation resul-
ted in poor yields. Hence, Boc-protection was used instead to obtain
the desired amino-substituted chalcones (9a-b). P-amino-
acetophenone (7) or m-aminoacetophenone (8) was Boc-protected,
reacted with benzaldehyde in the presence of sodium hydroxide as
a base catalyst, and then Boc-deprotected to yield 9a-b (62e90%
yield). These amino-substituted chalcones were reacted with cya-
noguanidine in the presence of concentrated HCl to form com-
pounds 10a-b. When this reaction was carried out under reflux
conditions, no product was formed. Hence, the temperature was
increased to 120 ꢀC using microwave-assisted synthesis. The reac-
tion was found to be incomplete after 15 min and gave 10a-b in low
yields of 8e10%, but longer reaction times resulted in more side-
products formed (TLC). Ring closure of these compounds with
acetone in the presence of concentrated HCl yielded the target
compounds 11a and 11b (53e63%). 11a was purified by recrystal-
lization from ethanol/acetone, while 11b was purified using column
chromatography.
2.2.2. Effects of merged dihydrotriazine-chalcone compounds on
in vitro DHFR and TrxR activities
The DHFR inhibition assay was carried out using a spectropho-
tometric assay against recombinant human DHFR. The rate of
NADPH consumption during the conversion of dihydrofolate to
tetrahydrofolate by DHFR in the presence of varying concentrations
of test compounds was monitored at 340 nm and calculated IC50
values are shown in Table 3. The compounds in which T1 was
merged with ring B of C1 (6a and 6b) failed to inhibit DHFR even at
a high concentration of 100 mM. Conversely, compounds in which
T1 was merged with ring A of C1 (11a and 11b) did not have DHFR
inhibitory activity abrogated. Notably, merging of T1 with ring A of
C1 at the para-position (11a) had led to a 6-fold reduction in DHFR
inhibitory activity as compared to that of the parent dihydrotriazine
T1. On the other hand, merging of T1 with ring A of C1 at the meta-
position (11b) gave rise to enhancement of DHFR inhibitory activity
with a 4-fold increase in potency as compared to that of T1.
The TrxR inhibitory activities of the test compounds were

68
H.-L. Ng et al. / European Journal of Medicinal Chemistry 115 (2016) 63e74
Fig. 4. Structures of T1, C1 and merged dihydrotriazine-chalcone compounds.
Scheme 1. Synthesis of compounds 6a and 6b. Reagents and conditions: a) acetophenone, NaOH, EtOH, H2O, rt; b) Fe/NH4Cl, EtOH, H2O, reflux; c) cyanoguanidine, conc. HCl,
acetonitrile, reflux; d) acetone, conc. HCl, EtOH, rt for 6a/microwave at 90 ꢀC for 1 h for 6b.
Scheme 2. Synthesis of compounds 11a and 11b. Reagents and conditions: a) Boc2O, dioxane, rt; b) benzaldehyde, NaOH, EtOH, H2O, rt; c) TFA, rt; d) cyanoguanidine, conc. HCl,
EtOH, microwave at 120 ꢀC for 15 min; e) acetone, conc. HCl, rt.
determined using the 5.5'-dithiobis-(2-nitrobenzoic acid) (DTNB)
assay following 30 and 60 min incubation of the compounds with

H.-L. Ng et al. / European Journal of Medicinal Chemistry 115 (2016) 63e74
69
Table 3
Phe34, Pro61 and Val115 (Fig. 5B).
IC50 values of T1, C1 and merged dihydrotriazine-chalcone compounds 6a-b and
11a-b against recombinant human DHFR and rat liver TrxR.
Molecular docking was also performed for compound 11b
against the predominant cytosolic isoform of TrxR, namely TrxR1,
through protomol generation by residues residing in the enzyme's
C-terminal active site. The residues involved with binding are
Ser495, Gly496, Cys497, Sec498, Ile478, Glu477, His472 and Phe406.
