Structural optimization of non-nucleoside DNA methyltransferase inhibitor as anti-cancer agent

Article abstract of DOI:10.1016/j.bmcl.2016.01.020

Inhibition of DNA methyltransferase 1 (DNMT1) can reverse the malignant behavior of cancer cells by restoring expression of aberrantly silenced genes that are required for differentiation, senescence, and apoptosis. Clinically used DNMT1 inhibitors decitabine and azacitidine inhibit their target by covalent trapping after incorporation into DNA as azacytidine analogs. These nucleoside compounds are prone to rapid enzymatic inactivation in blood, posing challenges to the development of purely epigenetic dosing schedules. Non-nucleoside compounds that suppress expression or function of DNMT1 may overcome this problem. Using a high-throughput PCR-based site specific chromatin condensation assay, we identified a compound that reactivated Cyclin-Dependent Kinase Inhibitor 2A (CDKN2A) in myeloma cells and suppressed expression of DNMT1 from a library of 5120 chemically diverse small molecules. Lead optimization was performed to generate 26 new analogs with lung cancer proliferation and DNMT1 expression as activity readout. Two of the new derivatives showed 2 fold improvement of growth inhibiting potency and also decreased DNMT1 protein levels in lung cancer cells.

Full text of DOI:10.1016/j.bmcl.2016.01.020

Bioorganic & Medicinal Chemistry Letters 26 (2016) 1272–1275
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Structural optimization of non-nucleoside DNA methyltransferase
inhibitor as anti-cancer agent
Bo Zhong a, Sergei Vatolin b, Nethrie D. Idippily a, Rati Lama a, Laila A. Alhadad a, Frederic J. Reu b,
,
⇑
Bin Su a,c,
⇑
a Department of Chemistry, College of Sciences and Health Professions, Cleveland State University, 2121 Euclid Ave., Cleveland, OH 44115, USA
b Department of Translational Hematology and Oncology Research, Taussig Cancer Institute, Cleveland Clinic, 9500 Euclid Ave., Cleveland, OH 44195, USA
c Center for Gene Regulation in Health and Disease, College of Sciences & Health Professions, Cleveland State University, 2121 Euclid Ave., Cleveland, OH 44115, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Inhibition of DNA methyltransferase 1 (DNMT1) can reverse the malignant behavior of cancer cells by
restoring expression of aberrantly silenced genes that are required for differentiation, senescence, and
apoptosis. Clinically used DNMT1 inhibitors decitabine and azacitidine inhibit their target by covalent
trapping after incorporation into DNA as azacytidine analogs. These nucleoside compounds are prone
to rapid enzymatic inactivation in blood, posing challenges to the development of purely epigenetic dos-
ing schedules. Non-nucleoside compounds that suppress expression or function of DNMT1 may over-
come this problem. Using a high-throughput PCR-based site specific chromatin condensation assay, we
identified a compound that reactivated Cyclin-Dependent Kinase Inhibitor 2A (CDKN2A) in myeloma cells
and suppressed expression of DNMT1 from a library of 5120 chemically diverse small molecules. Lead
optimization was performed to generate 26 new analogs with lung cancer proliferation and DNMT1
expression as activity readout. Two of the new derivatives showed 2 fold improvement of growth inhibit-
ing potency and also decreased DNMT1 protein levels in lung cancer cells.
Received 11 November 2015
Revised 6 January 2016
Accepted 8 January 2016
Available online 9 January 2016
Keywords:
DNMT1
Cancer
Non-nucleosides
Drug development
Published by Elsevier Ltd.
Inhibition of the maintenance DNA methyltransferase
1
were epigenetically effective and suppressed DNMT1 protein
expression at low micromolar concentrations in myeloma cells
(Fig. 1). In the current study we evaluated whether this could be
reproduced in lung cancer cells and if the lead could be improved
through medicinal chemistry using growth inhibition and DNMT1
suppression as the readout.
The structural modification initially focused on two aromatic
rings of the lead compound depicted in Figure 1. We attempted
to reveal the relationship between the biological activity of com-
pounds and substituent effects in the A and B position. The chlorine
atom at the position A was substituted by different groups such as
bromo, iodo, or even replaced by electron-donating groups such as
methyl and methoxy. 4-Hydroxyphenyl of the B position can be
(DNMT1) can reverse the malignant behavior of diverse cancer
cells by restoring expression of aberrantly silenced genes that are
required for differentiation, senescence, and apoptosis. The aza
nucleosides decitabine and azacitidine are currently the only clin-
ically used DNMT1 inhibitors and exert their epigenetic effects
after incorporation into DNA.1,2 However, poor stability, rapid
metabolism, non-specific incorporation into DNA and cell cycle
dependent effect have made design of optimal treatment schedules
difficult.3–5 In addition, the non-specific incorporation of these
nucleotide analogs may cause un-wanted side effects. Novel non-
nucleoside DNMT1 inhibitors may allow safer, more predictable
epigenetic treatment since direct DNA toxicity should be avoid-
able. To search for alternative epigenetic agents, we have
developed a novel high-throughput PCR-based site specific chro-
matin condensation assay.6 A library of 5120 chemically diverse
small molecules was screened with the assay, and we identified
C
O
a
lead compound {3-(4-chlorophenyl)-5-[(4-hydroxyphenyl)
methylidene] imidazolidine-2} from 15 non-nucleoside hits that
OH
NH
N
Cl
O
B
⇑
Corresponding authors. Tel.: +1 216 368 3378; fax: +1 216 368 1300 (F.J.R.); tel.:
+1 216 687 9219; fax: +1 216 687 9298 (B.S.).
A
E-mail addresses: reuf@ccf.org (F.J. Reu), B.su@csuohio.edu (B. Su).
Figure 1. The chemical structure of the lead compound.
http://dx.doi.org/10.1016/j.bmcl.2016.01.020
0960-894X/Published by Elsevier Ltd.

