Structure-based design of potent human dihydroorotate dehydrogenase inhibitors as anticancer agents

Article abstract of DOI:10.1039/c6md00179c

It has been proven that inhibiting human dihydroorotate dehydrogenase (hDHODH) restricts the growth of rapidly proliferating cells, thus hDHODH can be developed as a promising target for the treatment of immunological disease and cancer. Here, a succession of substituted hydrazino-thiazole derivatives were designed, synthesized, and biologically evaluated through structure-based optimization, of which compound 22 was the most potent inhibitor of hDHODH with an IC₅₀ value of 1.8 nM. Furthermore, 22 exhibited much better antiproliferative activity than brequinar, both in HCT-116 and BxPC-3 cancer cell lines. Flow cytometry analysis revealed that 22 induced S phase cell cycle arrest and promoted induction of apoptosis. All results established a proof that blocking the pyrimidine de novo synthesis pathway by inhibiting the rate-limiting enzyme hDHODH is an attractive therapy for cancer.

Full text of DOI:10.1039/c6md00179c

MedChemComm
View Article Online
View Journal
RESEARCH ARTICLE
Structure-based design of potent human
dihydroorotate dehydrogenase inhibitors as
Cite this: DOI: 10.1039/c6md00179c
anticancer agents†‡
Wenlin Song,§ Shiliang Li,§ Yi Tong,§ Jiawei Wang, Lina Quan, Zhuo Chen,
*
*
*
Zhenjiang Zhao, Yufang Xu, Lili Zhu, Xuhong Qian and Honglin Li
It has been proven that inhibiting human dihydroorotate dehydrogenase (hDHODH) restricts the growth of
rapidly proliferating cells, thus hDHODH can be developed as a promising target for the treatment of im-
munological disease and cancer. Here, a succession of substituted hydrazino-thiazole derivatives were
designed, synthesized, and biologically evaluated through structure-based optimization, of which com-
pound 22 was the most potent inhibitor of hDHODH with an IC50 value of 1.8 nM. Furthermore, 22
exhibited much better antiproliferative activity than brequinar, both in HCT-116 and BxPC-3 cancer cell
lines. Flow cytometry analysis revealed that 22 induced S phase cell cycle arrest and promoted induction of
apoptosis. All results established a proof that blocking the pyrimidine de novo synthesis pathway by
inhibiting the rate-limiting enzyme hDHODH is an attractive therapy for cancer.
Received 29th March 2016,
Accepted 7th June 2016
DOI: 10.1039/c6md00179c
www.rsc.org/medchemcomm
novo biosynthesis.5–7 It catalyzes the oxidation of dihydro-
orotate (DHO) to orotate (ORO) utilizing ubiquinone as the
Introduction
Pyrimidine bases play essential roles in cellular metabolism
and multiplication, and are considered as crucial precursors
in the biosynthesis of DNA (thymine and cytosine), RNA (ura-
cil and cytosine), phospholipids and glycoproteins.1 Moreover,
pyrimidines are linked by phosphodiester bridges to purine
nucleotides in double-stranded DNA in both the nucleus and
mitochondria.2 Generally, humans obtain pyrimidine nucleo-
tides through two routes, salvage pathway or de novo biosyn-
thesis pathway.3 It's energy-saving for resting and fully differ-
entiated cells to recycle existing pyrimidines (salvage pathway)
to maintain their survival, however, rapidly proliferating cells,
such as activated lymphocytes or cancer cells, depend heavily
on the de novo biosynthesis to meet their tremendous demand
for nucleic acids and other cellular components.1,4
terminal oxidant, which subsequently induces formation of
uridine and cytidine nucleosides.8 Taking into account the
significance of pyrimidine for cellular proliferation and me-
tabolism, blocking the pyrimidine de novo synthesis pathway
by inhibiting hDHODH would be a promising treatment for
cancer and immunological diseases, such as melanoma, mul-
tiple sclerosis and rheumatoid arthritis.9
Human dihydroorotate dehydrogenase (hDHODH), a flavin
mononucleotide (FMN) dependent and mitochondrial
membrane-associated enzyme, belongs to the family of class
2 enzymes and catalyzes the fourth step of pyrimidine de
State Key Laboratory of Bioreactor Engineering, Shanghai Key Laboratory of New
Drug Design, School of Pharmacy, East China University of Science and
Technology, Shanghai 200237, China. E-mail: zhulfl@ecust.edu.cn,
xhqian@ecust.edu.cn, hlli@ecust.edu.cn; Fax: +86 21 64250213;
Tel: +86 21 64250213
† The authors declare no competing interest.
