Structure and property based design, synthesis and biological evaluation of γ-lactam based HDAC inhibitors

Article abstract of DOI:10.1016/j.bmcl.2010.12.079

Histone deacetylases (HDACs) are involved in post-translational modification and gene expression. Cancer cells recruited amounts of HDACs for their survival by epi-genetic down regulation of tumor suppressor genes. HDACs have been the promising targets for treatment of cancer, and many HDAC inhibitors have been investigated nowadays. In previous study, we synthesized δ-lactam core HDAC inhibitors which showed potent HDAC inhibitory activities as well as cancer cell growth inhibitory activities. Through QSAR study of the δ-lactam based inhibitors, the smaller core is suggested as more active than larger one because it fits better in narrow hydrophobic tunnel of the active pocket of HDAC enzyme. The smaller γ-lactam core HDAC inhibitors were designed and synthesized for biological and property optimization. Phenyl, naphthyl and thiophenyl groups were introduced as the cap groups. Hydrophobic and bulky cap groups increase potency of HDAC inhibition because of hydrophobic interaction between HDAC and inhibitors. In overall, γ-lactam based HDAC inhibitors showed more potent than δ-lactam analogues.

Full text of DOI:10.1016/j.bmcl.2010.12.079

Bioorganic & Medicinal Chemistry Letters 21 (2011) 1218–1221
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Structure and property based design, synthesis and biological evaluation
of c-lactam based HDAC inhibitors
Eunhyun Choi a, Chulho Lee a, Jung Eun Park a, Jeong Jea Seo a, Misun Cho a, Jong Soon Kang b,
Hwan Mook Kim b, Song-Kyu Park b, Kiho Lee b, Gyoonhee Han a,c,
⇑
a Translational Research Center for Protein Function Control, Department of Biotechnology, Yonsei University, Seodaemun-gu, Seoul 120-749, Republic of Korea
b Bioevaluation Center, Korea Research Institute of Bioscience and Biotechnology, Yangcheong, Ochang, Cheongwon, Chungbuk 363-883, Republic of Korea
c Department of Biomedical Sciences (WCU Program), Yonsei University, Seodaemun-gu, Seoul 120-749, Republic of Korea
a r t i c l e i n f o
a b s t r a c t
Article history:
Histone deacetylases (HDACs) are involved in post-translational modification and gene expression. Can-
cer cells recruited amounts of HDACs for their survival by epi-genetic down regulation of tumor suppres-
sor genes. HDACs have been the promising targets for treatment of cancer, and many HDAC inhibitors
have been investigated nowadays. In previous study, we synthesized d-lactam core HDAC inhibitors
which showed potent HDAC inhibitory activities as well as cancer cell growth inhibitory activities.
Through QSAR study of the d-lactam based inhibitors, the smaller core is suggested as more active than
larger one because it fits better in narrow hydrophobic tunnel of the active pocket of HDAC enzyme. The
smaller c-lactam core HDAC inhibitors were designed and synthesized for biological and property opti-
mization. Phenyl, naphthyl and thiophenyl groups were introduced as the cap groups. Hydrophobic and
bulky cap groups increase potency of HDAC inhibition because of hydrophobic interaction between HDAC
Received 14 September 2010
Revised 7 December 2010
Accepted 16 December 2010
Available online 21 December 2010
Keywords:
Histone deacetylase
HDAC
Inhibitor
Isoform
and inhibitors. In overall,
c
-lactam based HDAC inhibitors showed more potent than d-lactam analogues.
Anticancer
In vitro activity
Docking study
ADME
Ó 2010 Elsevier Ltd. All rights reserved.