The crystal structure of rat TrxR1 (PDB: 3EAN) was taken from PDB,
and modified for docking calculations. During protein preparation,
water molecules were removed, H atoms were added and side
chains were fixed. Fig. 6A and B shows respectively the docking of
compound 11b into the C-terminal active site of TrxR1 and the
interaction between compound 11b and the hydrophobic surface
proximal to the active site of TrxR1. The docking score for this
interaction was found to be 4.33; the low docking score was likely
due to the inability of Surflex-dock to recognize covalent in-
teractions in its calculation. Nevertheless, the docking results show
a potential hydrogen bond formation between O of the carbonyl
group of compound 11b and the side chain amide group of Gln494
(O$$$HeN, 1.991 Å). In addition, non-polar amino acid residues
Phe406, Trp407, Gly496 and Ile478 at the C-terminal tail of TrxR
were found to be involved in hydrophobic interactions with com-
pound 11b. In summary, taking together the results obtained from
the molecular docking work that was performed, compound 11b
was shown to possess rational binding modes to both DHFR and
TrxR.
Compound
DHFR IC50
(
m
M)
TrxR IC50 (mM)
30 min Incubation
60 min Incubation
6a
6b
11a
11b
T1
>100
>100
56.4 15.3
2.4 0.5
9.4 0.7
>100
6.2 2.0
7.2 3.5
18.6 5.0
15.7 3.9
>100
5.4 1.1
4.2 1.3
12.6 3.0
10.1 2.3
>100
C1
MTX
46.4 9.9
>100
26.7 4.3
>100
0.0079 0.0002
IC50 values are presented as means SD from 3 independent experiments.
TrxR isolated from rat liver. DTNB is a substrate of TrxR that was
reduced to 2-nitro-5-thiobenzoate (TNB) in the presence of NADPH.
Rate of production of TNB, reflective of TrxR activity, was monitored
at 412 nm and calculated IC50 values are shown in Table 3. All of the
dihydrotriazine-chalcone hybrids exhibited more potent TrxR
inhibitory activity as compared to those of the parent chalcone C1.
Merging of T1 to ring B of C1 (6a-b) resulted in a greater
improvement in TrxR inhibitory activity as compared to merging of
T1 to ring A of C1 (11a-b) by 2e3 times.
A possible explanation for the improvement in TrxR inhibitory
activity of the merged dihydrotriazine-chalcone compounds over
C1 was the increased reactivity of the
a,b-unsaturated carbonyl
moiety, which had been postulated as the reactive pharmacophoric
motif in chalcones to react with the nucleophilic Sec residue within
the C-terminal active site of the TrxR enzyme. At physiological pH,
the N-3 atom of the 4,6-diamino-1,2-dihydro-1,3,5-triazine ring is
protonated. The resulting positive charge can be delocalized over
the N-3 atom, 4-amino group, 6-amino group and N-1 atom of the
ring. Since the N-1 atom of the dihydrotriazine ring is directly
attached to either ring A or ring B of the chalcone, the positively-
charged dihydrotriazine ring may exert an electron-withdrawing
2.2.4. Antiproliferative activities of merged dihydrotriazine-
chalcone compounds
Based on the obtained GI50 and LC50 values presented in Table 4,
one notable observation was that all of the merged dihydrotriazine-
chalcone compounds only exhibited growth inhibitory activities
but cytotoxic effects were generally lacking (LC50 values greater
than 100 mM). Merging of T1 with C1 at the para-position of ring A
(11a) or ring B (6a) apparently produced good growth inhibitory
potencies. Conversely, merging of T1 with C1 at the meta-position
of ring A (11b) or ring B (6b) was detrimental; the resulting merged
compounds possessed poor growth inhibitory activities against
MCF-7 and HCT116 cells (Table 4). A second notable observation
was the poor correlation between the anti-proliferative activities
and in vitro DHFR and TrxR inhibitory activities of the merged
compounds; the order of growth inhibitory potencies displayed by
the merged compounds did not correspond to a similar order of
inhibitory potencies against DHFR and TrxR. Taking together the
findings made of the anti-proliferative properties, as well as DHFR
and TrxR inhibitory properties of the merged compounds, the
following inferences were made: firstly, the merged
dihydrotriazine-chalcone compounds, particularly 11a and 11b,
brought about both DHFR and TrxR inhibition in an in vitro setting,
which not only demonstrated the feasibility of synthetically
combining a DHFR-targeting triazine and TrxR-targeting chalcone
pharmacophore into a single novel molecule, but also provided
experimental evidence that the resulting compounds indeed
possessed inhibitory character against both molecular targets of
interest. Secondly, the weak cell-killing efficacies among the
merged dihydrotriazine-chalcone compounds, as well as a poor
correlation between their anti-proliferative activities and in vitro
DHFR and TrxR inhibitory activities, had indicated that the current
“merged compound” design approach that involved structural
combination of the dihydrotriazine and chalcone scaffold through
an overlap of the common phenyl ring attached to N1 atom of the
dihydrotriazine ring could be improved on so as to produce novel
molecules that possessed potent anti-proliferative profiles corre-
lated to strong DHFR and TrxR inhibitory activities.