Fig. 1 Structures of representative hDHODH inhibitors of (1) leflunomide,
(2) teriflunomide (A771726), (3) brequinar, (4) 4SC-101, and (5) scaffold
of the (E)-4-phenyl-2-(((2-carboxyl)benzylidene)hydrazinyl)thiazole
derivatives in this study.
‡ Electronic supplementary information (ESI) available. See DOI: 10.1039/
c6md00179c
§ Authors contributed equally to this work.
This journal is © The Royal Society of Chemistry 2016
Med. Chem. Commun.

View Article Online
Research Article
MedChemComm
To date, most inhibitors of hDHODH are designed to oc-
cupy the ubiquinone binding tunnel to prevent the oxidation
of DHO through blocking the electron-transfer from FMNH2
to ubiquinone.9–14 Leflunomide, a well-known inhibitor of
hDHODH, was approved by the FDA for the treatment of
rheumatoid arthritis more than 10 years ago.15 It is an
isoxazole-based prodrug that is converted to active metabolite
teriflunomide (A771726) after administration (Fig. 1).15 Re-
cently, Hail Jr and co-workers demonstrated that
ortho-chlorine substituted phenyl ring at subsite S2 were fa-
vorable substituents for the hDHODH inhibition. Moreover,
the antiproliferative activity of compound 22 also exhibited a
drastic improvement against both HCT-116 and BxPC-3 cell
lines in comparison with compound 5. Cell cycle arrest and
apoptosis assays were performed subsequently and the re-
sults revealed that compound 22 induced S phase cell cycle
arrest and promoted apoptosis significantly.
teriflunomide promoted apoptosis in transformed human Results and discussion
keratinocytes
selectively,
because
the
transformed
Chemistry
keratinocytes need the pyrimidine de novo synthesis pathway
more than normal human epidermal keratinocytes.16 Zhu
et al. also reported that leflunomide induced apoptosis in
neuroblastoma cells and abolished tumor growth in xeno-
graft mouse because of the down-regulation of DHODH by
leflunomide.17 Moreover, a suppression of DHODH caused
an impaired mitochondrial respiration, which could be a
main mechanism for many cytotoxic agents, for example the
tumor suppressor p53 pathway could be triggered by the sup-
pression of DHODH.18,19 Brequinar (Fig. 1) is also a represen-
tative inhibitor of hDHODH, which was initially studied for
its anticancer profiles against a spectrum of human solid tu-
mors, but failed in phase II clinical trials.20,21 This com-
pound was also used as a potent immunosuppressant and
widely applied in the prevention of transplant rejection.22 In
addition, its use as an antiviral agent has also been investi-
gated, including in combination with ribavirin, and the
mechanism of action revealed that brequinar depleted the
intracellular pyrimidine pool and blocked RNA synthesis.23,24
4SC-101 (Fig. 1) is another hDHODH inhibitor as an immu-
nosuppressive agent in phase III clinical trials, which was
used in the treatment of systematic lupus and inflammatory
bowel disease.25,26 Although the specific mechanisms are not
really similar, it is indisputable that inhibiting hDHODH has
always been a validated therapy for cancer and immunologi-
cal diseases. Considering that some hDHODH inhibitors have
severe side effects and only a few have entered into clinical
study, developing new chemotherapeutic agents targeting
hDHODH remains a worthwhile goal for antiproliferation
and anticancer therapy.27–29
The strategies for the synthesis of compounds adopted in
this research are shown in Schemes 1 and 2. Reaction of
2-carboxybenzaldehyde with substituted hydrazino-thiazoles,
which were prepared via the hydrolysis of corresponding
acetone-protected hydrazine derivatives, gave compounds 5–9
or 11–17 in excellent yields. The intermediates of 5c–17c were
obtained through the cyclization of acetone-protected thio-
semicarbazide with substituted 2-bromoacetophenone or
2-bromopropiophenone or 2-bromobutyrophenone. To gener-
ate the intermediates 11e–17e, which were precursors of
11d–17d, appropriate alkyl halides were added dropwise into
the solution of acetone-protected hydrazino-thiazole deriva-
tives in DMF under alkaline conditions in an ice bath, then
the mixture was heated to about 80 °C overnight.
To get a series of 2-formyl-5-substituted benzoic acids, cor-
responding 2-bromo benzoic acids were treated with n-butyl
lithium and N,N-dimethylformamide at −78 °C. For com-
pounds 10 and 18–23, it was necessary to synthesize the
relevant 2-bromo-1-phenylalkyl-1-ones and substituted thio-
semicarbazones. Parts of compounds 5b–23b were commer-
cially available, while others were achieved from the reaction
of corresponding 1-phenylalkylones with bromine in an ice
bath. As for 2-methylthiosemicarbazide, it was primarily pre-
pared from methylhydrazine sulfate and ammonium thiocya-
nate in a rearrangement manner in ethanol. The structures
1
of all products were verified by H NMR, 13C NMR and HRMS
(ESI).