B. Zhong et al. / Bioorg. Med. Chem. Lett. 26 (2016) 1272–1275
1273
modified to other substituted aryl groups such as 4-bromophenyl,
4-iodophenyl, 4-methylphenyl, 4-methoxyphenyl, 2-naphthyl, and
3,4,5-trimethoxyphenyl. The combinatory strategy was used to
generate a series of novel analogs of the lead compound with
diversity at the A–B moieties for further biological investigation.
The synthesis of the proposed compounds is illustrated in
Scheme 1. Substituted aniline reacts with ethyl isocyanatoacetate
in chloroform and then cyclizes under acidic condition to form
an intermediate 3-substituted hydantoin. 3-Substituted hydantoin
condenses with aryl aldehyde to generate the final product in the Z
configuration. Variations in two aromatic rings can be achieved by
using different anilines and aldehydes, respectively.
Next, we tried to explore if more hydrophobic and bulky groups
at the A moiety could improve the biologic activity. Therefore, we
further introduced more hydrophobic and bulky substituent
groups on the aromatic ring at the A moiety. The synthesis of this
group of compounds is illustrated in Scheme 2.
At last, we explored modifications of the middle imidazole ring
at the C moiety (Fig. 1). We were wondering if a sulfur substituted
ring could impact the activity of the compound. The synthesis of
the sulfur atom substituted analogs is illustrated in Scheme 3.
Overall, 26 new derivatives were synthesized, purified and charac-
terized with NMR and MS for the following biological study.
The screening of the compounds was performed using a cell
proliferation assay with H292 non-small cell lung cancer cells.
We first examined all the compounds at 200 lM, and selected
the potent ones for further does-dependent study (Table 1). At
the A moiety, replacement of 4-chloro with 3- methoxy, iodo,
methyl, carboxylic and bromo all led to the loss of activity (com-
pounds 7–9, 14, 16), suggesting that 3 position of the A ring with
a small group is not favorable for the biological activity, regardless
whether it was electro-donating or electro-withdrawing. However,
bulky group substitution at 3 position of the A ring maintains anti-
proliferative effects (compounds 17–19). Several very bulky groups
such as N-benzyl and N-cyclohexylmethyl retained biological
Ar
N
Ar
N
ArNH2
O
O
substituted-benzaldehdye
piperidine, EtOH
O
1) CH3Cl
+
O
OEt
HN
OCN
2) HCl, EtOH, H2O
HN
O
R
Ar = 4-chlorophenyl
Ar = 4-bromophenyl
Ar = 4-iodophenyl
Ar = 4-chlorophenyl
Ar = 4-bromophenyl
Ar = 4-iodophenyl
Ar = 4-methylphenyl
Ar = 4-methoxyphenyl
Ar = 3-iodophenyl
Ar = 4-methylphenyl R=4-hydroxyl
R=4-chloro
R=4-bromo
Ar = 4-methoxyphenyl
Ar = 3-iodophenyl
Ar = 3-bromophenyl
Ar = 3-methylphenyl
Ar = 3-methoxyphenyl
Ar = 3-bromophenyl R=4-methoxyl
R=3-naphthalen
Ar = 3-methoxyphenyl R=3.4.5-trimethoxyl
Ar = 3-methylphenyl
Scheme 1. Synthesis of A and B moiety aromatic ring substituted analogs.
COOH
O
SOCl2
1) CH3Cl
OEt
+
C2H5O
N
O
OCN
2) HCl, EtOH, H2O
O
O
NH2
HN
Substitution at
3 or 4 position
Substitution at
3 or 4 position
4-hydroxylbenzaldehdye
piperidine, EtOH
CONR
CONR
COOH
O
COCl
O
R-NH2
K2CO3
4-hydroxylbenzaldehdye
piperidine, EtOH
N
O
N
O
N
N
O
O
O
O
HN
HN
HN
HN
Substitution at
3 or 4 position
Substitution at
3 or 4 position
R=N-benzyl
R=N-cyclohexylmethyl
R=pyrrolidine
HO
R=N-4-chlorophenyl
R=N-4-chlorophenyl
HO
Substitution at
3 or 4 position
Scheme 2. Synthesis of more bulky A moiety derivatives.
Ar
N
ArNH2
Ar
N
S
4-Hydroxybenzaldehdye
piperidine, EtOH
1) CH3Cl
O
S
+
2) HCl, EtOH, H2O
OEt
O
HN
SCN
HN
O
Ar= 4-chlorophenyl
Ar= 4-iodophenyl
Ar= 4-chlorophenyl
Ar= 4-iodophenyl
OH
Scheme 3. Synthesis of analogs with sulfur atom substituted imidazole ring.