effect on the chalcone. This intensifies the
d
þ charge on the
b-
carbon atom of the -unsaturated carbonyl moiety, thereby
a,b
increasing its reactivity. This effect is stronger when the dihydro-
triazine ring is attached to ring B than when attached to ring A,
which might explain the greater TrxR inhibitory activities of com-
pounds 6a and 6b as compared to those of 11a and 11b.
2.2.3. Molecular docking of compound 11b with DHFR and TrxR
In order to investigate the binding mode of the dihydrotriazine-
chalcone linked compounds to DHFR and TrxR, compound 11b was
selected for molecular docking experiments as it showed good
in vitro inhibitory activity against both DHFR and TrxR. The method
protomol generation by ligand was selected for the molecular
docking studies on compound 11b against DHFR using Surflex-
dock. Human DHFR complexed with MTX (PDB: 1U72) was ob-
tained from protein data bank (PDB) and modified for the docking
experiment. During protein preparation, the co-crystalized ligand
was first removed from the structure together with the water
molecules. H atoms were then added and side chains were fixed.
Fig. 5A depicts the docking of compound 11b into the active site of
human DHFR. As expected, compound 11b binds to the active site
with a similar binding mode compared to MTX [27]. In this binding
mode,11b demonstrated good binding to DHFR with a high docking
score of 7.17 (-log Kd). Several hydrogen bonds were observed be-
tween compound 11b and the crucial amino acid residues Val115
(NeH$$$O: 2.037 Å), Ile7 (NeH$$$O, 2.167 Å), Glu30 (NeH$$$OE2,
2.256 Å), and Ser59 (O$$$HeO, 2.090 Å). Compound 11b was also
revealed to be involved in strong hydrophobic interactions at the
active site with the hydrophobic residues such as Ile7, Val8, Phe31,

70
H.-L. Ng et al. / European Journal of Medicinal Chemistry 115 (2016) 63e74
Fig. 5. Molecular docking of compound 11 b at the active site of human DHFR using Surflex-Dock. (A) Selected docking pose of compound 11 b at the active site of human DHFR that
gave a high docking score. (B) Visualization of compound 11b interacting with the hydrophobic residues at the active site of DHFR.
3. Conclusion
synergy only at relatively low concentrations. In the approach
involving structural combination of the 4,6-diamino-1,2-dihydro-
1,3,5-triazine and chalcone scaffold into a single molecule (through
an overlap of the common phenyl ring attached to N1 atom of the
dihydrotriazine ring), four “merged” dihydrotriazine-chalcone
compounds were designed and synthesized. Two of the com-
pounds, 11a and 11b, were found to possess both in vitro DHFR and
TrxR inhibitory activities, which indicated the feasibility of
To exploit the possibility of harnessing the combined benefits of
simultaneous inhibition of molecular targets DHFR and TrxR so as
to achieve enhanced anti-tumor outcomes, two approaches had
been examined in this study. The approach involving the combined
treatment of MCF-7 and HCT116 cells with 4,6-diamino-1,2-
dihydro-1,3,5-triazines and chalcones at a 1:1 M ratio resulted in

H.-L. Ng et al. / European Journal of Medicinal Chemistry 115 (2016) 63e74
71
Fig. 6. Molecular docking of compound 11 b at the C-terminal active site of rat TrxR1 using Surflex-Dock. (A) Selected docking pose of compound 11 b at the C-terminal active site of
rat TrxR1. (B) Visualization of compound 11b interacting with the hydrophobic surface of the C-terminal active site of rat TrxR1.