Structural optimization and SAR interpretation
As reported in our previous study, in the process of identi-
fying new chemical entities to treat rheumatoid arthritis by
inhibiting hDHODH, compound 7 (Fig. 1) was discovered as
a hDHODH inhibitor with moderate activity.14 In this paper,
we aimed to explore more potent hDHODH inhibitors with
antiproliferative efficacy through structure-based optimiza-
tion of this thiazole-bearing compound. Three distinctive
subsites in the ubiquinone-binding pocket of hDHODH were
comprehensively explored by introducing diverse functional
groups to the hydrazino-thiazole scaffold. The most potent
compound 22 achieved an IC50 value of 1.8 nM, almost a 100-
fold increase in enzymatic inhibitory activity compared with
unsubstituted compound 5. Structure–activity relationship of
compounds in this series suggested that the methyl group on
the free NH of the hydrazone linker at subsite S1 and the
Based on the X-ray crystal structure of hDHODH in complex
with 7 (PDB ID: 4LS1), this paper is aimed at further improv-
ing the inhibitory activities of the thiazole-bearing deriva-
tives.14 According to the binding mode of 7 (Fig. 2A), the
tunnel-like ubiquinone-binding pocket of hDHODH could be
mainly divided into two sections: an inner polar region and
an outer apolar region. The carboxyl group of compound 7
could form a perfectly oriented salt bridge with the charged
guanidine group of Arg136, while the thiazole and chloro-
phenyl groups contact residues Met43, Leu46, Ala59, Phe62,
and Met111 at the entrance of the binding tunnel with favor-
able hydrophobic interactions. Nevertheless, there are still
subsites that remain to be explored further, for example the
subsite S1 formed mainly by Ala59, Phe98 and Met111, the
subsite S2 formed by Met43, Leu46 and Gln47, and subsite
Med. Chem. Commun.
This journal is © The Royal Society of Chemistry 2016

View Article Online
MedChemComm
Research Article
Scheme 1 Reagents and conditions: (i) acetone, EtOH, reflux; (ii) Br2, acetic acid, anhydrous ether, 0 °C to rt; (iii) EtOH, reflux; (iv) 80% hydrazine
hydrate, EtOH, reflux; (v) 2-carboxybenzaldehyde, EtOH, rt; (vi) alkyl halides, Cs2CO3, DMF, 0 °C to 80 °C, Ar; (vii) HCl/H2O, THF, 50 °C, Ar; (viii)
2-carboxybenzaldehyde, EtOH, reflux.
Scheme 2 Reagents and conditions: (i) 1. n-BuLi, THF, −78 °C; 2. DMF, −78 °C to rt; (ii) NH4SCN, EtOH, reflux, 72 h; (iii) EtOH, reflux; (iv)
substituted 2-bromoacetophenone or 2-bromopropiophenone, EtOH, reflux.
Fig. 2 (A) The X-ray crystal structure of 7 (PDB ID: 4LS1). Three hydrophobic subsites are highlighted with white dashes. (B) Diagram for the struc-
tural optimization strategy. Some designed compounds are aligned to the binding conformation of 7 by SHAFTS.30
S3 formed by Val134 and Val143. Therefore, the structural op-
timization in this study was conducted with the purpose of
strengthening the hydrophobic interactions between the
benzylidenehydrazinyl-substituted thiazole derivatives and
This journal is © The Royal Society of Chemistry 2016
Med. Chem. Commun.

View Article Online
Research Article
MedChemComm
those distinctive hydrophobic subsites as well as interpreting
the structure–activity relationship. A comprehensive modifi-
cation was performed in three phases (Fig. 2B): (i) strength-
ening the hydrophobic interactions around subsite S1 by in-
troducing small apolar groups; (ii) structural modification of
the hydrazone linker to detect the distinctive hydrophobic
cavity around subsite S2; (iii) enhancing the hydrophobic in-
teractions around subsite S3. The inhibitory activities of the
synthesized compounds against hDHODH are displayed in
Table 1. Brequinar was chosen as the positive control in all
assays.
methyl and ortho-chlorine simultaneously (8), showed im-
proved activities compared with 5. However, replacing the
methyl (8) with ethyl (9) on the thiazole ring led to a 5-fold
loss of activity, suggesting that the space of the hydrophobic
concavity around residue Met111 was limited, and bulky hy-
drophobic moieties cannot be accommodated properly.