1274
B. Zhong et al. / Bioorg. Med. Chem. Lett. 26 (2016) 1272–1275
Table 1
Growth inhibitory effects of the DNMT1 inhibitors on H292 cells
O
R2
R1
N
Entry
IC50 (lM)
NH
X
Lead
1
2
3
4
5
6
7
8
R1 = 4-Chlorobenzene, R2 = 4-hydroxylbenzene, X = O
R1 = 4-iodobenzene, R2 = 4-hydroxylbenzene, X = O
R1 = 4-Chlorobenzene, R2 = 4-Chlorobenzene, X = O
R1 = 4-Chlorobenzene, R2 = 4-bromobenzene, X = O
R1 = 4-Chlorobenzene, R2 = 4-methoxylbenzene, X = O
R1 = 4-methylbenzene, R2 = 4-hydroxylbenzene, X = O
R1 = 4-methoxylbenzene, R2 = 4-hydroxylbenzene, X = O
R1 = 3-methoxylbenzene, R2 = 4-hydroxylbenzene, X = O
R1 = 3-iodobenzene, R2 = 4-hydroxylbenzene, X = O
R1 = 3-methylbenzene, R2 = 4-hydroxylbenzene, X = O
R1 = ethyl 4-benzoate, R2 = 4-hydroxylbenzene, X = O
41.3 12.5
25.5 13.2
35.6 13.8
46.6 12.6
32.1 11.4
27.9 11.5
>200
>200
>200
>200
>200
>200
>200
>200
>200
69.9 24.4
>200
124.6 44.2
23.5 10.9
17.9 8.3
>200
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
R
1 = 4-carboxylbenzene, R2 = 4-hydroxylbenzene, X = O
R1 = 4-Chlorobenzene, R2 = 3-naphthalen, X = O
R1 = 4-Chlorobenzene, R2 = 3,4,5-trimethoxylbenzene, X = O
R1 = 3-carboxylbenzene, R2 = 4-hydroxylbenzene, X = O
R1 = 4-bromobenzene, R2 = 4-hydroxylbenzene, X = O
R1 = 3-bromobenzene, R2 = 4-hydroxylbenzene, X = O
R1 = 3-N-(4-chlorophenyl)benzamide, R2 = 4-hydroxylbenzene, X = O
R1 = 3-N-benzyl-benzamide, R2 = 4-hydroxylbenzene, X = O
R1 = 3-N-cyclohexylmethyl-benzamide, R2 = 4-hydroxylbenzene, X = O
R1 = 3-(pyrrolidine-1-carbonyl)phenyl, R2 = 4-hydroxylbenzene, X = O
R1 = 4-N-(4-chlorophenyl)benzamide, R2 = 4-hydroxylbenzene, X = O
R1 = 4-N-benzyl-benzamide, R2 = 4-hydroxylbenzene, X = O
R1 = 4-N-cyclohexylmethyl-benzamide, R2 = 4-hydroxylbenzene, X = O
R1 = 4-(pyrrolidine-1-carbonyl)phenyl, R2 = 4-hydroxylbenzene, X = O
R1 = 4-chlorobenzene, R2 = 4-hydroxylbenzene, X = S
>200
>200
>200
>200
28.5 14.3
19.7 6.8
R1 = 4-iodobenzene, R2 = 4-hydroxylbenzene, X = S
4
3
2
1
*
**
0
1
2
4
5
M
18
19
0u
DMSO
50μM
AD 10
LE
Figure 2. Compounds affect DNMT1 protein levels. H292 cells were treated with DMSO and the compounds for 48 h. Levels of DNMT1 and b-actin were analyzed by Western
blot of cell extracts with specific antibodies. Bands of DNMT1 (upper panel) and b-actin (lower panel) were quantified using ImageJ (NIH) and normalized to b-actin. The
⁄
experiments were repeated three times and the representative one is listed in the figure. **P <0.001 versus control, <0.01 versus control.
activity. On the other hand, these bulky groups at 4-positon of the
A ring all led to the loss of activity (compounds 21–23). Pyrrolidine
group erased all the activity no matter if it was connected to 3 or 4
position at A ring (compounds 20 and 24). Bulky groups to replace
the B ring decreased activity (compounds 12 and 13). The naph-
thalene and 2,3,4-trimethoxylbenzene all impaired the activity.
We identified 11 new derivatives from the 26 compounds pool
to perform dose-dependent cell proliferation investigation, and
summarized a brief structure activity relationship (SAR) based on
the results of the IC50 values. The 4-hydroxyl group of the B ring
is critical for activity. Replacement of the hydroxyl group with
other halogens such as chloro or bromo groups impaired activity
compared to the lead compound (compounds 2 and 3). A 4-meth-
oxyl group at the B ring also decreased the activity (compound 4).
Sulfur atoms did not affect the activity of the compound (com-
pounds 25 and 26). The IC50 values for the corresponding oxygen
substituted imidazole or sulfur substituted imidazole are in the
similar range, as demonstrated by lead compound versus com-
pound 25 and compound 1 versus compound 26.
The lead compound decreased DNMT1 protein levels in mye-
loma cells. We were wondering if the new analogs had this activity
in lung cancer cells. We first examined the compound with IC50