Transcriptional modification on histone controls epi-gene
expression.1 Histone acetyltransferase and histone deacetylase are
involved and play main roles in theses epigenetic modification. His-
tone acetyltransferases (HATs) cause hyper-acetylation of histones
and leads to transcriptional activation, whereas histone deacety-
lases (HDACs) suppress gene expression.2 Hypo-acetylation of his-
tones and proteins are known to be related to many pathogenic
mechanisms, especially carcinogenesis. HDACs cause epigenetic
suppression of tumor suppressor genes during carcinogenic patho-
mechanism and are considered as promising targets for treatment
of cancer.3 Many HDAC inhibitors effectively arrest of cancer cell
growth and induce apoptosis, which are considers as non-cytotoxic
chemotherapeutic agents.4 Currently, many compounds are in both
preclinical development and clinical trials.5 For examples, US FDA
approved SAHA (1, ZolinzaÒ, Vorinostat)6 and depsipeptide (2, Isto-
daxÒ, Romidepsin)7 for treatment of cutaneous T cell lymphoma
(CTCL) in 2006 and 2009, respectively. MGCD0103 (3)8 is a benzam-
ide class of HDAC inhibitor and it inhibits HDAC class I and IV selec-
tively (Fig. 1). In our prior studies, we reported design and synthesis
of d-lactam based HDAC inhibitors (4) and their pharmacological
evaluation for the potential anti-cancer chemotherapeutic agents.
The d-lactam core HDAC inhibitors showed very good potencies to
inhibition of both HDAC enzyme and cancer cell growth in vitro
and in vivo experiments.9
We also performed docking simulation and 3-dimensional
quantitative structure activity relationship (3D QSAR) of these
d-lactam analogues using Discover/Insight II modules and QSAR+/
Cerius2 modules.10,11 The d-lactam core inhibitors bind well in
the active site of HDAC and the result of 3D QSAR rationalize that
the big size of core groups are difficult to fit in the narrow tubular
hydrophobic pocket of the enzyme. It is also rationalized that the
bulky and hydrophobic moiety on cap group showed better HDAC
inhibitory activities by stabilizing interaction on the surface of
HDAC enzyme with the cap group of the inhibitors. We, however,
observed the large cap groups increase lipophilicity (i.e., c-log P)
and lead to decrease the microsomal stability of the molecules.9a,12
Based on these rationales, a smaller core is suggested as more ac-
tive than larger one because it fits better in narrow hydrophobic
tunnel of the active pocket of HDAC enzyme. Thus, we designed
and synthesized c-lactam analogues for potency as well as micro-
somal stability, and report here their preliminary biological
evaluation.
The synthetic process of
c-lactam analogues is shown in
Scheme 1. The secondary amines (6) were prepared by N-alkyl-
ation of synthetic or commercial amines (5) with ally bromide.
⇑
Corresponding author. Tel.: +82 2 2123 2882; fax: +82 2 362 7265.
E-mail address: gyoonhee@yonsei.ac.kr (G. Han).
0960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.12.079

E. Choi et al. / Bioorg. Med. Chem. Lett. 21 (2011) 1218–1221
1219
N
O
H
N
N
N
H
NH2
OH
H
N
N
H
N
O
O
SAHA (1)
MGCD-0103 (3)
H
N
O
O
O
OH
NH
NH
S
O
O
N
N
n
HN
R
H
S
O
O
cap group
Zn2+ binding moiety
lactam core
O
Depsipeptide (2)
δ-Lactam HDAC Inhibitors (4)
Figure 1. Chemical structures of HDAC inhibitors.