producing a single chemical entity that could result in simulta-
neous inhibition of DHFR and TrxR. However, the synthesized
merged compounds only exhibited growth inhibitory activities like
their parent dihydrotriazine component, and lacked cytotoxic ef-
ficacies like their parent chalcone component. In addition, their
anti-proliferative activities did not correlate with their DHFR and
TrxR enzyme inhibitory activities. Taken together, the findings of
this proof-of-concept study, while indicative of the synthetic
feasibility of producing a single chemical entity that could result in
simultaneous inhibition of DHFR and TrxR, had exposed the sub-
optimal structural design of combination of the two pharmaco-
phores in a single chemical entity. Other synthetic ways of deriving
compounds using the “designed multiple ligand” approach, such as
conjugation or fusion of the chemical frameworks containing the

72
H.-L. Ng et al. / European Journal of Medicinal Chemistry 115 (2016) 63e74
273 ꢀC (decomposed). 1H NMR (400 MHz, DMSO-d6)
d 9.86 (s, 1H,
Table 4
Antiproliferative activity of T1, C1 and merged dihydrotriazine-chalcone compounds
6a-b and 11a-b against MCF-7 and HCT116 cancer cell lines.
NH), 8.88 (s, 1H, NH), 8.72 (s, 1H, NH), 8.15e8.12 (m, 2H), 7.91e7.69
(m, 5H), 7.77 (br s, 1H, NH), 7.63e7.59 (m, 2H), 7.55e7.53 (m, 1H),
7.47 (t, 1H, J ¼ 7.8 Hz), 7.18 (br s, 1H, NH), 1.49 (s, 6H). 13C NMR
Compound
MCF-7 cells
GI50 M)
HCT116 cells
GI50 M)
(100 MHz, DMSO-d6) d 189.2, 157.5, 155.2, 143.6, 137.9, 137.4, 135.3,
(
m
LC50
(
mM)
(
m
LC50 (mM)
133.1, 129.3, 128.7, 128.4, 124.14, 124.09, 122.8, 122.6, 64.8, 28.9.
6a
6b
11a
11b
T1
2.0 0.6
51.5 12.9
3.0 0.6
34.4 2.0
12.7 2.6
12.4 2.0
0.024 0.007
>100
>100
>100
>100
>100
1.4 0.2
>100
>100
>100
>100
>100
>100
Purity: 99.6%, HPLC tR ¼ 3.55 min. ESI-MS m/z 348.1 (Mþ1)þ.
10.7 2.7
74.4 4.5
6.6 1.8
15.7 2.2
0.015 0.001
4.4. (2E)-1-[4-(4,6-diamino-2,2,dimethyl-1,3,5-triazin-1(2H)-yl)
phenyl]-3-phenylprop-2-en-1-one hydrochloride (11a)
C1
MTX
30.7 5.1
>100
22.5 1.5
>100
To a solution of 10a (0.15 g, 0.42 mmol) in 1 ml acetone and 1 ml
ethanol, 0.1 ml of concentrated HCl solution was added and the
reaction mixture was stirred at room temperature for 3 days. The
solvent was removed under reduced pressure and the residual solid
was purified by recrystallization from ethanol/acetone to give 11a
as a yellow solid, 0.086 g, yield 53.2%. mp: 175e177 ꢀC. 1H NMR
GI50 and LC50 values are presented as means SD from 3 independent experiments.
individual DHFR- and TrxR-targeting pharmacophores will need to
be explored.
(400 MHz, DMSO-d6)
d
9.22 (s, 1H, NH), 8.30 (d, 2H, J ¼ 8.4 Hz),
4. Experimental
8.02e7.78 (m, 4H), 7.60 (d, 2H, J ¼ 8.4 Hz), 7.55 (br s, 3H, NH),
7.49e7.48 (m, 3H), 6.53 (br s, 1H), 1.40 (s, 6H). 13C NMR (100 MHz,
4.1. General conditions for organic synthesis
DMSO-d6)
d 188.6, 157.5, 156.8, 144.7, 138.8, 138.1, 134.4, 130.8,
130.5, 130.3, 128.94, 128.87, 121.8, 69.8, 27.2. Purity: 99.3%, HPLC
All reagents were purchased from SigmaeAldrich, Tokyo
Chemical Industry, Alfa Aesar, or Merck and were used directly
without any purification. Microwave reactions were carried out
using a CEM Discover® SP System. Thin layer chromatography (TLC)
was performed using silica gel-coated aluminum plates with fluo-
rescent indicator and visualized with ultraviolet light at 254 nm.
Melting points were determined using a Gallenkamp melting point
apparatus and were uncorrected. 1H NMR and 13C NMR spectra
were recorded on a Bruker Ultrashield™ 400 Plus NMR spec-
tR ¼ 4.39 min. ESI-MS m/z 348.3 (Mþ1)þ.