Given the superior biological activity of compound 7 in
the preliminary optimization, we accepted the ortho-chlorine
substituted benzene ring as the hydrophobic tail, and then
the impact of substitutions on R2 was further assessed at
subsite S2. There are two main reasons why we conducted
structural optimization at this subsite. On the one hand, if
the nitrogen atom of the hydrazone linker is unsubstituted,
the hydrazine moiety may be unstable and inclined to cyclize
into a pyridazinone with the adjacent carboxyl group intra-
molecularly. On the other hand, there was still enough space
to explore around the “slim” hydrazone linker. According to
the overall structure of the tunnel-like ubiquinone binding
site, residues around the linker could compose two hydro-
phobic pockets, one is a small closed cavity contributed by
Met43, the backbone of Gln47 and the side chain of Leu46,
the other one is an open region consisting of residues Met43,
Leu50, Leu58 and Ala59, which opens towards the bulky
solvent.8,31
We first investigated the effects of substitutes R3 and R4
on the hydrophobic tail. Since the hydrophobic subsite S1 is
comparatively spacious, the unsubstituted thiazole and phe-
nyl ring may not occupy the concave pocket thoroughly, lead-
ing to a comparative poor binding affinity (compound 5,
Table 1). Therefore, with this scaffold, we determined to in-
troduce several hydrophobic groups to the thiazole ring and/
or the phenyl ring to further explore the subsite S1. It is
worth noting that some attempts on the phenyl ring had
been performed in our previous report, such as 2,5-di-Cl,
meta-OCH3, meta-carbamyl, and para-methyl, however, none
of them exhibited better activity than that of ortho-chlorine
substituted 7. In this study, we mainly focused on whether
small hydrophobic groups on the thiazole ring were favorable
for the potency. To our delight, introducing a methyl group
on R3 (6) or ortho-chlorine on R4 (7), or introducing both
Firstly, a methyl group was introduced to R2 (10), and the
inhibitory activity of compound 10 against hDHODH dramati-
cally improved to single digit nanomolar with an IC50 value
Table 1 Structure and activity profiles of compounds 5–23 against hDHODH enzyme, human cancer cell lines (HCT-116 and BxPC-3) and human nor-
mal cell line Wi-38
IC50 (μM)
Compd
R1
R2
R3
R4
hDHODHa
HCT-116b
BxPC-3c
Wi-38d
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
CH3
H
CH3
C2H5
H
H
H
H
H
H
H
CH3
H
H
H
H
H
CH3
H
H
0.139 0.001
0.081 0.000
0.029 0.001
0.066 0.000
0.356 0.006
0.008 0.000
0.059 0.002
0.009 0.000
0.093 0.001
0.425 0.009
0.024 0.000
0.019 0.001
0.015 0.000
0.028 0.000
0.011 0.000
0.087 0.001
0.008 0.000
0.002 0.000
0.014 0.001
0.008 0.000
0.880 0.414
6.422 0.165
0.713 0.029
0.375 0.019
0.922 0.011
0.031 0.004
0.299 0.006
0.276 0.005
0.108 0.012
0.763 0.045
0.244 0.073
0.588 0.016
0.181 0.054
0.301 0.052
0.268 0.002
2.004 0.071
0.054 0.003
0.023 0.000
0.087 0.004
0.158 0.004
18.924 0.407
19.778 0.407
3.414 0.014
2.969 0.020
2.791 0.013
0.080 0.003
0.199 0.003
0.263 0.020
0.312 0.019
4.319 0.136
0.536 0.002
0.988 0.042
0.801 0.003
1.610 0.022
1.590 0.008
2.975 0.034
0.203 0.033
0.088 0.001
0.300 0.000
0.940 0.025
34.193 2.465
21.490 1.128
39.098 0.198
44.643 0.149
22.396 0.411
9.580 0.015
9.982 0.666
10.834 0.019
18.836 2.199
48.747 0.990
8.302 0.174
29.132 0.295
10.139 0.119
61.046 0.217
6.572 0.208
60.863 1.033
17.020 1.462
29.738 2.676
41.835 2.322
33.954 0.491
2-Cl
2-Cl
2-Cl
2-Cl
2-Cl
2-Cl
2-Cl
2-Cl
2-Cl
2-Cl
2-Cl
2,5-di-Cl
3-Cl
2-Cl
2-Cl
2-Cl
H
H
CH3
C2H5
n-Pr
i-Pr
CH2CH2OH
2-C4H9
2-C5H11
2-C5H11
CH3
CH3
CH3
CH3
CH3
CH3
CF3
CH3
F
21
22
23
CH3
Brequinar
a
b
c
The IC50 values of compounds against hDHODH, in vitro assay. The IC50 values of compounds against HCT-116, in vitro assay. The IC50
values of compounds against BxPC-3, in vitro assay. d The IC50 values of compounds against Wi-38, in vitro assay.
Med. Chem. Commun.