O
O
ACHN (renal), HCT-15 (colon), NCI-H23 (non-small lung cancer),
NUGC-3 (gastric) and LOX-IMVI (melanoma). The HDAC inhibition
a
+
HO
O
N
H
n R
R
n NH2
of
c-lactam core inhibitors showed very promising activity at 0.01–
5
6
7
0.80
l
M in the range of IC50 which are better than d-lactam
b
analogues (Table 1 and Supplementary data). The most of these
analogues also showed good cancer cell growth inhibitory activi-
ties (GI50 = 0.26–10 lM, Table 1). HDAC inhibition of the analogues
O
O
O
R
n N
and their cancer cell growth inhibition are well correlated. Among
6 cancer cell lines, MDA-MB-231 and NUGC-3 cells are more sensi-
tive than others by comparing average GI50 values. For the struc-
tural rationale toward the activities, the carbon chain length is
affected on inhibitory activity; the longer chain is more active than
the shorter. The HDAC IC50 values of 10a, 10d and 10h are 0.80,
8
c
O
O
0.12 and 0.05
lM, respectively, and they are 0.15, 0.05, and
n N
0.01 M for 10b, 10e and 10i, respectively. This may be explained
l
O
R
that the structure of active pocket is narrow tunnel as longer chain
analogues showed tighter binding to zinc ion and then increase
inhibitory activity. Larger naphthyl groups on cap group (10b,
10f and 10i) are more potent than smaller phenyl group (10a,
10d and 10h), and 1-naphthyl (10f) is less potent than 2-naphthyl
(10e). These results also explained that larger hydrophobic cap
groups show extra stabilizing interaction on the surface of HDAC
enzyme. As a result, the hydrophobic and bulky groups on cap
group showed good potency. In addition to potent HDAC inhibi-
tion, most of these analogues increased acetylation level of histone
H3 and H4 in a dose-dependent manner.9d For HDAC isoform selec-
tivity, we examined inhibition of HDAC-1, 2, 3, 4 6 and 8 by 10i,
10j, 10k and trichostatin A (TSA), which are highly potent in whole
HDAC inhibition. The results of these assays indicated that these
analogues inhibited HDAC-6 more than other HDAC isoforms,
and HDAC-4 has no direct effect by three synthetic analogues (Ta-
ble 2). Among 3 compounds, 10i is most potent and selective to
HDAC-6.
The docking study of the most potent inhibitor, 10i and trichos-
tatin A to the human HDAC-2 (PDB code 3MAX) was performed be-
cause 10i inhibits HDAC-2 at IC50 of 16.6 nM which is the second
potent among the six HDAC isoforms. Although it inhibits HDAC-
6 very potently at IC50 of 0.8 nM, no crystal structure of HDAC-6
is available for docking study. The Docked structure of 10i and
HDAC-2 using Discovery Studio is demonstrated in Figure 2.14
The binding mode of 10i is similar to that of trichostatin A (TSA);
zinc ion in the active site is chelated with hydroxamic acids and
cap groups such as naphthyl moiety of 10i and N,N-dimethylamin-
ophenyl group of trichostatin A lies in hydrophobic surface of the
9
d
O
O
OH
n N
N
R
H
10
10a n = 1 R = Ph
10b n = 1 R = 2-Naphthyl
10d n = 2 R = Ph
10c n = 1 R = 6-Bromonaphthyl
10e
10g
10f
n = 2 R = 2-Naphthyl
n = 2 R = Thiophenyl
n = 2 R = 1-Naphthyl
n = 3 R = Ph
10h
10i n = 3 R = 2-Naphthyl
10j n = 3 R = Biphenyl
10k n = 3 R = 1-Bromonaphthyl
Scheme 1. Reagents and conditions: (a) allyl bromide, hunig’s base, CH3CN; (b)
EDC, DMAP, CH2Cl2; (c) Grubbs’ catalyst (2nd generation), CH2Cl2, reflux; (d) KONH2
(1.7 M in MeOH), MeOH, 0 °C.
EDC-medicated coupling of 6 with monoacid 7 afforded amides 8
in good yields. Metathesis reaction of amides 8 with 2–3 mol % of
Grubbs’ catalyst (the 2nd generation) yielded the
in high yields. The hydroxamic acid analogues 10 were obtained
by reacting with KONH2 in MeOH at low temperature. The pre-
-lactam analogues differ in carbon chain length between
-lactam core by extending from one to three. The
phenyl, naphthyl, and thiophenyl moieties were introduced as
c-lactam rings
pared
c
cap group and
c
the cap groups.