4.5. (2E)-1-[3-(4,6-diamino-2,2,dimethyl-1,3,5-triazin-1(2H)-yl)
phenyl]-3-phenylprop-2-en-1-one hydrochloride (11b)
To a solution of 10b (0.24 g, 0.69 mmol) in 1 ml acetone and 1 ml
ethanol, 0.1 ml of concentrated HCl solution was added and the
reaction mixture was stirred at room temperature for 9 days. The
solvent was removed under reduced pressure and the residual
brown sticky substance was purified by column chromatography to
give 11b as a light brown solid, 0.168 g, yield 63.3%. mp: 174e176 ꢀC.
trometer. Chemical shift values (d) were expressed as parts per
million (ppm) relative to tetramethylsilane as the internal standard.
Electrospray ionization mass spectrometry (ESI-MS) was recorded
on an AB SCIEX API 2000 Q Trap mass spectrometer. High perfor-
mance liquid chromatography (HPLC) was performed for purity
checking using Agilent 1100 Series HPLC on a Lichrosorb 10 RP-18
1H NMR (400 MHz, DMSO-d6)
d 9.36 (s, 1H, NH), 8.29 (d, 1H,
J ¼ 7.2 Hz), 8.17 (s, 1H), 8.02e7.92 (m, 3H), 7.83e7.70 (m, 3H),
7.50e7.48 (m, 3H), 7.43 (br s, 3H, NH), 6.51 (br s, 1H, NH), 1.40 (s,
6H). 13C NMR (100 MHz, DMSO-d6)
d 188.3, 157.7, 157.1, 144.8, 139.2,
250 ꢁ 4.0 mm 10
mm column.
135.3, 134.6, 134.5, 130.8, 130.6, 130.1, 129.6, 129.0, 128.8, 121.7, 69.7,
27.3, 27.2. Purity: 99.7%, HPLC tR ¼ 4.30 min. ESI-MS m/z 348.2
(Mþ1)þ.
4.2. (2E)-3-[4-(4,6-diamino-2,2,dimethyl-1,3,5-triazin-1(2H)-yl)
phenyl]-1-phenylprop-2-en-1-one hydrochloride (6a)
To a solution of 5a (1.15 g, 3.34 mmol) in 10 ml acetone and 10 ml
ethanol, 0.45 ml of concentrated HCl solution was added and the
reaction mixture was stirred at room temperature for 3 days. The
solvent was removed under reduced pressure and the residual solid
was purified by recrystallization from ethanol/water to give 6a as
orange crystals, 0.706 g, yield 55.1%. mp: 208e210 ꢀC. 1H NMR
4.6. In vitro DHFR inhibiton assay
To 1 ml of solution containing 50 mM dihydrofolate (DHF), 60 mM
NADPH, 0.15 M phosphate buffer (pH 7), 1.5 ꢁ 10ꢂ3 units of re-
combinant human DHFR (SigmaeAldrich) in a 1-ml quartz cuvette,
test compounds dissolved in DMSO vechicle were added to achieve
(400 MHz, DMSO-d6)
d 8.76 (s, 1H, NH), 8.16e8.13 (m, 2H), 8.03 (d,
final concentrations over the range of 0.01e100 mM. Upon addition
2H, J ¼ 8.4 Hz), 7.97 (d, 1H, J ¼ 15.6 Hz), 7.80 (d, 1H, J ¼ 15.6 Hz), 7.72
(tt,1H, J ¼ 7.4, 1.5 Hz), 7.62 (t, 2H, J ¼ 7.6 Hz), 7.53 (br s, 3H, NH), 7.46
(d, 2H, J ¼ 8.4 Hz), 6.47 (br s, 1H, NH), 1.40 (s, 6H). 13C NMR
of the compounds, the rate of consumption of NADPH during the
conversion of DHF to tetrahydrofolate (THF) was monitored.
Absorbance readings were taken at 340 nm every 30 s over 6 min
using a Hitachi U-1900 UV/visible spectrophotometer. The assay
was carried out at room temperature. A graph of absorbance
readings versus time was plotted and the linear slope of the graph
represents the rate of reaction. The percentage inhibition at each
concentration of inhibitor was calculated using the following
formula:
(100 MHz, DMSO-d6)
d 189.1, 157.7, 157.0, 142.6, 137.3, 136.5, 135.7,
133.2, 130.4, 128.8, 128.5, 123.4, 69.7, 27.2. Purity: 99.2%, HPLC
tR ¼ 3.24 min. ESI-MS m/z 348.3 (Mþ1)þ.