This journal is © The Royal Society of Chemistry 2016

View Article Online
MedChemComm
Research Article
of 7.5 nM, which is as good as, if not better than, the positive
control brequinar (IC50 = 8.4 nM). Based on this successful
modification experience, we next substituted the free NH
with various bulky groups (compounds 11–17). With the in-
crease of the length of straight-chain alkane to three carbon
atoms, the bioactivities of compounds 11 (ethyl group) and
12 (propyl group) decreased first, and then increased to the
same potent level as 10, informing us that the ethyl group
was unfavorable, probably due to the limited size of the
above described small closed hydrophobic cavity. In order to
verify this hypothesis, we then substituted the free NH with
branched alkanes like isopropyl (13), 2-butyl (15) and 2-pentyl
(16). Just as expected, 13 and 15 displayed about 2-fold de-
crease in potency compared with 11 and 12 respectively.
Meanwhile, along with the increasing length of the branched
alkanes, the inhibitory activities of compounds 13, 15 and 16
also exhibited an increasing trend, suggesting that the long
branch of each substituent could accommodate itself well to
the large opening hydrophobic pocket mentioned above and
make favorable VDW interactions with the hydrophobic resi-
dues. If the beneficial propyl group was replaced by polar
ethoxyl, the bioactivity of compound 14 decreased signifi-
cantly, again corresponding to the hydrophobic nature of the
binding cavity around the linker region. Therefore, it could
be concluded that methyl or more than two carbon atoms
longer straight-chain alkanes are preferred substituents for
the secondary nitrogen atom of the hydrazone linker. Addi-
tionally, we synthesized compounds 18 and 19 to verify if
ortho-chlorine was also the better position for the substituent
of the phenyl ring with the existence of a methyl on the sec-
ondary nitrogen of the hydrazone linker. Unfortunately, nei-
ther the 2,5-di-Cl substituted 18 nor the 3-Cl substituted 19
showed an improved inhibitory activity against hDHODH
compared with the 2-Cl substituted 10, confirming that
ortho-chlorine was still more favorable in this region to some
extent.
In the third round of optimization, we focused on the
small hydrophobic cavity capped by Val134 and Val143
(subsite S3), which can be occupied by a small hydrophobic
group R1 such as CF3, CH3 and F. Nevertheless, the introduc-
tion of CF3 on R1 (20) reduced the inhibitory activity sharply,
with an 11-fold decrease of IC50 value compared with 10. This
is probably due to a steric clash induced by trifluoromethyl.
Then we changed relatively small groups, methyl (21) and
fluorine (22), to replace the trifluoromethyl moiety on R1.
Both of them brought impressive boosts in potency com-
pared with 20. Specifically, compound 21 (IC50 = 8.4 nM) was
nearly equipotent with 10, and compound 22 was surprisingly
more potent than 10 with an IC50 value of 1.8 nM, which is
also the most potent hDHODH inhibitor reported so far. Fi-
nally, we revisited the hydrophobic entrance of the binding
pocket, using a methyl group on the thiazole ring (23) to re-
place the ortho-chlorine atom on the phenyl ring of com-
pound 21. However, the inhibitory activities against hDHODH
of the two compounds (23, IC50 = 13.6 nM vs. 21, IC50 = 8.4
nM) again demonstrated that the ortho-chlorine on the ben-
zene ring was better than the methyl on the thiazole ring to
interact with the surrounding hydrophobic residues.
Binding mode analysis
Molecular docking was conducted to further study the action
mechanism of this series of derivatives. The predicted bind-
ing modes of compounds 17 and 22 are shown in
Fig. 3A and B with a docking energy of −14.18 kcal mol−1 and
−14.86 kcal mol−1 respectively. It is apparent that the tunnel-
like binding pocket consists of a comparative polar narrow
end and a hydrophilic entrance. Like compound 7 in the crys-
tal structure, both of the carboxyl-substituted phenyl rings of
17 and 22 are located at the inner polar region of the binding
pocket with the carboxyl group forming a perfectly oriented
salt bridge with the guanidine group of Arg136. Meanwhile,
W659 bridges the hydrogen bond interactions of Gln47, the
sidechain oxygen atom of Thr360 and one oxygen atom of the
carboxyl group of 17 or 22. The binding modes of the thiazole
and chlorophenyl groups are identical to those of compound
7 in Fig. 2A.
For compound 17, the methyl group substituted at the thi-
azole ring could contact with residues Phe98, M111 and
Leu359 by hydrophobic interactions (Fig. 3A). Because of
these favorable hydrophobic interactions at subsite S1
(Fig. 2A), both compounds 6 and 7 displayed improved bioac-
tivities compared with 5. However, due to the limited space
around residue Met111, bulky hydrophobic moieties like
ethyl would not be accommodated properly, thus compound
9 is about 5-fold less potent than compound 8 (355.7 nM vs.