All of the prepared c-lactam analogues were evaluated on HDAC
inhibition and cancer cell growth inhibition. HDAC inhibition assay
was performed using HDAC Fluorescent Activity Assay protocol13
and cancer cell growth inhibitory activities were evaluated on six
human cancer cell lines; PC-3 (prostate), MDA-MB-231 (breast),
enzyme. Furthermore,
c-lactam core of the inhibitor 10i bound
in hydrophobic narrow pocket, which is stabilized by forming

1220
E. Choi et al. / Bioorg. Med. Chem. Lett. 21 (2011) 1218–1221
Table 1
HDAC enzyme and cancer cell growth Inhibition by
c-lactam analogues
a
a
Compound
IC50
(l
M)
GI50
(
lM)
HDAC
PC-3
MDA-MB-231
ACHN
HCT-15
NCI-H23
NUGC-3
LOX-IMVI
10a
10b
10c
10d
10e
10f
10g
10h
10i
0.80
0.15
0.07
0.12
0.05
0.36
0.30
0.05
0.01
0.04
0.02
0.18
30b
4.73
1.22
0.34
0.99
0.59
4.76
2.86
3.81
0.65
0.58
0.52
1.91
30b
1.39
NT
30b
30b
5.19
NT
NT
NT
NT
NT
0.80
5.23
NT
2.60
0.61
0.91
0.94
1.85
30b
4.88
1.16
4.43
0.34
5.76
30b
3.91
1.09
1.41
0.46
7.95
3.54
0.28
5.37
0.84
6.78
9.98
6.23
0.78
1.14
1.00
5.99
7.98
NT
1.08
0.63
6.41
4.50
4.24
0.65
0.77
0.76
6.01
4.29
0.50
5.13
30b
3.46
0.61
1.33
0.26
8.08
30b
0.61
3.71
30b
4.52
0.82
1.16
0.79
10.96
10j
10k
Ave.
a
Values are means of a minimum of three independent experiments. NT: not tested.
GI50 = 30 lM represents no activity in 10 lM.
b
Table 2
HDAC isoform inhibition of 10i, 10j and 10k and trichostatin A
Table 3
Physicochemical properties of
c
-lactam analogues
tPSAa
Compound
IC50 (nM)
Compound
Rotatable
log Da
bond numbera
HDAC 1 HDAC 2 HDAC 3 HDAC 4 HDAC 6 HDAC 8
10a
10b
10c
10d
10e
10f
10g
10h
10i
5
5
5
6
6
6
6
7
7
8
7
69.64
69.64
69.64
69.64
69.64
69.64
97.88
69.64
69.64
69.64
69.64
0.43
1.69
2.52
0.59
1.85
1.80
0.19
1.05
2.30
2.67
2.88
TSA
10i
10j
10k
3.9
25.6
76.5
84.3
7.6
16.4
24.0
23.2
5.2
80.1
82.3
7.690
NAa
0.6
0.8
6.6
2.2
73.1
119.7
248.6
331.1
NAa
119.8
NAa
a
NA: not active.
10j
10k
a
PREADME program, BMDRC.15
Compound 10i16 which showed the best activities of HDAC inhi-
bition and cancer cell growth inhibition was selected for further
in vitro and in vivo pharmacokinetic studies. The caco-2 cell per-
meability of 10i is good at 16.3 cm/s of Papp  106, and the result
of metabolic stability indicated 10i is moderately stable in mouse
liver microsomes (Table 4). The caco-2 cell permeability is similar
to d-lactam analogues but the microsomal stability is higher than
d-lactam inhibitors.8 The pharmacokinetic profiles of 10i are dis-
played in Table 5. It showed still high clearance (5.26 l/h/kg), short
half-life (t1/2 = 0.15, 0.64 h) and low oral bioavailability (F = 8.7%).
It may be cause by many rotatable bonds and high lipophilicity.
The bioavailability of 10i, however, is better compared to d-lactam
analogue. This rationale is an important consideration for further
optimization process.
Finally, in vivo tumor growth inhibitory activity of 10i using the
xenograft experiments was performed. To the female nude mice
Figure 2. Docked orientation of 10i (yellow) and TSA (green) to the catalytic site of
HDAC-2.
Table 4
The caco-2 cell permeability and metabolic stability of 10i
p
–
p
interaction between two aromatic side chains of Phe-145 and
Phe-210.