4.3. (2E)-3-[3-(4,6-diamino-2,2,dimethyl-1,3,5-triazin-1(2H)-yl)
phenyl]-1-phenylprop-2-en-1-one hydrochloride (6b)
Slopecompound ꢂ Slopeblank
To a solution of 5b (0.44 g, 1.28 mmol) in 3.5 ml acetone and
1.5 ml ethanol, 0.2 ml of concentrated HCl solution was added and
the reaction mixture was heated using microwave-assisted syn-
thesis at 90 ꢀC for 1 h. The solvent was removed under reduced
pressure and the residual solid was purified by recrystallization
from ethanol to give 6b as brown crystals, 0.199 g, yield 40.4%. mp:
Activity ð%Þ ¼
ꢁ 100
Slopecontrol ꢂ Slopeblank
Inhibition ð%Þ ¼ 100 ꢂ Activity
where Slopecompound is the slope of the graph for solutions

H.-L. Ng et al. / European Journal of Medicinal Chemistry 115 (2016) 63e74
73
containing DHFR, test compound, NADPH and DHF; Slopecontrol is
4.9. Cell viability assay
the slope of the graph for solutions containing DHFR, NADPH and
DHF; while Slopeblank is the slope of the graph for solutions con-
taining NADPH and DHF only. Graphs of percentage inhibition
against logarithmic concentration were plotted for each compound
and IC50 values were calculated using GraphPad Prism Version 5.01.
Breast adenocarcinoma MCF-7 and colon carcinoma
HCT116 cells were maintained in RPMI 1640 medium supple-
mented with 10% fetal bovine serum, 100 units/ml penicillin, and
100 m
g/ml streptomycin and incubated at 37 ꢀC in a humidified
atmosphere of 95% air and 5% CO2. Cells were seeded into 96-well
plates at a density of 1.0 ꢁ 104/well and allowed to adhere over-
night. Serial dilutions of the test compounds were freshly prepared
in medium from DMSO drug stocks for each assay. At the time of
addition of test compounds (T0) and after 72 h of incubation with
test compounds, cell viability was determined by MTT assay. MTT
4.7. In vitro TrxR inhibition assay (DTNB reduction assay)
In 96-well plates, test compounds of 1e100
with 2.1 units/ml of rat liver TrxR and 200 M of NADPH in a vol-
ume of 100 l of 50 mM TriseHCl and 1 mM EDTA, pH 7.5 (TE
buffer) at room temperature for 30 min or 60 min. A volume of
100 l of TE buffer containing 5 mM of DTNB and 200 M NADPH
was added (final concentrations: 2.5 mM DTNB, 200 M NADPH)
mM were incubated
m
was added into each well to achieve a final concentration of 400 mg/
m
ml and the plates were incubated at 37 ꢀC for 4 h. The supernatant
was aspirated and the insoluble formazan product was solubilized
m
m
in 150 ml of DMSO:glycine buffer, pH 10.5 (ratio 4:1). Absorbance at
550 nm was measured using the VersaMax Microplate Reader.
Cell viability was determined from the expression
m
and the increase in absorbance at 412 nm was measured over the
initial 2 min with a VersaMax Microplate Reader (Molecular De-
vices). Maximal velocity (Vmax) values were obtained from the
linear portion of the absorbance-over-time curve. TrxR activity was
calculated as a percentage of enzyme activity as compared to that of
vehicle-treated sample using the formula:
Absorbance of treated cells
Cell viabilityð%Þ ¼
ꢁ 100
Absorbance of control
To obtain GI50 and LC50 values, the absorbance readings were
plotted against logarithmic concentration on GraphPad Prism
Version 5.01. GI50 and LC50 values were interpolated from the
graphs as concentrations corresponding to the following absor-
bance values:
Vmax compound
Vmax control
Activity ð%Þ ¼
ꢁ 100
Graphs of percentage activity against logarithmic concentration
were plotted for each compound and IC50 values were calculated
using GraphPad Prism Version 5.01.