66.2 nM). The 2-pentyl group of compound 17 is occupying
simultaneously a small closed cavity contributed by Met43,
the backbone of Gln47 and the side chain of Leu46 and a
larger hydrophobic cavity consisting of residues Leu46, Leu50
and Leu58 around subsite S2. The small cavity described
above has limited space that could only accommodate small
groups, with methyl (10) performing better than ethyl (11).
The longer propyl group (12) could not be suitable for the
small cavity any more but for the large hydrophobic cavity
around Leu58, leading to an improved bioactivity compared
to 11. Furthermore, the isopropyl group (13) and the 2-butyl
group (15) present about 2-fold decrease in potency com-
pared with 11 and 12 respectively, proving again that a bulky
group like ethyl is not preferred in the small cavity around
Met43. However, the large hydrophobic cavity around Leu58
is spacious and could accommodate a bulky group like
2-pentyl. Therefore, compounds 16 and 17 displayed im-
proved bioactivities compared to 13 or 15.
For compound 22, besides the well accommodated methyl
at R2, the fluorine atom at R1 can fit pretty well in the small
hydrophobic cavity capped by Val134 and Val143 (Fig. 3B).
The bioactivities of the compounds decreased as the volume
of the hydrophobic substituents at R1 increased (22, 21 and
20), suggesting that small hydrophobic groups are preferred
at subsite S3. Combined with the favorable ortho-chlorine at
R4, compound 22 displayed the most potent hDHODH inhibi-
tory activity reported so far.
This journal is © The Royal Society of Chemistry 2016
Med. Chem. Commun.

View Article Online
Research Article
MedChemComm
Fig. 3 Proposed binding modes for representative compounds 17 (A) and 22 (B). Crystal structure of hDHODH (PDB ID: 4LS1) is shown as a light
blue cartoon, and some crucial parts are covered with a blue surface. Docked inhibitors are represented as purple sticks with a light cyan surface,
and the predicted binding conformation of each inhibitor is aligned to the binding conformation of 7 in the X-ray structure of 4LS1 (blue stick) at
the bottom right of each image. Key residues (thin sticks) are colored blue. Potential intermolecular hydrogen bonds are shown as black dashed
lines.
Antiproliferative activity
Cell cycle arrest and apoptosis assays
The synthesized compounds along with the standard bre-
quinar were evaluated for antiproliferative activity against
two human cancer cell lines named HCT-116 and BxPC-3,
and one human normal cell line called Wi-38. Dose–response
curves against the three cell lines were plotted with 7 differ-
ent analysis points i.e. with 7 different compound concentra-
tions and the IC50 values (Table 1), determined using Origin
8.0. It was observed that these compounds exhibit selective
antiproliferative inhibition against cancer cells rather than
normal cell. As depicted in Table 1, the antiproliferative activ-
ities exhibited a dramatic improvement compared with the
unsubstituted compound 5 after our structural optimization,
a 37-fold increase against HCT-116 (from 0.8797 μM to
0.0234 μM) and a 213-fold increase against BxPC-3 (from
18.924 μM to 0.088 μM), respectively. This suggested that the
substitutions on the three decorated hydrophobic regions
contributed greatly to the potency of synthesized compounds,
and the antiproliferative activities against Wi-38 were inferior
to HCT-116 and BxPC-3 from several to thousands fold. Fur-
thermore, the IC50 values of final compound 22 (0.023 μM
and 0.088 μM) were more efficient than the positive control
brequinar (0.158 μM and 0.940 μM) against the two cancer
cell lines.
With the most encouraging antiproliferative profile of com-
pound 22 in hand, we next investigated the possible mecha-
nism of cytotoxicity by exploring whether 22 could induce ap-
optosis and affect the cancer cell cycle arrest using flow
cytometry. The HCT-116 and BxPC-3 cells were treated with
It is shown in Table 1 that inhibitors with poor activities
against hDHODH enzyme also possess inferior anti-
proliferative activities against both HCT-116 and BxPC-3 cell
lines, and vice versa. The inhibitors' enzymatic inhibitory ac-
tivities and antiproliferative activities correlate well with each
other (Fig. 4), suggesting that the cytotoxicity of these com-
pounds was mainly attributed to the suppression of
hDHODH enzyme. Thus, this study provided a proof to sup-
port that hDHODH is a potential target for the treatment of
cancer.
Fig. 4 The correlation between the hDHODH inhibitory activity and
the antiproliferative activity against (A) HCT-116 cancer cell line and (B)
BxPC-3 cancer cell line.
Med. Chem. Commun.