The thiophenyl analogue (10g) showed that good HDAC inhibi-
tory activity at 0.050 M of IC50 but its cancer cell growth inhibi-
Compound
Caco-2 cell
% Remaining at 60 mina
Papp  106 (cm/s)
À NADPH
+ NADPH
Buffer
97.2
l
10i
16.3
71.1
43.1
100.8d
1.2e
tory activities are fairly lower compared to the related analogue
(10f). The polar surface area (tPSA) of 10g is higher at 97.88, and
lipophilicity (log D) is lower at 0.19 than other compounds (Table
3). Higher polar surface area and lower lipophilicity may reduce
cell permeability and leads to lower inhibition of cancer cell
growth. Based on the results of in vitro assays, the inhibitory activ-
Negative control
Positive control
1.5 0.1b
29.6 0.8c
96.2d
86.8e
a
Microsomal stability was determined by incubating 1 lM of test compounds
with female ICR mouse liver microsomes (0.5 mg protein/mL) for 60 min at 37 °C.
b
Ranitidine.
Metoprolol.
Atenolol.
Chlorpromazine.
c
d
ities of the prepared
study and QSAR rationale of d-lactam core HDAC inhibitors.
c-lactam analogues are relevant to docking
e

E. Choi et al. / Bioorg. Med. Chem. Lett. 21 (2011) 1218–1221
1221
Table 5
and are seemed more promising than d-lactam analogues. The Ap-
proach to discovery of a potent compound in vivo experiments and
optimization processes for orally active candidate are in progress.
The pharmacokinetics profiles of 10i in Balb/c mouse
a
Parameter
Unit
10i
IV
PO
Acknowledgments
Dose
Tmax
Cmax
(mg/kg)
(h)
5
20
—
—
890.2
5.62
0.85
0.15
0.25
813
308.6
—
—
0.64
8.7
This research was supported by National Research Foundation
(2009-0092966), Korea Research WCU grant (R31-2008-000-
10086-0) and Brain Korea 21 project, Republic of Korea.
(ng/ml)
(h ng/ml)
(l/h/kg)
(l/kg)
(h)
b
AUC0-inf
CL
Vss
t1/2
F
Supplementary data
(%)
a
PK parameters were based on mean plasma concentration–time profiles of
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmcl.2010.12.079.
three animals per time points. PK parameters were calculated by noncompart-
mental analysis using PK Solutions 2.0 (Summit Research Services, Montrose, CO,
USA).
AUC after iv administration was calculated from 0 to infinity, whereas the oral
AUC was calculated from 0 to 5 h.
b
References and notes
1. Magic, Z.; Supic, G.; Brankovic-Magic, M. J. BUON 2009, 14, S79.
2. Pontiki, E.; Hadjipavlou-Litina, D. Med. Res. Rev. 2010.
3. Budillon, A.; Di Gennaro, E.; Bruzzese, F.; Rocco, M.; Manzo, G.; Caraglia, M.
Recent Patents Anticancer Drug Disc. 2007, 2, 119.
4. Kang, M. R.; Kang, J. S.; Han, S. B.; Kim, J. H.; Kim, D. M.; Lee, K.; Lee, C. W.; Lee,
K. H.; Lee, C. H.; Han, G.; Kim, H. M.; Park, S. K. Biochem. Pharmacol. 2009, 78,
486.
5. Tan, J.; Cang, S.; Ma, Y.; Petrillo, R. L.; Liu, D. J. Hematol. Oncol. 2010, 3, 5.
6. Kavanaugh, S. M.; White, L. A.; Kolesar, J. M. Am. J. Health Syst. Pharm. 2010, 67,
793.
Table 6
In vivo tumor growth inhibition of 10i in mouse xenograft model
Dosea
(mg/kg)
Tumor
% Inhibition to
the vehicle
volume (mm3)
Vehicle
(n = 6)
10i (n = 6)
0
369.8
—
30
242.6***
34.4***
7. Hymes, K. B. Clin. Lymphoma Myeloma Leuk. 2010, 10, 98.
8. Prince, H. M.; Bishton, M. J.; Harrison, S. J. Clin. Cancer Res. 2009, 15, 3958.
9. (a) Yoon, H. C.; Choi, E.; Park, J. E.; Cho, M.; Seo, J. J.; Oh, S. J.; Kang, J. S.; Kim, H.