Absorbance at T0 þ Absorbance of control
Absorbance at GI50
Absorbance at LC50
¼
¼
2
Absorbance at T0
2
4.8. Molecular docking studies
Molecular docking of compound 11b was conducted using the
Surflex-Dock module [28,29] in SYBYLeX 1.1 (Tripos associate Inc.,
St. Louis, MO, USA) to evaluate the possible binding poses of the
compound at the binding sites of human DHFR and rat TrxR1. The
3D structure of 11b was minimized in SYBYL-X 1.1. The partial
atomic charge was calculated by the Gasteiger Huckel method and
energy minimization was performed using the Tripos force field
with a distance-dependent dielectric and the Powell conjugate
gradient algorithm convergence criteria. The 3D human DHFR
structure PDB ID 1U72 (resolution of 1.90 Å) and rat TrxR1 structure
PDB ID 3EAN (resolution of 2.75 Å) were obtained from the Protein
Data Bank. Water molecules co-crystallized with protein were
removed from the original 3D structure, hydrogen atoms were
added and side chains were fixed during protein preparation.
Surflex-Dock module docks ligands into the active site of the
enzyme automatically using a protomol-based method and an
empirically scoring function. The protomol in Surflex-Dock utilizes
various molecular fragments, which were tessellated in the active
site and optimized based on the scoring function. Docking calcu-
lation for DHFR 1U72 was performed through protomol generation
by ligand, while docking calculation referring to TrxR1 3EAN was
performed through protomol generation based on residues Phe405,
Phe406, Trp407, His472, Val474, Gly496, Cys497, Sec498 and
Gly499 in the C-terminal active site. Protomol generation param-
eters were threshold 0.5 and boat 0 for both enzymes. The docking
score of Surflex-Dock is expressed as elog Kd to represent binding
affinities which consist of hydrophobic, polar, repulsive, entropic
and solvation terms. During the automated docking process, the
ligand was docked into the active site in different ways, creating
different poses each with a docking score. The top 20 poses of each
ligand were obtained and analyzed for their interactions with the
protein.
Acknowledgments
The authors acknowledge the National University of Singapore
(postgraduate scholarship to H.L. Ng, Academic Research Fund Tier
1 grant R-148-000-153-112 to W.K. Chui and R-148-184-112 to E.H.
Chew) and National Medical Council (NMRC/NIG/0050/2009 to E.H.
Chew) for financial support.
References
[1] R. Morphy, Z. Rankovic, Designed multiple ligands. An emerging drug dis-
covery paradigm, J. Med. Chem. 48 (2005) 6523e6543.
[2] L.M. Espinoza-Fonseca, The benefits of the multi-target approach in drug
design and discovery, Bioorg Med. Chem. 14 (2006) 896e897.
[3] N.M. O'Boyle, M.J. Meegan, Designed multiple ligands for cancer therapy, Curr.
Med. Chem. 18 (2011) 4722e4737.
[4] E.M. Berman, L.M. Werbel, The renewed potential for folate antagonists in
contemporary cancer chemotherapy, J. Med. Chem. 34 (1991) 479e485.
[5] I.M. Kompis, K. Islam, R.L. Then, DNA and RNA synthesis: antifolates, Chem.
Rev. 105 (2005) 593e620.
[6] M. Visentin, R. Zhao, I.D. Goldman, The antifolates, Hematol. Oncol. Clin. North
Am. 26 (2012) 629e648 ix.
[7] E.J. Modest, G.E. Foley, M.M. Pechet, S. Farber, A Series of new, biologically
significant dihydrotriazines, J. Am. Chem. Soc. 74 (1952) 855e856.
[8] H.K. Lee, T.M. Rana, Microwave-assisted parallel synthesis of a 4,6-diamino-
2,2-dimethyl-1,2-dihydro-1-phenyl-s-triazine library, J. Comb. Chem. 6 (2004)
504e508.
[9] X. Ma, R.S. Woon, P.C. Ho, W.K. Chui, Antiproliferative activity against MCF-7
breast cancer cells by diamino-triazaspirodiene antifolates, Chem. Biol. Drug
Des. 74 (2009) 322e326.
[10] X. Ma, W.K. Chui, Antifolate and antiproliferative activity of 6,8,10-triazaspiro
[4.5]deca-6,8-dienes and 1,3,5-triazaspiro[5.5]undeca-1,3-dienes, Bioorg Med.
Chem. 18 (2010) 737e743.
[11] S. Gromer, S. Urig, K. Becker, The thioredoxin system e from science to clinic,
Med. Res. Rev. 24 (2004) 40e89.