This journal is © The Royal Society of Chemistry 2016

View Article Online
MedChemComm
Research Article
22 (0.2 μM) for 24 h, 36 h, and 48 h respectively, and then
the apoptosis ratio of cells were accessed by annexin V-FITC/
PI staining assay. The results showed that compound 22
caused apoptosis dramatically in HCT-116 and BxPC-3 cell
lines compared with the control group treated with DMSO,
and late apoptosis induction was pretty evident in both cell
lines (Fig. 5A). Furthermore, compound 22 was found to in-
crease the cell numbers in S phase after 12 h treatment with
different concentrations (5 nM, 10 nM, and 20 nM), accom-
panied by decreased cell numbers in G2/M phase in HCT-116
and BxPC-3 cell lines. As shown in Fig. 5B, the percentage of
HCT-116 cells in S phase increased from 42.24% to 64.92%
with the treatment of 20 nM 22, and BxPC-3 cells from
37.61% to 73.64%. Besides, the cell numbers of BxPC-3 cell
line in G0/G1 phase also reduced slightly after treatment of
22 at relatively high concentrations (10 and 20 nM), which
was not observed in HCT-116 cell line (Fig. 5B). These results
indicated that inducing apoptosis and cell cycle arrest may
be a possible reason for antiproliferation of compound 22.
hDHODH. By modifying this amphipathic scaffold, a series of
novel thiazole-bearing derivatives were designed, synthesized
and pharmacologically evaluated. As a result, most of the
compounds possessed comparable to remarkable inhibitory
activities against hDHODH compared with brequinar. Com-
bined with our previous study, molecular docking was
performed and led us extensively explore the structure–activ-
ity relationship of compounds in this series. Compounds
with an ortho-chlorine substituted phenyl ring on the hydro-
phobic tail displayed good activity. In addition, methyl or
more than two carbon atoms straight-chain alkanes at the
secondary nitrogen atom of the hydrazone linker also con-
tributed significantly to the potency of the synthesized com-
pounds. Furthermore, the most potent compound 22 was
chosen to perform flow cytometry analysis, suggesting that 22
could induce S phase cell cycle arrest and promote apoptosis.
Acknowledgements
This work was supported by the National Natural Science
Foundation of China (grants 21372078, 81302697, 81230090
and 81230076) (Z. Z., L. Z., H. L.), the Shanghai Committee of
Science and Technology (grant 14431902400) (H. L.), the Na-
tional S&T Major Project of China (grant 2013ZX09507004)
and the Twelfth Five-Year National Science & Technology
Conclusions
In summary, we successfully discovered a promising scaffold,
(E)-4-phenyl-2-(((2-carboxyl)benzylidene)hydrazinyl)thiazole, for
inhibiting the growth of cancerous cells selectively targeting
Fig. 5 (A) HCT-116 and BxPC-3 cells were treated with 0.2 μM compound 22 for 24 h, 36 h and 48 h respectively. The induction of apoptosis was
analyzed by flow cytometry using Annexin V-FITC/PI staining. (B) Compound 22 induced cell cycle arrest of HCT-116 and BxPC-3 cell lines in a
concentration-dependent manner. Cells were stained with PI after treatment with different concentrations (5 nM, 10 nM, 20 nM) of 22 for 12 h.
DNA content was determined by flow cytometry.
This journal is © The Royal Society of Chemistry 2016
Med. Chem. Commun.

View Article Online
Research Article
MedChemComm
Support Program (grant 2012BAI29B06) (H. L.). Honglin Li is
also sponsored by the Innovation Program of Shanghai Mu-
nicipal Education Commission (grant 13SG32) and Fok Ying
Tung Education Foundation (141035).
16 N. Hail Jr, P. Chen, J. J. Kepa and L. R. Bushman, Apoptosis,
2012, 17, 258–268.
17 S. Zhu, X. Yan, Z. Xiang, H.-F. Ding and H. Cui, PLoS One,
2013, 8, e71555.
18 S.-F. Chen, R. L. Ruben and D. L. Dexter, Cancer Res.,
1986, 46, 5014–5019.
19 A. A. Khutornenko, V. V. Roudko, B. V. Chernyak, A. B.
Vartapetian, P. M. Chumakov and A. G. Evstafieva, Proc.
Natl. Acad. Sci. U. S. A., 2010, 107, 12828–12833.
20 R. Cody, D. Stewart, M. DeForni, M. Moore, B. Dallaire, N.
Azarnia and J. Gyves, Am. J. Clin. Oncol., 1993, 16, 526–528.
21 D. L. Dexter, D. P. Hesson, R. J. Ardecky, G. V. Rao, D. L.
Tippett, B. A. Dusak, K. D. Paull, J. Plowman, B. M. DeLarco
and V. Narayanan, Cancer Res., 1985, 45, 5563–5568.