B. Zhong et al. / Bioorg. Med. Chem. Lett. 26 (2016) 1272–1275
1275
values below 100 lM via western blot in H292 lung cancer cells,
Acknowledgments
and identified compounds 1, 2, 4, 5, 18 and 19 that decreased the
DNMT1 protein level. Interestingly, the lead compound did not
affect DNMT1 expression in lung cancer cells but substitution of
the A ring with 4-iodobenzene or 3-N-benzyl-benzamide led to sig-
nificant DNMT1 suppression at doses 50% lower than the ones used
for parent compound (Fig. 2). This suggests that cell specific differ-
ences in networks controlling DNMT1 expression may exist and
that analogs of this class would have to be evaluated in multiple
cell lines before development into DNMT1 suppressing drugs
begins. Furthermore, results suggest that this class of compounds
can have anti-neoplastic effects that are independent of DNMT1
suppression which warrant further study.
In summary, our findings indicate that non-nucleoside inhibi-
tors of DNMT1 expression can be developed. Further research into
their mechanism of action and optimization of DNMT1 suppressing
potency may yield novel epigenetic therapies with improved
therapeutic index.
This work was supported by Center for Gene Regulation in
Health and Disease (GRHD) of Cleveland State University and Ohio
Department of Development (ODOD) to B. Su and Scott Hamilton
CARES grant to F.J. Reu.
References and notes
1. Chen, Z.; Liu, S.; Tian, L.; Wu, M.; Ai, F.; Tang, W.; Zhao, L.; Ding, J.; Zhang, L.;
Tang, A. Oncotarget 2015, 6, 38139.
2. Zhou, D.; Wan, Y.; Xie, D.; Wang, Y.; Wei, J.; Yan, Q.; Lu, P.; Mo, L.; Xie, J.; Yang,
S.; Qi, X. Exp. Mol. Med. 2015, 47, e182.
3. Wijermans, P. W.; Lubbert, M.; Verhoef, G.; Klimek, V.; Bosly, A. Ann. Hematol.
2005, 84, 9.
4. Momparler, R. L. Semin. Oncol. 2005, 32, 443.
5. Marcucci, G.; Silverman, L.; Eller, M.; Lintz, L.; Beach, C. L. J. Clin. Pharmacol.
2005, 45, 597.
6. Vatolin, S.; Khan, S. N.; Reu, F. J. PLoS One 2012, 7, e44690.