M.; Park, S. K.; Lee, K.; Han, G. Bioorg. Med. Chem. Lett. 2010, 20, 6808; (b) Kim,
H. M.; Lee, K.; Park, B. W.; Ryu, D. K.; Kim, K.; Lee, C. W.; Park, S. K.; Han, J. W.;
Lee, H. Y.; Han, G. Bioorg. Med. Chem. Lett. 2006, 16, 4068; (c) Kim, H. M.; Hong,
S. H.; Kim, M. S.; Lee, C. W.; Kang, J. S.; Lee, K.; Park, S. K.; Han, J. W.; Lee, H. Y.;
Choi, Y.; Kwon, H. J.; Han, G. Bioorg. Med. Chem. Lett. 2007, 17, 6234; (d) Kwon,
H. K.; Ahn, S. H.; Park, S. H.; Park, J. H.; Park, J. W.; Kim, H. M.; Park, S. K.; Lee, K.;
Lee, C. W.; Choi, E.; Han, G.; Han, J. W. Biol. Pharm. Bull. 2009, 32, 1723.
10. Yang, J. S.; Chun, T. G.; Nam, K. Y.; Han, G., unpublished result.
11. Kim, H. M.; Ryu, D. K.; Choi, Y.; Park, B. W.; Lee, K.; Han, S. B.; Lee, C. W.; Kang,
M. R.; Kang, J. S.; Boovanahalli, S. K.; Park, S. K.; Han, J. W.; Chun, T. G.; Lee, H.
Y.; Nam, K. Y.; Choi, E. H.; Han, G. J. Med. Chem. 2007, 50, 2737.
12. Kim, H. M.; Oh, S. J.; Park, S. K.; Han, G.; Kim, K.; Lee, K. S.; Kang, J. S.; Nam, M.;
Lee, K. Xenobiotica 2008, 38, 281.
a
Daily treatment (20 days); ip administration.
p <0.001.
***
(seven weeks), HCT-116 (human colorectal cancer) was trans-
planted (2 Â 107 cells/mL). After intraperitoneal administration of
10i at 30 mg/kg for 20 days, it displayed no body weight change
and 34.3% of tumor growth inhibition compared to vehicle (Table
6).
In conclusion, based on the SAR and property optimization
study of our previous work regarding on d-lactam based HDAC
inhibitors, we designed five-membered
13. HDAC Fluorescent Activity Assay/Drug Discovery Kit
International, Inc.
-
AK-500; BIOMOLÒ
c
-lactam core HDAC inhib-
itors for biological and property optimization. The
c-lactam core
14. Accelrys, Discovery StudioÒ 2.5.
has reduced ring size and results in better binding in narrow tunnel
of active site. Phenyl, naphthyl and thiophenyl groups were intro-
duced as the cap groups. Hydrophobic and bulky cap groups in-
crease potency of HDAC inhibition because of better hydrophobic
interaction between HDAC and inhibitors. The analogues with
longer chain length are more active than those with shorter chain.
15. BMDRC, PREADMET 2.0, Bmdrc, Seoul, Korea, 2007.
16. N-Hydroxy-3-[1-(3-naphthalen-2-yl-propyl)-2-oxo-2,5-dihydro-1H-pyrrol-3-yl]-
propionamide (10i): 1H NMR (CDCl3) d 10.54 (s, 1H), 7.71 (dd, 3H, JA = 7.9 Hz,
JB = 6.1 Hz), 7.54 (s, 1H), 7.41–7.33 (m, 2H), 7.22 (s, 1H), 6.64 (s, 1H), 3.69 (t, 2H,
J = 7.1 Hz), 3.40 (s, 2H), 2.69 (t, 2H, J = 6.7 Hz), 2.57 (s, 2H), 2.41 (s, 2H), 1.86 (s,
2H); 13C NMR (CDCl3) d 171.71, 169.96, 138.73, 137.72, 136.65, 133.51, 131.96,
127.90, 127.49, 127.36, 127.03, 126.30, 125.90, 125.16, 50.94, 42.16, 33.16,
31.06, 29.99, 21.98; ESI (m/z) 339.3 (MNa+).
In overall,
c-lactam based HDAC inhibitors showed more potent