22 X. Xu, J. W. Williams, J. Shen, H. Gong, D.-P. Yin, L. Blinder,
R. T. Elder, H. Sankary, A. Finnegan and A. S.-F. Chong,
J. Immunol., 1998, 160, 846–853.
23 M. A. Behnam, C. Nitsche, V. Boldescu and C. D. Klein,
J. Med. Chem., 2016, DOI: 10.1021/acs.jmedchem.5b01653.
24 K. L. Yeo, Y.-L. Chen, H. Y. Xu, H. Dong, Q.-Y. Wang, F.
Yokokawa and P.-Y. Shi, Antimicrob. Agents Chemother.,
2015, 59, 2086–2093.
25 O. P. Kulkarni, S. G. Sayyed, C. Kantner, M. Ryu, M. Schnurr,
M. Sárdy, J. Leban, R. Jankowsky, A. Ammendola, R. Doblhofer
and H.-J. Anders, Am. J. Pathol., 2010, 176, 2840–2847.
26 L. R. Fitzpatrick, L. Deml, C. Hofmann, J. S. Small, M.
Groeppel, S. Hamm, S. Lemstra, J. Leban and A.
Ammendola, Inflammatory Bowel Dis., 2010, 16, 1763–1777.
27 S. Cohen, G. W. Cannon, M. Schiff, A. Weaver, R. Fox, N.
Olsen, D. Furst, J. Sharp, L. Moreland and J. Caldwell,
Arthritis Rheum., 2001, 44, 1984–1992.
28 H. A. Burris III, E. Raymond, A. Awada, J. G. Kuhn, T. J.
O'Rourke, J. Brentzel, W. Lynch, S.-Y. P. King, T. D. Brown
and D. D. Von Hoff, Invest. New Drugs, 1998, 16, 19–27.
29 L. Makowka, L. S. Sher and D. V. Cramer, Immunol. Rev.,
1993, 136, 51–70.
Notes and references
1 M. Löffler, L. D. Fairbanks, E. Zameitat, A. M. Marinaki and
H. A. Simmonds, Trends Mol. Med., 2005, 11, 430–437.
2 M. Loeffler and E. Zameitat, Encycl. Biol. Chem., 2004, 3,
600–605.
3 V. Vyas and M. Ghate, Mini-Rev. Med. Chem., 2011, 11,
1039–1055.
4 F. Breedveld and J. Dayer, Ann. Rheum. Dis., 2000, 59,
841–849.
5 S. Nørager, K. F. Jensen, O. Björnberg and S. Larsen,
Structure, 2002, 10, 1211–1223.
6 J. Krungkrai, Biochim. Biophys. Acta, Biomembr., 1995, 1243,
351–360.
7 M. Hansen, J. Le Nours, E. Johansson, T. Antal, A. Ullrich,
M. Löffler and S. Larsen, Protein Sci., 2004, 13, 1031–1042.
8 S. Liu, E. A. Neidhardt, T. H. Grossman, T. Ocain and J.
Clardy, Structure, 2000, 8, 25–33.
9 H. Munier-Lehmann, P.-O. Vidalain, F. Tangy and Y. L.
Janin, J. Med. Chem., 2013, 56, 3148–3167.
10 N. A. Malmquist, R. Gujjar, P. K. Rathod and M. A. Phillips,
Biochemistry, 2008, 47, 2466–2475.
11 B. A. Palfey, O. Björnberg and K. F. Jensen, Biochemistry,
2001, 40, 4381–4390.
12 Y. Diao, W. Lu, H. Jin, J. Zhu, L. Han, M. Xu, R. Gao, X.
Shen, Z. Zhao, X. Liu, Y. Xu, J. Huang and H. Li, J. Med.
Chem., 2012, 55, 8341–8349.
13 J. Zhu, L. Han, Y. Diao, X. Ren, M. Xu, L. Xu, S. Li, Q. Li, D.
Dong, J. Huang, X. Liu, Z. Zhao, R. Wang, L. Zhu, Y. Xu, X.
Qian and H. Li, J. Med. Chem., 2015, 58, 1123–1139.
14 S. Li, G. Luan, X. Ren, W. Song, L. Xu, M. Xu, J. Zhu, D.
Dong, Y. Diao, X. Liu, L. Zhu, R. Wang, Z. Zhao, Y. Xu and
H. Li, Sci. Rep., 2015, 5, 14836.
30 X. Liu, H. Jiang and H. Li, J. Chem. Inf. Model., 2011, 51,
2372–2385.
31 R. Baumgartner, M. Walloschek, M. Kralik, A. Gotschlich, S.
Tasler, J. Mies and J. Leban, J. Med. Chem., 2006, 49,
1239–1247.
15 M. L. Herrmann, R. Schleyerbach and B. J. Kirschbaum,
Immunopharmacology, 2000, 47, 273–289.
Med. Chem. Commun.
This journal is © The Royal Society of Chemistry 2016