[12] E.S. Arner, A. Holmgren, Physiological functions of thioredoxin and

74
H.-L. Ng et al. / European Journal of Medicinal Chemistry 115 (2016) 63e74
thioredoxin reductase, Eur. J. Biochem. 267 (2000) 6102e6109.
Antioxid. Redox Signal 19 (2013) 1149e1165.
[13] K. Kahlos, Y. Soini, M. Saily, P. Koistinen, S. Kakko, P. Paakko, A. Holmgren,
V.L. Kinnula, Up-regulation of thioredoxin and thioredoxin reductase in hu-
man malignant pleural mesothelioma, Int. J. Cancer 95 (2001) 198e204.
[14] Y. Soini, K. Kahlos, U. Napankangas, R. Kaarteenaho-Wiik, M. Saily,
P. Koistinen, P. Paaakko, A. Holmgren, V.L. Kinnula, Widespread expression of
thioredoxin and thioredoxin reductase in non-small cell lung carcinoma, Clin.
Cancer Res. 7 (2001) 1750e1757.
[21] B. Zhang, D. Duan, C. Ge, J. Yao, Y. Liu, X. Li, J. Fang, Synthesis of xanthohumol
analogues and discovery of potent thioredoxin reductase inhibitor as poten-
tial anticancer agent, J. Med. Chem. 58 (2015) 1795e1805.
[22] C.I. Bliss, The toxicity of poisons applied jointly, Ann. Appl. Biol. 26 (1939)
585e615.
[23] S. Loewe, The problem of synergism and antagonism of combined drugs,
Arzneimittelforschung 3 (1953) 285e290.
[15] S.J. Kim, Y. Miyoshi, T. Taguchi, Y. Tamaki, H. Nakamura, J. Yodoi, K. Kato,
S. Noguchi, High thioredoxin expression is associated with resistance to
docetaxel in primary breast cancer, Clin. Cancer Res. 11 (2005) 8425e8430.
[16] J. Raffel, A.K. Bhattacharyya, A. Gallegos, H. Cui, J.G. Einspahr, D.S. Alberts,
G. Powis, Increased expression of thioredoxin-1 in human colorectal cancer is
associated with decreased patient survival, J. Lab. Clin. Med. 142 (2003)
46e51.
[17] W. Cai, L. Zhang, Y. Song, B. Wang, B. Zhang, X. Cui, G. Hu, Y. Liu, J. Wu, J. Fang,
Small molecule inhibitors of mammalian thioredoxin reductase, Free Radic.
Biol. Med. 52 (2012) 257e265.
[18] J. Lu, A. Holmgren, Thioredoxin system in cell death progression, Antioxid.
Redox Signal 17 (2012) 1738e1747.
[19] S. Urig, K. Becker, On the potential of thioredoxin reductase inhibitors for
cancer therapy, Seminars Cancer Biol. 16 (2006) 452e465.
[20] F.F. Gan, K.K. Kaminska, H. Yang, C.Y. Liew, P.C. Leow, C.L. So, L.N. Tu, A. Roy,
C.W. Yap, T.S. Kang, W.K. Chui, E.H. Chew, Identification of Michael acceptor-
centric pharmacophores with substituents that yield strong thioredoxin
reductase inhibitory character correlated to antiproliferative activity,
[24] T.C. Chou, P. Talalay, Quantitative analysis of dose-effect relationships: the
combined effects of multiple drugs or enzyme inhibitors, Adv. Enzyme Regul.
22 (1984) 27e55.
[25] Y. Grabovsky, R.J. Tallarida, Isobolographic analysis for combinations of a full
and partial agonist: curved isoboles, J. Pharmacol. Exp. Ther. 310 (2004)
981e986.
[26] N. Geary, Understanding synergy, Am. J. Physiol. Endocrinol. Metab. 304
(2013) E237eE253.
[27] V. Cody, J.R. Luft, W. Pangborn, Understanding the role of Leu22 variants in
methotrexate resistance: comparison of wild-type and Leu22Arg variant
mouse and human dihydrofolate reductase ternary crystal complexes with
methotrexate and NADPH, Acta Crystallogr. D. Biol. Crystallogr. 61 (2005)
147e155.
[28] A.N. Jain, Surflex: fully automatic flexible molecular docking using a molecular
similarity-based search engine, J. Med. Chem. 46 (2003) 499e511.
[29] R. Spitzer, A.N. Jain, Surflex-Dock: docking benchmarks and real-world
application, J. Comput. Aided Mol. Des. 26 (2012) 687e699.